RIGHT TO HEALTH CARE
Item
- Title
- RIGHT TO HEALTH CARE
- extracted text
-
RF_L_3_B_SUDHA
Articles
Effect of scaling up women's groups on birth outcomes in
three rural districts in Bangladesh: a cluster-randomised
controlled trial
Kishwar Azad, Sarah Barnett, Biplob Banerjee, Sanjit Shaha, Kasmin Khan, Z\rati Roseiyn Rego, Shampa Barua, Dorothy Flatman, Christina Pagel,
Audrey Prost, Matthew Ellis, Anthony Costello
Summary
Background Two recent trials have shown that women's groups can reduce neonatal mortality in poor communities.
We assessed the effectiveness of a scaled-up development programme with women’s groups to address maternal and
neonatal care in three rural districts of Bangladesh.
Methods 18 clusters (with a mean population of 27953 [SD 5953]) in three districts were randomly assigned to either
intervention or control (nine dusters each) by use of stratified randomisation. For each district, cluster names were
written on pieces of paper, which w ere folded and placed in a bottle. The first three cluster names drawn from the
bottle were allocated to the intervention group and the remaining three to control. All clusters received health services
strengthening and basic training of traditional birth attendants. In intervention clusters, a facilitator convened
18 groups ever); month to support participatory action and learning for women, and to develop and implement
strategies to address maternal and neonatal health problems. Women were eligible to participate if they were aged
15-49 years, residing in the project area, and had given birth during the study period (Feb 1, 2005, to Dec 31, 2007).
Neither study investigators nor participants were masked to treatment assignment. In a population of 229195 people
(intenention clusters only), 162 women's groups provided coverage of one group per 1414 population. The primary
outcome was neonatal mortality rate (NMR). Analysis was by intention to treat. This trial is registered as an
International Standard Randomised Controlled Trial, number ISRCTN54792066.
Lancet 2010; 375:1193 202
published u h
March 8.2010
D0l:10.1016/S01406736(10)60142-0
See Comment page 1142
See Articles page 1182
Perinatal Care Project, Diabetic
Association of Bangladesh,
Shahbagh, Dhaka, Bangladesh
(Prof K Azad FRCPCH. B Banerjee.
$ Shaha, K Khan. A R Rego,
S Barua); University College
London Centre for
International Health and
Development, institute of
Child Health (S Barnett PhD.
A Prost PhD
Prof A Costello FRCP), and
Clinical Operational Research
Findings We monitored outcomes for 36113 births (intervention dusters, n=17 514; control clusters, n=18 599) in a
population of 503163 over 3 years. From 2005 to 2007. there were 570 neonatal deaths in the intervention dusters and
656 in the control dusters. Cluster-level mean NMR (adjusted for stratification and clustering) was 33-9 deaths per
1000 livebirths in the intervention dusters compared with 36-5 per 1000 in the control dusters (risk ratio 0-93,
95% CI 0 -80-1 09).
Unit (C Pagel PhD), University
College London, London. UK;
Women and Children First,
London, UK (D Flatman MSc);
Health Foundation, London,
UK (D Flatman); and Centre for
Child and Adolescent Health,
interpretation For participatory women’s groups to have a significant effect on neonatal mortality in rural Bangladesh,
detailed attention to programme design and contextual factors, enhanced population coverage, and increased
enrolment of newly pregnant women might be needed.
University of Bristol, Bristol,
UK(Mi:ilis PhD)
Correspondence to;
Prof Kishwar Azad, Perinatal Care
Project, Diabetic Association of
Bangladesh (BADAS).
Funding Women and Children First, the UK Big Lottery Fund, Saving Newborn Lives, and the UK Department for
International Development.
Avenue, Shahbagh, Dhaka-1000.
introduction
pcp@dab-bd.org
BIRDEM-122 Kazi Nazrul Islam
Bangladesh
An estimated 3-7 million neonatal deaths occur
worldwide every year,1 98% of which are in developing
countries. In Bangladesh, the neonatal mortality rate
(NMR) declined from 63 per 1000 livebirths in 1985-89 to
34 per 1000 in 2002-06.2 4 A recent survey showed that
around 85% of births occur at home.’ Since around 45%
of deaths in children under 5 years of age in Bangladesh
occur in the first month of life, further progress in
reducing neonatal mortality is essential to achieve
Millennium Development Goal 4 (to reduce child
mortality by two-thirds by 2015).(’ This progress would
require community-based interventions to improve the
supply and demand for maternal and neonatal care.
We tested a low-cost, participatory, community-based
approach to improving birth outcomes in rural areas in
two cluster-randomised controlled trials: the first in
www thelancet.com Vol 375 April 3, 2010
Makwanpur, Nepal (2001-03),7 and the second in
Jharkhand and Orissa, India (2005-08)? Participatory
women’s groups reduced neonatal mortality by 30% in
Nepal over years 2 and 3 and by 32% in rural India over
the 3-year study period. We also assessed a similar
approach in a larger population within three rural
districts of Bangladesh." We recruited local female peer
facilitators who undertook twice as many meetings than
did facilitators in the Nepal trial (18 us nine) and covered
a population two to five times larger. We also introduced
a key informant system to monitor maternal and
neonatal mortality rates with detailed interview's or
verbal autopsies for all births, neonatal deaths, and
maternal deaths.'41 Our aim was to test the generalisability
and scalability of this community-based participatory
approach with women’s groups. A cluster-randomised
design was used in the trial because the women’s group
1193
Articles
Phase 1
Identify and prioritise difficulties
Introduce
project
Why
mothers
and babies
die
Problems
during
pregnancy
Problems
uuimg
during
delivery
Problems
after
delivery
Plan
Problems
community
for
newborn * visit
baby
Share
community
views
Phase 2
Prepare for community meeting
: Discuss strategies to address problems
i and plan involvement of community
Women were eligible to participate in the dtudy if they
were aged 15—49 years, residing in the project area, and
had given birth during the study period (Feb 1, 2005, to
Dec 31, 2007). The study population was an op'en cohort—
ie, women could enter the study at any time during the
trial period if they had given birth. Data was Obtained for
all eligible women throughout the study period.
Ethics approval was obtained from the ethics
committees of BADAS and the University College
London
Institute of Child Health. Women who chose to
Plan strategies
participate in the study during the baseline survey and
the period of prospective surveillance gave verbal consent
and were free to decline an interview at any time.
V
Randomisa
Community meeting
Each district constituted one stratum and each union a
cluster (see webappendix p 2). 18 unions (six per district)
Phase 3
Put strategies into practice
were selected. The total population vpthin these
18 unions was 503163 people, with union sizes ranging
Phase 4
from 15441 to 35110. Unions were randomly allocated to
Assess effect
either intervention or control groups by district in the
presence of four project staff (including the project
Figure 1: Description of women’s group meetings in the community action cycle
director and project manager) and two external
individuals (Nazmun Nahar, Department ojf Paediatrics.
intervention was implemented at a community rather Dhaka Medical College, Dhaka, and Azad Khan. BADAS,
Dhaka). For each district, cluster names wdre written on
than individual level.
pieces of paper, which were folded and placed in a bottle.
The first three cluster names drawn from the bottle
Methods
were allocated to the women’s group intervention and
Study design, location, and population
—
...
We assessed two interventions in the same study area the remaining three to control. The project manager
using a factorial design against a common background of drew the papers from the bottle. 1 he allocation sequence
b services strengthening:
°
- first, a community-based
] was decided upon by the project team bdfore drawing
health
the papers and was based on clusters1 rather than
intervention wiffi participatory women’s groups to
improve maternal and neonatal health outcomes; second, individuals. Clusters had been pre-identified by the
an intervention that involved training traditional birth team on the basis of previously mentioned criteria.
The control clusters included three tea garden estates
attendants in bag-valve-mask resuscitation of neonates
that had substantially worse health and socioeconomic
wiffi symptoms of birth asphyxia.
Three districts, Bogra, Faridpur, and Moulavib^zar, indicators than did the rest of the study area. In these
were selected by use of purposive sampling from three areas, surveillance started late because of entry
different divisions in Bangladesh on the basis of the restrictions. We did not know about the entry difficulties
’ ’ ’ mortality
” z rates before the recruitment and
districts having active Diabetic Association of Bangladesh and■ high
allocation ofclusters and therefore did noi exclude these
(BADAS)
offices
(webappendix
p
1).
Within
these
See uniine for webapperidix
districts, subdistricts (upazilas) and unions (the lowest- areas before allocation. Additionally, about 10% of
level administrative units in rural Bangladesh) were; also mothers in our study area were temporary residents and
purposefully sampled by use of recommendations from mainly came into the cluster areas to give birth, since the
BADAS representatives, the main criteria being perceived tradition is for women to go to their mothers’ home just
limited access to perinatal health care in those unions, before delivery. These temporary’ residents were not.
and a feasible travelling distance from BADAS district exposed to the women’s group intervention, and often
had returned to their marital homes outside the study
headquarters.
We approached community leaders and obtained their area before the post-natal interview.
In a second-level randomisation, the randomised
permission to establish women’s groups in the
clusters
were further randomly assigned by the same
intervention clusters in 2004. 451 community orientation
method
to traditional birth attendant intervention or
meetings were undertaken with the chairmen and
members of union councils and community members. control groups. Of the nine women’s group intervention
The team also undertook 451 village mapping exetcises clusters, five became traditional birth attendant
intervention clusters and four became controls. The nine
to identify the location of health facilities and social, and
women's group control clusters were randomised so that
religious meeting places.
J
www.thelancet.com Vol375 Aprils, 2010
1194
!
Articles
four received the traditional birth attendant intervention
and five became controls. Overall, for this second-level
randomisation, there were nine traditional birth attendant
intervention groups and nine control clusters
(webappendix p 2).
The randomisation process was done before the
collection and analysis of baseline data, and none of the
staff attending the randomisation process had any previous
knowledge of the health and socioeconomic status of the
chosen union clusters. Neither the study investigators nor
the participants were masked to group allocation.
Women's group intervention
Women’s group facilitators visited every tenth household
within the intervention clusters and invited married
women of reproductive age to join the groups. The
groups initially only included women of reproductive age
but others joined later because group members requested
that mothers-in-law, adolescents, and other women
should also attend.
Women's groups were facilitated by a local female peer
facilitator who acted as a catalyst for community
mobilisation. Every facilitator was responsible for
18 groups. Facilitators received five training sessions that
covered participatory modes of communication and
maternal and neonatal health issues. The role of the
facilitator was to activate and strengthen groups, to
support them in identifying and prioritising maternal
and neonatal problems, to help to identify possible
strategies, and to support the planning, implementation,
and monitoring of strategies in the community. Locally
recruited supervisors supported facilitators in preparing
for meetings and liaising with community leaders.
Groups took part in a participatory learning and action
cycle consisting of four phases (figure 1 and
webappendix p .3). Control and intervention clusters all
received health services strengthening and basic training
of traditional birth attendants.
■
18 dusters randomised with stratified allocation
503163 estimated population
27953 (range 15441-35110; SO 5953) mean duster population
*
9 dusters allocated to
intervention
9 clusters received the
intervention
I
I
9 dusters allocated to
control
9 dusters received
control
+
20943 births identified
17514 (84%) mothers
interviewed after birth
22 774 births identified
18599 (82%) mothers
interviewed after birth
Analysed for mortality
outcomes (ITT population)
9 clusters
17514 births (372 twins. 6 triplets)
16926 livebirths
588 stillbirths
570 neonatal deaths
63 maternal deaths
Analysed for mortality
outcomes (ITT population)
9 dusters
18 599 births (356 twins,
9 triplets)
17967 livebirths
632 stillbirths
656 neonatal deaths
35 maternal deaths
T
Excluded from secondary
outcome analyses
0 dusters
Temporary residents
1819 births
46 stillbirths
55 neonatal deaths
8 maternal deaths
4
Excluded from secondary
outcome analyses
0 clusters
Temporary residents
1977 births
60 stillbirths
44 neonatal deaths
3 maternal deaths
Tea garden residents
1365 births
51 stillbirths
55 neonatal deaths
T
Included in mortality subgroup
analysis
9 dusters
15695 births (348 twins.
6 triplets)
15153 livebirths
542 stillbirths
515 neonatal deaths
55 maternal deaths
Included in mortality subgroup
analysis
9 clusters
15257 births (310 twins,
9 triplets)
14 736 livebirths
521 stillbirths
557 neonatal deaths
32 maternal deaths
Traditional birth attendant intervention
In clusters assigned to the traditional birth attendant
intervention, 482 attendants were given basic training in
undertaking clean and safe deliveries, providing safe
delivery kits, recognising danger signs in mothers and
infants, making emergency preparedness plans,
accompanying women to facilities, and undertaking
mouth-to-mouth resuscitation. They also received
additional training in neonatal resuscitation with bag
valve-mask. A pre-lest and post-test questionnaire was
done for every traditional birth attendant at the start and
end of the initial training session and at subsequent
training sessions. Control clusters were given basic
training but no training in bag-valve-mask resuscitation.
Health service inputs
The project did not have resources to improve service
delivery intensively at all levels of government health
www.thelancet.corn Vol 375 April 3, 2010
Figure 2: Trial profile
ITT=intention to treat.
services. Activities undertaken in both intervention and
control clusters focused on improving referral systems,
links between the community and health services and
between different levels of health servrces, efficient use
of available resources, basic and refresher clinical
training relating to essential neonatal and maternal care,
in addition to information, education, and com
munication materials. The training was provided to
doctors, nurses, and paramedical staff working at district,
upazila, and union levels. This training consisted mainly
of refreshing knowledge about antenatal, natal, and
postnatal care, recognition of danger signs in the mother
and newborn baby, essential care of the newborn baby,
the five cleans (clean delivery surface, clean perineum,
1195
Articles
Intervention area
Control area
delivery- kits (soap, a blade, gauze, polythene, and a
thread), and safer motherhood.
Births
3162
3227
Surveillance
Livebirths
3054
3069
Stillbirths
42
73
Neonatal deaths
66
85
Neonatal mortality rate (per .1000 livebirths)’
21-6
26-9
22-6
A prospective monitoring system was developed to record
all births and their outcomes within the 18 control and
intervention clusters during the project. The System was
similar to the one implemented in the India trial and
consisted of two stages (webappendix p 4).'" First,
traditional birth attendants (ie, key informants) in the
study area identified all births, irrespective of whether
they attended them, and deaths in all women during
pregnancy or up to 6 weeks after delivery. Each traditional
birth attendant was responsible for about 200 households,
and was paid an incentive of 60 Taka (US$0-87 on June 8,
2009) for each accurate identification. Surveillance
monitors met with traditional birth attendants once a
month to gather the information. Second, Mien births
were identified, women were interviewed once 6 weeks
after delivery. Interviewers verified the births and deaths
identified by key informants and completed a
questionnaire that covered background characteristics
and the antenatal, delivery, and postpartum I periods. All
eligible women identified were also asked it they could
identify any other pregnant women. In the event of. a
stillbirth or neonatal death, a verbal autopsy was done
with the mother. In the event of a maternal death up to
6 weeks after delivery, a verbal autopsy was done with a
close friend or relative. Surveillance started in August,
2004, and covered all clusters by January, 2005.
Total
I
Stillbirth rale (per 1000 births)*
Socioeconomic characteristics
133
Household characteristics
1566 (49%)
Own agncultuial land
1558 (49%)
Own house
3095 (98%)
3110 (96%)
Own one of almyrah (wardrobe), radio or tape recorder, sewing
machine, or bicycle
941 (30%)
1429 (44%)
Own none of the appliances on the list
322 (10%)
248 (8%)
Use of sanitary latrine
1007 (32%)
1476 (46%)
Access to tubewellt water
3061 (97%)
3072(95%)
2794 (88%)
368 (12%)
2621(81%)
520(16%)
2044(65%)
417 (13%)
1930(60%)
591(19%)
851(26%)
None
1589(50%)
1560(48%)
Primary
101.1 (32%)
901 (28%)
Secondary or higher
560 (18%)
761 (24%)
853(27%)
1046 (32%)
Antenatal care by formal provider at last pregnancy
1148(36%)
1085 (34%)
Four or more antenatal check-ups by formal provider at last pregnancy
211(7%)
330(10%)
284 (9%)
406(13%)
Religion
Islam
Hinduism
601 (19%)
Maternal age (years)
<20
20-29
*30
Maternal education
NGO membershipt
|
Health-care-seeking and home-care practices
Health facility visit in case of illness during pregnancy
i Institutional delivery
226(7%)
302 (9%)
Horne delivery
2924 (92%)
2891 (90%)
Home delivery attended by trained TBA'j
529 (18%)
424 (15%)
Home delivery attended by untrained TBA§
1058 (37%)
1245 (43%)
Birth attendant washed hands’)
1959 (67%)
1847(63%)
1 Blade boiled during delivery^
1767 (60%)
1733(60%)
. Appropriate cord careS
2044 (70%)
1.891 (65%)
1405(48%)
1540(53%)
infant wrapped immediately after deliveryS
1640(56%)
1658(57%)
infant not bathed in first 24 h$
625(21%)
777(27%)
Health-care provider seen in first 24 h after delivery
978 (31%)
1434(44%)
Infant put to breast within 1 h
1611(51%)
1672(5?%)
; Infant wiped immediately after deliveryS
Data are number or nurnbei (%). 6389 mothers were interviewed. 3213 from intervention unions and 31/6 from
control unions. NGO-non-governmental organisation TBA-traditional birth attendant. "Baseline mortality rates
were lower than expected when compared with district and national estimates. Unlike the trial data, the baseline
i mortality estimates were based on retrospective recall rather than prospective identification of births and deaths,
which might account for under- reporting. 1 Well made by driving a tube into the earth to a stratum that bears water.
i tNGO membership is defined as belonging to a microcredit or savings organisation. §Home deliveries only.
Table 1: Baseline characteristics of mothers and their most recent birth in intervention and control areas
clean hands of attendant, clean blade to cut cord, clean
umbilical stump without anything applied to it), safe
1196
Primary and secondary outcomes
The primary outcome of the women’s group study was
NMR (deaths in the first 28 days per 1000 livebirths).
Secondary outcomes were maternal deaths (death of a
pregnant woman or within 42 days of cessation of
pregnancy from any cause related to the pregnancy or its
management, but not from accidental causes), stillbirths
(fetal death after 28 weeks of gestation but. before delivery
of the baby’s head), uptake of antenatal and delivery
services, home-care practices during and after delivery,
infant morbidity, health-care seeking behaviour (seeking
care for any maternal or newborn illness or complication),
perinatal mortality, and early and late NMR. We used the
International Classification of Diseased version 9
definition of stillbirth because it was appropriate for this
setting."
The primary outcome of the traditional bikth attendant
study was early NMR. This outcome is presented in this
report, but a more detailed analysis of the intervention
will be reported in a separate publication. Early neonatal
deaths refer to deaths within 6 completed days after
birth and late neonatal deaths from 7-28 completed days
after birth. Miscarriage was defined as cessation of a
presumptive pregnancy before 28 weeks of gestation.
Perinatal death describes either a stillbirtli or an early
neonatal death. We obtained background demographic
www.thelancet.com Vol 375 Aprils, 2010
■
Articles
Year 2(2006)
Year 1 (2005)
Year 3(2007)
Years 1-3 (2005-07)
All
Intervention
Control
Intervention
Control
Intervention
Control
Intervention
Births*
4620(4706)
5495(6183)
5296(5961)
5580(6625)
5421(7249)
15695(17514)
30952(36113)
4457(4538)
4586(4924)
4441(4770)
5250(6426)
Livebirths
5062(6200)
5400(6427)
Stillbirths
163(168)
145 (154)
199(222)
188(226)
180 (198)
5233(6997)
188(252)
542 (588)
521(632)
29889(34893)
1063(1220)
Neonatal deaths
138(139)
108(109)
30(30)
175(196)
187(215)
202 (233)
190 (216)
180(227)
128 (144)
159 (184)
158(184)
143 (167)
149 (186)
47(52)
271 (277)
273(298)
44(49)
346 (410)
47(49)
323(365)
31(41)
Perinatal deaths
28(31)
358(406)
Maternal deaths
14 (14)
11(11)
23(28)
9(11)
18(21)
Stillbirth rate per 1000 births
35-3(357)
31-6(31'3)
36-2 (35-9)
35-8(351)
NMR per 1000 livebirths
30-9 (30 6)
35-3 (36-1)
Early NMR per 1000
livebirths (0-6 days)
24'2(24'0)
39-4 (411)
28-9 (26-8)
Late NMR per 1000 livebirths
(7-2.8 days)
67(61)
Perinatal mortality rate per
1000 births
Maternal mortality ratio per
100 000 livebirths
Early (0-6 days)
Late (7-28 days)
■
Control
15257(18599)
15153(16926) 14736(17967)
557(656)
1072(1226)
845(974)
22?(252)
337(438)
105 (HO)
952 (1048)
435 (514)
122 (142)
956 (1146)
1908 (2194)
12 (13)
55(63)
32 (35)
87(98)
32-2 (29-9)
347 (34-8)
34-5 (33 6)
34-1 (33-9)
34-6 (33 8)
39-9 (37-5)
35-2 (33-6)
34(337)
30(30-9)
31-2(297)
26-5(26-4)
34-4 (32-4)
28-5(26-6)
27-0 (271)
37-8(36-5)
29'5(28-6)
28-3(279)
10-6 (10-9)
53(5'2)
8-7 (7-9)
8 7(7 6)
5'9(5 8)
6-9 (6-7)
8-3(.7'9)
7-6 (7-2)
58-6 (588)
59-5 (60-5)
651 (65-6)
65'9(63'8)
57'8(55'0)
62-2 (60 4)
60-6 (59-8)
62-6 (61-6)
624 (607)
3141 (308-5)
2477(230-6)
434-3 (4697)
177-8 (1774)
333-3 (326-7)
2293(1857)
363(372-2)
2171 (188-1)
291-1 (280-8')
515(570)
410 (460)
35'9 (351)
i Data in parentheses include temporary and tea garden residents. NMR»neonatal mortality rate. ‘Includes all births for which interviews were completed from Feb 1. 2005, to Dec 31,2007.
: Table 2: Births, deaths, and crude mortality rates in intervention and control clusters during the trial period (2005-07)
and socioeconomic information to investigate cluster
comparability.
Quality control of data
Data were double-entered in an electronic database.
Quality checks were undertaken by district-based
surveillance supervisors who manually checked
information provided by the traditional birth attendants
and monitors. The field surveillance manager, data
input officer, and data manager undertook manual and
systematic data checks. Additionally, we cross-checked
a subsample of our data with government records.
Statistical analysis
We undertook a cross-sectional baseline survey for the
women's group study from January to March, 2003, in
more than 6000 mothers who had delivered a baby
within the past year, to obtain data for household and
demographic characteristics, in addition to data for
pregnancy, delivery, and neonatal outcomes. Details of
the sampling method used for this survey have been
published elsewhere.” The baseline survey was
undertaken to gather data for neonatal care practices
and behaviour, but not to provide precise NMRs in view
of its limited sample size. We based our original sample
size calculations on the national estimate of neonatal
mortality from Bangladesh Demographic and Health
Survey data from 2004, which gave a value of 41 deaths
per 1000 livebirths for the 1999-2003 period. With an
estimated 1600 Livebirths per cluster over 3 years, a
k' value of 0-3. and a baseline NMR of 41 deaths per
1000 livebirths, the study had a power of 56% to detect a
www.thelancet.com Vol 375 April 3, 2010
30% reduction in NMR at the 95% significance level.
After the end of the trial, we undertook a retrospective
calculation to understand whether our inability to detect
an effect of the intervention could be caused by a lack of
power. From our study data, the harmonic mean of the
number of recorded livebirths per cluster over the study
period was 1467 (range 1081-2708). The stratum-average
intracluster correlation coefficient was 0-00056,
corresponding to a between-cluster coefficient (k) of
0-12 with the observed NMR in the control groups of
38 deaths per 1000 livebirths.!? On the assumption of a
baseline NMR of 38 deaths per 1000 livebirths, the study
had a power of 88% to detect a reduction in neonatal
mortality of 25% at the 95% significance level.
We did not expect the intervention to have adverse
effects at cluster or participant, level and therefore did not
have any stopping rules. A preliminary analysis was
undertaken in July, 2008, and findings were presented to
an independent data safety monitoring board. The board
recommended a final analysis of data for all births in the
study area between Feb 1, 2005, and Dec 31, 2007.
Analysis was by intention to treat (ITT) at cluster and
participant levels. Temporary and tea garden residents
were included in the analysis for mortality outcomes.
However, they were excluded from analyses for secondary
outcomes since they were unlikely to have been exposed
to the intervention.
We compared NMRs, stillbirth rates, and maternal
mortality ratios between control and intervention groups
by use of stratified duster-level analysis because of the
small number ofclusters in each group. These analyses
involved calculating risk ratios for each stratum and then
1197
Articles
Excluding temporary and tea garden residents
lntention-to-treat population
Intervention Control
cluster
cluster
Unadjusted* risk
ratio (95% Cl)
Intervention
cluster
Control
cluster
Unadjusted risk
ratio* (95% Cl)
Adjuster risk ratiot
(95% Cl)
NMR per 1000 livebirths
33-9
36-5
0-93 (6-80-1-09)
34-2
37-7
0-92 (0-75-112)
0-90(073-110)
Early NMR per 1000 livebirths (O -Sdays)
2.7-2
288
0-95 (0-78-1-16)
271
29-5
0-93 (0-75-1 15)
0-91 (0-72 114)
Late NMR per 1000 livebirths (7-28 days)
6-7
7-7
0 87(0 59-1-29)
70
81
0-87 (0! >4-138)
34-5
338
0-90 (0-57-1 41)
1-01 (0-82-1-21)
1-00 (O-p-1-21)
:
Stillbirth rate per 1000 births
33-6
343
0-97(0-82-115)
;
Perinatal mortality rate per 1000 births
59-9
62-2
0-96 (p-87-107)
60-7
62 3
0-97 (0-90 105)
0-96 (0-88 1 04)
j
!
Maternal mortality ratio per
100000 livebirths
3889
1891
2-02(111-3-68)
375-2
2'11-4
1-73 (0-98-305)
1 74(0-97-313)
i Data are mean rate. NMR-neonatal mortality rate. ‘Adjusted for stratification and clustering only, t Adjusted for stratification, clustering, maternal age (contint ous),
maternal education (categorical 1-5). and having no household assets.
; Table 3: Comparison of mortality rates in intervention and control clusters (2005-07)
an overall weighted mean of these, testing the null
hypothesis that the true overall risk ratio is 1 by use of a
stratified t test." We noted baseline differences in
maternal education, maternal age, and household assets
between intervention and control clusters: mothers in
the intervention clusters were slightly younger, less
educated, and had fewer household assets. We adjusted
for these covariates by use of the two-stage method
described by Hayes and co-workers" for cluster-level
analysis. First, a logistic regression model was fitted to
the individual-level data, which incorporated the stratum,
maternal age, education, and household assets, but
excluded any information about trial group. The resulting
regression model was used to calculate ratio residuals for
each cluster, which were then used in place of cluster
level observations for a stratified t test described above
for unadjusted analysis. Results are presented as risk
ratios with 95% Cis.
In the assessment of the traditional birth attendant
intervention, only home deliveries were included in the
analysis. Analysis was at cluster level and adjusted for
stratification as described for the women’s group
assessment, but in this analysis the strata were clusters
in which the women’s group intervention was
implemented, and clusters in which no women’s groups
were implemented. We undertook tests to check for
interactions between the traditional birth attendant and
women’s group interventions and did not find any. We
therefore analysed the women’s group data as if from a
single trial with two groups.
This study is registered as an International Standard
Randomised Controlled Trial, number ISRCTN54792066.
Role of the funding source
The sponsors of the study had no role in study design,
data collection, data analysis, data interpretation, pr
writing of the report. Representatives from the Big Lottery
Fund and Saving Newborn Lives visited the project during
the trial implementation. All authors had access to all tinie
data in the study. KA and AC had final responsibility for
the decision to submit for publication.
11198
Results
Figure 2 shows the trial profile. The estimated population
size was 503 163 people. All nine selected clusters had
the intervention. All women’s groups had finished their
first meeting by September, 2004. The traditional birth
attendant intervention started in March. 2005, with all
attendants completing their training in May, 2005. We
monitored births and deaths in the study area between
Feb 1, 2005, and Dec 31, 2007. Interviews were completed
for 17 514 births in the intervention clusters and foi
ls 599 births in control clusters (including temporary
and tea garden residents). These data correspond with
84% of 20943 births registered by key informants in
intervention areas and 82% of 22774 births in control
areas (figure 2). The main reason for failure to interview
was maternal migration.
Table 1 shows baseline characteristics of intervention
and control clusters gathered in a retrospective survey.
6389 mothers were interviewed. 3213 from intervention
unions and 3176 from control unions. The number of
mothers to be interviewed in each union wjs weighted
according to the total union population based on the 1991
Bangladesh census." Women who had delivered a baby
within the past 12 months were selected for interview by
use of random sampling. We noted differences in
maternal education, maternal age, and household assets
between intervention and control unions, with a greater
proportion of mothers in the intervention unions with no
education and no household assets. Mothers in
intervention unions were also more likely to be younger
than mothers in control unions. Further results from this
survey are reported elsewhere."
In a total population of 229195 people In the nine
clusters. 162 women’s groups provided a coverage of one
group per 1414 population. In 2007. 2363 (9%) women of
reproductive age in the intervention clusters (n=27614)
were group members. Almost half the members
(1158 women. 49%) were between 25 years and 34 years
old with fewer (378 women, 16%) younger members
(<24 years old). The groups held meetings once a month
and completed a cycle of 20 meetings. The mean
www.thelancet.com Vol 375 April 3, 2010
Articles
; attendance during the first ten meetings was 73%
' (1735 women) of registered members. Only 477 (3%) of
15 695 women who gave birth and were interviewed
during the study period reported attending a group.
Unadjusted NMRs per 1000 livebirths were .30-6 in 2005,
! 36 • 1 in 2006, and 33 • 6 in 2007 in the intervention clusters
; (including all residents), and 41-1, 37-5, and 32-4 in the
control clusters, respectively (table 2). The risk ratio for
neonatal mortality-, taking into account clustering and
stratification, for the 3-year period was 0-93 (95% CI
0-80-1-09; table 3). The difference in maternal mortality
ratio between intervention and control clusters was based
on fairly small numbers of deaths but reached significance
. when temporary and tea garden residents were included in
the analysis (table 3). Although this finding is of concern,
46 of the 55 maternal deaths in the intervention clusters
(excluding tea garden or temporary residents) were to
women who had neither heard of nor attended groups and
there were no maternal deaths in members of women’s
groups, which suggests that the intervention did not have a
direct adverse effect. Additionally, maternal mortality was a
secondary outcome of the study and the sample size only
gave us power to detect large differences.
In intervention clusters, neonatal mortality increased
during year 2 then stabilised in year 3. In control
clusters, neonatal mortality decreased over time,
although the differences were' not significant (figure 3).
Stillbirth rates did not differ between intervention and
control clusters.
No significant differences were noted in most home
care practices or health-care-seeking behaviours between
intervention and control clusters (table 4). However, we
did see higher frequencies of delayed bathing and
exclusive breastfeeding in the intervention clusters than
in the control clusters. The proportion of institutional
deliveries was slightly higher in control clusters than in
intervention clusters, which might have contributed to
the difference noted in maternal mortality.
To further examine the. effect of the women’s group
intervention, we compared birth outcomes and selected
home-care and health-care-seeking practices in women’s
group members and non-group members in the
intervention clusters (webappendix pp 5-6). The
proportion of women with hygienic delivery practices
was higher in group members than in non-group
members. Additionally, group members were more likely
to avoid bathing in the first 24 h and to undertake
exclusive breastfeeding for the first 6 weeks than were
non-group members.
Data from our process assessment showed that the
three main strategies implemented by women’s groups
were the creation of emergency funds managed by the
groups, the raising of awareness of maternal and
newborn health issues during meetings and in the
community by use of materials provided by the groups
(such as picture cards and flipcharts), and the fostering
of effective communication with health-care providers
www.thelancet.com Vol 375 April 3, 2010
Intervention clusters
Control clusters
^80'
!
§ 60-
i
1
S 40-
!
T
[
J
O’
2005
2006
Year
2007
2005 '
2006 '
Year
2007
Figure J: Mean overall neonatal mortality rates in intervention and control
clusters by year (2005-07)
Boxes include lower and upper quartiles, the lines within boxes represent the
median, and the error lines represent the range.
through meetings with group representatives. These
strategies changed over time and varied between groups.
In clusters in which selected traditional birth attendants
were trained in bag-valve-mask resuscitation, there were
12519 home births, of which 8618 were attended by any
traditional birth attendant and 2792 by a traditional birth
attendant trained in bag-valve-rnask resuscitation. In the
control clusters, there were 13195 home births, of which
9171 were attended by any traditional birth attendant and
2536 by a traditional birth attendant trained in mouth-tomouth resuscitation. Mean early NMRs did not differ
significantly between clusters in which traditional birth
attendants received training in bag-valve-mask
resuscitation (25 ■ 4 deaths per 1000 livebirths) and control
clusters (26-5 deaths per 1000 livebirths). Tlie risk ratio
for early neonatal death was 0-95 (95% Cl 0-75-1-21). A
more detailed assessment of this intervention will be
reported in a separate publication. We did not find any
interactions between the traditional birth attendant and
women's group interventions.
Discussion
Our study shows that participatory women’s groups did
not significantly reduce neonatal mortality in poor rural
populations of Bangladesh. This finding contrasts with
the large reductions in mortality reported in other trials
undertaken in Nepal and India.7” This trial monitored
birth outcomes in a large population and had fewer
clusters than did previous trials; however, the intracluster
correlation coefficient was small, thus our inability to
detect a difference in NMR between intervention and
control clusters cannot be attributed to lack of power.
Although neonatal mortality was lower in the intervention
clusters than in the control clusters over the 3 years of
the trial, there was also a decline in neonatal mortality in
the control, clusters over time. This finding might reflect
underlying secular trends in mortality seen in other parts
of Bangladesh, and merits further investigation.''’16
1199
Articles
Intervention
cluster
Control
cluster
Unadjusted risk
ratio* (95% Cl)
: Any antenatal care
58-7%
64-5%
0-88 (0-72-109) 0-91 (0-76-109)
>4 ANC visits
15- 8%
0-74 (0 39-1-39)
1 At y iion tablets
13- 1%
53.7%
57-5%
0-95 (0-69-1-30) 0-96 (0-70-1 31.)
i Maternal tetanus-toxoid injection
69-1%
69-4%
0-99 (0-86-114) 0-99 (O-86-1-14)
' Institutional deliveries
14- 6%
16- 2%
0-91 (0-67-1-24)
84-3%
82-4%
1-02 (0-96-1-09)
1-01 (0-96-1-06)
2-0%
4-3%
0-85 (O-63-114)
0-90 (0-72 1-14)
65-3%
18 4%
1-05 (0-86-1-29)
1-29 (O-77-2-16)
1-25(0-88-1-75)
1-28 (0-71-2-30)
Adjusted risk
ratiot (95% Cl)
1 Biffthst
Home deliveries
Attended by formal provider
1 doctor or nurse)
I
Attendant washed hands with soap
68-4%
■
Safe delivery kit used
271%
Plastic sheet used
46-7%
Cord cut with newer boiled blade
92-4%
681%
Appropriate cord care
0-79 (0-46-137)
■
0-97 (0-77.1-24)
41-4%
92-1%
1-11 (0-87-1-43)
1-12 (0-86-1-47)
1-01 (0-97-1-04)
67-2%
1-01 (0-80-1-27)
1-00 (0-97-1-03)
1-00 (0-80-1-26)
Ij'/ebirths (home deliveries)
Infant wiped within 30 min
78-1%
1-06 (0-85-1-32)
75-6%
72-7%
76-1%
1-06 (0-85-1-33)
Infant wrapped within 30 min
0-98 (0-75-1 28)
0-98 (0-76-1-27)
Infant not bathed in first 24 h
70-7%
60-4%
114(0-97-1-33)
1 15 (0-97-1-36)
Any of three infant illnesses (cough,
fever, diarrhoea)
27-4%
28-6%
0-93 (0-74-1-17)
0-92 (0-73-116)
Health-care-seeking behaviour in
event of infant illness
22-5%
24-3%
0-89 (0-70-113)
0 89 (0-71-113)
Exclusive breastfeeding for first
16 weeks
68-0%
61-5%
1-10 (0-98-1-24)
1-10 (0-98-1-23)
infants alive at 1 month
’ Data are %. Percentages based on cluster means. ANC=antenatal clinic. ‘Adjusted for clustering and stratification only.
1 Adjusted for clustering, stratification, maternal education, maternal age. and having no household assets. J Excludes
binths to tea garden and temporary residents, includes births between Feb 1. 2005. and Dec 31. 2007
'
|
Ta!ble4: Process indicators in intervention and control clusters
Maternal mortality was higher in intervention clusters
than in control clusters during the 3 years of the trial.
This difference only reached significance when tea
garden and temporary residents were included in the
analysis. There are no obvious population, health service,
or other contextual factors to explain this finding, and
there is no evidence that women’s groups discouraged
use of health services. Also, there were no deaths in!
mothers who attended women’s groups. We do not wish
to over-interpret differences in maternal mortality rates
on the basis of low numbers of maternal deaths. The
effects of women’s group membership on maternal
mortality might become evident with meta-analysis of
several trials.
Despite the absence of a significant effect of women's
groups on neonatal mortality, process data suggested that
good perinatal practices in intervention clusters were
slightly better than they were in control clusters, such as
use of a safe delivery' kit, exclusive breastfeeding for the
first 6 weeks, and avoidance ofearly bathing. Nonetheless,
none of these findings were significant, and the
continuing second phase trial, in which women’s group
coverage has been increased to one group per
400 population, should clarify questions about any
1200
significant effects of women's groups on maternal
mortality or perinatal care practices.
We believe that the lack of effect of the intervention in
this setting was caused by three main factors, first, the
population coverage of women's groups (one group per
1414 population) was less than a third of the coverage
achieved in the India trial (one group per 468 population)^
and less than half that in the Nepal trial (one group per
756 population). Findings from these three trials suggest
that population coverage and the proportion of newly
pregnant women enrolled in groups might need
threshold levels to have an effect on birth outcomes
(perhaps of the order of one group per 450-750 population,
and between 30% and 50% of newly pregnant women
attending groups, respectively). The enrolment of newly
pregnant women is likely to be a key determinant of the
effectiveness of women's groups.
Second, we believe that die quality of die intervention
might have been affected by several factors related to the
size of the study. The total population covered by die
project was much larger than that covered in the Nepal or
India trials, but facilitators in our study had to coordinate
more groups than did those in other trials. The project
had difficulties in retaining facilitators and supervisors,
which might have led to disruptions in meetings and
reduced support for community mobilisation. Facilitators
also had an increased workload, since diey arranged
18 meetings per mondi compared with nine peri month in
the trial in Nepal. Although the support structure and ratio
of supervisors to facilitators in this study were similar to
those in Nepal and India, in practice coordinators often
lived further away from women’s group facilitators than
did those in other sites, and they were not able to provide
refresher training and continuing support to facilitators.
Facilitators’ use of participatory techniques to stimulate
community mobilisation is the hallmark of this
intervention and lack of support for facilitators might have
damaged the quality of the intervention.
Third, local contextual factors could have had a role:
adverse climatic conditions affected the facilitators’ ability
to travel to meetings and one of the intervention unions
in Faridpur was entirely flooded in 2007. Additionally,
there are signs that gender-based barriers were strong in
some of the intervention unions, and might have
prevented some women from joining groups, seeking
care, or from implementing strategies if they bad joined
a group. For example, some women reported facing
problems when asking for their husbands' or in-laws'
permission to join a group. Other women's groups linked
to non-governmental organisations (NGOs) operated in
our study area, and women were regularly asked to
participate in NGO activities for which they could receive
financial incentives. Women might therefore have been
deterred from investing time in women's group meetings
for which no incentives were offered. Despite these
problems, all 162 groups continued to meet after the end
of the programme’s cycle.
www.thelancet.com Vol 375 April 3, 2010
Articles
Tlie purposive selection of districts and upazilas, and
the stratification of sampling, together with the restricted
number of clusters, might have limited the effectiveness
of randomisation procedures. The participatory' approach
assessed in this study contrasts with health-worker-led
interventions tested in other recent trials of community
based newborn care.’7'"* In Bangladesh, the Projahnrao
trial* compared home care, consisting of multiple
community health worker visits to pregnant mothers,
witfi a community mobilisation group and a control
group. The results showed no overall significant
differences between groups in NMR over the 3 years of
the trial, but there was a 34?^ reduction in NMR in the
' final 6 months of the trial in the home-care group.
However, the community mobilisation approach used in
the Projahnmo trial was not participatory and was less
intensive than the approaches used in the Nepal, Iridian,
and Bangladesh Perinatal Care Project trials. In the
Projahnmo community-care group, each female
community health worker was responsible for a
i population of 18000 people, which was divided so that
each geographical area of about. 225 people was visited
once every' 4 months; the male community health workers
visited each area every 10 months. Just as we propose for
this trial with women’s groups, we do not believe that
this level of coverage and intensity is sufficient to bring
about behaviour change and reduction in mortality.
Few government programmes have managed to provide
multiple prenatal or postnatal home visits to mothers
and infants when scaled up.'*’ By contrast, women’s
groups, if scaled to an adequate coverage, have the
potential to reach the poorest people arid bring about
substantial health and non-health benefits. Nonetheless,
a women’s group approach requires adequate human
, resources support for community mobilisation and
appropriate coverage. The threshold coverage or dose
effect for the women's group intervention needs to be
established for future scale-up programmes. This
threshold coverage raises an important, new research and
policy question: would scaling up the coverage of
women's groups in Bangladesh achieve the same effect
on mortality as did the intervention in India and Nepal,
or is the absence of effect caused by other delivery and
social context barriers? We are currently assessing the
effect of more intensive scale-up of women’s groups in
the intervention clusters to increase population coverage
and membership of pregnant women to levels similar to
those reported in the trial in India. Additionally, we are
improving support for facilitators and implementing
strategies to counter gender-based barriers, such as the
involvement of men and religious leaders. We are closely
monitoring contextual factors that might affect the
delivery of tlie intervention through a detailed process
assessment. The results of this assessment will establish
whether the success of women's groups is dependent on
population coverage, or whether specific contextual and
delivery factors reduce their effectiveness.
www.thelancet.com Vol375 Aprils, 2010
Contributors
All the authors contributed to the design of the study and criticised
drafts of die report. KA. SB. and AC were responsible for the conception
and overall supervision of the trial. KA and her team managed die
project, data collection, data entry, and administration with assistance
from BB and SS. SB. ME. and AC were technical advisers to the study.
AC, KA, and SB helped design the original trial protocol. SB designed
the data collection methods and epidemiological surveillance system.
AP, CP. and SB undertook the quantitative analysis. AC, AP. SB, and
KA wrote the first draft of the report and were responsible for
subsequent collation of inputs and redrafting. KA and AC are guarantors
for the report.
Conflicts of interest
We declare that we have no conflicts of interest.
Acknowledgments
This study was funded by Women and Children First, the U K Big Lottery'
Fund. Saving Newborn Lives, and the UK Department for International
Development. We thank all the mothers and relatives who gave their
time to be interviewed. We would also like to thank the staff of the
perinatal care project, in particular Golam Azam.
Rezaul Alam Chowdhury, Beni Madhab Chakrabarty. and
Mubinul Karim (district managers for Faridpur. Bogra, and
Moulavibazar, respectively). Nasim Mahmud Bhuiyan, Rashedul Islam,
Mahbubul Haque, Liakat Ali Miah, Apurbo Kumar Bala.
Rabat Alam Chowdhury (monitoring coordinators).
Manaj Kumar Biswas. Minara Begum Mina. Roksana Akter.
Asma Khanum, Silvia Gini Karmaker, Shomola Rani Das. Eitu Rani Das
(women's group coordinators). We would like to thank all the women's
group facilitators and monitors. We thank members of the trial steering
group, A K Azad Khan, Nazmun Nahar. Bhuiyan, T A Chowdhury.
Laila Arjumand Banu, and Md Shahidullah Bangabandhu Sheikh Mujib,
We also thank Sam Richmond (chair). Shams El Arifeen, and
Shamsun Nahar for joining the data safety and monitoring board.
We thank David Osrin. Prasanta Tripathy. Laila Younes. Tasmin Nahar,
Bedowra Haq. and Ruth Duebbert for their comments on this report.
We also wish to thank anonymous reviewers for suggestions that greatly
enhanced the reporting of this study. We acknowledge the important
contribution of M Kamruzzaman who was the project manager at the
start of the study, and sadly died in 2003.
References
WHO. Neonatal and perinatal mortality, country, regional and global
i
estimates 2006. http://whqlibdoc.who.int/publications/2006/
9241563206_eng.pdf (accessed Feb 26, 2010).
WHO. Country cooperation strategy: Bangladesh 2008-2013. http://
2
www.searo.who.int/LinkFiles/WHO_Country_Cooperation_
Strategy_-_Barigladesh_country_health.pdf (accessed Feb 20, 20'10).
WHO. Country profile on reproductive health in Bangladesh. 2003.
3
http://www.searo.who.int/LinkFiles/Reporductive_Health_Profi1e_
chp-bangladesh.pdf (accessed Feb 19. 2010).
Mercer A. Haseen F. Huq NL. Uddin N, Hossain Khan M. Larson CP.
4
Risk factors for neonatal mortality in rural areas of Bangladesh served
by a large NGO programme. Health Polity Plan 2006; 21: 432-4.3.
National Institute of Population Research and Training. Bangladesh
5
demographic and health survey 2007. Dhaka: National Institute of
Population Research and Training. 2007
United Nations Development Programme. Millennium
6
Development Goals. Mid-term Bangladesh progress report 2007
http://www.un.org.bd/pub/MDG%20Mid-term%20Progress%20
Report2007pdf (accessed Feb 20. 2010).
Manandliar DS. Osrm D. Shrcslha BP. cl al. Ell'cct of a participatory
7
intervention with women’s groups on birth outcomes in Nepal:
cluster-randomised controlled trial. Lancet 2004; 364: 970-09.
8 Tripathy, P, Nair N, Barnett S, et al. Effect of a participatory
intervention with women's groups on birth outcomes and maternal
depression in Jharkhand and Orissa, India: duster-randomised
controlled trial. Lancet 2010; published online March 8.
DOI.TO. 1016/80140-6736(09)62042-0.
Barnett S, Azad, K, Barua, S, et al. Maternal and newborn-care
9
practices during pregnancy, childbirth, and the postnatal period:
a comparison in three rural districts in Bangladesh.
J Health Popul Nutr 2006: 24: .394-402.
1201
Articles
10
11
12
13
14
15
1202
Barnett S. Nair N. Tripathy PK. Borghi J. Rath S. Costello AM.
A prospective key informant surveillance system to measure
maternal mortality'—findings from indigenous populations in
Jharkhand and Orissa, India. BMC Pregnancy Childbirth 2008; 8: 6.
WHO. International Classification of Diseases, ninth revision
(ICD-9). Geneva: World Health Organization. 1977.
Donner A. Klar N. Statistical considerations in the design and
analysis of community intervention trials. J Clin Epidemiol 1996;
49: 435-39.
Hayes R. Moulton LH. Morgan B|T, Keiding N, Van der Heliden,
Speed T. Cluster randomised controlled trials. New York. NY:
Chapman & Hall/CRC Press, 2009.
US Department of Commerce. Population trends Bangladesh.
http://www.census.gov/ipc/prod/ppt/Bangladesh93.pdf (accessed
Feb 20, 2010).
Ronsmans C, Chowdhury ME, Alam N, Koblinsky M, El Arifeen S.
Trends in stillbirths, early and late neonatal mortality in rural
Bangladesh: the role of public health interventions.
Paediatr Perinat Epidemiol 2008; 22: 269-79.
16
17
18
19
Chowdhury ME. Botlero R. Koblinsky M. Saha S. Dieltiens G,
Ronsmans C. Determinants of reduction in maternal mort; lity in
Matlab. Bangladesh: a 30-year cohort study. Lancet 2007.
370:1320 28.
Kumar V. Mohanty S. Kumar A. et al. Effect of community-,rased
behaviour change management on neonatal mortality in SI isvgarli.
Uttar Pradesh. Indiana cluster-randomised controlled trial. Lancet
2008;372:1151-62.
Baqui AH, El-Arifeen S, Darmstadt GL, et ai, for the Projahnmo
Study Group. Effect of coinmunily-bascd newburn-carc intcrverilion
package implemented titrough two service-delivery strategi ?s in
Sylhet district. Bangladesh: a cluster-randomised controller trial.
Lancet 2008; 371: 1936-44.
Haines A. Sanders D. Lehmann U. et al. Achieving child si rvival
goals: potential contribution of community health workers. Lancet
2007; 369: 2121-31.
www.thelancetxom Vol 375 April 3, 2010
Articles
@ Effect of a participatory intervention with women's groups
on birth outcomes and maternal depression in Jharkhand
and Orissa, India: a cluster-randomised controlled trial
Prasanta Tripathy, Nirmala Nair, Sarah Barnett, Rajendra Mahapatra, Josephine Borghi, Shibanand Rath, Suchitra Rath, Rajkumar Cope,
Dipnath Mahto, Rajesh Sinha, Rashmi Lakshminarayana, Vikram Patel, Christina Pagel, Audrey Prost, Anthony Costello
Summary
Lancet 2oio; 375: ii82 92
Published Online
March 8.2010
001:10.1016/501406736(09)62042-0
See Comment page 1142
See Articles page 1193
Ekjut, Chakradharpur,
Jharkhand, India
(I'Ttipathy MSc. N Nair MBBS.
r Maiiapatra MSc. sh Rath ma,
Su Rath MA. R Gope PGDRD,
D Mahto DCHM. R Sinha);
University College London
Centre for International Health
and Development, institute of
Child Health, University College
London, London, UK
(S Barnett PhD. A Prost PhD,
Prof A Costello FRCP); Effective
Intervention, Centre for
Economic Performance,
London School of Economics
and Political Science, London,
UK iR Lakshmjnarayana DNB);
London School of Hygiene and
Tropical Medicine, London, UK
(J Borghi PhD. Prof V Patel PhD);
Sangath, 841/I Alto Porvorim,
Goa, India (Prof V Patel); and
Clinical Operational Research
Unit, University College
London, London, UK
(C Pagel PhD)
Correspondence to
Di Prasanta Tripathy, Ekjut.Ward
Number 17. Plot 556B. Potka,
Po-Chakradharpur, District West
Singhbhum, Jharkhand, 8'33102.
India
prasanta.ekjut(t>>gmail.com
1182
Background Community mobilisation through participatory women’s groups might improve birth outcomes in poor
mral communities. We therefore assessed this approach in a largely tribal and rural population in three districts in
eastern India.
Methods From 36 clusters in Jharkhand and Orissa, with an estimated population of 228186, we assigned 18 clusters
to intervention or control using stratified randomisation. Women were eligible to participate if they were aged
15-49 years, residing in the project area, and had given birth during the study. In intervention clusters, a facilitator
convened 13 groups every month to support participatory action and learning for women, and facilitated the
development and implementation of strategies to address maternal and newborn health problems. The primary
outcomes were reductions in neonatal mortality' rate (NMR) and maternal depression scores. Analysis was by intention
to treat. This trial is registered as an International Standard Randomised Controlled Trial, number
ISRCTN21817853.
Findings After baseline surveillance of 4692 births, we monitored outcomes for 19 030 births during 3 years (2005-08).
NMRs per 1000 were 55-6, 37-1, and 36-3 during the first, second, and third years, respectively, in intervention
clusters, and 53-4. 59-6, and 64-3, respectively, in control dusters. NMR was 32% lower in intervention clusters
adjusted for clustering, stratification, and baseline differences (odds ratio 0 -68, 95% CI 0- 59-0 -78) during the 3 years,
and 45% lower in years 2 and 3 (0 • 55, 0 • 46-0 ■ 66). Although we did not note a significant effect on maternal depression
overall, reduction in moderate depression was 57% in year 3 (0 -43, 0-23-0-80).
Interpretation This intervention could be used with or as a potential alternative to health-worker-led interventions,
new opportunities for policy makers to improve maternal and newborn health outcomes in poor
populations.
and presents
Funding Health Foundation, UK Department for International Development, Wellcome Trust, and the Big Lottery
Fund (UK).
Introduction
Every year, an estimated 3-7 million children worldwide
die in the first month of life.” Global progress in
reduction of maternal and child mortality rates is
insufficient—only 16 of 68 countries are on track to
achieve Millennium Development Goal 4 (reduction of
mortality rate in children <5 years by two-thirds between
1990-2015).’ India accounts for 20% of maternal deaths
worldwide, 21% of all child (<5 years) deaths, and 25% of
all neonatal deaths.1 Urgent efforts are needed to reduce
these mortality rates quickly through cost-effective and
scalable interventions.
Large improvements were noted in birth outcomes in a
poor rural population in Makwanpur. Nepal, after a lowcost, potentially sustainable, and scalable participatory
intervention with women’s groups.’ Newborn mortality
rates were 30% lower in intervention areas than in control
areas (odds ratio 0-70, 9596 CI 0-53-0•94).’ Local female
facilitators assisted women's groups every month to
consider the causes and underlying problems leading to
maternal and newborn deaths, develop practical strategies
with community leaders and men, and implement and
assess the outcomes of these strategies. This community
action cycle, adapted from a programme developed in
Bolivia, created health and non-health benefits at low
cost?
Maternal depression is an increasing public health
concern in low-income countries because of its high
prevalence and wide-ranging implications for the health
of the mother and infant? Delivery' of appropriate
interventions to prevent or treat maternal depression
through health workers is a major challenge in countries
with under-resourced health systems, and community
groups assisted by non-health-care workers might have
some advantage in helping the poorest women?
Jharkhand and Orissa are two of the poorest states in
eastern India. About 4()% of their total combined
population lives below the poverty line.1' The average life
expectancy among women in both states is about 60 years,
and an estimated 63% are illiterate.1" Neonatal mortality
www.thelancet.com Vol 375 April 3, 2010
Articles
rate (NMR) per 1000 livebirths is 49 in Jharkhand and 45
in Orissa, and maternal mortality ratio per 100000
livebirths is 371 and 358, respectively."These are
disproportionately higher than India's national estimates
of 39 per 1000 for NMR and 301 per 100000 for maternal
mortality ratio."”
More than 20% of Jharkhand and Orissa’s population is
a fl ilia led with Scheduled Tribes (or Adivasi—ie, indigenous
groups), and about 12% with scheduled castes.” Despite
calls for inclusive development. Adivasi communities
remain underserved—their employment rate is roughly
half that of non-indigenous people, and nearly a third of
Adivasi children in Jharkhand and Orissa do not receive
primary' education.'J" Indigenous communities also have
higher mortality rates and poorer access to health services
than do the non-indigenous populations.’4
We hypothesised that a participatory' intervention with
women s groups could reduce neonatal mortality by at
least 25% in underserved tribal communities of eastern
India, and improve home-care practices and health
seeking behaviour of pregnant and postnatal women,
and their family members; and that the women's group
intervention could reduce maternal depression in the
intervention areas by 30%.
® Intervention clusters
s, Control clusters
e
i,
Saraikela
1 Kharswan J
India
West Singhbhum
Jharkhand
Orissa
Keonjhar
0
0^
Methods
300 miles
300 km
Study location and population
Our study was done in three contiguous districts of
Jharkhand and Orissa—Saraikela Kharswan, West.
Singhbhum. and Keonjhar (figure 1). The proportion of
Adivasis within the study clusters was 58-70%. Eligible
participants were women aged 15-49 years, residing in
the project area, and who had given birth during the
study (July 31. 2005, to July 30, 2008). Hie study
population was an open cohort—ie. women could enter
the study at any time during the trial period if they had
given birth. Women who chose to participate gave their
consent (written or left-thumb print) and were free to
decline an interview at any time. Women who were
identified by interviewers as having symptoms of
severe depression were referred to the nearest tertiary
mental health centre at Ranchi. Ethical approval was
obtained from an independent ethical committee in
Jamshedpur, India.
Figure 1; Map of districts and distribution of clusters
I 36 clusters (12 per district.)
I____
West Singhbum district
12 clusters
Saraikela district
12 clusters
Keonjhar district
12 clusters
j Step 1
7 clusters with pre-existing
women's groups
4 allpcated to intervention
3 allocated to control
3 clusters with pre-existing
women’s groups
1 allocated to intervention
2 allocated to control
8 clusters with pre-existing
women's groups
4 allocated to intervention
4 ajllocated to control
Step 2
5 clusters with no groups
2 allocated to intervention
3 allocated to control
!' T . ....
I
9 clusters with no groups
5 allocated to intervention
4 allocated to control
4 clusters with no groups
2 allocated to inteiverition
2 allocated to conrrol
Figure2: Randomisation process
Randomisation
We identified 12 rural clusters per district, with a mean
population of 6338 per cluster (range 3605-7467). The
estimated population in these 36 clusters was 228186 (on
the basis of the 2001 Indian census projections).’* In
18 clusters, existing women’s groups were involved in
savings and credit activities (seven in West Singhbhum,
three in Saraikela Kharswan, and eight in Keonjhar). In
the first district (West Singhbhum), an external
observer from a partner non-governmental organisation
(Professional Assistance for Development Action) drew
folded papers with numbers corresponding to clusters
www.thelancet.corn Vol 375 Aprils, 2010
with existing groups from a basket. The first four clusters
were allocated to the intervention group, the rest to the
control group. This process was repeated for clusters
without women’s groups and in the other two districts in
the presence of external observers (figure 2). We chose
this method because of simplicity and visibility, which
were necessary to convince local communities that the
process was transparent. Because of the nature of the
intervention, neither the intervention team nor the
participants! were masked to group assignment during
the trial.
1183
i
Articles
Key informer identifies all births and deaths
+__
I
| Deaths in women of reproductive age
Births
Interviewer interviews family member
regarding mother’s background cfiaractenstics.
antepartum, intrapartum and post-partum
information 6 weeks after delivery
Interviewer ascertains all information
r
livebirths
Interviewer ascertains all information
r
Neonatal
deaths
Stillbirths
Maternal
deaths
Pregnancyrelated deaths
ZT
Supervisor does verbal autopsy with family member
Late maternal
deaths
zz
I Supervisor does verbal autopsy with provider
T
Clinicians assign cause of death
Figure 3.' Surveillance system
Panel 1: Definitions
•
Miscarriage: cessation of a presumptive pregnancy before
•
22 weeks of gestation before delivery of the baby's head’ ’
Neonatal death: death of a liveborn infant within
28 completed days of birth
•
Early neonatal deaths: deaths arising within 6 completed
days
•
Late neonatal deaths: deaths arising from 7 to
28 completed days of birth
•
•
Perinatal death: a stillbirth or an early neonatal death
Maternal death: death of a woman while pregnant or
within 42 days of cessation of pregnancy from any cause
related to the pregnancy or its management, but not
from accidental causes
In September, 2004, we met with gram sabhas (village
councils), headmen, and representatives from panchayats
(elected representatives for basic governance) in the three
districts. These representatives granted permission to
start surveillance of births and deaths, and to work with
women’s groups.
age every month within their allocated area. The key
informant met with an interviewer once a month who
verified births and interviewed all identified mothers to
gather information about their sociodemographic
characteristics, pregnancy, delivery, and postnatal period
through a structured questionnaire about 6 weeks after
delivery. As additional checks for the identification
stage, all women identified were asked to locate any
others of reproductive age who had recently given birth
or died in the study area. Every district had 12
interviewers, one interviewer supervisor, and a district
manager. The monitoring manager super-vised field
based activities in all three districts and data entry at the
head office. The intervention and surveillance teams
were partitioned—interviewers and facilitators belonged
to different villages, their training was done separately,
and they had review meetings on separate days.
In the event of a stillbirth or neonatal death (panel 1),
the interviewer administered a questionnaire and did a
verbal autopsy with the mother and other individuals
present at the time of death; the verbal autopsy included
free text in which the mother was asked to narrate the
details of events leading to the death of the neonate. For
deaths among women of reproductive age, the
interviewer spoke to family members to ascertain the
age of the woman, cause of death, and whether she was
pregnant or had recently given birth. In the event of a
maternal death, the monitoring supervisors, initially
accompanied by a physician, did verbal autopsies with a
relative who was present at the time of the death.
Supervisors completed a standard questionnaire with
free text for elaboration of the sequence of events before
the death. Verbal autopsies for maternal deaths were
done by a physician (n=ll), interviewer supervisor
(n=96), or district manager (n=2).
Maternal depression was included as a trial outcome in
the second year of the study because of delays in
identification of a contextually appropriate scale. We used
the Kessler-1.0 item scale (K10), a questionnaire for the
detection of common mental disorders in community
settings, that has been used in India and World Mental
Health Surveys.18-20 A psychiatrist (RL) did three training
sessions of 2 days each for groups of 12-15 interviewers
that consisted of administration of the K10 questionnaire,
aspects of understanding depression and body language,
association between physical health and mental health,
active listening skills, and confidentiality.
Clusters and coverage of women's groups
Surveillance
A surveillance system with key informants was
established in the three districts. Figure 3 shows tins
system, and Barnett and colleagues"' describe it in detail.
One key informant, usually a traditional birth, attendant
or active village member, was responsible for about
250 households, and reported any births, maternal or
newborn deaths, and deaths in women of reproductive
1184
In the 18 intervention clusters, we used a participatory
action cycle with 172 existing groups and created an
additional 72 groups. Coverage of Ekjut groups was one
group per 468 population. In year 1, 546 (18%) of 3119
newly pregnant women, attended the groups, rising to
1718 (55%) of 3126 in year 3. We recorded 111006 group
attendances over 3 years. 74715 (67%) of these were from
married women of reproductive age, 15030 (14%) from
www.thelancet.com Vol 375 April 3, 2010
Articles
adolescent girls, 10452 (9%) from men. and 10809 (10%)
from elderly women.
Fvery group met monthly for a total of 20 meetings,
and a local woman, selected on the basis of criteria
(including speaking the local language and having the
ability to travel to meetings) identified by the
community, facilitated the meetings. After a 7-day
residential training course to review the cycle’s
contents, and to practice participatory communication
techniques, facilitators were given support through
fortnightly meetings with district coordinators.
Facilitators coordinated an average of 13 meetings every
month with as many groups.
Groups took part in a participatory learning and action
cycle (figure 4). Community members w'ho were not
regular group members were also encouraged to
participate in discussions. Information about clean
delivery practices and care-seeking behaviour was shared
through stories and games, rather than presented as key
messages. By discussion of case studies imparted
through contextually appropriate stories, group members
identified and prioritised maternal and newborn health
problems in the community, collectively selected relevant
strategies to address these problems, implemented the
strategies, and assessed the results. Although some
strategies were common, each group was free to
implement its own combination of strategies. The
intervention team adapted facilitation materials from
the study in Makwanpur, Nepal, to guide the meetings.’'
Groups used methods such as picture-card games, role
play, and story-telling to help discussions about the
causes and effects of typical problems in mothers and
infants, and devised strategies for prevention, home
care support, and consultations (figure 5).
I
I
I
/
Women's group intervention
/
/
/
4
x'"
■4
▼
Phase 1
Identify and
prioritise
difficulties
Phase 4
Assess effect
Comrnunity
meeting 2
Phase 2
Plan strategies
Phase 3
Put strategies
s 4 into practice
*
4
/ /
/
I
I
I
jk
X.
T'-'ffoos
%
.. \
\
I
X.
▼
Community
meeting 1
Figure 4: Meetings in women's group cycle
..
Health -service inputs
We formed health committees in all intervention and
control clusters so that community members would have
the opportunity to express their opinions about the
design and management of local health services. About
ten village representatives within every cluster met once
every 2 months and used a structured action cycle to
discuss maternal and newborn health entitlement issues.
As a result, committee members became more
knowledgable about the government health system and
assisted with the formation of village health committees
as part of the National Rural Health Mission programme. ’1
In addition to the creation of cluster-level health
committees, we provided workshops for appreciative
inquiry with frontline government health staff from
seven dusters per district in Jharkhand.22 Participants
assessed the progra m me qualitatively at the end of every
training session. We expected that any improvement in
performance or service quality would be equal in
intervention and control clusters.
www.thelancet.com Vol 375 April 3, 2010
Figure 5' Women's group meeting in Jharkhand, India
Individuals m the photo provided permission (written or thumb print.) for publication of niiage
Primary and secondary outcomes
The primary outcomes were reductions in NMR and
maternal depression scores. Secondary outcomes were
stillbirths, maternal and perinatal deaths, uptake of
antenatal apd delivery services, home-care practices
during and after delivery, and health-care-seeking
behaviour (seeking care from qualified providers in the
1185
I
Articles
36 clusters randomised with stratified allocation
(18 with existing women's groups)
228186 estimated population
6338 mean cluster population (range 3605-7467)
I
_______________________________________________ “
18 clusters allocated to intervention
(9 with existing women's groups)
4 __
18 clusters allocated to control
(9 with existing women's groups)
. . .. r...
18 clusters given intervention
9770 births (109 pairs of twins, two sets of triplets)
9469 livebirths
301 stillbirths
406 neonatal deaths
18 clusters not given intervention
9260 births (115 pairs of twins)
8980 livebirths
280 stillbirths
53'1 neonatal deaths
Excluded from analyses
0 clusters
2 mothers refused interview
Excluded from analyses
0 clusters
2 mothers refused interview
Excluded from adjusted analyses
Migrated
84 births (9 neonatal deaths. 3 stillbirths)
81 mothers (0 maternal deaths)
Excluded from adjusted analyses
Migrated
171 births (13 neonatal deaths. 10 stillbirths)
167 mothers (0 maternal deaths)
Analysed for mortality outcomes
18 dusters
8662 mothers
9686 births
397 neonatal deat hs
298 stillbirths
Analysed for mortality outcomes
18 clusters
8125 mothers
9089 births
518 neonatal deaths
270 stillbirths
Analysed for depression outcome
(years 2 and 3, excluding maternal deaths)
6452 mothers
Analysed for depression outcome
(years 2 and 3. excluding maternal deaths)
5979 mothers
I
I
Figure6:Trial profile
antenatal, delivery, and postnatal period, for checkups
and problems).
Quality control of data
Data were double-entered in an electronic database.
Surveillance supervisors manually checked information
provided by key informants and interviewers. The field
surveillance manager, data input officer, and data
manager undertook manual and systematic data
checks.
Statistical analysis
We did not expect the intervention to have adverse
effects at cluster or participant level, and therefore did
not have any rules for stopping the intervention. In
December, 2007, we presented findings from a
preliminary analysis to an independent data safety
committee. After an interim analysis in 2007, the
committee recommended that the trial continue for a
total of 3 years to enable comparison with the
Makwanpur study,’ in which the effect was measured
from 9 months after the beginning of the intervention
to allow exposure to the women's groups in pregnancy.
The data safety committee also noted that 3 years would
1186
allow analysis of possible seasonal variations in NMR.
The committee undertook a final review of the data in
December, 2008.
Our prospective surveillance from Nov 21, 2004, to
July 30, 2005, showed a baseline NMR of 58 per
1000 livebirths (261 deaths per 4509 livebirths) and
maternal mortality ratio of 510 per 100000 (23 deaths
per 4509 livebirths). The trial was planned for 3 years
and was originally powered, like the Makwanpur trial/
for a 2-year analysis of birth outcomes, after allowing a
period of up to 1 year for the women's groups to be
established and for pregnant women to be given the
intervention. We assumed a between-cluster correlation
coefficient of variation (k) of 0-15-0-25, and about
324 births per cluster during 2 years. On the basis of
10% loss to follow-up, a sample size of 18 clusters per
group resulted in 64-81% power to detect a 25% reduction
in NMR. With an estimated baseline prevalence of 15%
and k of 0-3, the study had 79-81% power to detect a
30% reduction in maternal depression over 1 year. We
used data for recorded births during the study to
estimate that the study had a power of 92% to detect a
25% reduction in NMR.
Analysis was by intention to treat at cluster and
participant levels. We excluded data from mothers who
migrated out of the region and their infants from
intention-to-treat analyses since many of these women
probably came into the clusters at the time of delivery
and would therefore not have been exposed to the
intervention in pregnancy. We aimed to do the tests of
significance for our primary and secondary outcomes on
the basis of previously agreed hypotheses about the likely
effect of the intervention. For comparison of mortality
outcome, we used multivariate logistic regression with
random effects on individual-level data in Stata (version
10.0)?3 We compared secondary indicators using
generalised estimating equations models with semirobust
SEs at the cluster level.” Generalised estimating equations
models were used for secondary indicators and categorical
scores of maternal depression because these outcomes
had high intracluster correlation coefficients (>()• 30). We
compared K10 scores grouped in three categories
(none/mild, moderate, or severe) during years 2 and 3 of
the trial. This method was chosen in favour of linear
regression to address the data’s strong positive skewness.
We adjusted for stratification by including strata as
variables in the regression analyses, and for multiple
hypothesis testing by adjusting the p values for the
primary outcomes using the Holm correction in the
results tables. We did not adjust for clustering at the level
of die mother. All results are presented as odds ratios
with 95% Cis.
Cost-effectiveness analysis
We used a similar method of cost-effectiveness analysis
as used by Borghi and colleagues/4 Costs were estimated
at 2007 prices, and were calculated separately for the
www.thelancet.com Vol 375 April 3, 2010
Articles
I
I
I
women’s group intervention and activities for health
service strengthening. These were the financial and
economic costs of setting up the intervention, and
running costs during the trial. Costs were estimated
|from the perspective of a provider to give an indication
of the potential costs of replication for the government
and interested agencies, and discounted at 3%.
Incremental cost effectiveness was measured in relation
I to a do-nothing alternative.
This study is registered as an International Standard
Randomised Controlled Trial, number ISRCTN21.817853.
Intervention area
Identified births
2457
Cbntrol area
_i--------------21235
Socioeconomic characteristics
Household assets
Radio, cassette tape, bicycle, or electricity
1752 (71%)
More costly itenjs (television, generator, battery, fan, fridge)
167(7%)
.771 (79%)
225 (10%)
Ownership of agricultural land
None
345(14%)
*364 (16%)
Own less than 2'bighas (<O-27 hectares)
1.157(47%)
Own between ^-4 bighas (O-27-O-54 hectares)
653(27%)
1969 (43%)
[ 593 (27%)
11557(70%)
Caste or tribal group
Role of the funding source
Scheduled tribe*
1849 (75%)
The funders had no role in the design of the study, data
collection, data analysis, interpretation, or writing up of
the findings, although they made a site visit early in the
study implementation. The corresponding author had
access to all the data and had final responsibility for the
decision to submit for publication.
Scheduled caste"
80(3%)
64(3%)
Other backward caste*
520 (21%)
606 (27%)
<20 years
147(6%)
20-29 years
1370(56%)
I 253(11%)
1385(62%)
230 years
345(14%)
348(16%)
Not known
933 (38%)
592(26%)
None
1908 (78%)
1533(69%)
Primary
143 (6%)
405(16%)
125 (6%)
Cannot read
1906 (78%)
1566(70%)
Can read
550 (22%)
669(30%)
Results
Figure 6 shows the trial profile. All 18 selected clusters
had the intervention. Loss to follow-up after birth as a
result of migration or refusal of interview was 86 (<1%)
of 9770 women in intervention clusters and 173 (2%) of
9260 in control clusters. In the study areas, 5661 (37%) of
15118 home deliveries were by a relative, friend, or
neighbour. 5368 (36%) by traditional birth attendants,
and 1913 (13%) by husbands.
"fable 1 shows the baseline characteristics of identified
births during 9 months of data gathering from Nov 21,
2004, to July 30, 2005. Numbers of births were similar in
intervention and control clusters, but differences were
rioted in household assets, maternal education, literacy,
and tribal membership, with women in the intervention
clusters being generally poorer and more disadvantaged
than those in the control clusters (table 1).
NMRs unadjusted for clustering decreased from year 1.
to year 3 in the intervention clusters compared with an
increase in the control clusters (table 2).
We noted a 32% reduction in NMR during the 3-year
trial when data were adjusted for clustering,
stratification, and baseline differences (table 3). NMR
was reduced by 45% in intervention clusters compared
with control clusters during the last 2 years (table 3).
The reduction in NMR was still significant when
migrated mothers and their babies were excluded from
the intention-to-treat analyses during the 3 years
(table 3). k, estimated from retrospective data from
control clusters, was 0 10 when we took stratification
into account, and corresponded to an intracluster
correlation coefficient of 0-0005.23 IIn the analysis
’ ’ of
neonatal mortality data at the cluster level, the risk ratio
was 0 71 (95% CI 0-57—0-90, p=0-0011) for years 1 to
$ when adjusted for muiltiple
’
hypothesis testing with
the Holm procedure and unadjusted for baseline
'-----e
differences.
<
www thelancet.com Vol 375 April 3, 2010
Maternal age
Maternal school education
Secondary or higher
577(26%)
Maternal literacy
Care-seeking behaviour and home-care practices
Any antenatal care
1460 (59%)
Three or more antenatal visits
539 (22%)
1532(69%)
701 (31%)
Any iron tablets during pregnancy
1571 (64%)
1497(67%)
Institutional delivery
277(11%)
2118(86%)
1858 (83%)
Home delivery
326(15%)
Delivery attended by traditional birth attendant I
778 (37%)
717 (39%)
Birth attendant washed handst
609 (29%)
471 (25%)
Birth attendant, used safe delivery kitt
195(9%)
163(8%)
197 (11%)
Birth attendant used plastic sheet !
Cord cut with new or boiled bladet
1493(70%)
1294(70%)
141(8%)
Cord tied with boiled thread!
276 (13%)
232 (13%)
Infant wiped within 30 mini
1247(61%)
1130(63%)
Infant wrapped within 30 mint
Infants alive at 1 month
78/(39%)
2202(90%)
2046 (92%)
Exclusive breastfeeding for 6 weeks§
1387(63%)
1168 (57%)
782 (43%)
Data are number or number (%). 'Standard terms used in Indian demographic surveys, tl lome deliveries only (stillbirths
were not exdudedlfrom the count). tStillbirths excluded. ^Number of infants alive at 6 weeks wpre 2.202 in intervention
dusters, and 2046! in control dusters
Table 1: Baseline characteristics of identified births in intervention and control areas
After year 1, NMRs were considerably lower in the
intervention clusters than in die control clusters (figure?).
In the third ylear of the study, NMR was almost half in the
intervention clusters compared with the control clusters
(odds ratio 0-53, 95% CI 0-41-0-68, adjusted for
clustering apd stratification only), much larger than the
findings of the Makwanpur trial in Nepal.5 Figure 8 shows
the duster-level changes in NMR between baseline and
1187
Articles
Year It
Baseline*
;
'
Inter
vention
Control
All
Births
2457
2347
2235
2162
4692
Livebirths
4509
Stillbirths
109
Neonatal deaths
145
73
116
183
261
100
80
180
116
Early (0-6 days)
Year 3i
Year 21
Years 1-3*
Years l-3t
All
Inter
vention
Control
All
9260
19030
9686
9089
8980
9388
298
8819
270
18775
18207
280
18449
581
531
937
397
518
380
253
368
151
639
298
144
150
294
660
1220
551
638
II89
Inter
vention
Control
9770
9469
301
Control
Inter
vention
Control
Inter
vention
Control
3171
3052
2960
3404
3286
3135
3073
3195
3110
98
92
118
3035
100
85
3073
2985
88
171
1.58
122
181
113
192
406
107
76
67
138
259
46
54
226
147
560
Inter
vention
568
915
621
45
36
81
55
51
46
135
46
209
16
153
362
199
30
235
18
7
12
109
109
327
309
30-1.
347
31-9
26-6
28-6
307
30-5
49
30-7
60
44-4
49
30-8
60
Stillbirth rate per
1000 births
23
390
194
22
152
7
214
20
297
307
Neonatal mortality rate
per 1000 livebirths
618
53-6
579
55 6
534
371
59-6
36-3
64 3
42-9
591
50-8
42-3
587
50-2
Early neonatal mortality
rate per 1000 livebirths
(0-6 days)
42-6
37-0
400
37-8
36 1
231
44-4
21-5
467
27-3
42-3
34-6
26-9
417
34-1
Late neonatal mortality
rate per 1000 livebirths
(7-28 days)
19-1
16-6
18-0
17-9
172
14-0
151
14-7
18-0
15-5
16 8
161
15-3
170
16-1
Perinatal mortality rate
per 1000 births
85-1
68-4
771
67-4
65-2
57-0
750
47-5
73-5
57-3
717
64-1
56-8
70-1
63-3
Maternal mortality ratio
per 100000 livebirths
6817
3238
510-1
650-8
1013-5
669-5
593-0
225-1
402-0
517-5
668-1
590-8
521-9
680-3
598-7
Late (7- 28 days)
Perinatal deaths
Maternal deaths
Data are unadjusted. ‘Excluding migrated mothers and infants, tlnduding migrated mothers and infants.
Table 2: Births and deaths in intervention and control clusters at baseline and during trial
Years 1-3 (including
migrated)*
p value
Years 1-3 (excluding p value
migrated)*
Years 1-3+
p value
Years 2 and 3+
p value
Neonatal mortality rate
per 1000 livebirthst
071 (0-61-0-83)
<0-0005
0-69 (0-60-0-81)
<0-0005
0-68 (0-59-078)
<0-0005
0-55 (0-46-0-66)
<0-0005
Early neonatal mortality
rate (0-6 days)
0-63 (0-54-0-75)
<0-0005
0-62 (0-53-0-74)
<0-0005
0-62 (0-52-0-73)
<0-0005
0-46 (0-37-0-57)
<0-0005
Late neonatal mortality
rate (7-28 days)
0-92 (0-67-176)
0-476
0-89 (0-65-172)
0-463
0-84 (0-64-1-12)
0736
0-80 (0-56-1 -14)
0-217
Stillbirth rate per
1000 births
1-02. (0-85-173)
0-833
1-04 (0-85-175)
0773
1-05 (0-86-178)
0-656
1-01 (0-80-178)
0-914
Perinatal mortality rate
per 1000 births
079 (070-0-90)
<0-0005
0-79 (O-69-O-9O)
<00005
0-79 (0-69-0-91)
<00005
0-68 (0-58-079)
<00005
Maternal mortality ratio
per 100 000 livebirths
0-80 (0-51-174)
0-180
0-80 (0-51-174)
0-180
070 (0-46-1-07)
0-104
0-50 (0-48-1-49)
0-563
Data are odds ratio (95% Cl). ‘Adjusted for stratification (by district and pre-existing women's groups) and clustering only, t Adjusted for stratification, clustering, maternal
education, assets, and any tribal affiliation, tp values adjusted for multiple hypothesis testing with Holm correction were <0-001.
TableJ: Comparison of mortality rates in intervention and control clusters
year 3—the NMRs fell below their baseline level in most
intervention clusters. Between 2005 and 2008, perinatal
mortality rates in the intervention clusters decreased
compared with those in the control clusters when
adjusted for clustering (table 2; table 3). Stillbirth rates
did not differ between intervention and control clusters
(table 2). Maternal mortality ratio was generally lower in
intervention than in control clusters, but the study was
not powered to detect significant differences (table 3).
Qualitative evidence from the assessment of the trial’s
process showed that community mobilisation through
1188
women's groups might have contributed to avoidance of
some maternal deaths (panel 2).
There was no detectable difference in maternal
depression K10 scores, when measured about 6 weeks
after delivery, between intervention and control clusters
in year 2 of the Study or overall (table 4). However, in
year 3, when 55% of all pregnant women in the
intervention clusters had joined a group, a 57% reduction
was noted in moderate depression among mothers in
the intervention clusters compared with control clusters
(table 4).
www thelancet.com Vol 375 April 3, 201.0
Articles
No significant differences were rioted
noted in health-care
health-care-
geeking behaviour between control and intervention
clusters (table 5). However, home-care practices showed
substantial improvements—in intervention clusters,
birth attendants were more likely to wash their hands,
Use a safe delivery kit and a plastic sheet, and boil the
thread used to tie the cord than were those in the control
Clusters. The proportion of infants exclusively breastfed
[.■■it 6 weeks was higher in intervention areas in adjusted
analyses for years 2 and 3.
Cause-specific differences in mortality rate as a
percentage of all causes—septicaemia, birth asphyxia,
hypothermia, and prematurity—during the 3 years were
not clearly discernable because there was a reduction in
I all causes (table 6). The incremental cost of the women’s
group intervention was US$910 per newborn life saved,
increasing to $1308 (in 2007 prices) when health-service
strengthening activities were included. The incremental
cost per life-year saved was $33 for the women’s group
intervention ($48 inclusive of health-service strength
ening activities). The women's group intervention in
this setting was therefore more cost effective than that
reported in Nepal5 as a result of the greater effect of
women’s groups on NMR combined with lower
operating costs in the current context.
Discussion
Control areas
100 -n
I 2i.
i
! Pg
I
■
Intervention areas
I
8060-
T
40-
T
20-
0
1
j
2
Year
3
2
Year
3
Figure 7: Boxplot of duster-level neonatal mortality rates by allocation and
study year
Control clusters
• Intervention clusters
100-i
80-
n
I!
I
60-
= £ 40-
20-
0
50
100
Baseline neonatal mortality rate per 1000 livebirths
ISO
Women's groups led by peer facilitators reduced NMR
and moderate maternal depression at low cost in largely Figure 8.- Scatterplot of duster-specific neonatal mortality rates in year 3
tribal, rural populations of eastern India. Our data show with rates at baseline
that mortality reduction in underserved rural settings
was not associated with increased care-seeking
Panel 2: Case-study effect of women's groups on
behaviour or health-service use. The most likely
strategies to avoid maternal deaths
mechanism of mortality reduction was through
improved hygiene and care practices. The availability of
A woman in the eighth month of her pregnancy, and her
safe delivery kits increased in both control and
mother-in-law attended a monthly women's group
intervention areas, but women’s groups seemed to
meeting where they participated in a drill based on what to
generate more demand in intervention clusters than in
do in the event of post-partum bleeding. After a month,
control clusters. In places where kits were not provided,
when the woman delivered at home and had severe
group members made them and provided information
bleeding, her mother-in-law remembered what had been
about their contents to mothers, then visited pregnant
said in the group, and, without wasting time, asked her
women during the eighth month of pregnancy to ensure
daughter-in-law to breastfeed the baby while she rushed to
that they had received kits and would use them. Birth
get money from the group and asked her son to arrange for
outcomes might have been affected by the fact that
a vehicle. The .daughter-in-law was immediately taken to
these community members attended the groups or
the district hospital, where she was given medicines,
were advised by group members, thus generating
intravenous fluid, and two pints of blood, and was
increased social awareness and support for clean
discharged after 1$ days.
delivery practices.
fhe most striking reduction in mortality rate was noted
in early neonatal deaths, which might be explained by the with improved care. "Flie reason for this combined
strong focus on intrapartum and early neonatal periods reduction of asphyxia, prematurity, and septicaemia
in several case studies and stories discussed during the could also Be improved intrapartum care. Potential
cycle. Attribution of cause of newborn death on the basis mechanisms for reduction of mortality' rate will be
of verbal autopsy is an imperfect science, and deaths further assessed in future analyses of verbal autopsies
might arise from several and overlapping causes. Early and seasonal mortality trends.
septicaemia could have been reduced with clean delivery
Our findings also show that a low-cost intervention
practices, and premature babies might have survived involving noh-health-care workers might affect maternal
www.thelancetxom Vol 375 April 3, 2010
1189
i
Articles
Year 2
Adjusted odds
ratio (95% Cl)*
Intervention
Control
6452
5979
2665(90%)
2-33 (125-4 38)
5277 (88%)
1-29(0-68-2-44)
154 (5%)
293 (10%)
0-43 (0-23-0-80)
5884(91%)
536 (8%)
676 (11%)
074(0-40-1-37)
4 (<1%)
5 (<1%)
070 (O-15-3-31)
32 (<1%)
26(<1%)
1-29 (0-46-3'64)
Control
Intervention
3120
2963
2612(87%)
0-91 (0-41-201)
2962(95%)
382 (13%)
104 (0-50-2-16)
21 (<1%)
1-53 (O-47-5O5)
Control
Mothers (n)
3332
3016
No or mild depression (10-1.5)
2922 (88%)
Moderate depression (16-30)
383 (11%)
28 (<1%)
Severe depression (31-50)
Years 2 and 3
Year 3
Adjusted odds
ratio (95% Cl)*
Intervention
Adjusted odds
ratio (95% Cl)*
I Data are number (%), unless otherwise indicated. "Results adjusted for clustering, stratification, maternal education, tribe affiliation, and household assets by use of generalised estimated equations with
! sernirobust SEs for individual-level data.
Table4: Kessler-10 depression scores in mothers in intervention and control clusters
Births*
Intervention
clusters
Control clusters
Odds ratio
(95% Cl) for years
1-3*
Odds ratio
(95% Cl) for years
1-31
1-60 (0-65-3 92)
0-69 (0-37-1-26)
Odds ratio (95% Cl)
for years 2 and 31
9468
8867
Any antenatal care
6990(74%)
6623 (75%)
0-97 (0-48-1.-97)
a3 antenatal care visits
3001 (32%)
3621 (41%)
0-63 (0-37-106)
Iron tablets
6997 (74%)
6293 (71%)
1-12(0-71 ■176)
7767(82%)
7377(83%)
0-90 (O-51-1-54)
1-31 (0-62-275)
1-39 (0-85-2-28)
1-34 (O-77"2'35)
Maternal tetanus-toxoid injection
Illness in pregnancy
5206 (55%)
4983 (56%)
1-03 (0-68-1-58)
1-10 (071-172)
101(0-67-1-52)
Visited health facility in case of illness during
pregnancy
945 (10%)
922 (10%)
0-78 (0-39-1'56)
0-86 (0-46-1-60)
0-80(0-39-l'65)
Institutional deliveries
1364 (14%)
1811(20%)
0-64 (0-39-104)
Birth attended by formal provider (doctor or nurse)
1490 (16%)
2067 (23%)
0-59 (0-37-1-94)
0-89(0-51-1-53)
0-81 (0-50-1-31)
0-82 (0-47-1-43)
1-86 (0-80-4-34)
0 68 (0 37-1 24)
1-40 (0-85-2-29)
0-94 (0-50-1-76)
8084
7034
Birth attended by traditional birth attendant
2692 (33%)
2676 (38%)
0-82 (O-43-1-6O)
0-84 (0-43-1-64)
Birth attendant washed hands with soap
1583(23%)
2-05 (1-14-373)
2- 07 (1-24-3-45)
0-85 (0-44-1'65)
2-50 (1-35-4'62)
Safe-delivery kit used
3291 (41%)
2594(32%)
1284 (18%)
2- 08 (1-25-3-44)
1-87 (1'11-314)
2 28 (1-27-4-09)
Plastic sheet used
2088 (26%)
560 (8%)
Cord tied with boiled thread
2559 (32%)
786 (11%)
3'85 (2-51-5-89)
3- 9 (1-82 -6-30)
3- 74 (2-48-5'65)
302 (1-61-5-65)
4-33 (2-O6-911)
Cord cut with newer boiled blade
6679 (83%)
7890
5570 (79%)
1-24(0-82-1-87)
1-35 (0-86-212)
1-55 (0'96-2-51)
6873
Horne deliveries
Livebirths (home deliveries)
2-98 (1-84-4-81)
Cord undressed or dressed with antiseptic
6600 (84%)
6115 (89%)
0-52 (0-24-112)
O-58 (0-27-1-26)
1-01 (0-39-2'62)
Infant wiped within 30 min
4741 (60%)
4227 (62%)
0-90 (0-38-2-14)
1-01 (0-43-2-36)
1-06 (0-44-2-57)
Infant wrapped within 30 min
2846(36%)
2980 (43%)
0-81 (0-37-1-80)
2107(27%)
1509 (22%)
0-74 (0-35 1-59)
1-06 (0-52-2 17)
078 (0-36 -1 -66)
Infant not bathed in first 24 h
0-95 (0-44-2-10)
1-22 (0-56-2-65)
8807
8119
1739 (20%)
2388 (29%)
0-62 (O-37-1O3)
0-67 (0-4O-1-12)
()61 (O-35-1O6)
Infants alive at 1 month
Any of three infant illnesses (cough, fever,
diarrhoea)
Care-seeking behaviour in event of infant illness
940(54%)S
1050(44%)$
1-53 (0-77-3-05)
0 88 (0'97-3-61)
1-55 (0-79-3-04)
Infant put to breast within 4 h
5390 (61%)
4942(61%)
1-01 (0-48-2-14)
0-90(0-38'- 3-11)
1 11 (0-45-2-76)
Exclusive breastfeeding for first 6 weeks
7022(80%)
5611 (69%)
1-82 (1 14-2-92)
1-44 (0-89-2-35)
1-74 (1-03-2-94)
I Data are number (%), unless otherwise indicated. ‘Adjusted for clustering and stratification only. i'Adjusted for clustering, stratification, maternal education, assets, and any
tribal affiliation. 1 Excludes births to migrated mothers and twins. ^Denominators are number of infants with any of three infant illnesses: 1739foOntervention clustersand
2388 for control clusters.
Table 5: Process indicators in intervention and control dusters
mental health. We hypothesise that the large reduction
in moderate depression seen in the third year could
have occurred through improvements in social support
and problem-solving skills of the groups. Adequate
social support reduces the risk of depression during
pregnancy and is an important social determinant of
mental health.'"' In meetings, information was shared
1190
about the difficulties encountered by mothers in the
community, and practical ways to collectively address
them were established. Group meetings also
strengthened problem-solving skills, a component of
psychotherapeutic interventions that has been shown to
affect depression in other settings.’" The intervention
seemed to have no effect on severe depression, perhaps
www,thelancet.com Vol375 Aprils, 2010
Articles
because it was more similar to primary prevention
Intervention
Control
rather than treatment, or because severe depression is
Years 1-3
less amenable to psychotherapeutic interventions. A
Years 2 and 3
Year 1-3
Complete analysis and discussion of these findings will
Early neonatal deaths
140
253
367
be presented in the future.
Birth asphyxia
92 (36%)
53(38%)
142 (39%)
Two potential effect modifiers in this trial, on the basis
Prematurity
85(34%)
46 (33%)
110(30%)
6f evidence, were differences in maternal education, and
Septicaemia
38(15%)
15(11%)
47(13%)
tribal membership and assets between the intervention
Hypothermia
16 (6%)
12 (9%)
26 (7%)
and control populations. These were taken into account
Other
22 (9%)
14(10%)
42 (11%)
in adjusted analyses and mainly provided an advantage
Data
are
nurnlxjr
or
number
(%).
for the control areas. Additionally, the high significance
of some of our results could be a result of an increase in
Table 6: Cause-specific mortality for early neonatal deaths during study
hiortality rate in tlie control areas between 2005-08. The
control and intervention clusters were in similar responsible for a population of about 500 and for
geographic areas, so factors that affected NMR should recruiting up to half of newly pregnant women. Costs are
have affected both groups equally, but further lower than for most other primary health-care
investigation is needed.
interventions, and these interventions can complement
We believe that the study had two main weaknesses. existing self-help groups in the community.
First, as in several other community-based randomised
Two other issues arising from our study are cost
control trials, the intervention and surveillance teams effectiveness and the effect on maternal mortality ratios.
were not unaware of allocation. However, there were no The interventions in the Ek jut trial were more cost
incentives or disincentives for over-reporting or under effective than those in the Makwanpur study5 because of
reporting births and deaths, and several process lower operating costs and greater effect of the
mechanisms were in place to detect errors. Second, intervention. In the Nepal trial, effect of women's groups
although migration out of districts was common, we on maternal mortality ratios was significant, although
cannot rule out some intercluster migration when the number of dbaths was small and maternal mortality
women married out of their home cluster. Our intention- ratio was not a stated primary outcome. In our trial the
m-treat analysis might have affected the results positively maternal mortalityz ratio) was higher in the intervention
J or negatively.
areas at baseline, and 20% lower after 3 years of
In the Shivgarh study,27 in Uttar Pradesh, India, the effect intervention, but tin's difference was not significant and
of an intensive behaviour-change programme involving the trial was not powered to measure differences in
community meetings and home visits by a new cadre of maternal mortality. Reduction in maternal mortality will
paid, non-governmental community workers in a depend mainly on improved access to health services
population of 104123 during 15 months resulted in a 54% and to life-saving drugs, but community mobilisation
reduction (relative risk 0-46, 95% CI 0-35-0-60) in NMR could help through improvement in hygiene at delivery
with changes in home-care practices, but no real change in and early care-Seeking behaviour for complications by
care-seeking behaviour. No overall differences in NMR addressing the first-delay component.
were noted during 30 months of intervention in the
This participatory intervention with women’s groups
Projahnmo trial,’8 in Bangladesh, but a 34% reduction could complement or be a potential alternative to health
(0-66,0-47-0 -93) was noted in the home-care group in the worker led interventions, two examples of which have
Iasi 6 months of the programme. Tlae investigators of the been discussed here. Our findings raise several important
Projahnmo study28 noted that “Availability of referral issues for policy makers in India. Could federal and state
services and a strong supervisory-- system were crucial to governments invest in this programme? Should
ffiis intervention and would be a necessary' feature of government pr non-government organisations be
scaling up the intervention.”
responsible for its scale-up? Could such a participatory
Interventions with health-worker home visits have intervention support and strengthen the National Rural
rarely achieved adequate coverage, quality, or effectiveness Health Mission's mandate of communitisation of health
when taken to scale in poor populations.20 Participatory and the Accredited Social Health Activist programme?”
groups have the advantage of helping the poorest, being Further assessments of this approach will involve a
gtoups
scalable at low cost, and producing potentially wide- scale-up in large populations with little access to health
ranging and long-lasting effects. By addressing critical services, and different delivery mechanisms of the
consciousness,*” groups have the potential to create intervention will need to be tested in partnership with
improved capability' in communities to deal with the government apd non-government organisations.
health and development difficulties arising from poverty Contributors
and social inequalities.11 The intervention requires a All authors contributed to the design of the study and criticised drafts of
training and support structure to manage facilitators in the report. PT, NN, SB. and AC were responsible for the conception and
charge of 12-14 groups per month, with every' group overall supervision of the trial. PT and NN managed the project, data
wv?w thelancet.com Vol 37s April 3, 2010
I
Years 2 and 3
264
104(39%)
77(29%)
29(11%)
22 (11%)
32(12%)
1191
I
Articles
gathering, data entry, and administration with assistance from ShR, SuR.
RM, RG, and DM. SB and AC were technical advisers for the study.
JB provided technical assistance with gathering and analysis of cost data.
AC, PT, and SB helped design the original trial protocol. SB designed the
methods for data gathering and epidemiological surveillance system.
VP and RL provided technical advice about the K10 scale and commented
on drafts of the report. RL trained interviewers to administer the K10
scale. NN, AP, CP, and SB did the quantitative analysis. AC, AP, PT. NN,
and SB wrote the first draft of the report, and were responsible for
subsequent collation of inputs and redrafting. PT and AC are guarantors
for the report.
Conflicts of interest
We declare that we have no conflicts of interest.
Acknowledgments
The study was funded primarily by the Health Foundation. Funds from
the UK Department for International Development contributed to field
expenses during the trial, and funds from the Wellcome Trust supported
the local team during the data analysis period in 2009. VP is supported
by a Wellcome Trust Senior Clinical Research Fellowship. We thank the
members of the partnering communities, women's group members, and
all mothers and relatives who gave their time to be interviewed and
without whom none of this study would have been possible; all the study
stall Peasant Behera. Birender Mahto, Gunjan Kumari,
Chaturbhuj Mahanta. Nibha Kumari Das. Lakhindar Sardar.
Sumitra Gagrai. Rajesh Singhdao. Uttam Mallick. and Lipton Sahoo, and
the interviewers, facilitators, and key informants; Asit Mishra.
Parameswar Pradhan. and A K Debdas for their assistance in assigning
cause of death from the verbal autopsies; H P S Sachdev.
Suranjeen Pallipamula. Raniesh Sharan, Haldhar Mahto. Almas Ali,
Alison Dernbo Rath, and Ashish Das for reviewing the data; David Osrin
for his support throughout the study, and for contributions to the data
monitoring process; Ruth Duebbert (Women and Children First) for
commenting on the report; Sambit Beura (Addsoft) and Aman Sen
(Mother and Infant Research Activities. Nepal) for their assistance with
the database; Sarah Ball for her continuous administrative support;
Sudharak Olwe for his photographs of Ekjut activities; Professional
Assistance for Development Action (PRADAN) for providing access to
the women's groups set up by them; the Health Foundation. UK
Department for International Development, and the Big Lottery Fund for
their financial support of the project; and the anonymous reviewers for
suggestions that greatly improved this report.
References
1
WHO. Neonatal and perinatal mortality; country, regional and
2
3
-1
5
6
7
8
9
1192
global estimates. 2006. http://whqlibdoc.who.int/
publications/2006/9241563206_eng.pdf (accessed Feb 26. 2010).
Lawn |, Cousens S. Zupan ), et al. 4 million neonatal deaths: when,
where? why? Lancet 2005; 365; 891-900.
UNICEF. Countdown to 2015; tracking progress in maternal, newborn
and child survival. The 2008 report. New York: UNICEF. 2008.
UNICEF. The state of Asia-Pacific's children. New York; UNICEF.
2008.
Manandhar DS. Osrin D. Shreslha BP, et al. Effect of a participatory
intervention with women's groups on birth outcomes in Nepal:
cluster-randomised controlled trial. Lancet 2004; 364: 970-79.
Morrison. I. "lamang S. Mesko N. et al. Women's health groups to
improve perinatal care in rural Nepal. BMC Pregnancy Childbirth
2005; 5: 6.
Engle PL. Maternal mental health; program and policy implications.
Am j Clin Nutr 2009; 89; 963S-66S.
Patel V, Kirkwood B. Perinatal depression treated by community
health workers. Lancet 2008; 372: 868-69.
Government of India. 2007 Poverty estimates for 2004-2005.
http://planningcommission.nic.in/news/prmar07.pdf (accessed
Jan 25, 2010).
World Bank. Jharkhand: addressing the challenges of inclusive
development. 2007. http://web.worldbank.org/WBSITE/
EXTERN AL/COUNTRIES/SOUTHASl.AEXT/0,,contentMDK:21370
355~pagePk.T467.36~piPK:146830~theSitePK;223547.00.htrnl
(accessed Feb 12. 2010).
11
National Family Health Survey. National final report. 2003.
http://www.nfhsindia.org/ (accessed Jan 25, 2010).
12 Government of India. Sample registration system bulletin
(2001-2003). New Delhi: Registrar General of India. 2006. 1-21.
http://censusindia.gov.in/Vital_Statistics/SRS_Bulletins/MMRBulletin-April-2009.pdf’(accessed Aug 24, 2009).
13 National Planning Commission. Scheduled castes and scheduled
tribe population census 2001. http://delhiplanning.nic.in/Economic
%20Survey/ES20Q7-08/T18.pdf (accessed Jan 22. 2009).
14 Subramanian SV, Smith GD. Subramanyam M. Indigenous health
and socioeconomic status in India. PLoS Med 2006; 3: e421.
15 Census of India: Census data 2001. Office of the Registrar General
and Census Commissioner. India, http://www.censusindia.gov.in/
(accessed Jan 26, 2010).
16 Barnett S, Nair N. Tripathy PK, Borghi J. Rath S. Costello A.
A prospective key informant surveillance system to measure
maternal mortality-findings from indigenous populations in
Jharkhand and Orissa. India. BMC Pregnancy Childbirth 2008;
8: 6.
17 WHO. International Classification of Diseases, tenth revision
(ICD-10), Geneva: World Health Organization. 2006.
18 WHO. The World Mental Health Survey Initiative.
http://www.hcp.med.harvard.edu/wmh/ (accessed Jan 25. 2010).
19 Kessler RC. Andrews G. Colpe LJ. et al. Short screening scales to
monitor population prevalences and trends in non-specific
psychological distress. Psychol Med 2002: 32: 956-76.
20 Patel V. Araya R. Chowdhary N, et al. Detecting common mental
disorders in primary care in India; a comparison of five screening
questionnaires. Psychol Med 2008; 38: 221-28.
21 National Rural Health Mission. Report of Common Review Mission:
Jharkhand. November 25th-6th December 2008. http;//www.
mohfw.nic.in/NRHM/CRM/Jharkharid_2nd_CRM_Report.pdf
(accessed Aug 18, 2009).
22 Whitney D, Trosten-Bloom A. The power of appreciative inquiry.
New York: Berrett-Koehler Publishers. 2002.
23 Hayes R, Moulton LH, Morgan BJT. et al. Cluster randomised
controlled trials. New York: Chapman and Hall/CRC Press. 2009.
24 Borghi J. Thapa B. Osrin D, et al. Economic assessment of a
women's group intervention to improve birth outcomes in rural
Nepal. Lancet 2005; 366: 1882-84.
25 Milgrom J, Gemmill AW. Bilztza JL. et al. Antenatal risk factors for
postnatal depression: A large prospective study, j Affect Disord 2008;
108: 147-57
26 Bell AC. D'Zurilla T|. Problem-solving therapy for depression:
a meta-aiialysis. Clin Psychol Rev 2009; 4: 348-53.
27 Kumar V, Mohanty S, Kumar A. et al. Effect of community-based
behaviour change management on neonatal mortality in Shivgarh,
Uttar Pradesh. India: a cluster-randomised controlled trial. Lan,et
2008; 372: 1151-62.
28 Baqui AH, El-Arifeen S, Darmstadt GL, et al. Effect of
community-based newborn-care intervention package
implemented tlnough two service-delivery strategies in Sylhet
district, Bangladesh: a cluster-randomised controlled trial.
Lancet 2008; 371:1936-44.
29 Haines A. Sanders D. Lehmann U, et al. Achieving child survival
goals: potential contribution of community health workers.
Lancet. 2007; 369: 2121-31.
30 Freire, P. Education for critical consciousness. New York:
Continuum International Publishing Group. 2005.
31 Sen A. Development as freedom. New York: Knopf. 1999.
32 National Rural Healtli Mission. Accredited Social Health Activists.
http://www.mohfiv.nic.in/NRHM/asha.htm (accessed Feb 2, 2010).
10
www.thelancet.com Vol 375 April 3, 2010
Articles
@ Effect of a participatory intervention with women's groups
on birth outcomes and maternal depression in Jharkhand
and Orissa, India: a cluster-randomised controlled trial
Prasonta Tripathy, Nirmala Nair, Sarah Barnett, Rajendra Mahapatra, Josephine Borghi, Shibanand Rath, Suchitra Rath, Rajkumar Gape.
Dipnath Mahto, Rajesh Sinha, Rashmi Lakshminarayana, Vikram Patel. Christina Pagel, Audrey Prost, Anthony Costello
Lancet 2010; 375= 1182-92
Published Online
March 8. 2010
Summary
Background Community mobilisation through participatory women’s groups might improve birth outcomes in poor
rural communities. We therefore assessed this approach in a largely tribal and rural population in three districts in
eastern India.
001:10.1016/50140 ■
6736(09)62042-0
See Comment page 1142
See Articles page 1193
Ekjut, Chakradharpur,
Jharkhand, India
(P Tripathy MSc, N Nair MBBS,
R Mahapatra MSc. Sh Rath MA,
Su Rath MA. R Gope PGDRD,
D Mahto DCHM. R Sinha);
University College London
Methods From 36 clusters in Jharkhand and Orissa, with an estimated population of 228186, we assigned 18 clusters
to intervention or control using stratified randomisation. Women were eligible to participate if they were aged
15-49 years, residing in the project area, and had given birth during the study. In intervention clusters, a facilitator
convened 13 groups every month to support participatory action and learning for women, and facilitated the
development and implementation of strategies to address maternal and newborn health problems. The primary
outcomes were reductions in neonatal mortality rate (NMR) and maternal depression scores. Analysis was by intention
to treat. This trial is registered as an International Standard Randomised Controlled Trial, number
ISRCTN21817853.
Centre for International Health
and Development, Institute of
Child Health, University College
London, London, UK
(S Barnett PhD. A Prost PhD.
Prof A Costello FRCP); Effective
Intervention, Centre for
Economic Performance,
Findings After baseline surveillance of 4692 births, we monitored outcomes for 19030 births during 3 years (2005-08).
NMRs per 1000 were 55-6, 37-1, and 36-3 during the first, second, and third years, respectively, in intervention
clusters, and 53 -4, 59-6, and 64-3, respectively, in control clusters. NMR was 32% lower in intervention dusters
adjusted for clustering, stratification, and baseline differences (odds ratio 0 • 68, 95% CI 0 • 59-0 • 78) during the 3 years,
and 45% lower in years 2 and 3 (0 • 55, 0 • 46-0 ■ 66). Although we did not note a significant effect on maternal depression
overall, reduction in moderate depression was 57% in year 3 (0 -43, 0 -23-0-80).
London School of Economics
and Political Science, London,
UK IR Lakshminarayana DNB);
London School of Hygiene and
Tropical Medicine, London, UK
(J Borghi PhD. Prof V Patel PhD);
Interpretation This intervention could be used with or as a potential alternative to health-worker-led interventions,
and presents new opportunities for policy makers to improve maternal and newborn health outcomes in poor
populations.
Sangath, 841/I Alto Porvorim,
Goa, India (Prof V Patel); and
Clinical Operational Research
Funding Health Foundation, UK Department for International Development, Wellcome Trust, and the Big Lottery
Fund (UK).
Unit, University College
London, London, UK
(C Pagel PhD)
Correspondence to:
Dr Prasanta Tripathy, Ekjut. Ward
Number 17. Plot 5568. Potka.
Po-Chakradharpur. District West
Singhbhum. Jharkhand, 833102,
India
prasanta.ekjut@gmail.com
1182
Introduction
Every year, an estimated 3-7 million children worldwide
die in the first month of life.1 ? Global progress in
reduction of maternal and child mortality rates is
insufficient—only 16 of 68 countries are on track to
achieve Millennium Development Goal 4 (reduction of
mortality rate in children <5 years by two-thirds between
1990-2015).'' India accounts for 20% of maternal deatlis
worldwide, 21% of all child (<5 years) deaths, and 25% of
all neonatal deaths.' Urgent efforts are needed to reduce
these mortality rates quickly through cost-effective and
scalable interventions.
Large improvements were rioted in birth outcomes in a
poor rural population in Makwanpur, Nepal, after a lowcost, potentially sustainable, and scalable participatory
intervention with women’s groups.1 Newborn mortality
rates were 30% lower in intervention areas than in control
areas (odds ratio 0-70, 95% CI 0-53-0 ■94).’ Local female
facilitators assisted women’s groups every month to
consider the causes and underlying problems leading to
maternal and newborn deaths, develop practical, strategies
with community leaders and men, and implement, and
assess the outcomes of these strategies. This community
action cycle, adapted from a programme developed in
Bolivia, created health and non-health benefits at low
cost.''
Maternal depression is an increasing public health
concern in low-income countries because of its high
prevalence and wide-ranging implications for the health
of the mother and infant.7 Delivery' of appropriate
interventions to prevent or treat maternal depression
through health workers is a major challenge in countries
witfi under-resourced health systems, and community
groups assisted by non-healtfi-care workers might have
some advantage in helping the poorest women.1*
Jharkhand and Orissa are two of the poorest states in
eastern India. About 40% of their total combined
population lives below the poverty line.” The average life
expectancy among women in both states is about 60 years,
and an estimated 63% are illiterate.”1 Neonatal mortality
www.thelancet.com Vol 375 April 3, 2010
Articles
I
|
I
rate (NMR) per 1000 livebirths is 49 in Jharkhand and 45
in Orissa, and maternal mortality ratio per 100000
livebirths is 371 and 358, respectively."12 These are
disproportionately higher than India's national estimates
of i39 per 1000 for NMR and 301 per 100000 for maternal
mortality ratio."12
More than 20% of Jharkhand and Orissa’s population is
affiliated with Scheduled Tribes (or Adivasi—ie, indigenous
grpups), and about 12% witli scheduled castes.13 Despite
calls for inclusive development. Adivasi communities
remain underserved—their employment rate is roughly
half that of non-indigenous people, and nearly a third of
Adivasi children in Jharkhand and Orissa do not receive
primary' education.'11 Indigenous communities also have
higher mortality rates and poorer access to health services
than do the non-indigenous populations.14
We hypothesised that a participatory' intervention with
women s groups could reduce neonatal mortality by at
least 25% in underserved tribal communities of eastern
India, and improve home-care practices and health
seeking behaviour of pregnant and postnatal women,
and their family members: and that the women's group
intervention could reduce maternal depression in the
intervention areas by 30%.
• ±7
L
Saraikela
1 Kharswan _ I
India
West Singhbhum
Jharkhand
Orissa
/r
/
0
h
0
Methods
I____ 300 miles
300 km
Study location and population
Our study was done in three contiguous districts of
Jharkhand and Orissa—Saraikela Kharswan, West
Singhbhum, and Keonjhar (figure 1). The proportion of
Adivasis within the study clusters was 58-70%. Eligible
participants were women aged 15-49 years, residing in
the project area, and who had given birth during the
Study (July 31. 2005, to July 30, 2008). The study
population was an open cohort—ie. women could enter
(the study at any time during the trial period if they had
(given birth. Women who chose to participate gave their
consent (written or left-thumb print.) and were free to
decline an interview at any time. Women who were
identified by interviewers as having symptoms of
severe depression were referred to the nearest tertiary
mental health centre at Ranchi. Ethical approval was
obtained from an independent ethical committee in
Jamshedpur, India.
Figure 1: Map of districts and distribution of clusters
I 36 dusters (12 per district)
Saraikela district
12 clusters
Keonjhar district
12 clusters
i Step 1
7 dusters with pre-existing
women's groups
4 allocated to intervention
3 allocated to control
3 clusters with pre-existing
women's groups
1 allocated to intervention
2 allocated to control
8 dusters with pre existing
women's groups
4 allocared to intervention
4 allocated to control
I Step 2
5 clusters with no groups
2 allocated to intervention
3 aitocateill to control
9 clusters with no groups
5 allocated to intervention
4 allocated to control
4 dusters with no groups
2 allocated to intervention
2 allocated to control
West Sinqhbum district
12 clusters
+
+
Figure 2: Randomisation process
Randomisation
We identified 12 rural clusters per district, with a mean
population of 6338 per cluster (range 3605-7467). The
estimated population in these 36 clusters was 228186 (on
the basis of the 2001 Indian census projections).1' In
18 clusters, existing women’s groups were involved in
savings and credit activities (seven in West Singhbhum,
three in Saraikela Kharswan, and eight in Keonjhar). In
the first district (West Singhbhum), an external
observer from a partner non-governmental organisation
(Professional Assistance for Development Action) drew
folded papers with numbers corresponding to clusters
www.thelaricet.corn Vol 375 April 3, 2010
with existing groups from a basket. The first four clusters
were allocated to the intervention group, the rest to the
control group. This process was repeated for clusters
without women’s groups and in the other two districts in
the presence of external observers (figure 2). We chose
this method because of simplicity and visibility, which
were necessary to convince local communities that the
process was transparent. Because of the nature of the
intervention, neither the intervention team nor the
participants were masked to group assignment during
the trial.
1183
Articles
I Key informer identifies all births and deaths
I
£
. ..................... ;____
Births
Deaths in women of reproductive age
Interviewer interviews family member
leoarding mother's background characteristics,
antepartum, intrapartum and post-partum
information 6 weeks after delivery
............... T.
Interviewer ascertains all information
Interviewer ascertains all information
r
Livebirths
r
Z3
Neonatal
deaths
Stillbirths
Maternal
deaths
Pregnancyrelated deaths
Late maternal
deaths
zr
Supervisor does verbal autopsy with provider
Supervisor does verbal autopsy with family member
I
I
Clinicians assign cause of death
Figure 3- Surveillance system
Panel 1: Definitions
•
Miscarriage: cessation of a presumptive pregnancy before
•
Neonatal death: death of a liveborn infant within
22 weeks of gestation before delivery of the baby's head’-'
28 completed days of birth
•
Early neonatal deaths: deaths arising within 6 completed
days
•
Late neonatal deaths: deaths arising from 7 to
28 completed days of birth
Perinatal death: a stillbirth or an early neonatal death
•
Maternal death: death of a woman while pregnant or
•
within 42 days of cessation of pregnancy from any cause
related to the pregnancy or its management, but not
from accidental causes
In September, 2004, we met with gram sabhas (village
councils), headmen, and representatives from panchayats
(elected representatives for basic governance) in the three
districts. These representatives granted permission to
start surveillance of births and deaths, and to work with
women's groups.
age every month within their allocated area. The key
informant met with an interviewer once a month who
verified births and interviewed all identified mothers to
gather information about their sociodemographic
characteristics, pregnancy, delivery, and postnatal period
through a structured questionnaire about 6 weeks after
delivery. As additional checks for the identification
stage, all women identified were asked to locate any
others of reproductive age who had recently given birth
or died in the study area. Every district had 12
interviewers, one interviewer supervisor, and a district
manager. The monitoring manager supervised field
based activities in all three districts and data entry at the
head office. The intervention and surveillance teams
were partitioned—interviewers and facilitators belonged
to different villages, their training was done separately,
and they had review meetings on separate days.
In the event of a stillbirth or neonatal death (panel 1),
the interviewer administered a questionnaire and did a
verbal autopsy with the mother and other individuals
present at the time of death; the verbal autopsy included
free text in which the mother was asked to narrate the
details of events leading to the death of the neonate. For
deaths among women of reproductive age, the
interviewer spoke to family members to ascertain the
age of the woman, cause of death, and whether she was
pregnant or had recently given birth. In the event of a
maternal death, the monitoring supervisors, initially
accompanied by a physician, did verbal autopsies with a
relative who was present at the time of the death.
Supervisors completed a standard questionnaire with
free text for elaboration of the sequence of events before
the death. Verbal autopsies for maternal deaths were
done by a physician (n=ll), interviewer supervisor
(n=96), or district manager (n=2).
Maternal depression was included as a trial outcome in
the second year of the study because of delays in
identification of a contextually appropriate scale. We used
the Kessler-10 item scale (K10), a questionnaire for the
detection of common mental disorders in community
settings, that has been used in India and World Mental
Health Surveys.”’-20 A psychiatrist (RL) did three training
sessions of 2 days each for groups of 12-15 interviewers
that consisted of administration of the K10 questionnaire,
aspects of understanding depression and body language,
association between physical health and mental health,
active listening skills, and confidentiality.
Clusters and coverage of women's groups
Surveillance
A surveillance system with key informants was
established in the three districts. Figure 3 shows this
system, and Barnett and colleagues1'' describe it in detail.
One key informant, usually a traditional birth attendant
or active village member, was responsible for about
250 households, and reported any births, maternal or
newborn deaths, and deaths in women of reproductive
1184
In the 18 intervention clusters, we used a participatory
action cycle with 172 existing groups and created an
additional 72 groups. Coverage of Ekjut groups was one
group per 468 population. In year 1, 546 (18%) of 3119
newly pregnant women attended the groups, rising to
1718 (55%) of 3126 in year 3. We recorded 111006 group
attendances over 3 years. 74715 (67%) of these were from
married women of reproductive age, 15 030 (14%) from
www.thelancet.com Vol 375 April 3, 2010
Articles
adolescent girls, 10452 (9%) from men, and 10809 (10%)
from elderly women.
/
Women's group intervention
F'very group met monthly for a total of 20 meetings,
and a local woman, selected on the basis of criteria
(including speaking the local language and having the
ability to travel to meetings) identified by the
community, facilitated the meetings. After a 7-day
residential training course to review the cycle's
contents, and to practice participatory communication
techniques, facilitators were given support through
fortnightly meetings with district coordinators.
Facilitators coordinated an average of 13 meetings every
month with as many groups.
Groups took part in a participatory learning and action
cycle (figure 4). Community members who were not
regular group members were also encouraged to
participate in discussions. Information about clean
delivery practices and care-seeking behaviour was shared
through stories and games, rather than presented as key
messages. By discussion of case studies imparted
IHrough contextually appropriate stories, group members
identified and prioritised maternal and newborn health
problems in the community, collectively selected relevant
stkitegies to address these problems, implemented the
strategies, and assessed the results. Although some
strategies were common, each group was free to
implement its own combination of strategies. The
intervention team adapted facilitation materials from
the study in Makwanpur. Nepal, to guide the meetings.’'
Groups used methods such as picture-card games, role
play, and story-telling to help discussions about the
causes and effects of typical problems in mothers and
infants, and devised strategies for prevention, home
care support, and consultations (figure 5).
www.thelancet.com Vol 375 April 3, 2010
4
it
^9
Phase 4
Assess effect
<
4
/ .z
Phase 1
Identify and
prioritise
difficulties
Community
meeting 2
Phase 3
Put strategies
s A into practice
Phase 2
Plan strategies
'
%
%
%
4
4
///
t
•9
1
X
%
X
‘A
T.
%
v
Community
meeting 1
Figure^: Meetings in women's group cycle
'1
Health-service inputs
We formed health committees in all intervention and
control clusters so that community members would have
the opportunity to express their opinions about the
design and management of local health services. About
ten village representatives within every cluster met once
every' 2 months and used a structured action cycle to
discuss maternal and newborn health entitlement issues.
As a result, committee members became more
wnowledgable about the government health system and
assisted with the formation of village health committees
as part of the National Rural Health Mission programme.’’
Id addition to the creation of cluster-level health
committees, we provided workshops for appreciative
inquiry with frontline government health staff from
seven clusters per district in Jharkhand.’2 Participants
assessed rhe programme qualitatively at the end of every
training session. We expected that any improvement in
performance or service quality would be equal in
intervention and control clusters.
/
/
'Avj
;•
■
F 5
' %
<
Ifm.
'fl I
Wfci-.....
Figure 5-' Women's group meeting in Jharkhand, India
Individuals m the photo provided permission (written 01 thumb print! for publication of image
Primary and secondary outcomes
The primary outcomes were reductions in NMR and
maternal depression scores. Secondary outcomes were
stillbirths, maternal and perinatal deaths, uptake of
antenatal and delivery services, home-care practices
during and after delivery, and health-care-seeking
behaviour (seeking care from qualified providers in the
1185
Articles
36 clusters randomised with stratified allocation
(18 with existing women's groups)
228186 estimated population
6338 mean cluster population (range B6O5--7467)
1
I
1
18 clusters allocated to control
(9 with existing women's groups)
18 clusters allocated to intervention
(9 with existing women's groups)
+
18 clusters given intervention
9770 births (109 pairs of twins, two sets of triplets)
941’9 livebirths
301 stillbirths
406 neonatal deaths
18 clusters not given intervention
9260 births (115 pairs of twins)
8980 livebirths
280 stillbirths
531 neonatal deaths
Excluded from analyses
0 clusters
2 mothers refused interview
Excluded from analyses
0 clusters
2 mothers refused interview
Excluded from adjusted analyses
Migrated
84 births (9 neonatal deaths, 3 stillbirths)
81 mothers (0 maternal deaths)
Excluded from adjusted analyses
Migrated
171 births (13 neonatal deaths, 10 stillbirths)
16/ mothers (0 maternal deaths)
Analysed for mortality outcomes
18 clusters
8662 mothers
9686 births
397 neonatal deaths
298 stillbirths
Analysed for mortality outcomes
1.8 clusters
8125 mothers
9089 births
518 neonatal deaths
270 stillbirths
Analysed for depression oulcome
(years 2 and 3. excluding maternal deaths)
6452 mothers
Analysed for depression outcome
(years 2 and 3. excluding maternal deaths)
5979 mothers
Figure 6: Trial profile
antenatal, delivery, and postnatal period, for checkups
and problems).
Quality control of data
Data were double-entered in an electronic database.
Surveillance supervisors manually checked information
provided by key informants and interviewers. The field
surveillance manager, data input officer, and data
manager undertook manual and systematic data
checks.
Statistical analysis
We did not expect the intervention to have adverse
effects at cluster or participant level, and therefore did
not have any rules for stopping the intervention. In
December, 2007, we presented findings from a
preliminary analysis to an independent data safety
committee. After an interim analysis in 2007, the
committee recommended that the trial continue for a
total of 3 years to enable comparison with the
Makwanpur study,’ in which the effect was measured
from 9 months after the beginning of the intervention
to allow exposure to the women’s groups in pregnancy.
The data safety committee also noted that 3 years would
1186
allow analysis of possible seasonal variations in NMR.
The committee undertook a final review of the data in
December. 2008.
Our prospective surveillance from Nov 21, 2004, to
July 30, 2005, showed a baseline NMR of 58 per
1000 livebirths (261 deaths per 4509 livebirths) and
maternal mortality ratio of 5.10 per 100000 (23 deaths
per 4509 livebirths). The trial was planned for 3 years
and was originally powered, like the Makwanpur trial?
for a 2-year analysis of birth outcomes, after allowing a
period of up to 1 year for the women’s groups to be
established and for pregnant women to be given the
intervention. We assumed a between-cluster correlation
coefficient of variation (fc) of 0-15-0-25, and about
324 births per cluster during 2 years. On the basis of
10?4 loss to follow-up, a sample size of 18 clusters per
group resulted in 64-81% power to detect a 25% reduction
in NMR. With an estimated baseline prevalence of 1596
and k of 0-3, the study had 79-81% power to detect a
30% reduction in maternal depression over 1 year. We
used data for recorded births during the study to
estimate that the study had a power of 92% to detect a
25% reduction in NMR.
Analysis was by intention to treat at cluster and
participant levels. We excluded data from mothers who
migrated out of the region and their infants from
intention-to-treat analyses since many of these women
probably came into the clusters at the time of delivery
and would therefore not have been exposed to the
intervention in pregnancy. We aimed to do the tests of
significance for our primary and secondary outcomes on
the basis of previously agreed hypotheses about the likely
effect of the intervention. For comparison of mortality
outcome, we used multivariate logistic regression with
random effects on individual-level data in Stata (version
10.0).n We compared secondary indicators using
generalised estimating equations models with semirobust
S Es at the cluster level.Generalised estimating equations
models were used for secondary indicators and categorical
scores of maternal depression because these outcomes
had high intracluster correlation coefficients (>0- 30). We
compared K10 scores grouped in three categories
(none/mild, moderate, or severe) during years 2 and 3 of
the trial. This method was chosen in favour of linear
regression to address the data’s strong positive skewness.
We adjusted for stratification by including strata as
variables in the regression analyses, and for multiple
hypothesis testing by adjusting the p values for the
primary outcomes using the Holm correction in the
results tables. We did not adjust for clustering at the level
of ffie mother. All results are presented as odds ratios
with 95% Cis.
Cost-effectiveness analysis
We used a similar method of cost-effectiveness analysis
as used by Borghi and colleagues/4 Costs were estimated
at 2007 prices, and were calculated separately for the
www.thelancet.com Vol 375 April 3, 2010
I
Articles
I
I
women’s group intervention and activities for health
service strengthening. These were the financial and
economic costs of setting up the intervention, and
running costs during the trial. Costs were estimated
frdm the perspective of a provider to give an indication
oflthe potential costs of replication for the government
an|d interested agencies, and discounted at 3%.
Incremental cost effectiveness was measured in relation
to a do-nothing alternative.
This study is registered as an International Standard
Randomised Controlled Trial, number ISRCTN21817853.
Intervention area
Control area
2457
2235
Radio, cassette tape, bicycle, or electricity
1752(71%)
1771 (79%)
More costly items (television, generator, battery, fan, fridge)
167(7%)
225 (10%)
Identified births
Socioeconomic characteristics
Household assets
Ownership of agricultural land
None
345(14%)
364 (16%)
Own less than 2 bighas (<O-27 hectares)
1157(47%)
969 (43%)
Own between 2-4 bighas (O-27-O-54 hectares)
653(27%)
593 (27%)
Caste or tribal group
Role of the funding source
Scheduled tribe*
1849(75%)
1557 (70%)
The funders had no role in the design of the study, data
collection, data analysis, interpretation, or writing up of
tlife findings, although they made a site visit early in the
study implementation. The corresponding author had
access to all the data and had final responsibility for the
decision to submit for publication.
Scheduled caste*
80(3%)
520 (21%)
64 (3%)
606 (27%)
<20 years
147(6%)
20-29 years
1370(56%)
253 (11%)
1385(62%)
a.30 years
345(14%)
348(16%)
Not known
933 (38%)
597 (26%)
Results
Other backward caste*
Maternal age
Maternal school education
Figure 6 shows the trial profile. All 18 selected dusters
None
1908 (78%)
1533 (69%)
had the intervention. Loss to follow-up after birth as a
Primary
143(6%)
12S (6%)
result of migration or refusal of interview was 86 (<1%)
Secondary or higher
405(16%)
577(26%)
of 9770 women in intervention clusters and 173 (2%) of
Maternal literacy
9^60 in control clusters. In the study areas, 5661 (37%) of
Cannot read
1906 (78%)
1566 (70%)
1|118 home deliveries were by a relative, friend, or
66]) (30%)
Can read
550 (22%)
neighbour. 5368 (36%) by traditional birth attendants,
Care-seeking behaviour and home-care practices
and 1913 (13%) by husbands.
Any antenatal care
1460(59%)
15.32 (69%)
Table 1 shows the baseline characteristics of identified
Three or more antenatal visits
539(22%)
70(1 (31%)
births during 9 months of data gathering from Nov 21,
Any iron tablets during pregnancy
1571 (64%)
1497(67%)
2004, to July 30, 2005. Numbers of births were similar in
Institutional delivery
277(11%)
326(15%)
intervention and control clusters, but differences were
Home delivery
2118 (86%)
1858 (83%)
noted in household assets, maternal education, literacy,
Delivery attended by traditional birth attendant I'
778 (37%)
717 (39%)
and tribal membership, with women in the intervention
Birth attendant washed handst
609 (29%)
471 (25%)
clusters being generally poorer and more disadvantaged
Birth attendant used safe delivery kitt
195(9%)
197(11%)
man those in the control clusters (table 1).
Birth attendant used plastic sheet I
163
(8%)
141(8%)
NMRs unadjusted for clustering decreased from year 1
Cord cut with newer boiled bladet
12^4 (70%)
1493
(70%)
to year 3 in the intervention clusters compared with an
Cord tied with boiled tlhreadt
276 (13%)
2^2(13%)
increase in the control clusters (table 2).
Infant wiped within 30 mint
1247(61%)
11 jo (63%)
We noted a 32% reduction in NMR during the 3-year
Infant wrapped within 30 mint
787(39%)
782 (43%)
tjrial when data were adjusted for clustering,
Infants alive at 1 month
2202 (90%)
2046(92%)
stratification, and baseline differences (table 3). NMR
Exclusive breastfeeding for 6 weeks§
1387(63%)
1168 (57%)
was reduced by 45% in intervention clusters compared
I
i
with control clusters during the last 2 years (table 3).
Data are number or number (%). 'Standard terms used in Indian demographic, surveys. tHome deliveries only (stillbirths >
The reduction in NMR was still significant when
were not excluded from the count). iStillbirths excluded. $ Number of infants alive at 6 weeks were 2201? in intervention i
dusters, and 2046 in control clusters
migrated mothers and their babies were excluded from
the intention-to-treat analyses during the 3 years
Table 1: Baseline characteristics of identified births in intervention and control areas
(table 3). k. estimated from retrospective data from
control clusters, was 0-10 when we took stratification
After year 1, N^4Rs were considerably lower in the
into account, and corresponded to an intracluster intervention clusters than in die control clusters (figure 7).
correlation coefficient of 0-0005.25 In the analysis of In the third year of the study, NMR was almost half in the
heonatal mortality data at the cluster level, the risk ratio intervention clusters compared with the control clusters
was 0-71 (95% CI 0-57-0-90, p=0-0011) for years 1 to (odds ratio 0-53), 95% CI 0-41-0-68, adjusted for
3 when adjusted for mu..dtiple hypothesis testing with clustering and stratification only), much larger than the
the Holm procedure and unadjusted for baseline findings of the Majkwanpur trial in Nepal.’ Figure 8 shows
differences.
' -------,
the cluster-level changes in NMR between baseline and
ww.thelancet.com Vol 375 April 3, 2010
1187
Articles
Year If
Basci inc’
Inter
vention
Control
All
Inter
vention
Year 31
Year 21
Years 1-3*
Years 1-31
Inter
vention
Control
Inter
vention
Control
Inter
vention
3195
3110
3073
9770
2985
9469
Control
All
Inter
vention
Control
9260
19 030
9686
9089
18775
898O
18449
9388
18207
915
Control
All
Births
2457
2235
4692
3171
3052
3404
3135
Livebirths
2347
2162
4509
3073
2960
3286
3035
Stillbirths
109
183
98
92
118
100
301
280
581
298
261
171
158
122
181
192
406
531
937
397
518
Early (0-6 days)
145
100
85
113
88
Neonatal deaths
73
116
8819
270
80
180
116
107
76
135
67
138
259
380
639
253
368
621
Late (7-28 days)
45
36
81
55
51
46
46
46
54
147
151
298
144
150
294
209
16
153
362
214
199
194
235
152
226
560
660
1220
551
638
1189
7
23
20
30
22
18
7
12
49
60
109
49
60
109
Stillbirth rate per
1000 births
44-4
32-7
390
30-9
30-1
347
319
26-6
28-6
30-8
30-2
30-5
30-7
297
307
Neonatal mortality rate
per 1000 livebirths
618
536
579
556
534
371
596
36-3
64 3
429
59-1
508
42-3
587
507
Early neonatal mortality
rate per 1000 livebirths
(0-6 days)
42-6
370
400
37-8
361
231
44-4
215
46 2
27-3
42-3
34-6
269
417
341
Late neonatal mortality
rate per 1000 livebirths
(7-28 days)
19-1
16-6
18-0
17-9
177
14-0
15-1
14-7
18-0
15-5
16 8
16-1
15-3
170
161
Perinatal mortality rate
per 1000 births
85-1
68-4
77-1
67-4
657
57-0
750
47-5
73-5
57-3
71-2
64-1
56-8
70-1
63-3
Maternal mortality ratio
per 100 000 livebirths
6817
323-8
510-1
650-8
1013-5
669-5
593-0
225-1
402-0
517-5
668-1
590-8
5219
680-3
5987
Perinatal deaths
Maternal deaths
568
' Data are unadjusted. 'Excluding migrated mothersand infants, tlnduding migrated mothersand infants.
Table 2: Births and deaths in intervention and control clusters at baseline and during trial
Years 1-3 (including p value
migrated)*
Years 1-3 (excluding p value
migrated)*
Years l-3t
p value
Years 2 and 3t
p value
Neonatal mortality rate
per 1000 livebirthsi
0-71 (0-61-0-83)
<0-0005
0-69 (0-60-0-81)
<00005
0-68 (0-59-078)
<0-0005
0-55 (0-46-0-66)
<0-0005
Early neonatal mortality
rate (0-6 days)
0-63 (0-54-0-75)
<0-0005
0-62 (0-53-0-74)
<0-0005
0 62 (0-52-0 73)
<0-0005
0-46 (O-37-O-57)
<0-0005
Late neonatal mortality
rate (7-28 days)
0-92 (0-67-176)
0-476
0-89 (0-65-1-22)
0-463
0-84 (0-64-1 12)
0736
0-80 (0-56-1-14)
0717
Stillbirth rate per
1000 births
1-02 (0-85-173)
0-833
104 (0-85-1-25)
0-773
1-05 (0-86-1-28)
0-656
.1-01 (0-80-178)
0-914
Perinatal mortality rate
per 1000 births
079 (070 -0-90)
<0-0005
079 (0-69-0-90)
<00005
0-79 (0-69-0-9D
<0-0005
0-68 (0-58-079)
<0-0005
Maternal mortality ratio
per 100000 livebirths
0-80 (0-51-174)
0-180
0-80 (0-51-1-24)
0-180
070 (0-46-1-07)
0-104
0-50 (0-48-1-49)
0-563
Data are odds ratio (95% Cl). "Adjusted for stratification (by district and pre -existing women's groups) and clustering only, t Adjusted for stratification, clustering, maternal
education, assets, and any tribal affiliation, ip values adjusted for multiple hypothesis testing with l-lolrn correction were <0-001.
Table3: Comparison of mortality rates in intervention and control dusters
year 3—the NMRs fell below their baseline level in most
intervention clusters. Between 2005 and 2008, perinatal
mortality rates in the intervention clusters decreased
compared with those in the control clusters when
adjusted for clustering (table 2; table 3). Stillbirth rates
did not differ between intervention and control clusters
(table 2). Maternal mortality ratio was generally lower in
intervention than in control clusters, but the study was
not powered to detect significant differences (table .3).
Qualitative evidence from the assessment of the trial’s
process showed that community mobilisation through
1188
women's groups might have contributed to avoidance of
some maternal deaths (panel 2).
There was no detectable difference in maternal
depression K10 scores, when measured about 6 weeks
after delivery, between intervention and control clusters
in year 2 of the study or overall (table 4). However, in
year 3, when 55% of all pregnant women in the
intervention clusters had joined a group, a 57% reduction
was noted in moderate depression among mothers in
the intervention clusters compared with control clusters
(table 4).
www.thelancet.com Vol 375 April 3, 2010
Articles
No significant differences were rioted in health-careseeking behaviour between control and inteivention
clusters (table 5). However, home-care practices showed
substantial improvements—in intervention clusters,
birth attendants were more likely to wash their hands,
use a safe delivery' kit and a plastic sheet, and boil the
thread used to tie the cord than were those in the control
clusters, "flie proportion of infants exclusively breastfed
at 6 weeks was higher in intervention areas in adjusted
analyses for years 2 and 3.
Clause-specific differences in mortality rate as a
percentage of all causes—septicaemia, birth asphyxia,
hypothermia, and prematurity—during the 3 years were
not clearly discernable because there was a reduction in
all causes (table 6). The incremental cost of the women’s
group intervention was US$910 per newborn life saved,
inareasing to $1308 (in 2007 prices) when health-service
strengthening activities were included. The incremental
cojft per life-year saved was $33 for the women’s group
intervention ($48 inclusive of health-service strength
ening activities). The women’s group intervention in
this setting was therefore more cost effective than that
reported in NepaP as a result of the greater effect of
wdmen’s groups on NMR combined with lower
operating costs in the current context.
Control areas
Intervention areas
100
a.
h
If1
11s
:
I I
80-
T
40 -
T
20-
I
0
2
Year
3
2
Year
3
Figure?: Boxplot of cluster-level neonatal mortality rates by allocation and
study year
lOO-i
Control clusters
• Intervention clusters
80-
£ £
11
i§
O
60-
40-
I
20"150
0
50
100
Discussion
Baseline neonatal mortality rate per 1000 livebirths
Wknen's groups led by peer facilitators reduced NMR
and moderate maternal depression at low cost in largely Figure 8: Scatterplot of cluster specific neonatal mortality rates in year 3
tribal, rural populations of eastern India. Our data show with rates at baseline
that mortality reduction in underserved rural settings
was not associated with increased care-seeking
Panel 2: Case-study effect of women's groups on
behaviour or health-service use. The most likely
strategies to avoid maternal deaths
mechanism of mortality reduction was through
iniproved hygiene and care practices. The availability of
A woman in the eighth month of her pregnancy, and her
safe delivery kits increased in both control and
mother-in-law attended a monthly women's group
intervention areas, but women’s groups seemed to
meeting where they participated in a drill based on what to
generate more demand in intervention clusters than in
do in the event of post-partum bleeding. After a month,
control clusters. In places where kits were not provided,
when the woman delivered at home and had severe
group members made them and provided information
bleeding, her mother-in-law remembered what had been
about their contents to mothers, then visited pregnant
said in the group, and, without wasting time, asked her
women during the eighth month of pregnancy to ensure
daughter-in-law to breastfeed the baby while she rushed to
that they had received kits and would use them. Birth
get money from the group and asked her son to arrange for
outcomes might have been affected by the fact that
a vehicle. The daughter-in-law was immediately taken to
these community members attended the groups or
the district hospital, where she was given medicines,
wbre advised by group members, thus generating
intravenous fluid, and two pints of blood, and was
increased social awareness and support for clean
discharged after 15 days.
delivery practices.
The most striking reduction in mortality rate was noted
iri early neonatal deaths, which might be explained by the with improved care. Tile reason for this combined
strong focus on intrapartum and early neonatal periods reduction of asphyxia, prematurity, and septicaemia
ill several case studies and stories discussed during the could also be improved intrapartum care. Potential
cycle. Attribution of cause of newborn deada on the basis mechanisms for reduction of mortality rate will be
of verbal autopsy is an imperfect science, and deaths further assessed in future analyses of verbal autopsies
might, arise from several and overlapping causes. Early and seasonal mortality trends.
septicaemia could have been reduced with clean delivery
Our findings also show that a low-cost intervention
pkactices, and premature babies might have survived involving non-heallh-care workers might affect maternal
www.thelancet.com Vol 375 April 3, 2010
1189
I
Articles
Year 2
Mothers (n)
No or mild depression (10-15)
Moderate depression (16-30)
Severe depression (31-50)
Years 2 and 3
Year3
Adjusted odds
ratio (95%CI)‘
Intervention
Control
6452
5979
2665 (90%)
233 (l'25-438)
5884 (91%)
5277 (88%)
1-29 (0-68-2-44)
293 (10%)
0-43 (O-23-O-8O)
536 (8%)
676 (11%)
0-74 (0-40-1-37)
5 (<1%)
0-70 (015-3 JI)
32(<1%)
26 (<1%)
1-29 (0-46-3-64)
Adjusted odds
ratio (95% Cl)'
Intervention
Control
3120
2963
2612(87%)
0-91 (0-41-2-01)
2962 (95%)
382 (13%)
21 (<1%)
1-04 (0-50-216)
154 (5%)
4 (<1%)
Intervention
Control
3332
2922 (88%)
3016
383 (11%)
28 (<1%)
153 (0-47-5-05)
Adjusted odds
ratio (95% Cl)*
i Data are number (%), unless otherwise indicated. ‘Results adjusted for clustering, stratification, maternal education, tribe affiliation, and household assets by use of generalised estimated equations with
; semirobust SEs for individual-level data.
Table4: Kessler-10 depression scores in mothers in intervention and control clusters
Intervention
clusters
Control clusters
Odds ratio
(95% Cl) for years
1-3*
Odds ratio
(95% Cl) for years
i-3t
Odds ratio (95% Cl)
for years 2 and 3+
9468
8867
Any antenatal care
6990 (74%)
6623 (75%)
0-97 (0-48-1-97)
1-60 (0-65-3-92)
1-86 (0-80 •4-34)
s3 antenatal care visits
3001 (32%)
3621(41%)
0-63 (0-37-106)
0-69(0-37-1-26)
0-68 (0-37-1 24)
Iron tablets
6997 (74%)
6293 (71%)
1-12 (071 1-76)
1-31 (0-62-2-75)
1-34 (0-77-2-35)
Maternal tetanus-toxoid injection
7767 (82%)
7377 (83%)
0-90(0-51-l-54)
1-39 (0-85-2-28)
1-40 (0-85-2-29)
Illness in pregnancy
5206 (55%)
4983 (56%)
103 (0-68-1-58)
1-10 (0-71-1-72)
1-01 (0-67-1-52)
Visited health facility in case of illness during
pregnancy
945 (10%)
922 (10%)
0-78 (0-39-1-56)
0-86 (0-46-1-60)
0-80 (0-39-1-65)
Births*
Institutional deliveries
1364 (14%)
1811 (20%)
0-64(0-39-104)
0-89 (0-51-1-53)
0-94 (0-50-1-76)
Birth attended by formal provider (doctor or nurse)
1490 (1.6%)
2067 (23%)
0-59 (0-37-1 -94)
0-81. (0-50-1-31)
0-82 (0-47-1-43)
8084
7034
Birth attended by traditional birth attendant
2692 (33%)
2676 (38%)
0-82 (0-43-1-60)
0-84 (0-43-1-64)
0-85 (0-44-1-65)
Birth attendant washed hands with soap
3291 (41%)
1583(23%)
2-05 (1 14-3-73)
2- 07 (1-24-3-45)
2-5O(l-35-4-62)
Safe-delivery kit used
2594(32%)
1284(18%)
2- 08 (1-25-3-44)
1-87 (1-11-3-14)
2-28 (1-27-4 09)
Plastic sheet used
2088 (26%)
560 (8%)
3- 85 (2-51-5-89)
3- 74(2-48-5-65)
2-98 (1-84-4-81)
Cord tied with boiled thread
2559 (32%)
786(11%)
3-9 (1-82-6-30)
3-02 (161-5-65)
4-33 (2-06--9-11)
Cord cut with new or boiled blade
6679 (83%)
5570(79%)
1-24 (0-82-1-87)
1-35 (0-86-2-12)
1-55 (0-96-2-51)
7890
6873
1-01 (0-39-2-62)
Home deliveries
Livebirths (home deliveries)
Cord undressed or dressed with antiseptic
6600 (84%)
6115(89%)
0-52 (0-24-1-12)
0-58 (0-27-1-26)
Infant wiped within 30 min
4741 (60%)
4227(62%)
0-90 (038-214)
1-01 (0-43-2-36)
1-06 (0-44-2-57)
Infant wrapped within 30 min
2846 (36%)
2980(43%)
0-74 (O-35-1-59)
0-78 (0-36-1-66)
0-81 (0-37-1-80)
1-06 (0-52-2-17)
0-95 (0-44-2-10)
1-22 (0-56-2-65)
Infant not bathed in first 24 h
Infants alive at 1 month
Any of three infant illnesses (cough, fever,
diarrhoea)
2107(27%)
1509 (22%)
8807
8119
1739 (20%)
2388 (29%)
0-62 (0-37-1-03)
0-67 (0-40-112)
()-61 (O'35-I 06)
Care-seeking behaviour in event of infant illness
940(54%)S
1050 (44%)$
1-53 (0-77-3-05)
0-88 (0-97-3-61)
1-55 (0-79-3-04)
Infant put to breast within 4 h
5390(61%)
4942(61%)
1-01 (0-48-2-14)
0-90 (0-38-311)
1-11 (0-45-2-76)
Exclusive breastfeeding for first 6 weeks
7022(80%)
5611 (69%)
1-82 (1-14-2-92)
1-44 (0-89-2-35)
1-74 (1-03-2-94)
Data are number (%). unless otherwise indicated. ‘Adjusted for clustering and stratification only. I Adjusted for clustering, stratification, maternal education, assets, and any
tribal affiliation. I Excludes births to migrated mothers and twins. §Denominators are number of infants with any of three infant illnesses: 1739 for intervention clusters and
2388 for control clusters.
Table S'- Process indicators in intervention and control clusters
mental health. We hypothesise that the large reduction
in moderate depression seen in the third year could
have occurred through improvements in social support
and problem-solving skills of the groups. Adequate
social support reduces the risk of depression during
pregnancy and is an important social determinant of
mental health?'1 In meetings, information was shared
1190
about the difficulties encountered by mothers in the
community, and practical ways to collectively address
them were established. Group meetings also
strengthened problem-solving skills, a component of
psychotherapeutic interventions that has been shown to
affect depression in other settings.7" The intervention
seemed to have no effect on severe depression, perhaps
wwwthelancet.com Vol 375 April 3, 2010
Articles
because it was more similar to primary prevention
Intervention
Control
rather than treatment, or because severe depression is
lesi amenable to psychotherapeutic interventions. A
Years 1-3
Years 2 and 3
Year 1-3
complete analysis and discussion of these findings will
Early neonatal deaths
253
140
367
be presented in the future.
Birth asphyxia
92 (36%)
53 (38%)
142 (39%)
lU'o potential effect modifiers in this trial, on the basis
Prematurity
85 (34%)
46 (33%)
110(30%)
ol evidence, were differences in maternal education, and
Septicaemia
38(15%)
15(11%)
47 (13%)
tribal membership and assets between the intervention
Hypothermia
16(6%)
12 (9%)
26(7%)
and control populations. These were taken into account
Other
22 (9%)
14(10%)
42 (11%)
in adjusted analyses and mainly provided an advantage
Data
are
number
or
number
(%).
fori the control areas. Additionally, the high significance
of some of our results could be a result of an increase in
Table 6: Cause-specific mortality for early neonatal deaths during study
mortality rate in the control, areas between 2005-08. The
control and intervention clusters were in similar responsible for a population of about 500 and for
geographic areas, so factors that affected NMR should recruiting up to half of newly pregnant women. Costs are
have affected both groups equally, but further lower than for most other primary health-care
investigation is needed.
interventions, and these interventions can complement
we believe that the study had two main weaknesses. existing self-help groups in the community.
First, as in several other community-based randomised
Two other issues arising from our study are cost
control trials, the intervention and surveillance teams effectiveness and the effect on maternal mortality ratios.
were not unaware of allocation. However, there were no The interventions in the Ekjut. trial were more cost
incentives or disincentives for over-reporting or under effective than those i|n the Makwanpur study' because of
reporting births and deaths, and several process lower operating costs and greater effect of the
mechanisms were in place to detect errors. Second, intervention. In the Nepal trial, effect of women’s groups
although migration out of districts was common, we on maternal mortality ratios was significant, although
cannot rule out some intercluster migration when the number of deaths was small and maternal mortality
women married out of their home cluster. Our intention- ratio was not a stated primary outcome. In our trial the
10-treat analysis might have affected the results positively maternal mortality ratio was higher in the intervention
or negatively.
areas at baseline, and 20% lower after 3 years of
In the Shivgarh study,’7 in Uttar Pradesh, India, the effect intervention, but this difference was not significant and
of an intensive behaviour-change programme involving the trial was not powered to measure differences in
community meetings and home visits by a new cadre of maternal mortality. Reduction in maternal mortality will
paid, non-governmental community workers in a depend mainly on improved access to health services
population of 104123 during 15 months resulted in a 54% and to life-saving drugs, but community mobilisation
reduction (relative risk 0-46, 95% CI 0-35-0-60) in NMR could help through improvement in hygiene at delivery
with changes in home-care practices, but no real change in and early care-seeking behaviour for complications by
care-seeking behaviour. No overall differences in NMR addressing the first-delay component.
were noted during 30 months of intervention in the
Iffiis participatory intervention with women’s groups
Prpjahnmo trial,w in Bangladesh, but a 34% reduction could complement oir be a potential alternative to health
(0-66. 0-47-0-93) was noted in the home-care group in the worker led interventions, two examples of which have
last 6 months of the programme. Tlie investigators of the been discussed here. Our findings raise several important
Prpjahnmo study* rioted that “Availability of referral issues for policy makers in India. Could federal and state
services and a strong supervisory system were crucial to governments invest in this programme? Should
tills intervention and would be a necessary feature of government or non-government organisations be
scaling up die intervention."
responsible for its scale-up? Could such a participatoryInterventions with health-worker home visits have intervention support and strengthen the National Rural
rafely achieved adequate coverage, quality, or effectiveness Health Mission's mandate of communitisation of health
when taken to scale in poor populations.7'’ Participatory and the Accredited Social Health Activist programme?’7
groups have the advantage of helping the poorest, being Further assessmentfs of this approach will involve a
scalable at low cost, and producing potentially wide- scale-up in large populations with little access to health
ranging and long-lasting effects. By addressing critical services, and different delivery mechanisms of the
consciousness,’0 groups have the potential to create intervention will need to be tested in partnership with
injproved capability in communities to deal with the government and non-government organisations.
health and development difficulties arising from poverty Contributors
and social inequalities.” The intervention requires a All authors contributed to the design of the study and criticised drafts of
training and support structure to manage facilitators in the report. PT, NN, SB. and AC were responsible for the conception and
charge of 12-14 groups per month, with every' group overall supervision of the trial. PT and NN managed the project, data
'4wthelancet.com Vol375 Aprils, 2010
I
Years 2 and 3
264
104(39%)
77(29%)
29 (11% 1
22 (8%)
32 (12%)
ligi
II
Articles
gathering, data entry, and administration with assistance from ShR, SuR.
RM, RG, and DM. SB and AC were technical advisers for the study.
| B provided technical assistance with gathering and analysis of cost data.
AC, PT, and SB helped design the original trial protocol. SB designed the
methods for data gathering and epidemiological surveillance system.
VP and RL provided technical advice about the K10 scale and commented
on drafts of the report. Rl. trained interviewers to administer the K10
scale. NN, AP, CP. and SB did the quantitative analysis. AC, AP. PT NN.
and SB wrote the first draft of the report, and were responsible for
subsequent collation of inputs and redrafting. PT and AC are guarantors
for the report.
Conflicts of interest
We declare that we have no conflicts of interest.
Acknowledgments
The study was funded primarily by the Health Foundation. Funds from
the UK Department for International Development contributed to field
expenses during the trial, and funds from the Wellcome Trust supported
the local team during the data analysis period in 2009. VP is supported
by a Wellcome Trust Senior Clinical Research Fellowship. We thank the
members of the partnering communities, women’s group members, and
all mothers and relatives who gave their time to be interviewed and
without whom none of this study would have been possible; all the study
staff Peasant Behera, Birender Mahto, Gunjan Kumari,
Chaturbhuj Mahanta, Nibha Kumari Das. Lakhindar Sardar.
Sumitra Gagrai. Rajesh Singhdao. Uttam Mallick. and Lipton Sahoo, and
the interview'ers, facilitators, and key informants; Asit Mishra.
Parameswar Pradhau, and A K Debdas for their assistance in assigning
cause of death from the verbal autopsies; H P S Sachdev.
Suranjeen Pallipamula. Ramesh Sharan, Haldhar Mahto. Almas Ali.
Alison Dernbo Rath, and Ashish Das for reviewing the data: David Osrin
for his support throughout the study, and for contributions to tire data
monitoring process; Ruth Duebbert (Women and Children First) for
commenting on the report: Sambit Beura (Addsoft) and Aman Sen
(Mother and Infant Research Activities. Nepal) for their assistance with
the database; Sarah Ball for her continuous administrative support;
Sudharak Olwe for his photographs of Elqut activities; Professional
Assistance for Development Action (PRADAN) for providing access to
the women’s groups set up by them; the Health Foundation. UK
Department for International Development, and the Big Lottery Fund for
their financial support of the project; and the anonymous reviewers for
suggestions that greatly improved this report.
References
1
WHO. Neonatal and perinatal mortality: country, regional and
global estimates. 2006. http://whqlibdoc.who.int/
publications/2006/9241563206_eng.pdf (accessed Feb 26, 2010).
Lawn J. Cousens S, Zupan J. et ai. 4 million neonatal deaths: when,
2
where? why? Lancet 2005: 365: 891-900.
UNICEF. Countdown to 2015: tracking progress in maternal, newborn
3
and child survival. The 2008 report. New York: UNICEF. 2008.
4
UNICEF. The state of Asia-Pacific's children. New York: UNICEF,
2008.
Manandhar DS. Osrin D. Shreslha BP, el al. Effect of a participatory
5
intervention with women's groups on birth outcomes in Nepal:
cluster-randomised controlled trial. Lancet 2004; 364: 970-79.
6
Morrison, J. Tamang S. Mesko N. et al. Women's health groups to
improve perinatal care in rural Nepal. BMC Pregnancy Childbirth
2005; 5: 6.
7
Engle PL. Maternal mental health: program and policy implications.
Am J Clin Nutr 2009; 89: 963S-66S.
Patel V, Kirkwood B. Perinatal depression treated by community
8
health workers. Lancet 2008; 372: 868-69.
9
Government of India. 2007. Poverty estimates for 2004-2005.
http://planningcommission.nic.in/news/prmar07.pdf (accessed
Jan 25. 2010).
1192
io
World Bank. Jharkhand: addressing the challenges of inclusive
development. 2007. http://web.worldbank.org/WBSITE/
EXTERNAL/COUNTRlES/SOUTHASlAEXT/0,,contentMDK:21370
355~pagePk:146736~piPK:146830~theSitePK:223547.00.htrnl
(accessed Feb 12. 2010).
n National Family Health Survey. National final report. 2003.
http://www.nfhsindia.org/ (accessed Jan 25, 2010).
12 Government of India. Sample registration system bulletin
(2001-2003). New Delhi: Registrar General of India, 2006: 1-21.
http://censusindia.gov.in/Vital_Statistics/SRS_Bulletins/MMRBulletin-April-2009.pdf (accessed Aug 24, 2009).
13 National Planning Commission. Scheduled castes and scheduled
tribe population census 2001. http://delhiplanning.nic.in/Economic
%20Survey/ES2007-08/TI8.pdf (accessed Jan 22. 2009).
14 Subramanian SV, Smith GD. Subramanyam M. Indigenous health
and socioeconomic status in India. PLoS Med 2006: 3: e421.
15 Census of India: Census data 2001. Office of the Registrar General
and Census Commissioner. India, http://www.censusindia.gov.in/
(accessed Jan 26, 2.010).
16 Barnett S, Nair N. Tripathy PK, Borghi J. Rath S. Costello A.
A prospective key informant surveillance system to measure
maternal mortality-findings from indigenous populations in
Jharkhand and Orissa. India. BMC Pregnancy Childbirth 2008:
8: 6.
17 WHO. International Classification of Diseases, tenth revision
(ICD-10). Geneva: World Health Organization, 2006.
18 WHO. The World Mental Health Survey Initiative.
http://www.hcp.med.harvard.edu/wmh/ (accessed Jan 25. 2010).
19 Kessler RC. Andrews G. Colpe 1_J. et al. Short screening scales to
monitor population prevalences and trends in non-specific
psychological distress. Psychol Med 2002: 32: 956-76.
20 Patel V, Araya R. Chowdhary N. et al. Detecting common mental
disorders in primary care in India: a comparison of five screening
questionnaires. Psychol Med 2008; 38: 221-28.
21 National Rural Health Mission. Report of Common Review Mission:
Jharkhand. November 25th-6th December 2008. http://www.
rnohfw.nic..in/NRHM/CRM/J harkhand_2nd_CRM_Report.pdf
(accessed Aug 18, 2009).
22 Whitney D, Trosten-Bloom A. The power of appreciative inquiry.
New York: Berrett-Koehler Publishers. 2002.
23 Hayes R, Moulton LH, Morgan BJT. et al. Cluster randomised
controlled trials. New York: Chapman and Hall/CRC Press. 2009.
24 Borghi J. Thapa B. Osrin D, et al. Economic assessment of a
women's group intervention to improve birth outcomes in rural
Nepal. Lancet 2005; 366: 1882-84.
25 Milgrom J. Gernmill AW. Bilztza ] L. et al. Antenatal risk factors for
postnatal depression: A large prospective study. ] Affect Disord 2008;
108: 147-57.
26 Bell AC. D'Zurilla T|. Problem-solving therapy for depression:
a meta-analysis. Clin Psychol Rev 2009; 4: 348-53.
27 Kumar V, Mohanty S, Kumar A. et al. Effect of community-based
behaviour change management on neonatal mortality in Shivgarh,
Uttar Pradesh, India: a cluster-randomised controlled trial. Lancet
2008; 372: 1151-62.
28 Baqui AH, El-Arifeen S. Darmstadt GL, et al. Effect of
community-based new'born-care intervention package
implemented through tw'o service-delivery strategies in Sylhet
district, Bangladesh: a duster-randomised controlled trial.
Lancet 2008; 371:1936-44.
29 Haines A. Sanders D. Lehmann U. et al. Achieving child survival
goals: potential contribution of community health workers.
Lancet 2007; 369: 2121-31.
30 Freire, P. Education for critical consciousness. New York:
Continuum International Publishing Group, 2005.
31 Sen A. Development as freedom. New York: Knopf, 1999.
32 National Rural Health Mission. Accredited Social Health Activists.
http://www.mohfw.nic.in/NRHM/asha.htin (accessed Feb 2, 2010).
www.thelancet.com Vol 375 April 3, 2010
I
Series
Alma-Ata: Rebirth and Revision 5
Community participation: lessons for maternal, newborn,
and child health
Mikey Rosato, Glenn Laverock, Lisa Howard Grabrnan, Prasanta Tripathy. Nirrnala Nair, Charles Mwansambo, Kishwar A/.ad, Joanna Morrison,
Zulfiqar Bhutto, Henry Perry, Susan Rifkin, Anthony Costello
Lancet 2008; 372:962-71
See Editorial page 863
This is the fifth in a Series of
eight papers about Alma-Ata:
rebirth and revision
Centre for International Health
and Development, Institute of
Child Health, UniversityCollege
London, UK (M Rosato MSc.
J Morrison MSc,
Prof A Costello FRCP);
Department of Social and
Community Health, School of
Population Health, University
of Auckland, Auckland, New
Primary health care was ratified as the health policy of WHO member states in 1978.' Participation in health care was
a key principle in the Alma-Ata Declaration. In developing countries, antenatal, delivery, and postnatal experiences for
women usually take place in communities rather than health facilities. Strategies to improve maternal and child
health should therefore involve the community as a complement to any facility-based component. The fourth article
of the Declaration stated that, “people have the right and duly to participate individually and collectively in the
planning and implementation of their health care”, and the seventh article stated that primary health care “requires
and promotes maximum community and individual self-reliance and participation in the planning, organization,
operation and control of primary health care”. But is community participation an essential prerequisite for better
health outcomes or simply a useful but non-essential companion to the delivery of treatments and preventive health
education? Might it be essential only as a transitional strategy: crucial for the poorest and most deprived populations
but largely irrelevant once health care systems are established? Or is the failure to incorporate community participation
into large-scale primary health care programmes a major reason for why we are failing to achieve Millennium
Development Goals (MDGs) 4 and 5 for reduction of maternal and child mortality?
Zealand
(ProfG Laverack PhD);
Independent Consultant
Introduction
Soon after the Alma-Ata Declaration, arguments for
s Rifkin PhD); Ekjut project, selective rather than comprehensive primary health care
Chakradhapur, Jharkhand,
dominated health system debates.2 Policy makers in favour
India (PTripathy MSc,
of selective primary health care argued that community
N Nair MBBS); Department of
Paediatrics, Kamuzu Central
interventions such as oral rehydration solution, immun
Hospital, Lilongwe, University
isation, or vitamin A capsules could be targeted effectively
of Malawi
at poor, albeit passive, recipients with immediate benefit.
(C Mwansambo FRCPCH);
They
recognised that community participation was
Perinatal Care Project,
Bangladesh Diabetic important in supporting the provision of local health
Association, Dhaka,
services and in delivering such interventions at scale, but
Bangladesh (Prof K Azad MSc);
believed that pilot programmes showing long-term benefits
Aga Khan Department of
from more comprehensive community mobilisation had
Paediatrics, Karachi, Pakistan
(prof z Bhutta PhD); and Future been much less successful when governments tried to take
Generations, Franklin, WV, USA
them to scale.
(H Perry PhD)
More recently the lack of progress with the Millennium
Correspondence to
Development Goals (MDG) and primary health care in
Prof Anthony Costello
many poor countries has encouraged those in favour of
UCL Centre for International
Health and Development, comprehensive primary health care to question whether
Institute of Child Health, London
the failure to address community care and participation
WC1N 1EH, UK
effectively within health programmes is a major reason
a.costello@ich.ud.ac.uk
for poor sustainability and ineffective scaling-up of
selective interventions of proven efficacy. The review of
the WHO Integrated Management of Childhood Illness
strategy reinforced these questions: “Delivery systems
that rely solely on government health facilities must be
expanded to include the full range of potential channels
(L Howard Grabman MA.
962
families have a direct effect on maternal and child health?
If so, how do these interventions work most effectively,
and how can they be taken to scale?
What are participation, mobilisation, and
empowerment?
The closely related concepts of participation, mobilisation,
and empowerment require definition. Participation has
been used to indicate active or passive community
involvement. In the past, mobilisation consisted of
communities responding to directions given by
professionals to improve their health. This process usually
took the form of mass campaigns for immunisations
where communities were passively involved as the setting
where the interventions were implemented or the target
of the specific intervention. More recently, health and
development workers have begun to act as facilitators
focusing on the process of health improvements as well
as the outcomes. In this approach the facilitators support
local communities to become actively involved—to
participate—in both activities and decisions that affect
their own health, either as a resource that can provide
assets to address a health problem or an agent of change
that uses its own supportive and developmental capacities
to address its needs. In this paper we will discuss this
more recent form of community mobilisation, which we
define as "a capacity-building process through which
in a setting ana strong community-based approaches.
community individuals, groups, or organizations plan,
The focus on process within child health programmes
must change to include greater accountability for
intervention coverage at population level.”*
A crucial policy question is whether specific community
participation interventions aimed at women and their
carry out, and evaluate activities on a participatory and
sustained basis to improve their health and other needs,
either on their own initiative or stimulated by others".4
Health programmes today often identify empowerment
rather than participation as an objective. Empowerment
www.thelancet.com Vol 372 September 13, 2008
Series
can be defined as the process and outcome of those
without power gaining information, skills, and confidence
and thus control over decisions about their own lives,5
and can lake place on an individual, organisational, and
community level. Community mobilisation, by our
definition, is a way to support this empowerment process
and reach this empowerment outcome (figure 1).
What evidence led to Alma-Ata?
The Alma-Ata Declaration arose from evidence generated
by the Joint WHO/UNICEF Study of Alternative
Approaches to Meeting Basic Health Needs of Populations
in Developing Countries under the leadership of Halfdan
Mahler (Assistant Director-General at WHO, 1970-73)
and Kenneth Newell (Director of Research in
Epidemiology and Communications Science at WHO,
1962-72), who were influenced by the work of the
Christian Medical Commission in Geneva and its
growing commitment to community-oriented primary
health care as the most appropriate approach to
addressing the health needs of poor people.6,7* This
evidence revealed the successes of national health
programmes in China, Cuba, Sri Lanka, Tanzania, and
Venezuela as well as in subnational programmes in
Guatemala, India, Indonesia, Iran, Kenya, and Niger,
which all used community participation as a fundamental
component of primary health care.
The Jamkhed Project in the state of Maharastra in India
and the Kakamega Project in Western Kenya are examples
of successful smaller-scale subnational pilot programmes
where community mobilisation was a key intervention
(panel I).10" Communities were assisted to identify their
own problems, collect their own data, and implement
their own solutions. These demonstration projects
provided clear evidence of a dramatic effect on health but
could not be easily replicated by governments on a larger
scale. Once part of a national programme, bureaucratic
rules and top-down directives changed the nature of
community participation and heavy donor support
emphasised performance targets rather than the
unhurried process necessary for engagement with
communities. Miriam Were, director of Kakamega,
lamented that “officials and international experts could
not understand that successes had arisen from the
process, not from the setting of performance targets, and
that the forward momentum had been generated from
within the community and not from external
financing’’."
The failure to scale-up Jamkhed, Kakamega, and other
similar projects through national governments con
tributed to a move away from participatory approaches to
primary health care. Since 1990 the focus of child survival
efforts has been on increasing the coverage of health
commodities with proven effectiveness—such as oral
rehydration solution for diarrhoea,,2,13'4 cotrimoxazole for
childhood pneumonia,15 '6 vitamin A supplementation,17,1*
insecticide-treated bednets,,,,,2" and vaccinations.2' At the
www.thelancet com Vol 372 September 13, 2008
I
Increasing empowerment
i
Information sharing
Consultation
Collaboration
Full responsibility
Figure 1: From passive to active community participation
same time, maternal survival efforts also moved away
from community approaches focusing on traditional
birth attendants, which lacked clear evidence of
effectiveness, to efforts entirely focused on strengthening
district hospital midwifery and obstetric care services and
health systems.22 21
These approaches to the diseases of poverty proved
more saleable to policy makers for two main reasons.
Firstly, the clear-cut and rapid public health gains shown
by these approaches fitted well within the new culture of
evidence-based medicine. Secondly, the scalability of
distribution of these approaches seemed intrinsically
easier and less expensive than more long-term
comprehensive primary health care approaches involving
community mobilisation despite strong evidence
supporting their effectiveness and affordability.24
What is the effect of community mobilisation
on maternal, newborn, and child health ?
Progress towards MDGs 4 and 5 in the poorest countries
has remained slow in high-mortality settings,.25,26 Between
1990 and 2005 there was no substantial change in
maternal mortality in sub-Saharan Africa, and of the 68
priority countries targeted for child survival
improvements, 41% were deemed to have made
insufficient progress and 38% made no progress.26
Additionally, in 11 African countries there were reversals
in under-5 mortality rates in the same period.26 The
evident ineffectiveness of existing programmes and
conclusion that this may in part be due to the lack of
community involvement has led to a renewed focus on
community mobilisation strategies for maternal,
newborn, and child survival.1
Most studies of community mobilisation interventions
have investigated
investigated the
the effectiveness
effectiveness of specific
have
interventions targeted at a passive recipient community—
the old style of community mobilisation (for example,
breastfeeding promotion, diarrhoea prevention and
treatment, growth
growth promotion,
promotion,27,2
27a* *-29,3
29-"3<’ promotion of
treatment,
complementary feeding after 6 months of age,” treatment
of severe acute malnutrition12 and pneumonia prevention
and treatment’1’4'”). Far fewer studies have investigated
the effectiveness
of community
mobilisation
interventions, either on their own or in combined
packages with other interventions, where the community
provides the resources and is the active agent of change
(table). In Ethiopia a cluster randomised controlled trial
(cRCT) showed that mobilising women’s groups to
effectively recognise and treat malaria at home led to a
963
Series
Panel 1: Projects in rural India and Kenya which influenced
Alma-Ata
Jamkhed Project (1970 to date)
In 1970, Raj and Mabelle Arole, two doctors, started a primary
health care programme in Jamkhed, a rural area in
Maharashtra state in India.910The project used a participatory
approach to bring villages together and establish farmers
clubs. These clubs identified problems facing the community
and chose to focus on improvements to water supplies and
sanitation. As the clubs evolved they became women's
development organisations and implemented solutions such
as: identifying women to be trained as health workers; funds
for women with a household health emergency or food crisis;
keep village clean drives; literacy programmes; advocacy for
encounters with bureaucracy; and micro-credit schemes.The
programme expanded to other villages, eventually covering a
population of more than 250 000. Over the first 20 years
(1972-1992) the project showed a reduction in infant
mortality rate from 176 to 19 per 1000, and a birth rate
decline from 40 to 20 per 1000.10 Additionally, rates of
antenatal care, safe delivery, and immunisation are nearly
universal and rates of malnutrition have declined from 40%
to less than 5%-’0 In parallel, the women's groups have
developed a greater sense of their potential for agency, and
caste barriers among women have gradually diminished.
Kakamega Project (1974 to 1982)
The Kakamega project led by Miriam Were was established in
western Kenya in 1974-11 Women in communities were
supported to identify their own problems, collect their own
data and select their own community health workers with
open community involvement. Among otherthings
communities set up village funds and bank accounts and
established transport schemes enabling access to secondary
care. The project achieved improvements in primary care,
immunisation, water supplies, family planning, and malaria
control. It also increased community support and self-
reliance. As the women became empowered the visits from
developed strategies to address them, and then
implemented and assessed the strategies in co-operation
with local leaders, men, and health workers.19 The
mobilisation intervention had been developed in Bolivia
under the Warmi programme
(figure 2). The Warmi
programme had seen a large reduction in perinatal
mortality rate using before and after analysis of a small
population, and the larger Makwanpur cRCT showed a
30% reduction in neonatal mortality rate, as well as
significantly fewer maternal deaths (although the
numbers of maternal deaths were few and maternal
mortality ratio had not been a primary outcome for the
trial).”
Two more recently published studies are the Hala and
Projahnmo community effectiveness trials in Pakistan
and Bangladesh, which combine demand and supply
side interventions, with different results.42,41 The Hala
trial was a pilot non-randomised controlled trial in which
Lady Health Workers (government health workers
responsible for about 200 families each) received training
in home-based neonatal care and local traditional
midwives (dais) received voluntary training. In addition,
village health committees were established for maternal
and newborn health. Compared with baseline rates the
trial showed a 35% decline in perinatal mortality rate and
a 28% decline in the neonatal mortality rate in the
intervention villages. The control villages showed no
decline.42 The Projahnmo cRCT assessed the effectiveness
of specially trained community health workers, who
provided a home-care package including assessment of
newborn infants on the first, third, and seventh days after
birth, and referral or treatment of sick neonates. The
study showed a 34% reduction in neonatal mortality rate
in the final 6 months of the trial compared with the
comparison group.41 However, unlike the studies outlined
above, the third community care arm, in which
community mobilisers held community meetings with
women in villages, showed no effect on neonatal mortality
compared with the control arm.41
outside facilitators became less frequent.
What are the current controversies surrounding
964
40% reduction in under-5 mortality.16 For newborn care,
the SEARCH Project in India showed the value of a
complex home-based newborn care package (which
included community delivery of injectable antibiotics,
health promotion, training of traditional birth attendants,
and physician visits) within a programme where
communities had been mobilised over an extended
period.17 Bang and colleagues1* ascribe 36% of the
reduction in neonatal mortality rate to sepsis
management; assessing the contribution of community
community mobilisation interventions?
mobilisation within the intervention compared with
impossible.44
control villages is more difficult, although important.
In Makwanpur district, Nepal, women’s groups, led by
a locally recruited woman facilitator, were supported
through a community mobilisation action cycle where
they discussed maternal and newborn health problems,
community mobilisation is less effective than a home
care strategy in reducing neonatal mortality rate in
communities with a weak health system and low health
care use. Several other trials testing different combina
tions of interventions, with mobilisation as a core
Community mobilisation versus home care visits
Although increasing evidence favours the effectiveness
of community mobilisation interventions, a comparison
of the Makwanpur and Projahnmo trials is central to this
policy dilemma. The Makwanpur trial suggests that
community mobilisation through women's groups is a
cost-effective approach to reduce neonatal mortality rate
in remote villages where developing and maintaining a
programme of home visits by outreach workers has been
Projahnmo, by contrast, suggests
that
www.thelancet.com Vol 372 September 13, 2008
g
1I
□
Design
Type of interventions
Sample
Results
O'Rourke"
Before and
after analysis
Women's groups using
Newborn mortality
community action cycle.
rate;
Intervention focused on initiating perinatal mortality
and strengthening women's
rate
organisations, developing
women's skills in problem
identification and prioritisation,
and training community
members in safe birthing
techniques
Remote Bolivian
mountain villages, around
15000 population.
Evaluated by comparing perinatal mortality rate and obstetric
behaviour among 409 women before and after the
intervention. Perinatal mortality rate decreased from 117 deaths
per 1000 births before the intervention to 43 8 deaths per 1000
births after. The proportion of women receiving prenatal care
and initiating breast-feeding on the first day after birth was also
significantly larger.Number of infants attended to immediately
after delivery increased, but the change was not statistically
significant
SEARCH,
Gadchiroli
district, India
Bang'7”
1999 and
2005
Home-based newborn care
Still birth rate;
Controlled
consisting of sepsis management, newborn mortality
trial (not
supportive care of low
rate; perinatal
randomised)
birthweight newborn babies,
mortality rate;
Baseline
phase (1993- asphyxia management, primary
newborn mortality
rate; infant
1995) ,
prevention, health education and
observational training of traditional birth
mortality rate
phase (1995- attendants. Built upon a preexisting community mobilisation
1996) , ano
the 7 years of orogramme
intervention
(1996-2003)
The baseline population
was 39 312 in
39 intervention villages
and 42 617 in 47 control
villages Livebirths in
10 years were 8811 in
inten/ention villiages and
9990 in control villages
Newborn mortality rate in the control area showed an increase
from 58 in 1993-1995 to 64 in 2001-2003-The rate fell by 70%
(95% Cl 59-81%) compared with the control area. Early
newborn mortality rate decreased by 64% and late newborn
mortality rate by 80%. Still birth rate decreased 49% and the
perinatal mortality rate by 56%. Newborn mortality rate did not
change, and the infant mortality rate decreased by 57%, (95% Cl
46-68%) Cause-specific newborn mortality rate (1995-1996 vs
2001-2003) for sepsis decreased by 90%, for asphyxia by 53%.
and for prematurity by 38%
Tigray,
KidaneJi
2000
Cluster
randomised
controlled
trial (cRCT)
Mother coordinators trained to
teach other local mothersto
recognise symptoms of malaria in
their children and to promptly
give chloroquine
Under-5 mortality
rate—
Total population of 70
506 in 37 tabias (cluster of
villages) in two district
were paired according to
under-5 mortality rates.
24 tabias with the highest
malaria morbidity were
selected
190 of 6383 (29-8 per 1000) children younger than 5 years died
in the intervention tabias compared with 366 of 7294 (50-2 per
1000 children) in the control tabias. Under-5 mortality was
reduced by 40% in the intervention localities (95% Cl 29 2-50-6;
paired t test, p<0 003). Of 190 verbal autopsies, 13 (19%) of 70
in the intervention tabias were consistent with possible malaria
compared with 68 (57%) of 120 in the control tabias.
MIRA
Makwanpur
distict, Nepal
Manandhar”
2004
cRCT
Women's groups through
community action cycle
Newborn mortality
rate
24 clusters of mean 7000
people per cluster
Newborn mortality rate was 26-2 per 1000 (76 deaths per
2899 livebirths) in intervention clusters compared with 36 9 per
1000 (119 deaths per 3226 livebirths) in controls (adjusted odds
ratio 0-70 [95% Cl 0-53-0-94]). Stillbirth rates were similar in
both groups.Maternal mortality ratio was 69 per
100000 (two deaths per 2899 livebirths) in intervention clusters
compared with 341 per 100 000 (11 deaths per 3226 livebirths)
in control clusters (0 22 [O O5-O-9O]).
cRCT
Health promotion intervention
includes a health education
campaign, participatory
discussion groups, training of
village health
workers and midwives, and
improved coordination of
antenatal services. The
intervention group will also have
subsidised access to pregnancyrelated health care services at
non-public health centres
Newborn mortality
rate
464 villages
Ongoing
1998
o
I
NJ
kn
Q
“C
5
?
o
I
Comment
Authors
Warmi Project,
Bolivia
I Ethiopia"
CHAMPION
Boone
trial,
2007
Mahabubnagar
district, Andhra
Pradesh, India
Primary outcomes
Total reduction in
newborn mortality rate
during intervention
(1996-2003) was
ascribed to sepsis
management (36%);
supportive care of low
birthweight newborn
babies (34%); asphyxia
management (19%);
primary prevention
(7%), and management
of other illnesses or
unexplained (4%). The
contribution of
community mobilisation
in the intervention areas
is difficult to estimate
Final data analysis due in
2010
UO
O
rs'
in
(Continues on the next page)
o
un
CTA
<T>
Authors
Design
Type of interventions
Primary outcomes
Sample
Results
Comment
cRCT
Female community workers doing
home care visits (1 per 4000) and
community mobilisers running
health education groups every
4 months vs control comparison
areas
Newborn mortality
rate
24 clusters of 20 000 per
cluster in. 14769 livebirths
in the home-care groups,
16325 in the community
care groups, and 15350 in
the comparison groups
Newborn mortality rate was reduced in the home-care arm by
34% (adjusted relative risk 0-66;
arm involved fairly
95% Cl O-47-O-93) during the last 6 months of the trial vs that in
the comparison arm. No mortality reduction was noted in the
community-care arm (0.95; 0-69-1 31)-
The community care
Before and
after analysis
Lady Health Workers and
traditional birth attendants
providing health education and
maternal and child health
services
Still birth rate;
newborn mortality
rate; health service
use
315 villages with a total
population of 138 600
Still birth rate reduced from 66 to 43 per 1000 and newborn
mortality rate 57 to 41 per 10000 before and after in
intervention sites. Skilled birth attendance from 18 to 30%
Preliminary results to be
confirmed in an
adequately powered trial
cRCT
Women's groups through
community action cycle
Newborn mortality
rate; maternal and
newborn
morbidity; care
practices and
health care use
48 urban slum clusters of
1000-1500 households
each 24 clusters
randomly allocated to
receive the community
intervention. 24 clusters
will act as control groups,
but will benefit from
health service quality
improvement
Ongoing
Final data analysis due in
2010
(Continued from previous page)
Prcjahnmo
Project, Shylet
district,
Bangladesh
Baqui41
2008
Mala, Pakistan
Bhutta41
2008
SNEHA Project, More
Mumbai,
2008
India
I
i
I
MaiMwana
Project,
Mchinji
district,
Malawi
Unpublished
cRCT
Womens groups through
community action cycle and
volunteer infant care and feeding
counsellors
Newborn mortality
rate; still birth rate;
perinatal mortality
rate; maternal
mortality ratio
48 clusters (3000 people
per cluster)
Ongoing
Final data analysis of
20 000 births due in
November 2008
Perinatal Care
Project DAB,
Bangladesh
Unpublished
cRCT
Womens groups through
community action cycle and
management of birth asphyxia in
the home using traditional birth
attendants trained in bag and
mask resuscitation
Newborn mortality
rate; still birth rate;
perinatal mortality
rate; maternal
mortality ratio
18 union clusters in
3 districts (25000 people
per cluster)
Ongoing
Final data analysis of
32 000 births due in
November 2008
Ekjut,
Jharkhand and
Orissa, India
Unpublished
cRCT
Women's groups through
community action cycle
Newborn mortality
rate; still birth rate;
perinatal mortality
rate; maternal
mortality ratio
36 clusters across 3
districts. 7000 people per
cluster
Ongoing
Final data analysis of
20000 births due in
December 2008
MIRA,
Dhanusha
district, Nepal
Unpublished
cRCT
Womens groups through
community action cycle and
newborn sepsis management in
the home using female
community health volunteers
Newborn mortality
rate; still birth rate;
perinatal mortality
rate; maternal
mortality ratio;
nutrition indicators
60 clusters, 7000 people
per cluster
Ongoing
Final data analysis due in
2009
Ifakara,
Tanzania
Unpublished
cRCT
Community-based package
focussed around interpersonal
communication through home
visits in pregnancy and the early
neonatal period by a village
based "agent of change" linked
to existing village health
volunteers.
Newborn mortality
rate
24 divisions in 5 districts
Ongoing
Data analysis due in 2011
r
3
I
UJ
KJ
tn
o
“C
I
£
I
S’
infrequent contact with
community groups
(once every 4 months)
and provided health
education rather than
community
mobilisation
Table: Published and currently ongoing trials of interventions involving community mobilisation for maternal and child health
LH
Series
component, are currently in progress (table). The
interpretation of the findings of these trials must be
considered carefully to guide policy makers. For example,
the community mobilisation component of Projahnmo
was less intensive than in Makwanpur. Thus, an
important question to ask of these trials might be, what
is the necessary level of intensity and coverage of
community mobilisation and home-care interventions,
to produce the most cost-effective effect? Other important
questions include which are the most effective models of
these interventions, can they be scaled up in the poorest
communities, and what are the institutional and financial
barriers to scale-up?
Community health workers
The use of so-called barefoot doctors in China inspired
primary health care. This model involved local community
residents—community health
workers—liberating
communities by providing first line health care and
facilitating others to embrace changes brought about by
the new government.45 This model was adopted by many
governments and non-governmental organisations after
the Alma-Ata Declaration and in many cases became the
definition of primary health care. However, by the 1990s
many government programmes for community health
workers had vanished because of problems in integrating
them into national programmes.4'’ People also questioned
whether community health workers actually empowered
or oppressed as a result of the existing, socioeconomic
political structures, bureaucracies, and lack of support
from health professionals.47
Recently, community health workers have generated
renewed interest, in part because they are seen as a cheap
way of scaling up primary health care, and also because
HIV/AIDS programmes demand more care at community
level. The pandemic has claimed the lives of many health
workers especially in Africa. The current interest lies
mainly in community health workers as care providers
but this can be problematic as large-scale government
training programmes often lack standards, supervision,
and resources. Furthermore, the evidence suggests that
community health workers are most effective when they
also facilitate change at the community level4ti■4,, and
participatory approaches promoted by the online journal
Participatory Learning and Action have provided
structures and frameworks that support this role.5
Overall, community health workers are most successful
when they have the respect and support of governments,
public service workers, and the communities they serve.
The whole community meets
a number of times to:
• evaluate progress,
achievements and
challenges in relation to the
group, the priority
problems being addressed
and the solutions
• plan for the future Of the
group, the priority problems
and the solutions
The whole community
meets a number of times to:
• implement the solutions
• monitor the progress of
Groups meet a number of
Identifying
Evaluating
together
and
prioritising
problems
4
1
3
2
together
Implementing
Planning
solutions
together
solutions
together
the solutions
times to:
• identify health problems
affecting mothers and
childten in the community
• identify the mot causes of
these pfoblems
• select the problems they
consider to be most
important and need to be
addressed
Groups meet a number of
times to:|
• identify feasible solutions
to the priority problems
that mike the best use of
locally available resources
• plan the solutions with
the help of the whole
community
Figure 2: Women's groups community mobilisation action cycle
The Warmi project in Bolivia developed a model for community mobilisation using this community action cycle
Women's groups discuss and prioritise their problems, develop strategies to solve them, and, after engaging with
other community members, implement and evaluate these solutions. The completed Makwanpur (Nepal) trial and
ongoing trials in Mumbai (India), Jharkhand and Orissa (India), Mchinji (Malawi), Dhanusha (Nepal), and
Bangladesh are assessing the effect of different women’s group models, developed from this model, on mother
and child health (table).
steep power hierarchies) can impair control or capacity
and the respectful relationships that enable good
maternal and child health.5"51 Community mobilisation
initiatives reported to improve the socioenvironmental
causes of ill health have addressed a range of concerns
including alcohol related violence, breast cancer
treatment, and safety in public environments.5‘)(>l"‘' The
impetus to address these causes of ill-health began when
there was sufficient support to form a community of
interest. This community started a process of capacity
building—community empowerment—toward gaining
more control over the decisions for resource allocation
such as the award of a grant or to decision making such
as the development of policy or legislation (panel 2). The
key to the success of community empowerment was the
moment when the community engaged with the problem
posing, problem-solving process and recognised that
they could collectively change their circumstances.
However, effect can vary greatly depending on decisions
about the goal, who constitutes the community, who is
facilitating and supporting the process, the social and
political context, the duration of external or donor
support, and the cost-effectiveness of the programme/’2
Different forms of community mobilisation might simply
Does community mobilisation empower people to
address socioenvironmental causes of ill-health?
Health, particularly in marginalised groups, is indirectly
but powerfully affected by the social environment in
which personal behaviours are embedded. Risk factors
(such as isolation, lack of social support, low self-esteem)
and risk conditions (such as poverty, discrimination,
www.thelancet.com Vol 372 September 13, 2008
mobilise communities to initiate localised actions based
on their immediate needs rather than broader social and
political actions.
What is not known is to what extent peoples’ involvement
can actually increase resources to support health care,
whether participation can create a genuine social learning
partnership between people and professionals, whether
96/
I
Series
Panel 2: Building community empowerment
Community empowerment is a synergistic interplay between
individual empowerment,52 organisational empowerment,53
and broader social and political actions.54 Empowered
community-based organisations are at the heart of community
empowerment, since they link empowered individuals and
effective political action.55This interplay can be conceptualised
as a continuum of five progressively more organised and
broadly based forms of social and collective action (main
bullets).5657These five forms can be further subdivided into
domains (sub-bullets), which represent the means through
which individuals and groups can organise themselves to
harness the interpersonal elements of empowerment and
address the broader determinants of their health.58
•
Personal action
• Community participation
•
Small mutual groups
•
•
•
•
Problem assessment
•
Local leadership
Community organisations
•
Local leadership
•
•
Organisational structures
Resource mobilisation
Partnerships
•
Organisational structures
•
Resource mobilisation
•
Links to others
•
Asking why
Social and political action
•
•
Links to others
Asking why
•
•
Role of outside agents
Programme management
community mobilisation can really change a commitment
to social justice and democracy, and whether community
mobilisation can actually accelerate progress at scale
toward achievement of MDGs 4 and 5 in high-mortality,
resource-poor settings.
What are the mechanisms through which community
mobilisation brings about improved health outcomes?
Some observers feel that community mobilisation works
simply by bringing about changes in behavioural risk
factors such as home care practices and decisions about
care seeking. Although undoubtedly one important
mechanism through which community mobilisation
works, studies of health education suggest that simply
providing key messages to improve maternal and newborn
care cannot possibly account for all the effect these
approaches have on morbidity and mortality.6164 A large
proportion of this effect is thought to be due to community
mobilisation bringing about changes in socioenvironmental risk factors by developing the capacities of
communities, the choices they make, and their ultimate
968
empowerment. This mechanism is enshrined in the
Ottawa Charter (1986) and the Jakarta Declaration (1997),
which equated health promotion with goals of
empowerment and a more long term and fundamental
shift in village, family, and gender power relations.65
Women’s groups in Malawi and Nepal are increasing
the important capacities within communities, such as
the ability to identify maternal and neonatal health
problems and their root causes; the ability to mobilise
resources necessary for improving the health of mothers
and newborn infants; the internal and external social
networks they can draw on when needed; and the
development of strong local leaders who have the
motivation and drive to improve maternal and neonatal
health in the community.66-67 The women’s groups are
also drawing on these social capacities to make
fundamental choices to improve their health, such as
about the equitable sharing of resources needed for better
maternal and neonatal health; about planning feasible
strategies to address maternal and neonatal health
problems; about planning, implementation, evaluation,
finances and reporting of programmes; and about which
people and organisations to approach to address
problems. Detailed longitudinal exploration of these
processes is crucial to provide answers to policy makers
about how community mobilisation works, to inform
programme design, and to build the case for government
investment.
Is community mobilisation less important than facility
based medical interventions?
Many safer motherhood analysts, such as policy makers
and academics, would consider community mobilisation
a peripheral component of a package to reduce maternal
mortality, which is far more dependent on specific facility
based interventions than is child survival.27 However, the
evidence supports a more central role for community
mobilisation. Firstly, numerous interventions such as
family planning, nutritional support for women, and the
treatment of haemorrhage, sepsis, and unsafe abortion
are all potentially amenable to interventions in the
community/*8 Secondly, the so-called first delay
(recognising a maternal problem in the home and
deciding to seek care) is a key problem for safer
motherhood programmes and solving it requires the
participation of communities (panel 3). Thirdly, poverty
and disadvantage are the underlying causes of many
neonatal and maternal deaths; 99% of maternal and
neonatal deaths occur in low-income and middle-income
families and in poor countries, and maternal mortality is
often more than twice as high in the poorest compared
with the richest economic quintile household.''”" The
link between social disadvantage and mortality is subtle
and indirect but maternal and newborn survival and good
health are ultimately the result of a society that values
women and children irrespective of their race, social,
economic, and political status and provides unimpeded
www.thelancet.com Vol 372 September 13, 2008
Series
access to information and health services from the
household to the hospital. Community mobilisation, in
addressing inequality rather than only improving health
services, is thus a priority strategy for improving survival
of mothers and newborn infants.6970
Although maternal survival requires improvements in
comprehensive and basic obstetric care at hospitals and
health centres, community mobilisation has an important
role in improving care practices, increasing the use of
safer motherhood services, promoting timely referral
when problems arise, and reducing social disadvantage.
Some of the ongoing trials cited in the table could have
the statistical power to add to this debate by exploring the
extent to which community approaches reduce maternal
mortality directly compared with indirectly by promoting
deliveries in hospitals.
■
Panel 3: Case study of how women's groups are addressing first-delay in! maternal
and child care: Jharkhand, India
Sini Koda comes from Tipusai, a remote hamlet of Baraibir village in West Singhbhum
district of Jharkhand state. It is 25 kms from a private facility where emergency obstetric
care is available and receives infrequent visits from Auxiliary Nurse Midwives. She, her
husband, mother-in-law, and other members of the family regularly attend women's
group meetings, facilitated by Rani Kayam who was trained and is employed by Ekjut
Project, a local non-governmental organisation. The group meets monthly ijnd engages in
participatory learning and action activities focusing on maternal and child health. During
one of these meetings they engaged in a "woman in labour-emergency drill" role-play
session. In this session they learnt how to mobilise quickly at the time of labour and avoid
delays. When it came time to deliver, Sini's in-laws tried to perform traditional rituals that
would delay her from getting to the health facility. However, her husband and other
women's group members used what they had learnt to collect 5000 rupees from other
community members for transport and hospital costs. As a result, Sini was able to get to
the facility with the minimum of delay where she delivered normally and successfully.
How can community mobilisation be taken to scale?
Scale-up of health interventions might involve increasing
coverage by geographical expansion, adding technical
interventions to an existing programme, advocacy to
change policies, and strengthening capacity with more
resources, new alliances, and technical skills.7’ But how
can governments, even in partnership with civil society
organisations, achieve scale-up ofcommunity mobilisation
interventions in these ways? Several approaches have
been used including: government directed and
implemented programmes;72 partnerships between
government and non-governmental organisations;7’7' '75 socalled living universities and centres of learning;7276
dissemination of methods and results through manuals,
training packages, internet, radio, video, TV, and
university classes; and organic spread from community
to community through word-of-mouth or direct
observation. These approaches have succeeded in massive
scale-up of community mobilisation interventions in
countries such as Bangladesh, China, Cuba, Sri Lanka,
and Tanzania. Thus, these interventions, due to their
dependence only on facilitation and community resources,
seem to be no more difficult to scale-up than others such
as immunisation programmes, which depend on cold
chains, drugs, technology, and a large network of paid
health workers. However, in the poorest countries the
capacity and commitment for scale-up remains weak and
extensive coverage alone is insufficient to ensure that the
most vulnerable populations benefit in the long-term.
Case studies, trials, and large-scale programmes have
shown that, when given the opportunity, communities
can develop effective strategies to address their needs
and reduce mortality and morbidity. These strategies are
often highly innovative, practical, and culturally
acceptable. What is scaled-up is not solutions but a
process to support communities to develop their own
solutions. As a result, programmes must be flexible
enough to respond to variations between, and within,
communities and must allow adequate time for this
process of capacity building. Also, a favourable
www.thelancet.com Vol 372 September 13, 2008
environment for scaling up can be created if national
policies are in place which support community
mobilisation. Programmes are more successful if they
communicate from the same belief system. This success
can be achieved by seeking to understand and take into
account the social norms and local cultural context
around health, community participation, gender roles,
use of health services, and household decision making.
Importantly, programmes should not cut out or limit
essential steps such as problem identification,
prioritisation, and strategy formulation by communities.
Irrespective of whether the facilitating agent is a
representative of a non-governmental organisation,
member of a community based organisation, government
fieldworker, or volunteer they must have: credibility in
the communitifes; language skills and cultural sensitivity;
knowledge of community structures and protocols;
interest in being a facilitator and in maternal and
newborn health; affiliation with and support from an
organisation; good interpersonal communication skills;
and availability of time to do the work.
The main programme cost is building human and
community capacity, which needs adequate investment.
This process means prioritising investment in ongoing
training, facilitation, and capacity strengthening and the
use of cost-effective methods such as cascade-training
structures. Costs within the programme can in part be
covered by contributions from the community but this
must be done carefully while respecting roles and
responsibilities and keeping in mind programme
principles of community ownership and sustainability.
Partnerships of government, non-governmental
organisations, private sector, and community-based
organisations arc essential, but can face differences in
organisational cultures and values, competition for
resources, and varying levels of capacity. Successful
programmes define roles and responsibilities clearly,
allocate resources fairly, and establish operational
969
I
Series
guidelines, communication systems, parameters for
implementation, and mechanisms for dealing with
problems or disputes. Finally, new technologies such as
community radio, mobile phones, internet, and digital
and video cameras have rapidly become more accessible
and could present new opportunities for communication,
gathering information, organising, coordinating, and
increasing participation.
Conclusion
There is evidence that community mobilisation is an
effective method for promoting participation and
empowering communities among a wide range of other
non-health benefits. The experience of pilot programmes
before the Alma-Ata Declaration, and subsequent trial
evidence, also suggests that community mobilisation
can bring about cost-effective and substantial reductions
in mortality and improvements in the health of newborn
infants, children, and mothers. Nonetheless community
mobilisation is not a feature of most large-scale primary
health care programmes, because it is characterised by
several fundamental controversies. What form should it
take to be most effective? Does it effectively address the
socioenvironmental risk factors that underpin health
problems and mortality? How does it work? What part
does it have to play in interventions for maternal sur
vival? How can it be scaled-up effectively? Continuing
studies and future research, particularly focusing on
process, are needed to address these controversies and
fully unlock the potential that community mobilisation
approaches have to improve health and reduce
mortality.
References
1
Declaration of Alma-Ata: international conference on primary
2
3
4
5
6
7
8
9
10
970
health care. Alma-Ata. USSR. Sept 6-12.1978. http://www.who.
int/hpr/NPH/docs/cleclaration_almaata.pclf (accessed July 16.
2008).
Walsh JA, Warren KS. Selective primary health care: an interim
strategy for disease control in developing countries. N Engl J Med
1979; 301: 967-74.
Bryce J, Victora CG. Habicht JP, on behalf of the MCE-1MC1
Technical Advisors. Programmatic pathways to child survival:
results of a multi-country evaluation of Integrated Management of
Childhood Illness. Health Policy Plan 2005; 20 (suppl 1); 15-17.
Howard-Grabman L, Snetro G. How to mobilise communities for
health and social change. Baltimore, MD: Health Communication
Partnership/USAID, 2003.
Rifkin S. Pridmore P. Partners in Planning. London: MacMillan.
2001.
Bryant JH, Richmond JE. The story of Alma-Ata—past, present,
future—in a changing world. In: Encyclopedia for Public Health.
Elsevier Press (in press).
Litsios S. The long and difficult road to Alma-Ata: a personal
reflection. Int J Health Services 2002; 32: 709-32.
Litsios S. The Christian Medical Commission and the development
of the World Health Organization's primary health care approach.
Am J Public Health 2004; 94: 1884-93.
Arole M, Arole R. A comprehensive rural health project in Jamkhed
(India). In: Newell KW. ed. Health by the People. Geneva: World
Health Organization. 1975: 70-90.
Arole M. Arole R. Jamkhed, India—the evolution of a world
training center. In: Taylor-Ide D. Taylor CE. Just and lasting change:
when communities own their futures. Baltimore: Johns Hopkins
University Press. 2002: 150-60.
Were M. Kakamega, Kenya: a promising start derailed. In: TaylorIde D, Taylor CE. Just and lasting change: when communities own
their futures. Baltimore: Johns Hopkins University Press, 2002:
168-77.
12 Chowdhury AMR, Cash R. A simple solution: teaching millions to
treat diarrhoea at home. Dhaka, Bangladesh: University Press
Limited, 1996.
13 Levine R, What Works Working Group. Case 7: preventing diarrheal
deaths in Egypt. In: Case studies in global health: millions saved.
Sudbury, MA: Jones and Bartlett Publishers. 2007.
14 Werner D. Sanders D. Questioning the solution: the politics of
primary health care and child survival. Health Wrights. 1997.
15 Sazawal S. Black RE. Ed'ect of pneumonia case management on
mortality in neonates, infants, and preschool children: a meta
analysis of community-based trials. Lancet Infect Dis 2003; 3: 547-56.
16 Bang AT, Bang RA, Tale O, et al. Reduction in pneumonia mortality
and total childhood mortality by means of community-based
intervention trial in Cadchiroli. India. Lancet 1990; 336: 201-06.
17 Fawzi WW, Chalmers TC, Herrera MG. Mosteller F. Vitamin A
supplementation and child mortality: a meta-analysis. JAMA 1993;
269: 898-903.
18 Gottlieb J. Case 4: reducing child mortality with vitamin A in Nepal.
In: Levine R, What Works Working Group. Case studies in global
health: millions saved. Sudbury, MA: Jones and Bartlett Publishers.
2007.
19 Binka FN, Kubaje A, Adjuik M. et al. Impact of permethrin
impregnated bednets on child mortality in Kassena-Nankana
district. Ghana: a randomised controlled trial. Trap Med Int Health
1996; 1:147-54.
20 Lengeler C, Insecticide-treated bed nets and curtains for preventing
malaria. Cochrane Database Syst Rev 2004; 2: CDOOO363.
21 UNICEF. State of the World’s Children 2008. New York: United
Nations Children's Fund, 2007: 28-29.
22 Maine D. Detours and shortcuts on the road to maternal mortality
reduction. Lancet 2007; 370: 1380-82.
23 Freedman LP, Graham WJ, Brazier E. et al. Practical lessons from
global safe motherhood initiatives: time for a new focus on
implementation. Lancet 2007; 370: 1383-91.
24 Husein K, Adeyi O. Bryant J. Cara NB. Developing a primary
health care management information system that supports the
pursuit of equity, effectiveness and affordability. Soc Sci Med 1993:
36: 585-96.
25 Hill K. Thomas K, AbouZahr C, et al. Maternal Mortality Working
Group. Estimates of maternal mortality worldwide between
1990 and 2005: an assessment of available data. Lancet 2007;
11
370:1311-19.
26
27
28
29
30
31
32
33
Bryce J, Daelmans B. Dwivedi A, et al. Countdown to 2015 Core
Group. Countdown to 2015 for maternal, newborn, and child
survival: the 2008 report on tracking coverage of interventions.
Lancet 2008; 371: 1247-58.
Bhandari N, Mazumder S, Bahl R, Martines J, Black RE, Bhan MK.
Infant Feeding Study Group. An educational intervention to
promote appropriate complementary feeding practices and physical
growth in infants and young children in rural Haryana, India.
J Nutr 2004; 134: 2342-48.
Coutinho SB, de Lira PI. de Carvalho Lima M. Ashworth A.
Comparison of the effect of two systems for the promotion of
exclusive breastfeeding. Lancet 2005; 366: 1094—100.
Morrow A, Matin L. Community-based strategies for breastfeeding
promotion and support in developing countries. Geneva: World
Health Organization. 2003.
Britton C, McCormick FM, Renfrew MD, Wade A, King SE. Support
for breastfeeding mothers. Cochrane Database Syst Reviews 2007;
1: CD001141.
Dewey KG. Adu-Afarwuah S. Systematic review of the efficacy and
effectiveness of complementary feeding interventions in developing
countries. Matern Child Nutr 2008; 4 (suppl 1): 24-85.
Ashworth A. Efficacy and effectiveness of community-based
treatment of severe malnutrition. Food Nutr Bull 2006;
27 (suppl 3): S24-48.
Wafula F.M. Kinyanjui MM. Nyabola L. Tenambergen ED. Effect of
improved stoves on prevalence of acute respiration infection and
conjunctivitis among children and women in a rural community in
Kenya. East Aft Med J 2000; 77: 37-41.
www.thelancet.com Vol 372 September 13, 2008
Series
34
B5
36
37
Ali M, Emch M.Tofail F, Baqui AH. Implications of health care
provision on acute lower respiratory infection mortality in
Bangladeshi children. Soc Sci Med 2001; 52: ICl-Tk
Hadi A. Management of acute respiratory infections by community
health volunteers: experience of Bangladesh Rural Advancement
Committee (BRAC). Bull World Health Organ 2003; 81: 183-89.
Kidane G, Morrow RH. Teaching mothers to provide home
treatment of malaria in Tigray. Ethiopia: a randomised trial. Lancet
2000;356: 550-55.
Bang AT. Reddy HM, Deshmukh MD, Baitule SB, Bang RA.
Effect of home-based neonatal care and management of sepsis on
neonatal mortality: field trial in rural India. Lancet 1999;
354: 1955-61.
Bang AT, Reddy HM, Deshmukh MD, Baitule SB, Bang RA.
Neonatal and infant mortality in the ten years (1993 to 2003) of the
Gadchiroli field trial: effect of home-based neonatal care. J Perinatol
2005; 25 (Suppl 1): S92-107.
39 Manandhar DS. Osrin D. Shrestha B. et al. Effect of a participatory
intervention with women's groups on birth outcomes in Nepal:
cluster-randomised controlled trial. Lancet 2004; 364: 970-79.
40 Howard-Grabman L, Seone G, Davenport C. MotherCare, Save the
Children. The Warmi Project: a participatory approach to improve
maternal and neonatal health. An implementor's manual. Westport:
John Snow International. Mothercare Project. Save the Children.
2002.
41 O'Rourke K. Howard-Grabman L. SeoaneG. Impact of community
organization of women on perinatal outcomes in rural Bolivia.
Rev Panam Salud Publica 1998; 3: 9-14.
42 Bhutta ZA, Memon ZA. Soofi S. Salat MS. Cousens S. Martines J.
Implementing community-based perinatal care: results from a pilot
study in rural Pakistan. Bull World Health Organ 2008; 86: 452-59.
43 Baqui AH. El-Arifeen S. Darmstadt GL, et al. Projahnmo Study
Group. Effect of community-based newborn-care intervention
package implemented through two service-delivery strategies in
Sylhet district, Bangladesh: a cluster-randomised controlled trial.
Lancet 2008; 371:1936-44.
44 Borghi J, Thapa B, Osrin D, et al. Economic assessment of a
women's group intervention to improve birth outcomes in rural
Nepal. Lancet 2005; 366: 1882-84.
45 Sidel VW, Sidel R. The health care delivery system of the People's
Republic of China. In: Newell KW. ed. Health by the people.
Geneva: World Health Organization. 1975: 1-12.
46 Walt G, ed. Community health workers in national programmes: just
another pair of hands? Milton Keynes: Open University Press, 1990.
47 Werner D. The village health worker: lackey or liberator? http://
www.healthwrights.org/articles/lackey_or_liberator.htm (accessed
July 16, 2007)
48 Bhattacharyya K. Winch P. LeBan K. Tien M. Community health
worker incentives and disincentives: how they affect motivation,
retention, and sustainability. Arlington. Virginia: Basic Support for
Institutionalizamg Child Survival Project (BASICS II) for the
United States Agency for International Development. 2001.
49 Gryboski N. Yinger N. Dios R. Worley H. Fikree F. Working with
the community for improved health. Health Bulletin Number 3.
Population Reference Bureau: Washington. DC. 2006.
50
Labonte R. Health promotion and empowerment: practice
frameworks. Toronto: University of Toronto, 1993.
51
Labonte R. A community development approach to health
promotion: a background paper on practice tensions, strategic
models and accountability requirements for health authority work
on the broad determinants of health. Edinburgh: Health Education
Board of Scotland. 1998.
52 Rissel C. Empowerment: the holy grail of health promotion?
Health Promotion Int 1994; 9: 39-47
53 Israel B. Checkoway B, Schultz A, Zimmerman M. Health
education and community empowerment: conceptualising and
measuring perceptions of individual, organisational and
community control. Health Educ Q1994; 21: 149-70.
54
55
56
57
58
59
38
www.thelancet.com Vol 372 September 13, 2008
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
Wallerstein N. Powerlessness, empowerment and health:
implications for health promotion programs. Am J Health Promotion
1992; 6: 197-205.
Laverack G. Health promotion practice: power and empowerment.
London: Sage, 2004.
Jackson T, Mitchell S. Wright M. The community development
continuum. Community Health Stud 1989; 8: 66-73.
Labonte R. Empowerment: notes on professional and community
dimensions. Can Rev Soc Policy 1990; 26: 64-75.
Laverack G. Addressing the contradictions between discourse and
practice in health promotion. Doctoral dissertation. Melbourne:
Deakin University, 1999.
Bang A, Bang R. Gadchiroli, India: addiction as a barrier to
development. In: Taylor-Ide D, Taylor CE. Just and lasting change:
when communities own their futures. Baltimore: Johns Hopkins
University Press. 2002:161-67.
Klawiter M. Breast cancer in two regimes: the impact of social
movements on illness experience. Soc Health Illness 2004;
26: 845-74.
Gee D. Park Scheme seeks more 'friends'. The Christchurch Press.
http://www.bush.prg.nz/library/934.html (accessed April 3. 2008).
Howard-Grabman L (2007). Demystifying community mobilisation:
An effective strategy to improve maternal and newborn health.
Baltimore. MD; JHPIEGO/USAID.
Renkert S. Nutbeam D. Opportunities to improve maternal health
literacy through antenatal education: an exploratory study.
Health Promol Int 2001; 16: 381-88.
Bolam A, Manandhar DS. Shrestha P. Ellis M. Costello AM. The
effects of postnatal health education for mothers on infant care and
family planning practices in Nepal: a randomised controlled trial.
BMJ 1998; 316: SpS-ll.
The Jakarta Declaration on health promotion in the 21st century.
Health Millions 1998; 24: 29-30.
Rosato M, Mwansambo CW, Kazembe PN, et al. Women's groups'
perceptions of maternal health issues in rural Malawi. Lancet 2006;
368:1180-88.
Morrison J, Tamang S, Mesko N, et al. Women’s health groups to
improve perinatal care in rural Nepal. BMC Pregnancy Childbirth
2005; 5: 6.
Costello A, Azad K, Barnett S. An alternative strategy to reduce
maternal mortality. Lancet 2006; 368:1477-79.
Lawn JE, Cousens S, Zupan J. 4 million neonatal deaths: when?
Where? Why? Lancet 2005; 365: 891-900.
Ronsmans C. Graham WJ, Lancet Maternal Survival Series steering
group. Maternal mortality: who, when, where, and why. Lancet 2006;
368: 1189-200.
Uvin P. Miller D. Paths to scaling up: Alternative strategies for local
nongovernmental organizations. Human Organization 1996:
55: 344-54.
Taylor-Ide D. Taylor CE. |ust and lasting change: when communities
own their futures. Baltimore: Johns Hopkins University Press.
2002.
Gonzales F, Artejaga E. Howard-Grabman L. Scaling up the Warmi
Project: lessons learned mobilizing Bolivian communities around
reproductive health. http://www.coregroup.org/resources/Saveboli.
pdf (accessed Aug 28, 2008).
Coe AB. Health, rights and realities: an analysis of the ReproSalud
Project in Peru, working paper. Takoma Park, MD: Center for
Health and Gender Equity, 2001.
The CORE Group. “Scale" and "Scaling Up": A CORE Group
background paper on "scaling up" maternal, newborn and child
health services, http://www.coregroup.org/resources/meetings/
april05/Scaiing_Up_Background_Paper_7-13.pdf (accessed
Aug 28, 2008).
Marchione T, ed. Scaling up, scaling down; overcoming
malnutrition in developing countries. Amsterdam: Gordon and
Breach,1999.
971
Right to Health
Health systems and the right to health: an assessment of
194 countries
iniilia BuckniG-i IViu! Hi
fjjcit khoslu, Camila laramillo -Stmuss. Belachcw Mekuria Fikre. Caroline Rumble. David Pevalin, [)avid Acurio Paez.,
isuficl'k-. i/umska Fuieo. Mitra Motlagh Dana Farcasanu, CristianVIudescu
60 years ago, the Universal Declaration of Human Rights laid the foundations for the right to the highest attainable
standard of health. This right is central to the creation of equitable health systems. We identify some of the right-tohealth features of health systems, such as a comprehensive national health plan, and propose 72 indicators that reflect
some of these features. We collect globally processed data on these indicators for 194 countries and national data for
Ecuador, Mozambique, Peru, Romania, and Sweden. Globally processed data were not available for 18 indicators for
any country, suggesting that organisations that obtain such data give insufficient attention to the right-to-health
features of health systems. Where they are available, the indicators show where health systems need to be improved to
better realise the right to health. We provide recommendations for governments, international bodies, civil-society
organisations, and other institutions and suggest that these indicators and data, although not perfect, provide a basis
for the monitoring of health systems and the progressive realisation of the right to health. Right-to-health features are
not just good management, justice, or humanitarianism, they are obligations under human-rights law.
l.ctrwet 2008: 372:2647 -85
Published ■ ■■ bi .•
December 10, 2008
□01:10.1016/501406736(O8)61781-X
See Editorial page 2001
See Comment pages 2005.
2007. 2008. and 2010
See I’eis.pec'i-.'Cs page 2015
Nordic School of Public Health,
Gothenburg, Sweden
(G Backman MSc); Human
Right*. Centre (P Hunt MJur,
Introduction
December, 2.008, marks the 60th anniversary of the
Universal Declaration of Human Rights.1 The declaration
provides the foundation for the international code of
human rights.-’ This code gives an internationally agreed
set of standards to guide and assess the conduct of
governments across a wide range of sectors and has a
direct, close bearing on medicine, public health, and the
strengthening of health systems.1
The international code of human rights consists of
legally binding international components. Among the
most important of these components for health systems
are the International Covenant on Economic, Social, and
Cultural Rights (ICESCR)'5 and the Convention on the
Rights of the Child (CRC).6 Both these human-rights
treaties are legally binding for those countries that have
ratified them. Most states have ratified the ICESCR, and all
but two (Somalia and the USA) have ratified the CRC. Ihe
right of everyone to enjoy the highest attainable standard
of physical and mental health - -sometimes known as the
right to the highest attainable standard of health or the
right to health—is an integral part of both of these
international treaties. All countries have ratified one or
more binding treaty that includes the right to health, such
as die International Convention on the Elimination of All
forms of Racial Discrimination.7 Also, many countries
include this right in their national constitutions.” Ihe
Constitution of WHO.'1 the Declaration of Alma-Ata,10 the
Ottawa Charier for 1 lealth Promotion," the Bangkok
Charter for Health Promotion in a Globalized World,’-' and
other important documents agreed by the health
community also recognise this fundamental human right.
In recent years, national and international policy makers,
courts, non-governmental organisations, and other
stakeholders have adopted and applied features of the right
to the highest attainable standard of health. Uganda’s
review of its health policy expressly uses a right-to-health
analysis" as does WHO in, for example, its publication on
www.thelancet.com Vol 3/2 December 13, 2008
human rights, health, and poverty reduction.14 Courts, too,
are explicitly relying on the right to health in their decisions,
most recently in a landmark judgment of the Colombian
Constitutional Court.15-1'’ On the basis of a detailed under
standing of the right to health, this court effectively ordered
a phased restructuring of the country’s health system by
way of a participatory and transparent process based on
current epidemiological information?" Civil-society guides
to the right to health are increasing in number, and many
civil-society organisations use these in their work?1-24 Both
the UN General Assembly and Human Rights Council
have discussed numerous reports on the right to health,
covering a wide range of issues, such as neglected
diseases,25 sexual and reproductive health,25 maternal
mortality,2'1 mental disability,27 the Millennium Development
Goals (MDGs),2” medicines,2(1 and water and sanitation.2’’
Recognition that a strong health system is an essential
element of a healthy and equitable society is growing.
However, according to a recent WHO publication, health
systems in many countries are failing and collapsing.’"
Too many health systems are inequitable, regressive, and
unsafe.'" WHO also confirms that sustainable
development, including achievement of the MDGs,
depends on effective health systems."’
As with a fair court system, an effective health system
is a core social institution and, for this reason, crucially,
Although
both systems are protected by human rights.1112 Although
many human rights are important to a well-functioning
court system, the key one is the right to a fair trial."
Through human-rights treaties, national laws and
policies, judicial decisions, and so on, the right to a fair
trial has helped to identify the key features of a fair court
system, such as an independent judiciary and trials
without undue delay. The right to a fair trial has not only
identified unfair judicial processes but also led to
welcome reforms in many countries.
By analogy, the right to the highest attainable standard
of health can help to establish health systems that are
R Khosla LLM,
B Mekuria Fikre LLM,
C Rumble MBChB) and School of
Health and Human Sciences
(David Pevalin PhD) University
of Essex, Colchester, UK;
Office of the Mayor of Bogata,
Colombia
(C Jaramillo-Strouss LLM);
Foundation for Alterative
Social Development, Cuenca,
Ecuador
(David Acurio Paez MPH,
M Armijos Pineda MA); Health
Team National Coordinator,
CARE Peru, Lima, Peru
(Ariel Frisancho MHPPF);
National Health Council
Secretariat, Ministry of Health,
Lima, Peru (DTarco MD);
WHO-Western Pacific Regional
Office, Manila, Philippines
(M Motlagh LLM); Centre of
Health Policies and Services,
Bucharest, Romania
(D Farcasanu MPH); Department
of Public Health, Victor Babes
University of Medicine and
Pharmacy, Timisoara, Romania
(C Vladecu PhD)
Correspondence to
Gunilia Backman. Nordic School
of Public Health, Gothenburg
SE-402 42. Sweden
gunillabackrnan@yahoo.com
204/
Right to Health
I
reasonably equitable. However, to make this happen, the
right-to-health features of health systems need to be
identified. This process will take time, just as our
understanding of the right to a fair trial has developed
over many years. Once identified, the right-to-health
features will not provide a neat blueprint or formula for a
health system. There will be many grey areas, just as
there are in relation to the right to a fair trial and court
systems.'4 lhe right to a fair trial does not provide detailed
prescriptions, rather it insists upon key principles, such
as fairness, independence and impartiality, and several
important features that a court system must have if it is
to be fair. The right to health has a similar role.
Of all the important human rights that bear upon
health systems, the right to the highest attainable
standard of health is the cornerstone of both an effective
health system and the growing movement for health and
human rights.1'1
In this Report, we aim to assess the degree to which the
health systems of 194 countries include some of the
features that arise from the right to health. We introduce
the right to health and identify some (of" the
’ right-to-health
'J
’ ‘
features of health systems. These features are not justt a
matter of good management, justice, or humanitarianism—•theyare
f
a matter of human-rights law. We set
out our methods and their limitations and identify
72 indicators of right-to-health features of health systems.
We present some of the findings and results arising from
the data on the indicators, and discuss these data and
make recommendations for a range of stakeholders.
What is the right to health?
•>
protection.'6 Furthermore, the right to health requires that
there are indicators and benchmarks to monitorj progressive
realisation’6 and that individuals and communities have
opportunities for active and informed participation in
health decision making that affects them.'6 Under
international human rights law, developed countries have
some responsibilities towards the realisation of the right to
health in developing countries.16 Because the right to health
gives rise to legal entitlements and obligations, effective
mechanisms of monitoring and accountability are
needed.'6
Although the right to health adds power to campaigning
and advocacy, it is not just a slogan, it has a Loncise and
constructive contribution to make to health policy and
practice. Health workers can use the right to devise
equitable policies and programmes that strengthen health
systems and place important health issues higher up
national and international agendas.'7'"
Medicine, public health, and human rights have much
common ground. To one degree or another, each field
stresses the importance of the underlying determinants of
health and good-quality medical care, looks beyond the
health sector, struggles against discrimination and
disadvantage, demands respect for cultural diversity, and
attaches importance to public information and education.
The right to health cannot be realised without the
xo and the
interventions and insights of health workers;
classic, long-established objectives of public health and
medicine can benefit from the newer, dynamic discipline
of human rights. A few enlightened people understood
these relations when the WHO Constitution was drafted
in 1946” and when the Declaration of Alma-Ata was
adopted in 1978,"' affirming the right to the highest
attainable standard of health.
The right to the highest attainable standard of health
encompasses medical care, access tto safe
' drinking
’ ‘ '
water,
adequate sanitation, education, health-related informa
However, until recently, the right to health was <only
\
tion, and other underlying determinants of health;'6 it dimly understood and attracted limited support from
includes freedoms, such as the right to be free from civil society or any other sector. The understanding and
discrimination and involuntary medical treatment, and practice
t
of health and human rights has improved since
entitlements, such as the right to essential primary health the
conference,
— Alma-Ata
- ----- One vital jhart of this
care."’ Like other human rights, the right to health has
process has been a deepening understanding of the right
particular concern for disadvantaged people and to health. But it was not until 2.000 that an authoritative
populations, including those living in poverty. The right understanding of the right to health emerged when the
to health requires an effective, responsive, integrated
UN Committee on Economic, Social, and Cultural Rights,
health
system of good quality that is accessible to all.'71
'
working in close collaboration with WHO and many
International human-rights law recognises that the right others, drafted and adopted general comment 14."’
to the highest attainable standard of health cannot be
o__ neither complete, perfect, 11W1
Although
nor binding,
realised overnight; it is expressly subject to both progressive general comment 14 is compelling and groundbreaking.
realisation and resource availability.4 Put simply, progressive The comment shows a substantive understanding of the
realisation means that a country has to improve its
right to health that can be made operational and improved
human-rights performance steadily; if there is no progress, in the light of practical experience. The influence of
the government of that country has to provide a rational Alma-Ata
Alma-Ata on
on g_.
general comment 14 is explicit and clear.
and objective explanation. Because oftheir greater resource Although much1 more work is needed to grasp all the
availability, more is expected of high-income than of lowimplications of the right to the highest attainable standard
income countries. However, the right to health also of health, the general comment confirms that the right
imposes some obligations of immediate effect, such as cannot be dismissed as a rhetorical device. General
non-discrimination,4 and the requirement that a state at comment 14 provides a common right-to-health language
least prepares a national plan for health care and for talking about health issues and sets out a way of
>048
www.thelancet.com Vol 372 December 13, 2008
Right to Health
analysing the right to health, making it easier for policy
makers and practitioners to use.27 Panel 1 summarises
general comment 14, including the requirement that
health facilities and services be available, accessible, and
culturally acceptable.
The right-to-health analysis can be used to identify and
expose, for example, the lack of available mental-health
facilities properly serviced by trained staff.44 Health-related
facilities and services, including mental-health facilities
with properly trained staff, must be available in adequate
number throughout a country. Of course, the need is
subject to resource availability: more and better facilities
are required of Canada than of Chad. Few nations,
however, devote adequate funds to mental health.44 45 On a
routine basis, mental-health facilities are neglected,
workers untrained, and patients uncared for.44 Poor
mental health gives rise to other profound problems, not
least discrimination and stigmatisation, important to the
right to health.
The test of availability can also be applied to
harm-reduction initiatives.*’ Provision of injecting drug
users with comprehensive and integrated treatment,
counselling, and clean needles and syringes is good for
public health, reduces avoidable suffering, saves lives, and
is cost-effective.47 An appropriate harm-reduction initiative
is also a right-to-health initiative. However, most countries
do not provide harm-reduction services for people who
use drugs, and those that do, such as Sweden, provide a
limited and scattered service.41* The right to health requires
all countries to have an effective, national, comprehensive
harm-reduction policy and plan, delivering essential
services. A high-income country such as Sweden is
expected to provide more than the essential services.
Health-related facilities and services can be available
within a country but inaccessible to all those who need
them. For example, access to essential medicines is an
indispensable part of the right to health with several
dimensions.v> First, medicines must be accessible in
remote rural areas as well as in urban centres, which has
major implications for the design of medicine supply
systems. Second, medicines must be affordable to all,
including those living in poverty, which has obvious
implications for funding and pricing arrangements.
Third, given the fundamental human-rights principles of
non-discrimination and equality, a national medicines
policy must be designed to ensure access for dis
advantaged individuals and communities, such as
women and girls, people living with HIV/AIDS, elderly
people, and people with disabilities. Because equal access
is not always secured by equal treatment, a state must
sometimes take measures in favour of disadvantaged
people. As far as possible, data must be disaggregated to
identify marginalised groups and monitor their progress
towards equal access. Fourth, reliable information about
medicines must be accessible to patients and health
workers so they can take well-informed decisions and
use medicines safely.
www.thelancet.com Vol 372 December 13, 2008
Panel 1: Some important points from general comment 14
Article 12 of the International Covenant on Economic, Social, and Cultural Rights very
briefly sets out the right to the highest attainable standard of health. General comment 14
provides the UN Committee on Economic, Social, and Cultural Rights' interpretation of
article 12. Although not legally binding, the comment is highly authoritative.
•
Encompassing physical and mental health, the right to health places obligations on
governments in relation to health care and the underlying determinants of
health -these obligations include provision of clean water, adequate sanitation,
nutritious food, adequate shelter, education, a safe environment, health-related
information, and freedom from discrimination.
•
Governments have, for example, obligations regarding maternal, child, and
reproductive health; healthy natural and workplace environments; the prevention,
treatment, and control of diseases; health facilities, services, and goods.
•
•
Governments have an obligation to give particular attention to marginal individuals,
communities, and populations, creating a need for as much disaggregation of data as
possible.
Within a country, health facilities, services, and goods must be available in sufficient
quantity, accessible (including affordable) to everyone without discrimination,
culturally acceptable (eg, respectful of medical ethics and sensitive to gender and
culture), and of good quality.
•
•
The right to health is subject to progressive realisation and resource availability.
Nonetheless, governments must take deliberate, concrete, and targeted steps to ensure
the progressive realisation of the right as expeditiously and effectively as possible
•
However, core obligations are subject to neither progressive realisation nor resource
availability. Expressly taking into account the Declaration of Alma-Ata, they include
obligations to ensure access to health facilities, goods, and services to everyone,
including marginal groups, without discrimination; to ensure everyone is free from
hunger; to ensure access to basic shelter, housing and sanitation, and an adequate
supply of safe and potable water; to provide essential drugs, as defined under the Wl 10
action programme on essential drugs; to ensure equitable distribution of all health
facilities, goods, and services; and to adopt and implement a national public-health
strategy and plan of action, by way of a participatory and transparent process.
•
The right to health requires opportunities for as much participation as possible by
•
individuals and communities in health-related decision making.
Governments have an obligation to ensure that non-state stakeholders are respectful
•
of the right to health (eg, do not discriminate).
Developed states, and others in a position to assist, should provide international
assistance and cooperation in health to developing countries (eg, economic and
technical assistance to help developing countries fulfil their core obligations). All states
"have an obligation to ensure that their actions as members of international
•
organizations take due account of the right to health".
Monitoring, accountability and redress are essential. Given progressive realisation,
•
The right to health is closely related to, and dependent upon, numerous other human
•
rights, such as the rights to life, education, and access to information.
In narrowly defined circumstances and as a last resort, the enjoyment of some human
indicators and benchmarks are indispensable if governments are to be held to account.
rights may be interfered with to achieve a public health goal. For example, quarantine
for a serious communicable disease, such as ebola fever, may, under certain
circumstances, be necessary for the public good, and lawful under human rights, even
though it limits an individual's freedom of movement.
Health-related facilities and services may be available
and accessible but be insensitive to culture and gender.
For example, improving the access to sexual and
reproductive health care is not simply about scaling up
2049
Right to Health
technical interventions or making them affordable. A
Peruvian project that studied indigenous communities
with very high maternal mortality found an acute
reluctance within the population to use the health facilities
offered by the state, partly because they did not take
account of local cultural conceptions of health and
sickness. In close consultation with the indigenous
communities, culturally sensitive facilities and services
were introduced, such as sturdy ropes in delivery rooms
so that women could give birth squatting and gripping
the rope, as they were accustomed to. These changes led
an increase in deliveries in local health centres,50 and the
success of these local initiatives helped to generate a
corresponding change in national health policy on
deliveries in all primary health-care facilities.51
include both medical care and public health but not secure
fair access, or there might be a health information system
but key data might not be suitably disaggregated.
A major challenge for human rights is to apply or
integrate the right to health across the six building blocks.
The right-to-health analysis provided by general com
ment 14 has to be systematically and consistently applied to
health services, health workforce, health information,
medical products, financing, and stewardship-—that is, all
the elements that together constitute a functioning health
system. Panel 3 identifies some of the issues that arise
when the right-to-health analysis is applied to the second
WHO building block—the health workforce. The
right-to-health analysis of availability, accessibility, cultural
acceptability, quality, participation, international assist
ance and cooperation, monitoring and accountability, and
Right-to-health features of health systems
so on, can also be applied to health systems to identify
The Declaration of Alma-Ata identifies some vital compo some of the right-to-health features of health systems,
nents of an effective health system. The declaration is encompassing what health systems do (for example,
especially instructive because of its public-health, providing access to essential medicines and safe drinking
medicine, and human-rights aspects (panel 2), and it water) and the way in which they function (for example,
provides compelling guidance on the core obligations of transparently, in a participatory process, and without
the right to health.'6
discrimination). Health systems run the risk of being
Other attempts have been made to identify what impersonal, top-down, and dominated by experts, but the
constitutes a functioning health system.57 WHO identifies right to health places the wellbeing of individuals,
six essential building blocks that make up health systems: communities, and populations at the centre.” Irrespective
health services (medical and public health); health of which of the many definitions of a health system is
workforce; health information system; medical products, used, io.w.s-, ay
foliowing features should be part of any
vaccines, and technologies; health financing; and leader health system.
ship, governance, and stewardship.'" Although debatable,
Legal recognition—Countries should give recognition to
these building blocks provide a useful way of looking at the right to health in national law and by ratifying relevant
health systems and can be thought of as building blocks human-rights treaties.'6 In some countries legal provisions
for the realisation of the right to health. However, a health on the right to the highest attainable standard of health are
system might have all these building blocks but still not generating significant case law.” For example, Hogerzeil
serve human rights. For example, the system might and colleagues56 analysed 71 court cases from 12 countries
and concluded that in 59 cases access to essential medicines
was enforced through the courts as part of the right to
Panel 2; The Declaration of Alma-Ata (1978)
health. Legal recognition is just one of the first steps on a
Principal themes
long and difficult journey to realising the right to health.
• The importance of equity
Without follow-up from social movements, health workers,
• The need for community participation
progressive government ministers and public officials,
• The need for a multisectoral approach to health problems
activist courts, and international support, ih addition to
• The need for effective planning
governmental respect for the rule of law, legal recognition
• The importance of integrated referral systems
is likely to be an empty promise.
• An emphasis on health-promotional activities
Standards—Although important, legal recognition of the
• The crucial role of suitably trained human resources
right to health is usually confined to a general formulation
• The importance of international cooperation
that does not set out in any detail what is required of those
with responsibilities for health. For this reason, countries
Essential health interventions
must not only recognise the right to health in national law,
• Education concerning prevailing health problems
but also ensure that there are more detailed provisions
• Promotion of food supply and proper nutrition
clarifying what society can expect by way of health-related
• Adequate supply of safe water and basic sanitation
services and facilities. For example, provisions are needed
• Maternal and child health care, including family planning
for quality and quantity of drinking water,, blood safety,
■ Immunisation against major infectious diseases
essential medicines, the quality of medical care, and so on.
• Prevention and control of locally endemic diseases
Such clarifications may be provided by laws, regulations,
• Appropriate treatment of common diseases and injuries
protocols, guidelines, and codes of conduct. WHO has
• Provision of essential drugs
published important standards on various health issues.57’5”
2050
www.thelancet.com Vol 3/2 December 13, 2008
Right to Health
Many others have also contributed; for example, the Sphere
Project provides minimum standards for responses to
disasters.6" Clarification is important for providers, so they
know what is expected of them and also for those for whom
the service or facility is intended, so they know what they
can legitimately expect.
Participation-—Health systems must also include
institutional arrangements for the active and informed
participation in strategy development, policy making,
implementation, and accountability by all relevant
stakeholders, including disadvantaged individuals, com
munities, and populations.’6 Examples of such parti
cipation include conferences to develop national health
plans in Brazil and Peru; a legislative requirement of
Maori participation in New Zealand’s District Health
Boards; village health teams in Uganda; and the
participatory transfer of an HIV/Al DS clinic from
Medecins Sans Frontieres to the Guatemalan Ministry of
Health.61 Participation improves health outcomes.62
Transparency—Tempered by the confidentiality of
personal data, this requirement applies to all those working
in health-related sectors, including countries, international
organisations, public-private partnerships, business
enterprises, and civil-society organisations.16 The Medicines
Fransparency Alliance, funded by the UK Government, is
an alliance of governments, international agencies,
pharmaceutical companies, and civil-society organisations,
committed to increasing transparency of information on
the quality, availability, and pricing of essential medicines
in the public, private, and non-profit sectors.6’
Equity, equality, and non-discrimination—Health systems
must be accessible to all, including those living in poverty,
minority groups, indigenous people, women, children,
people living in slums and rural areas, people with
disabilities, and other disadvantaged individuals,
communities, and populations.16 Additionally, health
systems must be responsive to the particular health needs
of women, children, adolescents, elderly people, and so
on.’6 Outreach programmes are needed to ensure that
disadvantaged people have the same access as more
privileged people. Several European governments, for
example, have established Roma health mediator
programmes.64 As members of the Romani community
themselves, the mediators aim to improve community
health by mediating between patients and health workers
during consultations and communicating with Romani
communities on behalf of the public health system.
Although the programmes have limitations, mediators
have greatly assisted some Romani.64
The right-to-health principles of equality and
non-discrimination are akin to the health concept of
equity. All three concepts have a social-justice component.
In some respects, equality and non-discrimination, being
reinforced by law, are more powerful than equity.6’’ For
example, if a government or other body does not take
effective steps to tackle discrimination, it can be held to
account and required to take remedial measures.6667
www.thelancet.com Vol 372 December 13, 2008
Panel 3: Some issues arising when the right to health is applied to health workforces
•
General comment 14 requires a comprehensive national health plan
(eg, paragraphs 43[6] and 55) encompassing human resources. So is there an up-to-date
plan for human resources in preventive, curative, and rehabilitative health,
•
encompassing physical and mental health?
Is there a role for midlevel providers who can increase access to health care, such as
assistant medical officers and surgical technicians, and public-health professionals?
•
Are there outreach programmes for the recruitment of health workers from
marginalised communities and populations, such as indigenous peoples, to reduce
•
Are effective measures in place to achieve a gender balance among health workers in
•
all fields to ensure equality, non-discrimination, and respect for cultural difference?
Because health-related services must be available in sufficient quantity, subject to resource
non-discrimination and improve respect for cultural difference?
availability, are effective measures in place to ensure that the number of domestically
trained health workers is commensurate with the health needs of the population?
•
Is health information about the number of health workers by category (eg, nurses and
public health professionals) collected, centralised, and made publicly available on a
•
regular basis?
Are human rights, including respect for cultural diversity, as well as the importance of
treating patients and others with courtesy, a compulsory part of the training for all
•
health workers?
General comment 14 (paragraph 44(51) requires appropriate training for health
personnel, so are opportunities for further professional training in place for all health
•
workers without discrimination?
Are health workers receiving domestically competitive salaries as well as other reasonable
terms and conditions of employment? A lack of reasonable terms and conditions of
employment, one of the causes of the skills drain, is likely to undermine a health system
•
Are incentives in place to encourage the appointment, and retention, of health
workers in underserved areas to improve access, especially of marginal communities
and populations?
Respect for cultural difference- From the right-to-health
perspective, health systems must be respectful of cultural
difference. ih 'sw< Health workers must be sensitive to issues
of culture, ethnicity, and sex, strategies must be in place to
enable indigenous people to study medicine and public
health, and so on.1’9
Quality—All health-related services and facilities must
be of good quality. For example, water quality regulations
and standards consistent with the WHO guidelines for the
quality of drinking water should be in place/7 The good
quality requirement also extends to the way patients and
others are treated: health workers must treat patients and
others politely and with respect. Because medicines may
be counterfeit, states must establish appropriate regulatory
systems."'’ In Nigeria, for example, there is evidence that
the National Agency for Food and Drug Administration
and Control’s dual strategy of strengthening the regulatory
environment, while encouraging intolerance of counterfeit
drugs through public enlightenment campaigns is
improving medicine safety and quality.7"
Planning— Some important implications arise from the
right to health being subject to progressive realisation and
resource availability. The crucial importance of planning is
recognised in the Declaration of Alma-Ata,1" general
2051
Right to Health
comment 14, and elsewhere.71 States must have
comprehensive national health plans, encompassing both
the public and private sectors, for the development of
health systems; because the plans have to be evidence
based, a situational analysis with disaggregated data is
needed before the plan is drafted. Health research and
development should also inform the planning process.72-7’
According to general comment 14, the plan must include
certain features, such as clear objectives (and how these
are to be achieved), timeframes, effective coordination
mechanisms, reporting procedures, a detailed budget,
financing arrangements (national and international),
assessment arrangements, indicators and benchmarks to
measure achievement, and one or more accountability
devices.1'' Indicators and benchmarks are already common
place features of many health systems, but they rarely have
all the elements that are important from a human-rights
perspective, such as appropriate disaggregation.'6
The identification of indicators and benchmarks to
measure the progressive realisation of the right to health
is a national and international process that involves coun
tries, international organisations, the UN Committee on
Economic, Social, and Cultural Rights, and others. A
wealth of data is available at the global level, some of
which is highly relevant to the right to health. But are
international bodies making other data important to the
right-to-health perspective available? If not, countries may
wrongly assume that these other data, and the issues to
which they relate, are less important. Many countries look
to UN bodies for technical assistance, ideas, and
leadership. Whether or not UN bodies are making data
that are highly relevant to the right to health available at a
global level is an important issue.
A fair, transparent, participatory, and inclusive process
for prioritising competing health needs is required—*one
that takes into account explicit criteria, such as the well
being of those living in poverty, and not just the claims of
powerful groups with vested interests.77 The process of
prioritisation should give particular attention to the core
obligations identified in general comment 14 because they
are required of all countries, whatever their stage of econ
omic development.'6 The list of core obligations is illu
strative rather than exhaustive (panel I).74 One of the core
obligations is to adopt and implement a national public
health strategy and plan of action, on the basis of epidemiological evidence, addressing the health concerns of the
whole population.16
Before the finalisation of the plan, key elements must
undergo impact assessment to ensure that they are likely
to be consistent with national and international legal
obligations, including those relating to the right to the
highest attainable standard of health.75 In addition, the
present realisation of the right to health must be
maintained, although this might be waived in exceptional
circumstances.’6
Progressive realisation does not mean that a government
is free to choose whatever measures it wishes to take so
I
2052
long as they reflect some degree of progress. General
comment 14 requires that governments take deliberate,
concrete, and targeted steps to ensure progressive
realisation as quickly and effectively as possible.'6
Progressive realisation, maximum available resources,
and core obligations need closer conceptual and
operational attention. Some courts have rejected the idea
of core obligations and required that government policies
are reasonable.76 Other courts have taken the same
position as the UN Committee on Economic, Social, and
Cultural Rights in general comment 14 and found that
some health-related responsibilities are so fundamental
that they are subject to neither progressive realisation
nor resource availability.15-77 This position most closely
matches the right to health: progressive realisation is an
important concept with a crucial role, but only up to the
boundaries of core obligations.
Referral systems—Health systems should have a mix of
primary (community-based), secondary (district-based),
and tertiary (specialised) facilities and services, providing
a continuum of prevention and care.’2 The system also
needs an effective process by which health workers assess
whether patients will benefit from additional services
and patients are referred from one facility or department
to another. Referrals are needed between alternative
health systems (eg, traditional health practitioners) and
mainstream health systems. The absence of an effective
referral system is inconsistent with the right to health.
Coordination—Health systems and the right to health
depend on effective coordination across a range of public
and private stakeholders (including non-governmental
organisations) at the national and international levels.
Effective coordination between various sectors and depart
ments, such as health, environment, water, sanitation,
education, food, shelter, finance, and transport is important
for health systems, which also require coordination within
sectors and departments, such as ministries of health. The
need for coordination extends to policy making and
delivery of services.1652
’6 57 Uganda has recently added several
interventions, such as de-worming ofchildren, supplementation with vitamin A, and health promotion information,
to its Child Health Days. Now known as Child Health Days
Plus, these days depend on, and reinforce, improved
coordination between and within sectors and national and
international partners, including civil society, 7X.79
International cooperation—Health systems have inter
national dimensions, including the control of infectious
diseases, the dissemination of health research, and
regulatory initiatives, such as the International Health
Regulations58 and the WHO Framework Convention on
Tobacco Control.80 The international dimension of health
systems is reflected in countries’ human-rights
responsibilities of international assistance and cooperation
that can be traced through the Charter of the UN, the
Universal Declaration of Human Rights, and some morerecent international human-rights declarations and
binding treaties.8'-82 At least, all countries have a
www.thelancet.com Vol 372 December 13, 2008
Right to Health
human-rights responsibility to cooperate on transboundary
health issues and to do no harm to their neighbours.8’
High-income countries have an additional responsibility to
provide appropriate international assistance and coopera
tion in health for low-income countries. High-income
countries should especially help others fulfil their core
obligations."’ The Swedish International Development
Cooperation Agency (SIDA), for example, supports several
stakeholders with crucial roles in relation to the right to
health in Uganda. The agency has given funds to various
organisations: the Ugandan Government: WHO for its
human-rights work in Uganda; the Uganda Human Rights
Commission; and civil-society organisations, including
Straight Talk, which aims to increase understanding of
adolescence, sexuality, and reproductive health.8,1 For their
part, low-income countries have a responsibility to seek
appropriate international assistance and cooperation to
help them strengthen their health systems.85
General comment 14 confirms that the human-rights
responsibility of international assistance and cooperation
in health extends to countries’ actions as members of
international organisations.1" Scandinavian countries, for
example, have proposed a trust fund for justice and
human rights in the World Bank.85-87
Legal obligation—Crucially, the right to the highest
attainable standard of health gives rise to legally binding
obligations. The health system must have, for example, a
comprehensive national health plan; outreach programmes
for the disadvantaged; a minimum package of healthrelated services and facilities; effective referral systems;
arrangements to ensure the participation of those affected
by decision making in health; respect for cultural
difference; and so on. One of the distinctive contributions
of the right to the highest attainable standard of health is
that it reinforces good health practices with legal obligation
and accountability. States are legally obliged to take all
appropriate steps to implement the right-to-health features
of health systems. Of course, some governments imple
ment these features without reference to the right to
health. But many governments do not ensure that these
features are in place, and, in these cases, the right to health
has an especially important role.
Monitoring and accountability—Individuals and com
munities should have the opportunity to understand how
those with responsibilities have discharged their duties and
provide those with responsibilities the opportunity to
explain what they have done and why.88 Where mistakes
have been made, accountability requires redress. Account
ability is not a matter of blame and punishment but a fair
and reasonable process to identify what works, so it can be
repeated, and what does not, so it can be revised.88
Something as complex and important as health
systems needs effective, transparent, accessible, and
independent
accountability
mechanisms—health
commissioners, national human-rights institutions,
democratically elected local health councils, public
hearings, patients' committees, impact assessments,
www,thelancet.com Vol 372 December 13, 2008
and judicial proceedings. The media and civil-society
organisations also have crucial roles.8”
Accountability in many health systems is extremely
weak. In some countries, the same body provides and
regulates health services, as well as holding those
responsible to account. Accountability can also be little
more than a device to check that health funds were spent
as they should have been. Human-rights accountability is
concerned with ensuring that health systems are
improving, and the right to the highest attainable standard
of health is being progressively realised, for all, including
disadvantaged individuals, communities, and populations.
In some countries, although playing an important part,
the private health sector is largely unregulated. The
requirement of human-rights accountability extends to
both the public and private health-related sectors8” and to
international bodies working on health-related issues.
Accountability mechanisms are urgently needed for all
bodies—public, private, national, and international-working on health-related issues. The design of appro
priate and independent accountability mechanisms
needs creativity and leadership, such as recently shown
by the Uganda Human Rights Commission with the
launch of its new Right to Health Unit in Kampala.w
I
Scope and objectives
We begin to assess the degree to which the health systems
of 194 countries include features arising from the right to
the highest attainable standard of health.
From the start, this project did not aim to give a
weighting to indicators nor to rank countries in an index,
although we are aware that ranking can appeal to
politicians and sometimes might enhance monitoring
and accountability, leading to improved health and respect
for human rights.”1”2 Ranking in league tables is also
problematic with technical difficulties and problems of
interpretation.”' However, indicators and benchmarks are
needed to measure the present condition of a country’s
health system and to monitor its progress over time. We
hope that this project will be repeated periodically so that
the progress of individual countries, in relation to health
systems and the right to health, can be monitored.
Although much more work has to be done to help govern
ments identify the minimum package of health-related
services and facilities needed by the right to the highest
attainable standard of health, that vital task is not our aim
here. In this Report, we do not attempt to provide a list of
essential services and facilities needed for a well-functioning
health system. Rather, we attempt to identify several
additional, and commonly neglected, features arising from
the right to health and informed by good practices that are
required of all health systems.
Methodology
Development and selection of indicators
Our aim was to assess how much the health systems of
all countries include some of the features that arise from
2053
_
Right to Health
the right to health. To meet this aim, we identified the
following objectives: to promote awareness of the
complementary relation between a health system and the
right to health; to select a manageable set of indicators to
assess the degree to which a health system includes some
of the right-to-health features; to assess if sufficient
information is available about these features both
nationally and internationally; to increase monitoring
and accountability in relation to health systems and the
right to the highest attainable standard of health; to
deepen the understanding of the important role of health
data and indicators in relation to the progressive
realisation of the right to health; to consider the
limitations of data for health and human rights in relation
to the progressive realisation of the right to health; to
provide a basis to monitor, over time, health systems and
the progressive realisation of the right to the highest
attainable standard of health.
We developed indicators to reflect right-to-health
features of health systems. The features arise from
general comment 14,including core obligations, and
reflect many of the themes of the Declaration of
Alma-Ata,‘M and elements of the WHO building blocks
of a health system.711 We also referred to article 24 of the
Convention of the Right of the Child," general
comments 3 and 4 of the Committee on the Rights of
the Child,,,5% and general recommendation 24 of the
Committee on the Elimination of the Discrimination
Against Women.97 We also relied on the framework of
structure, process, and outcome indicators on the right
to the highest attainable standard of health,9X and the
requirement that health facilities and services should be
available, accessible, culturally acceptable, and of good
quality.’6
Assessment, monitoring, accountability, redress
for national, international, public, and private individuals and organisations
To ensure that a similar project had not allready been
done, we reviewed existing projects (both published and
in draft form) relying heavily on indicators, such as the
Human Development Reports,99 the World Health
Report 2000 on health systems,” the WHO and Office of
the High Commissioner of Human Rights (OHCHR)
indicators joint project of 2008, lM' the UN Millennium
Development Goals,"" the poverty-reduction indicators
from OHCHR,102 the WHO essential-medicine
indicators,and indicators of UNICEF,"M| UNAIDS,’05
and the World Bank.106 '"7
The development and selection of indidators was a
lengthy process with numerous stages. We selected
indicators according to the following criteria: scientific
robustness, usefulness, representativeness, understand
ability, and importance.9' Data availability was not a
determining factor. We also selected indicators that would
be accessible to a broad group of professions, including
policy makers in both health and human rights.
In an ongoing process over 18 months, we consulted
different individuals who helped in the selection of
indicators, including academics (eg, from political
science, law, health, and sociology). UN bodies, national
and local non-governmental organisjations and
associations, health practitioners, lawyers, economists,
and anthropologists. Individuals were consulted from
Africa, Latin America, Europe, North America, and
Asia-Pacific, with balance between sexes. We also
consulted people from indigenous communities. We
used purpose sampling to address specific questions and
right-to-health features. We also asked delegates at two
health and human-rights conferences (Italy and
Zimbabwe) for their views, and consulted Maori and
non-Maori people in New Zealand. However, our
Formal
recognition of
the right to health
Initiatives to ensure the
public knows its entitlements and howto vindicate them
(eq, educational campaigns of national human-rights
Standards
for public and private individuals and organisations (eg,
regulations, guidelines, ethical codes of conduct, and irotoccils)
Research and development
including economic, social, and cultural issues I
institutions)
Health situational analysis
Coordination mechanisms
between and within health-related sectors and
departments, in relation to policy making
and service delivery
Medicines
Health services, facilities
and goods
Comprehensive national health plan
objectives, timeframes reporting ptocpduti"., wh<>
responsible lor what, indicators, benchmarks; before
plan finalised, undertake impact assessment
People
Communities
Populations
Health promotion
leg, public-health awareness campaigns;
Health information
eitective iiealtli-inrormalion system, including
appiopi lately disaggregated data; respect for confidential
Underlying determinants of
health (eg, water, sanitation, food,
shelter, and education)
personal medical data
Non-discrimination
consider gender and marginalised groups;
need outreach programmes and
disaggregated data
Transparency
Respect for cultural difference
Quality
Health workers
including traditional health workers (eg, numbers commensurate
with health needs of the population; good terms, conditions, and
training, including human rights, upgrading the skills of the
established health workers; polite and respectful attitude; skills drain)
Participation
in policy formulation, implementation, and accountability
National financing
equitable and evidence-informed
International dimension
(eg, low-income countries to seek, and high-income countries
to provide, internatioihal assistance and cooperation in health,
the "do-no-harm" principle; south -south cooperation;
Figure I: Right-to-health features of a health system underpinned by legal obligations based on general comment 14: preliminary working model
2054
www.thelancet.com Vol 372 December 13, 2008
Right to Health
consultation process could have been better and we
suggest that, when our selection of indicators is revisited,
more consultations should take place.
We used five steps in the process of indicator selection.
First, we reviewed the right-to-health features of health
systems and WHO building blocks and, after numerous
consultations, we created a preliminary working model as
a way to assist development and selection of indicators
(figure 1). We focused on the wellbeing of individuals,
communities, and populations. We also recognised the
importance of health-related services, facilities, and goods,
including underlying determinants of health, water,
sanitation, food, shelter, and education. Additionally, we
identified a selection of features that were of particular
importance to health systems. We revised and refined this
preliminary model because we were aware that it had
shortcomings—eg, the model did not clearly convey that
several of the features, such as non-discrimination, are
recurrent.
Second, we looked closely at health-related services
and the underlying determinants of health (middle
section of figure 1) in the context of the right-to-health
requirements—ie, that these should be available,
accessible, culturally acceptable, and of good quality.
However, of these four requirements, we focused on the
first (availability) and second (accessibility). From the
right-to-health perspective, access is crucial because of
its relation with non-discrimination, equality, and equity.
Health care (eg, antenatal, mental-health, cancer care,
and access to medicines), underlying health determinants
(eg, drinking water), and government spending on
health care were also taken into account (webtable 1).
Third, we considered the features shown in the perimeter
of figure 1. Formulation of indicators for standards proved
difficult. Formal recognition of the right to health is
important but it is usually confined to a few vague
sentences. A more detailed elaboration is needed by way of
legislation, protocols, guidelines, codes of conduct, and
others. We therefore considered formulating indicators
that questioned whether countries had adopted
international standards on blood safety and water quality,
but we were unable to identify indicators conforming to
our criteria.
We recognise that planning is only a means to an end.
Nonetheless, general comment 14 underlines the import
ance of planning. An appropriate plan, prepared with a
suitable process, is a vital vehicle for realising the right to
health. Ihus, we devoted great attention to the identification
of appropriate indicators in relation to devising a
comprehensive national health plan (webtable 2).
We struggled to identify indicators of participation that
conformed to our criteria. Especially challenging were
indicators capturing whether a country has appropriate
institutional mechanisms for participation. Participation
should not be confined to the development of the national
health plan; it should also extend to the national
health-workforce strategy, national medicine policy,
www.thelancet.com Vol 372 December 13, 2008
implementation measures, accountability, and so on.
Furthermore, participation should not be confined to
marginalised groups. However, the right to health has a
special focus on disadvantaged people. But, if
marginalised groups are participating, we can be
confident that non-marginalised groups are too. In the
end, we addressed participation in the context of national
health planning.
Indicators for research and development in health were
also challenging. We wanted to identify indicators that
showed whether adequate and appropriate research and
development are being undertaken in a country, but we
failed to identify an indicator that conformed to our criteria,
and so this issue is not present in our final selection.
Fourth, we aimed to merge the two sets of draft
indicators already identified (webtables 1 and 2) and revise
them where appropriate. Also, we wanted to add some
new indicators. For example, we were looking for an
indicator that reflected the right-to-health requirement
that health-related services, facilities, and goods should be
culturally acceptable. General comment 14 emphasises
that special attention should be paid to indigenous people
and so, prompted by the indicators joint project of WHO
and the OHCHR,"10'"”* we added a new draft indicator—ie,
the proportion of people covered under indigenous or
alternative systems of health care (webtable 3). However,
placing indigenous and alternative health systems in the
same indicator was confusing; therefore, this indicator
was not included in our final list. Other indicators
approach aspects of cultural acceptability, such as the
indicator about participation of marginalised groups.
Also, the definition of a comprehensive national health
plan extends to the whole population, including
indigenous people, and incorporates public and private
sectors, including traditional and indigenous health
practices and medicines.
Fifth, we made a final selection of 72 indicators
(panel 4), divided into 15 groups. Some of the groups
overlap; for example, the participation indicator overlaps
with the planning indicators. Of course, many different
health workers are crucially important, but some
indicators use doctors and nurses as proxies. After
consultations, we created a new group—additional
safeguards—for indicators that did not fit neatly into any
other group. By placing indicators on monitoring,
assessment, accountability, and redress at the end of the
list, we are not saying that such issues only arise at the
end of a process. On the contrary, these issues must be
seen as recurrent elements in a continuous process.
As prioritised by general comment 14, several indicators
focus on maternal and child health. Women and children
are among those groups that are often marginalised. In
the past 2-3 years, maternal mortality has increasingly
been recognised as a human-rights issue."”-"1
We prepared explanatory notes for each indicator, along
the lines of the meta-sheets used in the recent indicators
joint project of WHO and the OHCHR."10 Explanatory
!
See Olmito: for webtables 1-3
2055
Right to Health
I
ISee !
see
We initially selected six countries for national data
collection because one of us, PH (then the UN special
rapporteur on the right to the highest attainable standard
of
had recently
x,* health)
—----- ;-------------, -been to them on mission and
prepared formal UN reports on each
f
before this .project
.
began. Only five could commit to data collec|tion in the
allotted time: Sweden, Mozambique, Romania, Peru,
and Ecuador. Although not globally representative, these
high-income, middle-income, and low-incom|e countries
provided
range of different political background,
provided aa range
history, geographical location, and cultural contexts.
Countries such as China, Bangladesh, USA, and India
would have been interesting to study, but they fell outside
our selection criteria. In each of the five countries
selected, we chose data collectors who, at the relevant
time, were independent from the government (although
they might have collaborated closely with the government)
and had a good knowledge of the country, of health and
human rights, and preferably of the right to health.
For every indicator, the information obtained was the
response to the indicator, the source of the c|ata, and the
date of last update and of access (if from an internet
source). Relevant comments to explain the answer were
documented, in addition to exact legal provisions for
indicators related to the law and exact quotes for
indicators related to plans or policies. In addition to
table 1 on global data (194 countries) and table 2 on
national data (five countries), webtables 4 and 5 include
extended tables with sources, comments, and other
information. The International Committee of Medical
Journal Editors Uniform Requirements wa-f used as the
basis for the referencing system."' For internet sources,
the date of access and of last update was of particular
External review
After selection of draft indicators, we sent them ^nd importance to document. The working currency for all
’ i was US dollars at the exchange rate at the
explanatory notes to 40 experts, who had not previously monetary data
time
of
data
collection.
been part of this project, for their comments and review.
We
clearly
defined terms used for indicators that were
22 experts responded, including lawyers, human-rights
professionals, clinicians, public-health practitioners, strictly adhered to during data collection; although this
academics, and policy makers. We recognise the rigidity might have led to reduced data availability. All
advantages of random sampling, but constraints of time relevant definitions can be found in the explanatory notes
and resources prevented this approach. However, expert online (webappendix 1).
Data were accepted in any of the official UN languages:
comments were very constructive in deciding and
Arabic,
Chinese, English, French, Russian, and Spanish,
devising the final version of our list of indicators.
with translations into English done by the team where
necessary. National data were received in English,
Data collection
To fulfil die project’s objectives, we needed to consider data Spanish, or Swedish, and translated where necessary.
Primary data were not collected; therefore, formal ethical
availability at the global level and, in relation to five coun
tries, at the national level as well.” "2 For a few
f indicators, approval was not necessary for this research project. Only
only the global perspective was necessary (eg, the number secondary data were used on both global and national
of international and regional treaties recognising the right levels. All the information was in the public domain,
to health ratified by a country). If information was not defined to be any document that is in print and should be
present about a particular indicator, it was marked as not easily accessible, such as in a library or on the web.
To ease collection of global data, we created the so-called
available in the tables 1 and 2. If an indicator was not
one-click
rule, which defines that the limits of the search
applicable to a particular country, it was marked as such.
‘
The time period for data collection was from August,
2007, f°r data should only be no more than one mouse click
August, 2008, with the intention of obtaining data for away from the global source. For example, if, while
to 1
navigating through the WHO website, a link to a national
the same indicators at about the same time.
notes for every indicator include: definitions, rationale,
method of computation, data source, periodicity,
comments, and limitations (webappendix 1). Each data
source was assessed for its quality and any potential bias
was noted.
To reduce the number of indicators, disaggregation of
all appropriate indicators on all relevant grounds was
not required, although we■ acknowledge that such data
should be available. Several of the selected indicators,
for webappendix 1
however, address discrimination, including the
indicator on civil registration requiring disaggregation
on five priority grounds: sex, ethnic origin, rural or
urban residence, socioeconomic group, and age. We
identified these priority grounds of disaggregation
through a process of consultation between lawyers and
health workers, including representatives from the
British Medical Association, WHO, and academic
institutions. Ideally, all appropriate data should be
disaggregated, at least, by these five priority grounds.
The civil registration system should be one of the most
comprehensive data-collection systems in a country,
and therefore this indicator was chosen as the proxy
measure.
We did not identify indicators of all right-to-health
features conforming to our selection criteria (eg,
indicators related to coordination and research and
development). Although several indicators exist
for webtables 4 and 5 addressing different dimensions of access, such as
indicators 24 (access to clean water), 27 (antenatal care),
and 46 (catastrophic health expenditure), we accept that
issues of access demand more attention.
www.thelancet.com Vol 3/2 December 13, 2008
2056
f
Right to Health
Panel 4 Indicators of right to health
Recognition of the right to the highest attainable standard of health
1
Number of international and regional human-rights treaties recognising the right to health ratified by the state
2
Does the state's constitution, bill of rights, or other statute recognise the right to health?
Non-discrimination
3 Number of treaty-based grounds of discrimination that the state protects out of: sex; ethnic origin, race, or colour; age;
disability; language; religion; national origin; socioeconomic status, social status, social origin, or birth; civil status; political
4
status, or political or other opinion; and property
Number of non-treaty-based grounds of discrimination that the state protects out of: health status (eg, HIV/ Al DS); people living
5
in rural areas; and sexual orientation
General provisions against discrimination
Health information
6 Does the state law protect the right to seek, receive, and disseminate information?
7
Does the state law require registration of births and deaths?
8
9
Does the state have a civil registration system?
Does the state disaggregate data in the civil registration system on grounds of: sex, ethnic origin, rural or urban residence,
socioeconomic status, or age?
10 What proportion of births is registered?
11 Does the state regularly collect data, throughout the territory, for the number of maternal deaths?
12 Does the state centralise these data for the number of cases of maternal deaths?
13 Does the state make publicly available these data for the number of cases of maternal deaths?
14 Does the state regularly collect data, throughout the territory, for the number of neonatal deaths?
15 Does the state centralise these data for the number of cases of neonatal deaths?
16 Does the state make publicly available these data for the number of cases of neonatal deaths?
National health plan
17 Does the state have a comprehensive national health plan encompassing public and private sectors?
18 Has the state undertaken a comprehensive national situational analysis?
19 Before adopting its national health plan, did the state undertake a health impact assessment?
20 Before adopting its national health plan, did the state undertake any impact assessment explicitly including the right to health?
21 Does the state's national health plan explicitly recognise the right to health?
22 Does the state's national health plan include explicit commitment to universal access to health services?
Participation
23 Is there a legal requirement for participation with marginalised groups in the development of the national health plan?
Underlying determinants of health
24 What percentage of the rural and urban population has access to clean water?
25 What are the CO, emissions per capita?
26 Prevalence rate of violence against women
Access to health services
27 Proportion of women with a livebirth in the last 5 years who, during their last pregnancy, were seen at least three times by a
health-care professional, had their blood pressure checked, had a blood sample taken, and were informed of signs of complications
Medicines
28 Is access to essential medicines or technologies, as part of the fulfilment of the right to health, recognised in the constitution or
national legislation?
29 Is there a published national medicines policy?
30 Is there a published national list of essential medicines?
31 What is the public per capita expenditure on medicines?
32 What is the average availability of selected essential medicines in public-health facilities?
33 What is the average availability of selected essential medicines in private-health facilities?
34 Percentage of 1-year-old children immunised against measles
35 Percentage of 1 -year-old children immunised against diphtheria, tetanus, and pertussis
(Continues on next page)
www.thelancet.com Vol 3/2 December 13, 2008
205/
Right to Health
(Continued from previous page)
Health promotion
36 Does state law require comprehensive sexual and reproductive-health education during the compulsory school years for boys and
girls?
37 Proportion of 15-24-year old boys and girls with comprehensive HIV and AIDS knowledge
Health workers
38 Does the state have a national health-workforce strategy?
39 Does the state law include provision for adequate remuneration for doctors?
40 Does the state law include provision for adequate remuneration for nurses?
41 Do the state's workforce policies or programmes include a plan for national self-sufficiency for doctors?
42 Do the state's workforce policies or programmes include a plan for national self-sufficiency for nurses?
43 Do the state's workforce policies or programmes provide incentives to promote stationing in rural areas of doctors?
44 Do the state's workforce policies or programmes provide incentives to promote stationing in rural areas of nurses?
National financing
45 Is the per capita government expenditure on health greater than the minimum required for a basic effective public-health
system?
46 What is the proportion of households with catastrophic health expenditures?
47 Total government spending on health as percentage of gross domestic product (GDP)
48 Total government spending on military expenditure as percentage of GDP
49 Total government spending on debt service as percentage of GDP
50 Proportion of national health budget allocated to mental health
International assistance and cooperation
51 Does the state's international development policy explicitly include specific provisions to promote and protect the right to health?
52 Does the state's international development policy explicitly include specific provisions to support the strengthening of health
systems?
53 Proportion of net official development assistance directed to health sectors
Additional safeguards
54 Does the state law require protection of confidentiality of personal health data?
55 Does the state law require informed consent to treatment and other health interventions?
56 Does the constitution protect freedom of expression?
57 Does the constitution protect freedom of association?
58 Does the state have a patients' rights charter?
59 Is the patients' rights charter available in all official languages?
Awareness raising about the right to the highest attainable standard of health
60 Does the state have a national human-rights institution with a programme of budgeted activities to raise awareness of the right
to health among the public?
61 Does the state have a national human rights institution with a programme of budgeted activities to raise awareness of the right
to health among doctors?
62 Does the state have a national human- rights institution with a programme of budgeted activities to raise awareness of the right
to health among nurses?
63 Are human rights a compulsory part of the national curriculum forthetraining of doctors?
64 Are human rights a compulsory part of the national curriculum for the training of nurses?
Monitoring, assessment, accountability, and redress
65 Infant mortality rate
66 Mortality rate of children younger than 5 years
67 Maternal mortality ratio
68 Life expectancy
69 Does the state have a national human-rights institution with a mandate that includes the right to health?
70 Number of judicial decisions, nationally, that considered the right to health during 2000-05
71 Does the state have a national human-rights institution with a mandate to monitor international assistance and cooperation?
72 In the past report submitted by the state to the UN in relation to the International Covenant on Economic, Social, and Cultural
Rights, was there a detailed account of the international assistance and cooperation in health that the state is providing?
2058
www.theiancet.com Vol 372 December 13, 2008
Right to Health
website was found, then the link would be opened. If the
required information was accessible on that webpage, then
it would be judged as globally processed data and therefore
acceptable. If, however, there was need to follow more than
one link, open tabs, or for any navigation other than
scrolling up and down, the information was not judged as
globally available. This rule only applied to links to
webpages outside the global source, with no limitations
imposed on navigation within the global source. We
decided that this rule was needed to provide a practical
limit to what could otherwise become an eternal internet
search, but still allowing adequate and useful data
collection.
We generated a list of global websites (webappendix 2).
If any data were unavailable in any of these websites, we
did a more detailed search. We identified suitable search
terms where necessary; these are documented in the
explanatory notes that accompany each indicator
(webappendix 1).
Sometimes, the presence of a right-to-health feature
was not apparent at the global level, which could be
interpreted as the data were not existing. However, in
these situations we reported that the information was not
available, which meant that it might exist and be
documented elsewhere. For example, if there was no
mention of a national health-workforce strategy in the
information available at the global level, then the data
were listed as unavailable for that country. However, if
the information stated that there was no health-workforce
strategy, then that was recorded with a ‘no'.
Methods for data collection nationally were slightly
different and defined as accessible information in the
public domain. These data were needed to be publicly
available and therefore included information available on
the internet, published reports, or information publicly
available on request. We requested documentation of the
source. Data that were only available to selected groups or
information acquired through interviews were not
allowed. National teams of data collectors could decide
how best to search for the information as long as it
fulfilled the methodological requirements. National teams
searched relevant websites and published documents
unrestricted by the one-click rule. However, they were
restricted to national sources and could not access global
databases, such as UN sources. We relied on the same
explanatory notes (webappendix 1) to ensure that
definitions and criteria were consistent with the global
data and the same referencing guidelines were applied.
We are aware of some limitations of the methods.
For some indicators, a year was included from which
data were collected (identified in webappendix 1). The
project started in 2007 and data were often not yet
available for 2007, or even 2006, and therefore data
from 2005 were often the most up-to-date (further
information is available in webtables 4 and 5 and
explanatory notes in the webappendix 1). Furthermore,
the indicator relating to judicial decisions was restricted
www.thelancet.com Vol 3/2 December 13, 2008
to cases within a 5-year period (2000-05) because we
thought that this was a manageable recent timeframe.
Several indicators we selected are commonly used (eg,
indicator 68 on life expectancy), and information is
available in published material about their usefulness,
validity, and limitations.
Because of the many differing lists of states worldwide,
we opted to use a list of 194 countries generated from the
WHO member-state list of 2000r’? and those countries listed
by the UN Development Programme,"2 acknowledging that
other lists may differ. Several indicators were only applicable
to donor states, and the list of states to be considered for
these indicators was compiled from Organisation for
Economic Co-operation and Development (OECD)
members"4 and the International Development Association
of the World Bank list of donors from March, 2008"5
(webpanel). However, the indicator may be related to an
event that happened before a state was a donor, such as the
past report in relation to the International Covenant on
Economic, Social, and Cultural Rights.
!
I
See ! :
. for webappendix 2
SeeOi'HiK for webpanel
Key findings
We discuss some of the key findings and results arising
from the data collected for the 72 indicators, giving special
attention to three of our objectives. Do countries’ health
systems have the relevant right-to-health features? Are
the relevant data available at the global level? Do the data
provide a basis to monitor, over time, health systems and
the progressive realisation of the right to the highest
attainable standard of health.
We did not try to find directional relations between
variables, such as treaty ratification and health outcomes.
Several indicators in our list illustrate that, no matter how
sophisticated they are, indicators never provide a complete
picture and they need to be supplemented with qualitative
information. Table 1 summarises the global data from
194 countries and table 2 the national data from
five countries. All data are available in webtables 4 and 5.
Recognition of the right to health
Recognition of the right to the highest attainable standard
of physical and mental health is a right-to-health feature of
a health system. Although recognition can have various
forms, we focused on international (indicator 1) and
national (indicator 2) recognition (panel 4). Figure 2 shows
the number of countries that have ratified three interna
tional human-rights treaties that include the right to health.
The step after ratification of treaties is the recognition of
the right to health in the national constitution or other
statute, but more than two-thirds of countries do not have
this recognition. Only 56 countries that have ratified the
International Covenant on Economic, Social, and Cultural
Rights include the right to health in their constitution or
other statute. International recognition of the right to health
(indicator 1) is substantially more widespread than national
recognition (indicator 2), probably because international
accountability is weaker than national accountability.
2059
Right to Health
1
i
Although legal recognition of the right to the highest
attainable standard of health can mean commitment
towards the realisation of the right to health, this doeji
not capture the actual process or success of
implementation. Other indicators attempt to do this and
are discussed later. Legal recognition is important
because it can increase accountability of stakeholders
with responsibilities to, and within, a health system.
Although eight indicators explicitly mention the right
to health, different countries use different terminology
for this human right. Some countries use terminology
that does not match our wording, and negative results
were recorded in these cases (webappendix 1). Online
documents were sometimes translations of original
documents, introducing another possible reason for
different terminology.
Non-discrimination
We aimed to record aspects of non-discrimination,
equality, and equity—key right-to-health features of health
systems. Indicator 3, for example, lists 11 treaty-based
grounds of discrimination, and indicator 4 lists three
non-treaty-based grounds of discrimination (panel 4). The
treaty-based ground of discrimination most commonly
protected by law was ethnic origin (122 countries), whereas
the least-protected was age (13 countries; figure 3).
However, 95 countries protect only five or less treaty-based
grounds of discrimination, and none protects all 11.
We addressed non-discrimination asking whether data in
the civil registration system were disaggregated on the five
priority grounds of sex, ethnic origin, rural or urban status,
socioeconomic group, and age (indicator 9). None of the
five countries studied nationally disaggregate these data by
ethnic origin, and therefore they cannot show any inequity
between ethnic groups. Disaggregation of data on the basis
of ethnic origin is a controversial issue and, although such
information can be used in a positive way, it can also be
used in a negative way (eg, to reinforce stigmatisation).
Therefore, article 8 of the EU Data Directive prohibits the
"processing of personal data revealing racial or ethnic
origin”, but with important exemptions related to data
processed by health professionals.""
For non-treaty-based grounds of discrimination,
protection was even less widespread than for treaty-based
grounds (figure 3). For example, according to our
approach, only three countries (Fiji, South Africa, and
Ecuador) protected against discrimination on the ground
of sexual orientation.
People with mental illnesses are frequently neglected
and discriminated against, and this might lead to
inadequate financial provision for mental health*
Therefore, we took into account the proportion of the
national health budget allocated to mental health
(indicator 50). Of 98 countries for which data were
available, almost half allocated 2% or less of their national
budget to mental health. Sweden and Ecuador did not
allocate a specific budget for mental health. The teank
2060
gathering data in Sweden remarked that this is “partly a
consequence of the objective to not stigmatise the group"
(webtable 4). We do not agree that a specific budget
allocation for mental health could stigmatise those with
mental-health problems or that it is inconsistent with the
integration of mental-health care across health systems.
By contrast, the absence of a specific budget allocation
might maintain the marginalisation and neglect
experienced by many people with mental disabilities.
Later, we consider the indicator on access to clean water
(indicator 24). Disaggregated on the basis of urban or
rural residence, this indicator confirms the disadvantage
of rural dwellers in most countries.
Health information
Because of its crucial importance in relation to both the
right to health and the WHO building blocks, health
information is prominent in our profile of
indicators (indicators 6-16). We focused on maternal and
neonatal deaths, and the civil registration system. We
questioned whether countries obtained data for the
number of maternal deaths throughout their territory.
On the basis of global data and our approach, 69 countries
obtained, centralised, and made publicly available these
data; whereas 88 countries did not. Data for the remaining
37 countries were unavailable at the global level, including
those for Ecuador and Peru. Compared with data for
maternal deaths, data for neonatal deaths were available
at a global level for even fewer countries. Nationally, data
were gathered, centralised, and made publicly available
by Ecuador, Romania, and Sweden (panel 5j; however,
Mozambique did not do this for maternal deaths (they
only include those deaths occurring in institutions) and
Peru did not for neonatal deaths. Overall, on dhe basis of
our approach, 88 countries do not seem to have in place
an adequate health information system fok maternal
deaths, suggesting that their health systems afe seriously
deficient in terms of both the right to health and relevant
WHO building blocks. Also, despite their importance,
global data for maternal and neonatal deaths are
inadequate.
As with equity, human rights have a particular concern
for marginalised individuals, groups, and populations.
Several indicators in the profile take into account
disadvantaged groups, such as indicators on discrimination
(indicators 3-5), participation with marginalised groups
(indicator 23), and whether or not the patients’ rights
charter is available in all official languages (indicator 59).
Disadvantage cannot be monitored without data that are
disaggregated on key grounds."’ We questioned whether a
country disaggregates data from the civil registration
system on the priority grounds of sex, ethnic drigin, rural
or urban residence, socioeconomic group, and age
(indicator 9). On the basis of our approach, no global data
were available for any country. However, research in the
five selected countries showed that national data were
available.
www thelancet com Vol 372 Decerpber 13, 2008
Right to Health
All five countries disaggregated on two of the five
grounds (sex and age); only one country (Romania)
disaggregated on four of the five grounds, and none
disaggregated on the ground of ethnic origin, which
makes the design of appropriate interventions that
address ethnic disadvantage very difficult for policy
makers. Lack of disaggregated data also makes it difficult
to hold countries accountable for accessibility of their
health systems. From the right-to-health perspective, this
shortcoming is important.
Overall, our data confirmed that disaggregated
information, which is crucial for right to health, even when
available nationally, is not always made available globally.
National health plan
According to general comment 14, the adoption of a
national public-health strategy and plan of action is a
core obligation.’6
An essential precondition for the development of a
comprehensive national-health plan is a national health
situational analysis (indicator 18). Global data showed that
57 countries had done health situational analyses, although
all were done as a part of the WHO country cooperation
strategy development process."” However, global data were
not available for the other 137 countries. Data were more
readily available nationally than they were globally. For
example, although global data are not available for
Romania, national data confirmed that they have not done
a health situational analysis. Indicator 18 does not capture
whether the analysis was used to develop the national
health plan or the quality of the analysis, as shown in
panel 6 for Mozambique. Nonetheless, the existence of a
health situational analysis is an important precondition
for a national health plan and a step in the right direction.
Our research confirmed that, despite the importance of
health situational analyses, global data collection neglects
this right-to-health feature.
Assessments of health and human-rights effects are
also needed, together with a comprehensive national
health plan. There is a growing trend to undertake health
impact assessments before a health initiative is finalised,
adopted, and implemented.” We asked whether countries
undertook either a health impact assessment (indicator 19)
or any impact assessment that included the right to
health (indicator 20) before adopting their national health
plan. Of course, indicators have limitations; for example,
even with an impact assessment, any negative findings
might be ignored and the plan implemented without
revision. No global data were available with our approach
for any country regarding either indicator. Nationally,
none of the five countries did a health impact assessment
before adopting their national health plans. We confirmed
that, despite the importance of such assessments, global
data collection ignores this important right-to-health
feature of a health system.
We asked whether a country has a comprehensive
national health plan encompassing public and private
www.thelancet.com Vol 372 December 13, 2008
sectors (indicator 17). Our explanatory notes (web
appendix 1) identified the essential criteria of a plan, such
as clear objectives, timeframes, indicators, benchmarks,
and reporting procedures. For 181 countries, we were
unable to gather global data with our approach, and
13 countries do not have a comprehensive national health
plan—ie, their health systems lack this important rightto-health feature. However, the indicators for national
health plans highlighted limitations of the internet as a
resource. Dates of last updates are commonly unavailable,
and therefore information might be out of date. Although
we made every effort to gather data as completely and
accurately as possible, with such an extensive database
there might be inaccuracies. Nonetheless, despite the
importance of a comprehensive national health plan,
global data collection seems to neglect this important
right-to-health feature of a health system.
More-detailed national data were available for the
comprehensive national health-plan indicators, showing
that Mozambique, Romania, and Sweden have com
prehensive national health plans, whereas global data
were not available. At the time of data collection, Peru
was preparing a national health plan but it did not
include all the features of a comprehensive health plan
(panel 7).
We asked whether the national health plan includes an
explicit commitment to universal access to health services,
defined as access to primary, secondary, and tertiary
physical and mental care (indicator 22). We regarded a
commitment to basic or essential care as inadequate. A
low-income country might not be in a position to deliver
universal access to health services, but a comprehensive
national health plan should include a commitment to
reach this aim.'6 Such a commitment is the minimum
expected from all countries, whatever their stage of
economic development. A developing country’s commit
ment to universal access gives an important message to
health workers, the public, and donors. When a country
cannot provide universal access, it must have fair,
transparent, rational, evidence-informed processes (eg,
protocols and guidelines) in place to ensure that reasonable
decisions are made when determining who has access to
health-related facilities and services, and on which terms.
On the basis of global data and our approach, national
health plans of 15 countries (Antigua and Barbuda,
Bahrain, Botswana, Chile, North Korea, Dominican
Republic, Honduras, Libya, Mauritius, Mozambique,
Peru, Seychelles, Timor-Leste, Uruguay, and Yemen)
include an explicit commitment to universal access to
health services, whereas plans of 14 countries
(Afghanistan, Argentina, Bangladesh, Bolivia, Costa
Rica, Croatia, Egypt, El Salvador, Lesotho, Malawi, Nepal,
Papua New Guinea, Romania, and Tanzania) do not.
However, global data are not available for 165 countries.
Again, our research suggests that global data collection
neglects information that is important for the right to
health.
I
5
2061
Right to Health
Recogni
Non-
tion
4.__
Health information
National health
discrimin
2
Afghanistan
4/5
N
Albania
5/6
Y
3
4
6
5
8
7
lOt
N
7
Algeria
7/8
Y
4
Andorra
5/6
Y
6
1
Y
1
Y
N
Y
N
Y
Angola
5/8
N
5
1
Y
Antigua and
4/7
N
5
1
Y
11
12
13
14
16
15
6
N
N
N
N
N
N
98
N
N
N
N
N
N
N
N
N
99
N
29
N
N
N
N
N
N
N
18
17
Y
N
7/7
Y
Armenia
6/6
N
Australia
5/5
N
Austria
6/6
N
Azerbaijan
6/6
The Bahamas
3/7
10
Y
Bahrain
5/5
N
Bangladesh
5/5
N
Barbados
6/7
7
4/6
Y
Belgium
6/6
Y
Belize
3/7
N
Benin
8/8
N
Bhutan
2/5
N
Bolivia
7/7
Y
Bosnia and
4/6
N
Y
91
Y
Y
Y
Y
Y
Y
Y
Y
96
N
N
N
N
N
N
5
1
1
5
1
5/8
N
4
7/7
Y
4
Brunei
3/5
N
Bulgaria
6/6
N
Burkina Faso
8/8
Y
Burma
2/5
N
Burundi
7/8
Y
Cambodia
5/5
N
Cameroon
6/8
Y
1
Y
97
Y
N
Y
N
Y
Y
Y
Y
10
N
N
N
N
Y
Y
Y
Y
N
93
N
N
N
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Comoros
7/8
N
,Congo
7/8
N
Cook Islands
2/5
N
N
N
N
N
N
N
N
N
N
N
N
Y
Y
Costa Rica
7/7
N
N
N
58
N
N
N
N
N
N
89
N
N
N
N
N
N
Y
N
N
N
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
N
64
N
N
N
N
N
N
65
N
N
N
N
N
N
60
N
N
N
N
N
N
66
N
N
N
N
N
N
N
N
70
N
N
N
N
9
N
Y
96
Y
Y
N
N
Y
N
Y
1
N
Y
Cote d'Ivoire
7/8
N
Croatia
6/6
Y
8
1
Y
Cuba
4/7
Y
4
1
N
Y
Cyprus
6/6
N
7
1
N
Y
N
N
N
Y
N
Y
Y
53
85 3
3-7762
N
N
86
98
6-2916
N
N
Y
100
23-8656
•
N
N
Y
711
81 3
0-2469
62
100
100
4-3607
N
Y
Y
100
100
6-5892
N
Y
Y
Y
N
9-722
N
2-9395
N
54-3
76-1
0-2902
N
Y
Y
60
86
0-6649
N
Y
Y
55-8
95 3
0-774
N
Y
Y
94-9
99-8
3-9936
90-5
24-0912
N
N
Y
5-4598
N
Y
Y
63 1
91
0-0812
4-1
N
75 2
0-2052
13-7
N
77-9
91-4
0-0291
N
Y
333
62-4
0-039
N
Y
Y
45-2
86-3
0-2205
N
Y
Y
99
100
20-0095
N
N
N
64-1
0-5553
N
Y
Y
N
N
55 7
881
0-0614
N
40-9
81-4
0-0127
Y
Y
N
Y
Y
Y
89-4
N
N
29
74
N
Y
N
37
Y
N
N
N
100
N
Y
54-4
2-3693
1-8001
N
N
N
99-7
N
Y
15
97
N
N
N
100
N
Y
8-5
100
100
N
N
53
82
N
66
99
3 8712
85-5
99-5
3-8393
N
Y
62
N
Y
N
Y
Y
49-4
N
N
Y
90
N
N
N
N
N
N
63-8
99-1
1-2102
N
83
Y
N
N
N
N
N
N
74-2
935
0-1132
44-1
N
Y
N
N
N
37-4
956
1-0034
25-5
N
N
N
N
Y
Y
Y
81
N
(Brazzaville)
1
N
N
N
N
Y
0
N
Y
N
5
N
N
N
N
1
N
Y
N
Y
N
6
Y
N
82
Y
N
N
Y
100
N
Y
N
Y
Y
Y
N
Y
7
Y
Y
N
Y
Y
N
Y
Y
Y
N
Y
Y
70
Y
N
Y
N
Y
Y
N
Y
N
7/7
Y
Y
Y
N
Y
Colombia
Y
N
Y
N
1
N
1-2052
Y
Y
N
3 6951
Y
Y
Y
N
98 3
99
Y
Y
Y
6/7
78-8
75
N
N
5/5
N
N
Y
Chile
N
5-0556
16-272
1
China
0-5051
95
84628
1
6
63-8
89
100
6
N
39-9
N
N
100
7
6/8
89-6
100
100
N
Chad
5-994
82-7
100
100
Y
1
N
N
Y
5/7
5
00288
1-172
Y
8/8
N
Y
56-6
99 3
Y
Cape Verde
7/8
Y
27-4
92 5
Y
Canada
CAR
30
Y
N
7
29
Y
N
1
28
Y
N
7
|
Y
N
1
27t
Y
N
7
urban
Y
Y
7
26t
rural
Y
Y
1
N
26t
Y
Herzegovina
Brazil
Y
25*
Y
N
1
N
24t
urban
Y
N
7
N
24t
rural
Medicines
I
Y
N
5
Botswana
Acc-
ess
Y
N
4
Belarus
Y
1
1
22
N
1
5
21
Y
Y
Barbuda
Argentina
Underlying determinants
plan
ation
Y
55
Y
N
Y
100
N
Y
N
Y
N
Y
N
Y
Y
Y
Y
Y
Y
Y
N
N
N
Y
Y
Y
Y
Y
Y
N
Y
Y
875
985
2 0445
N
N
92
100
1-506
N
N
Y
87-8
95 2
0-2825
12-7
N
N
Y
51765
70-3
N
N
77-7
949
2-2956
N
N
Y
100
100
8-1634
N
Y
Y
2062
www.thelancet.com Vol 372 December 13, 2008
I
Right to Health
315
32U
3311
34t
35t
57
45
61
52
91
92
53
34
37m
t
37ft
Health workers
National financing
38
45
43
44
Y
42-7
35 2
Y
67**
68
tt
Y
165
257
1800
42
NA
Y
15
17
92
71
NA
Y
Y
33
38
180
71
NA
NA
Y
Y
3
4
82
NA
NA
NA
Y
Y
154
260
41
NA
3
NA
N
Y
10
11
73
NA
N
N
14
17
77
75
N
Y
Y
21
24
76
69
NA
5
6
4
82
N
NA
Y
NA
Y
NA
NA
4
6
2-9
5-8
1-5
6- 5
Y
3-5
Y
6-3
N
1-8
Y
4-8
41
6-8
10-2
1
5-8
NA
N
5-4
2-9
2-8
5
NA
N
33102
Y
8-8
1-8
10
4-2
N
171 98
Y
10-2
0-9
NA
7 97
Y
N
3-9
2-3
2
Y
6-7
0-6
11
90
Y
3-8
3-6
88
N
2-8
1
1-3
1
Y
68
0-8
31
12
3612
Y
6-6
1-5
2-3
NA
NA
500-02
Y
96
11
6
2-21
Y
4-9
1-4
20-7
1
N
5-4
11
1-6
4
Y
72
1 57
88
71
N
94
NA
NA
30-55
57-36
94
81
N
3-7
419
NA
NA
83
83
72
79
96
99
N
64
72
Y
64
88
Y
90
97
87
81
10-6
81
18
33
Y
75-4
37-55
Y
57
83
83
75
74
2-6
77
78
1 25
71
75
23
15
Y
3-6
1-21
009
10-27
13
14
16
74
NA
NA
Y
Y
9
10
32
75
NA
NA
N
Y
52
69
570
63
NA
N
Y
11
12
16
75
NA
Y
Y
6
8
18
69
NA
N
Y
4
5
8
79
Y
NA
N
Y
14
16
52
69
NA
NA
Y
Y
88
148
840
55
NA
63
70
440
64
NA
29
0-5
1
7-9
1-4
7-9
3
Y
Y
N
NA
Y
NA
NA
N
Y
Y
50
61
290
66
NA
Y
Y
13
15
3
75
N
N
Y
90
124
380
52
NA
Y
Y
19
20
110
72
N
NA
9
13
Y
Y
10
12
11
73
NA
Y
Y
122
204
700
47
NA
74
104
380
60
NA
Y
109
181
1100
49
NA
Y
Y
65
82
540
62
NA
Y
Y
87
149
1000
51
Y
Y
5
6
7
81
NA
Y
Y
25
34
210
70
NA
NA
Y
Y
114
174
980
48
NA
NA
Y
Y
124
209
1500
46
NA
N
6-7
1-1
0-9
NA
NA
N
2-2
1-3
1
NA
3-4
6-2
4-9
NA
NA
Y
6-4
1-1
0-5
NA
NA
5-2
13
4-7
0
NA
9-7
1-1
6-43
0-4
23
20
N
3-7
0-9
1-1
Y
5-4
3-7
6-7
2
4-7
1-9
1-2
7-3
4
73
N
82
81
Y
73
70
66
68
15
92
87
1-19
69
70
6-26
N
Y
N
Y
0-2
Y
NA
1
Y
Y
Y
8
9
16
78
NA
2
N
N
N
Y
20
24
45
73
N
8-3
0
N
Y
Y
17
21
130
74
N
1
NA
NA
N
Y
51
68
400
65
NA
1-4
23
NA
NA
Y
Y
79
126
740
54
NA
NA
NA
Y
Y
16
19
73
NA
7-1
0
3
8
NA
Y
Y
11
12
30
78
NA
3-9
1-5
2-8
0
NA
Y
Y
90
127
810
53
NA
74
1-9
12-8
Y
Y
5
6
7
76
N
Y
Y
5
7
45
78
NA
Y
Y
3
4
10
80
N
4-6
0-12
N
N
1-9
Y
Y
N
3
N
NA
8
77
NA
1-1
Y
Y
8-3
4
11-54
N
NA
N
NA
NA
NA
2-7
48
NA
64
5-7
85
N
82
1-8
52
4032
89
8-8
75
Y
73
3
84
10
Y
6-9
8-45
19-2
N
NA
3-4
15-17
N
NA
0
0-7
157-91
80
1-6
5-6
8
4
NA
Y
21
4
21-7
0-09
1400
72
4
2-4
5-02
59
Y
0
Y
13-6
N
58
N
0-8
279-34
313
Y
3-9
N
27 2
Y
7-7
2.00
N
3-6
1-9
Y
55
2
N
34-3
715
Y
40
Y
56
5 77
70
66||
1
5- 2
Y
54
Monitoring, assessment,
accountability, and redress
6511
1-3
N
53t
57
50t
48t
Additional safeguards
56
49t
47t
46t
IAC
Y
7-6
Y
6
5
1-4
7
NA
N
NA
N
(Continues on next page)
www.thelancet.com Vol 3/2 December 13, 2008
2063
Right to Health
iI f
Recogni
tion
Nondiscrimin
ation
2
Health information
4
3
5
6
1
Y
7
8
National health
plan
lOt
11
12
13
14
15
16
Y
Y
Y
Y
Y
Y
N
N
N
18
17
21
Underlying determinants
22
24+
rural
24+
urban
Access
Medicines
27+
28
29
30
11-4759
61-6
N
Y
Y
0037
75 7
Y
25*
26+
rural
26+
urban
(Continued from previous page)
Czech
Republic
6/6
Y
8
DRC
6/8
N
6
Denmark
6/6
N
2
Djibouti
Y
N
1
5/8
N
N
N
Y
Y
Y
89
N
N
N
78
N
N
N
34
Y
Y
N
N
N
Y
Y
Y
N
N
N
Y
Y
Y
N
N
N
Dominica
5/7
Dominican
Republic
6/7
N
Ecuador
7/7
Y
8
1
Y
Y
N
N
N
N
N
N
Egypt
7/8
N
3
1
N
Y
N
N
N
N
N
N
1
N
Y
N
N
N
N
N
N
N
N
N
N
N
N
El Salvador
7/7
N
4
Equatorial
Guinea
6/8
N
3
2
Y
N
32
28 5
N
Eritrea
6/8
N
7
1
N
N
N
N
N
N
N
N
Estonia
6/6
Y
7
1
Y
Y
Y
Y
Y
Y
Y
N
8
1
N
N
N
N
N
N
N
Y
N
1
N
Y
N
N
N
N
1
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
N
N
N
N
N
N
Ethiopia
7/8
N
FSM
2/5
Y
Fiji
3/5
N
Finland
France
Gabon
2
9
6/6
7
6/6
N
3
Y
7/8
N
2
N
The Gambia
7/8
Georgia
5/6
Y
8
Germany
6/6
N
7
Ghana
7/8
Greece
6/6
7
1
1
6
N
1
Grenada
4/7
N
5
Guatemala
7/7
Y
2
Guinea
7/8
Y
Guinea
Bissau
4/8
Y
4
Guyana
5/7
N
10
Haiti
4/7
Y
1
1
N
Y
Y
Y
N
Y
Y
2 1063
73 7
897
96-4
2-2658
40-5
N
96
100
2-2116
N
67-1
96
42-4
453
Y
Y
Y
Y
0-9378
N
N
Y
11-4748
N
Y
537
717
0-1735
99
140496
N
39 7
96-4
0-1037
N
N
86-1
98-8
N
N
100
46-9
94-9
71
N
Y
559
N
Y
N
7-4
N
Y
Y
N
1-301
N
Y
Y
12-5782
N
Y
N
6-1608
N
N
Y
1-0796
N
Y
Y
N
N
Y
Y
Y
Y
N
N
771
94-6
0-1821
N
N
N
95-5
99-8
0-866
Y
Y
Y
Y
Y
Y
100
100
N
N
N
N
N
N
Y
Y
Y
Y
Y
Y
N
N
N
N
N
N
N
N
Y
N
N
98
100
N
N
N
646
N
41-8
N
9-7881
N
0-326
66-5
88-4
87275
N
2 0693
N
N
N
93
97
0-9857
N
Y
N
Y
Y
43
N
N
N
N
N
N
98-9
89-8
01515
N
39
N
N
N
N
N
N
33 8
79-1
0 1752
N
N
Y
N
Y
N
Y
Y
N
Y
N
Y
N
97
N
N
N
N
N
N
81
N
N
N
N
N
N
94
N
N
N
2
Y
Y
Y
Y
Y
7
1
Y
Y
Y
Y
Y
Iceland
6/6
N
7
1
Y
Y
Y
Y
Y
Y
India
5/5
N
4
1
Y
N
Indonesia
5/5
N
1
N
Iran
3/5
N
2
Iraq
4/5
Y
5
Ireland
6/6
N
Israel
5/5
N
Y
Italy
6/6
Y
6
Jamaica
6/7
N
4
Japan
5/5
N
5
Jordan
5/5
N
Kazakhstan
4/5
Y
Kenya
7/8
N
2064
1-5636
91-4
N
Y
1
Y
100
72-9
N
Y
1
Y
90
Y
N
6/6
8
Y
0-4639
N
6/7
1
N
70-5
N
Hungary
1
N
N
51-5
N
Honduras
1
N
N
N
1
Y
N
93
Y
N
Y
N
9-8013
55
51
Y
N
1
89
Y
Y
1
7
Y
83-8
Y
Y
41
N
N
N
N
N
N
55
N
N
N
N
N
N
Y
Y
Y
N
N
N
Y
Y
N
Y
N
Y
N
N
N
N
N
N
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
N
N
N
N
N
N
Y
Y
Y
Y
Y
95
89
Y
Y
Y
Y
Y
Y
Y
Y
N
Y
N
N
N
N
N
N
N
Y
99
N
N
N
N
N
N
N
Y
48
N
N
N
N
N
N
491
79 2
1-9547
N
832
83-4
0-1919
N
543
54 6
1-1362
80-4
93-6
5-6543
99.9
100
7-6103
N
Y
85
98
1-6945
N
Y
Y
72-1
88-1
6-3139
N
Y
N
86
99
2-9739
N
N
Y
51-2
97-5
10-4119
N
Y
N
10-8377
N
7-6908
N
Y
Y
3-974
N
N
Y
N
Y
N
N
Y
Y
Y
Y
X
71-6
N
1-2023
100
100
11 8
87-4
95-9
9-8434
Y
100
100
3-0658
94-4
98-8
13-2574
64-9
N
Y
96
100
0-3054
30
N
15
www.thelancet.com Vol 372 December 13, 2008
sgor
8OOZ '£t -taqiuaaao Z/E |°A uios ja^uepqj MMM
(oGed jxau no sanuijuo})
VN
SS
09s
IZl
6Z
A
N
VN
0
El
Z-l
S-fr
N
IL
EZ
V-ZZ
ZZE
VN
V9
OVl
6?
9Z
A
A
VN
Z
l-EZ
I
6£
A
IS
IL
QL
0
VN
IZ
29
SZ
iz
N
VN
VN
8^
8-V
SOI
A
86
56
08
8ZZ
A
E8
9
V
E
A
A
N
Ell
s
I
2-8
A
VN
ZZ
OZl
ZE
92
A
N
VN
S
N
18
E
E
A
A
A
8WI
S6
9S
SZZ
N
18
A
N
N
N
08
I
P
P
A
A
A
VN
95
OOE
LP
LI
A
A
VN
SlL\
S
N
N
N
A
ESI
EgS
S6Z
90
ZV
6-1
6-8
9
6Z
SZ
Z
90
2’8
A
8E
IV
A
09
09
101
981
LI ll
A
A
SEO
180
A
ISE
VN
IZ
0>l
SE
OE
A
VN
E
El
9-V
8Z
VN
89
OZV
Pt
92
A
A
VN
I
E-9
Zl
IZ
VN
E9
OSV
9Z
ZS
A
A
VN
Z
E
8Z
S
N
18
P
£
Z
A
A
N
A
VN
0
S-6
EO
A
N
EZ
9
Z
9
A
A
A
A
8
SlZ
M
8Z
ZO
A
VN
OZ
08?
ZZ
EZ
A
A
VN
Z
9V
90
SZ
A
EG
58
VN
19
0Z9
08
09
A
A
VN
VI
2'9
N
tP
VS
VN
V9
OLP
Z9
SP
A
A
VN
Zt7
80
68
Z6
8ZE
VN
8V
0011
ooz
Gil
A
N
VN
Z
801
GE
ZS
N
8E
OZ
zzo
VN
ES
016
191
86
A
VN
VN
GV
Z
9’5
N
IS
OS
ZIO
VN
89
OGZ
IV
IS
A
A
N
VN
I
Si
EO
ZS
A
ZZ
SZ
VN
89
OZ
Zl
A
N
VN
01
9Z
N
08
E
P
P
A
N
A
VEl
VN
8E
101
VN
ZS
095
OZl
9Z
A
A
VN
I
N
08
V
S
V
N
A
E6Z
VN
OZ
99
ZE
82
A
A
A
N
VN
VN
6S
069
Pll
V8
A
A
VN
VN
85
OZS
16
09
A
A
VN
N
18
8
s
V
N
A
A
VZE
8
N
GZ
Z
£
E
A
A
A
SSS
VN
69
OlZ
81
91
A
A
VN
69
IV
££
A
VN
95
OIL
EZl
ZZ
K
VN
SZ
SZ
9
S
A
A
VN
E9
OSV
VZ
8P
A
A
A
A
A
A
089
90Z
VZl
A
A
OZl
SZ
ZZ
A
A
N
89
OEl
SE
GZ
A
A
VN
EZ
OlZ
vz
IZ
A
A
VN
OZ
OSl
GZ
SZ
A
A
VN
VZ
Si
El
A
VN
95
0S9
OEl
98
N
GZ
£
V
E
VN
ZV
0011
SOZ
N
ZZ
V
44
89
.«Z9
ZZ
OZ
A
A
GZI
A
A
V
E
A
A
1199
1159
ZS
95
N
N
A
A
A
A
GZ
EE
9’8
VN
E-9
90
0
VI
El
SZ
Zll
10-0
A
VI
SZ
vvo
A
90
El
IV
SEI
A
80
9Z
GV
N
ZZ
19 12
Z-9V
PSL
VzS
SOZ
GOt?
VN
VN
N
SOS
VS El
A
ZlZ
A
*8
S8
N
08
EZ
ZS
N
Z8
68
It
A
8£
SS
8SE
E-OV
9VV
GZi
SEO
EOV
ZZO
9V-S0S
VlZE
A
Z6
Z6
SZ
99
VS
88
GZS
8E
80011
VN
VN
IZl
9’1
S
VN
VN
IZ
IVZ
ZE
N
Zl
A
EE
IS
VN
8<
90
Z
A
68
08
6
82
GZ
1-9
V6
Z6
PL-9
VN
Vll
9Z
ES
A
58
58
SO-E
90
VS
A
59
89
5- 9
A
ZZ
ZZ
10
VN
N
6S
N
Z
IZ
SS
El
VN
9V
N
£00
VN
N
ZZ
2-9
A
N
ZZ
96
85
ZOl
N
A
90
88
VI
Z
VN
A
A
VS
A
N
90
VN
VN
A
N
9C-1
VN
VN
VN
I
£
VN
E
9
9Z
8Z
A
6- 9
ZEO
A
El
1-6
ZOO
A
E
IZ
ZV
E
8V
z
l-Z
000
A
IOS
■16V
+8V
IZV
+9t'
SV
VN
VN
VN
VN
A
N
ZV
VN
Zill
A
IEO
Z6
ZE
A
VN
VN
N
S6
GZ
2'69
8 ES
80Z
-
N
IE
£69
8E6SV
8V
591
+
ssajpaj pue 'Xiijiqe^unoaae
’luatussasse 'Buijojiuoyy
qilBBH ouqSiy
6S
85
SS
VS
lES
spjenGajes (euoiiippy
2)VI
Gupueuy |euoi;eN
VV
EV
8E
sja>|JOM q;|ean
IJZE
lu/E
+SE
+VE
HEE
HZE
SlE
I
Right to Health
Recogni
tion
Non
discrimin
ation
1*
3
2
4
Health information
6
5
7
National health
plan
lOt
8
11
12
14
15
16
N
N
N
13
21
18
17
Underlying determinants
22
Acc
ess
Medicines
27t
28
29
30
N
Y
Y
24t
rural
24t
urban
25*
51-9
96-4
0-3244
N
37-9684
N
26+
rural
26t
urban
(Continued from previous page)
Kiribati
2/5
N
5
Kuwait
4/5
N
5
Kyrgyzstan
4/5
Y
6
Laos
5/5
N
4
Latvia
1
1
N
N
1
N
Y
Y
1
Lebanon
5/5
N
Lesotho
7/8
N
Liberia
6/8
N
Libya
8/8
Y
Y
8
1
1
5
N
Y
N
N
Liechtenstein
4/6
N
1
6/6
N
6
Luxembourg
6/6
N
1
Y
9
6
Malta
6/6
N
Marshall
Islands
2/5
Y
7
Mauritania
7/8
N
4
Mauritius
7/8
N
1
5
Mozambique
7/8
N
1
8/8
N
4
1/5
N
5
Nepal
5/5
N
5
Netherlands
6/6
N
4
New Zealand
5/5
N
6/7
Y
7
Niger
7/8
Y
4
8/8
N
1
6
N
Palau
1/5
N
6
Panama
7/7
N
6
PNG
Paraguay
3/5
N
7/7
Y
5
Y
Y
Y
4 1019
N
74-5
0-1401
N
721
10-331
N
N
Y
Y
Y
Y
Y
N
N
N
N
N
N
N
N
N
Y
N
Y
3-8686
N
Y
N
N
Y
Y
N
Y
Y
N
Y
Y
100
100
24-9271
34 3
78 3
0-1506
77
99
0-081
18 2
N
N
Y
96
100
7-0494
48-8
N
Y
Y
N
Y
Y
N
Y
Y
N
Y
Y
N
N
N
N
N
N
N
N
79-9
98-8
2-4981
N
N
N
N
N
54-4
83-6
00501
Y
Y
Y
Y
Y
100
6-1299
N
N
N
Y
Y
55
Y
Y
N
N
N
N
N
N
Y
Y
Y
Y
Y
Y
84-1
0-8866
12-5
N
Y
Y
42-3
54-2
2-598
57-6
N
N
Y
55-8
Y
Y
Y
Y
Y
Y
100
100
4-2387
Y
Y
Y
Y
Y
Y
Y
Y
92 1
94-4
1-9578
N
Y
Y
98
Y
Y
Y
34-7
90-9
33455
N
Y
98
Y
Y
Y
Y
N
Y
85
Y
Y
Y
959
Y
35
N
N
5
98
71
N
N
N
N
N
N
100
N
Y
N
N
N
N
N
N
N
N
Y
N
N
N
N
N
N
Y
N
N
N
N
N
N
Y
N
N
N
N
N
N
N
N
N
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
56 9
97 3
1-3654
24-4
73 3
0-1079
88-6
99-8
1-2394
N
36
39 9
14-1681
Y
N
N
197
N
Y
Y
N
N
Y
79
93
0-1146
100
100
8-7349
N
Y
Y
7-7946
N
N
N
Y
12-9
N
81
Y
Y
Y
N
N
N
56-7
88
0-743
N
Y
32
N
N
N
N
N
N
36
80-8
0-0947
N
Y
Y
33
N
N
N
N
N
N
313
64-9
0-8263
N
Y
Y
N
N
N
Y
N
N
N
N
N
N
Y
Y
Y
Y
Y
Y
Y
Y
1
Y
N
N
6
N
N
N
N
4/5
Y
Y
N
68-4
52 9
73
Y
1
Y
Y
3-0657
47
N
1
Y
Y
Y
N
N
Pakistan
94
75
N
1
N
49-4
Y
N
N
Y
N
N
N
100
96-3
N
6-4
N
Y
1/5
4/5
N
Y
100
78 3
■■
0-2296
N
N
4/5
Oman
N
N
Y
Niue
6/6
N
N
N
North Korea
Norway
N
N
82-6
N
N
5
Nicaragua
Nigeria
1
1
N
N
N
Nauru
N
N
9
Namibia
N
N
N
2
N
N
N
6
N
N
N
1
N
Y
Y
N
N
Y
Y
Y
Y
Y
N
Y
5/8
N
N
N
Y
Morocco
N
N
Y
8
5/6
N
N
Y
3
Montenegro
N
N
Y
N
Y
N
N
Y
Y
4/5
N
N
Y
6/6
Mongolia
N
Y
Y
1-1114
47-8
N
26
Y
7/7
N
Y
N
Moldova
5/6
Y
Y
Mexico
Monaco
Y
Y
N
8/8
Y
Y
6
Mali
Y
Y
Y
Y
N
N
N
Y
7/8
5/5
N
N
Y
Madagascar
Maldives
N
N
Y
Y
N
N
N
Y
7
N
N
N
Y
Y
3/5
N
N
Y
6/6
8/8
Y
Y
Macedonia
Malaysia
Y
94
N
1
Y
59
Y
N
Lithuania
Malawi
Y
N
Y
5/6
Y
Y
Y
Y
100
100
2-1941
N
N
100
100
3-3645
N
Y
Y
19 0086
N
Y
Y
N
Y
Y
N
Y
N
N
N
N
N
N
N
Y
100
100
12-4662
Y
N
N
N
N
N
N
Y
84
95-6
0-809
8-5
N
1
N
Y
N
N
N
94-5
78
11-9017
1
Y
Y
N
N
N
N
Y
Y
Y
86
88
1-7827
N
Y
Y
N
N
N
N
N
N
100
100
0-4126
N
Y
Y
Y
Y
Y
Y
N
N
N
67-7
94-3
0-7215
N
Y
Y
1
1
N
Y
Y
N
42-3
Y
2066
www.thelancet.com Vol 372 December 13, 2008
Right to Health
31S
3251
3351
34t
35+
12
0
96
98
37 m
t
37ft
Health workers
National financing
38
45
43
44
163
0-23
N
80
62
NA
NA
83 8
0-27
47t
481
49t
Additional safeguards
sot
53+
54
2
NA
55
81
31
5-2
8
NA
Y
Y
36
41
150
66
NA
3-6
21
6
NA
NA
N
Y
59
75
660
60
NA
6
Y
Y
8
9
10
71
NA
0-62
Y
Y
2 75
6-4
1-7
NA
Y
517
8-7
4-5
161
NA
NA
Y
Y
27
31
150
70
NA
85
83
N
NA
NA
N
5- 5
23
37
7
NA
N
Y
102
132
960
42
NA
41
27
N
6- 4
1-2
0-2
NA
NA
N
Y
Y
157
235
1200
44
NA
92
96
Y
3-2
1-8
NA
NA
N
N
N
17
18
97
72
NA
Y
Y
7
9
11
71
NA
92
98
59
61
15 7
19-4
85
86
36
23 5
92
91
49
40
Y
Y
15
9
Y*
Y
Y
13
1-42
Y
Y
N
Y
3
4
12
80
N
Y
Y
15
17
10
73
N
NA
Y
Y
72
115
510
59
NA
2
NA
Y
Y
76
120
1100
50
NA
7-2
2
NA
N
Y
Y
10
12
62
72
NA
4-4
NA
NA
Y
N
Y
26
30
120
72
NA
1-7
0
NA
N
Y
119
217
970
46
NA
10
NA
N
Y
5
6
8
79
NA
0
NA
Y
Y
50
56
63
NA
Y
Y
78
125
820
58
NA
12
15
15
73
NA
7-7
0-8
Y
7-8
2-2
4-1
N
32
1-1
1-5
1
N
12 2
1-6
4-6
Y
4-2
2-4
Y
12-4
N
5-8
23
Y
8-4
0-7
Y
15-4
2-7
3-6
3-6
1
NA
4-3
0-2
4-5
0
NA
34
Y
1-54
6-4
0-4
5-7
1
92
N
7-5
0-4
8-6
Y
4-6
N
4-3
11
59
91
86
N
Y
89
N
Y
3
5
1-6
2-4
5
N
95
2 81
77
72
10-41
80
79
52 5
33
NA
NA
12
N
20
Y
N
Y
Y
45-31
85
89
103-2
N
017
0-11
29
35
60
74
N
NA
Y
Y
16
19
22
68
NA
NA
N
Y
3
4
82
NA
Y
Y
35
42
66
NA
Y
Y
9
10
74
NA
NA
Y
Y
3-9
53
NA
Y
Y
34
37
240
72
NA
4-3
0-9
1-4
NA
Y
Y
96
138
520
50
NA
5-3
31
NA
NA
Y
Y
45
61
210
61
NA
NA
NA
N
Y
25
30
61
NA
0
NA
Y
Y
46
59
830
62
NA
4
5
6
80
N
Y
10-3
N
5-8
1-9
Y
9-2
1-5
7
14-44
N
Y
Y
Y
8-9
11
11
345
N
Y
Y
5
6
9
80
N
8-3
0-7
3-5
1
NA
Y
Y
29
36
170
71
NA
Y
Y
148
253
1800
42
Y
Y
99
191
1100
48
34
42
1-6
5-09
N
0-2
47
39
N
3-8
1
1-1
NA
62
54
N
3-9
0-6
9
NA
Y
14-5
NA
NA
N
3-5
NA
NA
0
4-65
21
18
639
92
96
Y
472 S3
98
3-3
313
63
2-05
0-28
9
1-6
2-5
11-8
4-1
3-4
2-2
98
Y
65
N
21
Y
9-6
N
NA
49-44
NA
Y
3 22
76
47
5-39
67
61
N
Y
46
5-3
83
36-4
Y
Y
86
26 2
Y
Y
N
8-2
90
Y
NA
1-2
1-28
89
NA
Y
7
10-1
5-9
134
Y
84
57
1-92
N
6
N
NA
84
0
78
N
7 99
80
NA
4
90
0-85
100
65
11
88
18-43
4 95
64
9
67**
19-6
438
70
47
Y
Y
59
4-7
449
0-5
72
Y
N
58
2-2
Y
25
70
66||
57
12 7
82-29
13-4
68
tt
65||
56
Y
NA
0-08
Monitoring, assessment,
accountability, and redress
Y
N
28
Y
0
461
I AC
2-35
N
N
3-51
N
N
N
Y
Y
NA
1
NA
70
NA
N
Y
42
55
370
66
NA
Y
N
3
4
7
80
N
Y
Y
10
11
64
74
NA
78
97
320
63
NA
69
NA
0
NA
2
NA
Y
Y
10
11
7-3
0
13-5
NA
NA
Y
Y
18
23
130
76
NA
4-2
0-6
7-9
1
NA
N
Y
54
73
470
62
NA
7-3
0-8
6-7
0
NA
Y
Y
19
22
150
75
NA
N
(Continues on next page)
www thelancet.com Vol 372 December 13, 2008
2067
4’
Right to Health
Recogni
tion
Non
discrimin
ation
2
3
4
Health information
6
5
8
7
National health
plan
lOt
11
12
13
14
16
15
17
18
21
Underlying determinants
22
24t
rural
24t
urban
25+
26+
26t
rural
urban
69
51
Access
Medicines
27t
28
29
30
vN
Y
Y
(Continued from previous page)
Peru
7/7
Y
Y
Y
93
Philippines
5/5
Y
Y
Y
83
Poland
Y
6/6
1
Portugal
6/6
Y
7
Qatar
3/5
N
4
5/5
N
South Korea
Romania
Y
3
Y
N
N
N
N
Y
Y
Y
Y
40-2
91
11682
82
87
0-9716
100
8-0331
100
100
Y
N
Y
Y
Y
Y
Y
N
N
N
71-4
97
9-7652
N
Y
Y
Y
N
Y
Y
Y
Y
16
91
4-1621
N
Y
Y
Y
N
Y
Y
Y
Y
95 2
99-7
10-5393
N
N
N
N
N
N
67-6
90-9
0-0632
N
N
Y
N
N
N
99
99
2-5697
N
98
2-2997
N
1 6708
N
0-8236
N
N
Y
0-6106
N
N
Y
N
N
Y
Y
Y
82
N
Y
N
Y
Saint Lucia
3/7
N
N
Y
Y
Y
Y
Saint Vincent
and the
Grenadines
98
4/7
N
Y
Y
Y
Y
Y
93
Y
N
N
N
N
N
N
SaoTome
and Principe
3/8
Saudi Arabia
4/5
N
Senegal
8/8
N
Serbia
N
Y
N
8/8
Y
7/8
N
Y
69
N
1
3
Y
5/6
Seychelles
10
1
2/5
N
1
N
N
N
55
N
N
N
N
N
N
99
Y
Y
Y
N
N
N
Y
48
N
N
N
N
N
N
Y
Y
Y
Y
Y
Y
Y
-
Y
Y
Y
N
N
N
Y
N
1
3
N
Y
1
4
N
N
1
1
N
N
Y
Slovakia
6/6
N
8
1
Y
Y
Y
Y
Y
Slovenia
6/6
Y
7
1
Y
Y
Y
Y
Y
Solomon
Islands
N
4/5
3/8
N
6
South Africa
7/8
Y
10
6/6
Y
5/5
N
7
Sudan
5/8
N
4
Suriname
7/7
6/8
6/6
N
2
N
9
Tajikistan
4/5
Y
7/8
N
91-6
N
1
Y
5/5
9
Y
8
7/8
Y
3/5
N
Trinidad and
Tobago
5/7
N
lunisia
6/8
N
4
1
1
95
N
Y
Y
6-7367
N
Y
8-1175
N
Y
Y
N
Y
Y
65
94
0-3819
63-4
Y
Y
Y
Y
Y
Y
Y
100
99-9
7-7227
56-9
N
Y
Y
N
N
N
58-4
Y
Y
Y
Y
Y
N
Y
N
Y
N
N
Y
12-2252
N
Y
95
8
N
N
N
99
N
N
N
96
N
Y
35
Y
Y
N
N
9-1927
Y
Y
Y
6-4395
0-1843
100
Y
Y
Y
87
Y
N
N
N
Y
78
100
Y
N
N
Y
N
74-8
N
Y
N
Y
53
45-6
N
N
8-5
N
Y
Y
13-3811
04353
N
53
N
100
93 2
N
Y
Y
64-3
653
Y
64
N
Y
Y
Y
Y
Y
N
N
4/5
logo
1
7
N
Tonga
1
3
N
Y
N
1
N
Tanzania
Timor-Leste
1
Y
N
4/6
Thailand
1
7
Switzerland
5/5
1
Y
Sweden
Syria
N
2
5
Sn Lanka
Swaziland
87-7
99
N
5
Somalia
Spam
Y
56-4
N
Sierra Leone
Singapore
N
N
8
1
Y
N
N
Y
7
N
Y
Y
6
N
5-6253
69-221
Y
Y
2/5
Y
Y
7/8
5/6
Y
Y
4/5
Samoa
N
N
Y
Rwanda
San Marino
Y
23
Y
Y
N
Y
N
N
Y
Y
3/7
Y
N
N
6/6
8
N
Y
1
Russia
Samt Kitts
and Nevis
N
N
Y
N
Y
N
Y
N
N
N
N
Y
74-6
96-2
0 6058
N
N
N
Y
63-4
79-4
0-287
Y
Y
Y
72-5
98-1
5-0805
N
N
N
Y
66-2
86-8
0-8589
Y
Y
Y
N
N
N
Y
39
N
Y
Y
N
Y
Y
N
Y
Y
N
Y
Y
Y
100
100
5-894
N
Y
100
100
5-4731
N
Y
Y
83
97-5
3-7207
Y
Y
Y
N
N
N
47
93
0-7738
N
Y
Y
N
N
N
42-5
78-9
0-116
56
41
N
N
Y
47
41
N
N
N
N
Y
97-9
985
4-2849
N
N
N
Y
62
803
0-1737
N
Y
N
N
N
34-2
76-4
0-3805
N
Y
N
N
N
100
100
1-1852
N
Y
Y
Y
Y
89
93
24-6802
N
Y
N
N
N
Y
78 5
95 7
2 2895
N
N
N
36 7
Y
Y
N
2068
www.thelancet.com Vol 372 December 13, 2008
Right to Health
315
32U
33H
34+
35+
37m
37f+
Health workers
National financing
38
45
+
43
44
46t
47+
48+
49+
50+
IAC
Additional safeguards
53+
54
61-5
60-9
87
85
Y
3-21
4-3
1- 5
7
2
NA
N
8-47
15-4
26 5
80
79
N
0-78
3-2
0-9
10
0
NA
N
6-2
2- 1
11-2
NA
10-2
2
76-15
Y
99
98
90
91
Y
N
NA
NA
N
N
187-3
59-35
86
Y
87
Y
Y
8 55
88
26-67
83
Y
2-71
Y
4-1
5-9
2-6
5-5
2
7
3
Y
5-2
3-7
5-5
NA
N
7-2
1-9
11
1
Y
5-5
10-6
Y
5-9
4
Y
6
55
Y
4.9
5-5
Y
7-3
Y
9-8
72
1
NA
79
N
N
Y
Y
9
11
12
77
NA
N
Y
5
5
14
79
N
Y
Y
14
16
24
73
NA
N
Y
Y
10
13
28
66
N
Y
Y
97
160
1300
52
NA
NA
17
19
71
NA
4
NA
12
14
75
NA
5
NA
17
20
70
NA
NA
NA
NA
NA
NA
NA
Y
Y
N
NA
N
3-7
2
2-1
NA
NA
N
Y
NA
4-7
7
1-7
Y
0-06
8-5
1-5
2 48
N
4-3
N
7-9
355
Y
0-03
8-7
1-6
Y
0-48
8-2
1
1 25
12-6
2-6
1-9
3-8
4-4
1-4
60
79
Y
5- 3
96
97
Y
6- 3
402
1-6
NA
63
96
61
NA
N
N
21
26
18
70
NA
Y
Y
60
116
980
59
NA
Y
Y
7
8
73
NA
12
13
72
NA
159
269
2100
40
NA
Y
Y
3
3
14
80
NA
Y
Y
7
8
6
74
N
NA
N
Y
Y
3
4
6
78
N
N
Y
55
72
220
67
NA
90
145
1400
55
56
69
400
51
4
4
4
81
N
NA
4- 1
NA
82
N
NA
N
68
3
Y
2
N
28
5
1-2
91
23
3
6
NA
44
-
NA
1
51
Y
9
4-7
67
N
11
Y
NA
NA
4
NA
N
68
3
3
63
73
230
Y
7-9
2-3
240
32
Y
21
1-4
25
24
Y
6-8
8
5-4
21
75
Y
3-4
3-5
462-44
70
8
8
055
0.00
6-76
68
7
Y
12-5
67**
6
96
76
66||
Y
87
36
6511
Y
N
38
59
Y
93
Y
77 2
N
Y
78
82
58
N
3
13-8
Y
51-4
NA
0-09
237-13
25
1
1-73
38-34
1-3
8-91
Y
Y
Y
2
Y
Y
74
76
139-76
57
97
37-74
0-81
56
++
1-81
224
55
Monitoring, assessment,
accountability, and redress
NA
3-67
N
2
NA
N
Y
Y
Y
Y
Y
Y
Y
Y
Y
N
NA
1
NA
Y
Y
11
13
58
72
NA
NA
Y
Y
62
89
450
60
NA
4
NA
Y
Y
29
39
72
68
NA
0
NA
N
112
164
390
42
NA
Y
0-18
9-2
1-5
11
5 62
Y
Y
0-57
11-4
1
NA
6-1
Y
N
Y
Y
3
4
3
81
N
Y
Y
4
5
5
82
N
98 2
80
83
N
4-2
5-3
0-8
NA
Y
N
12
13
130
72
NA
035
75
85
89
86
N
5
2-2
3-4
NA
Y
Y
56
68
170
64
NA
0-8
23 4
47-9
80
86
N
5-1
1-1
1-1
7
NA
74
118
950
50
NA
96
98
3-5
11
11
3
NA
NA
55
57
58
64
1-6
0-8
0
NA
6-01
84
9.6
1
NA
Y
N
20
24
29-3
58
81
NA
Y
Y
33
38
71
96
NA
Y
Y
19
23
44-05
64-3
95-1
49
44
Y
Y
NA
NA
0-8
N
13-7
N
5-3
5
1- 9
Y
4-5
2- 6
Y
5-5
NA
1-6
7-2
NA
N
Y
Y
Y
7
8
110
72
Y
Y
47
55
380
66
NA
69
107
510
57
NA
NA
71
45
69
100
72
-
NA
NA
(Continues on next page)
www.thelancet.com Vol 372 December 13, 2008
2069
■ Right to Health
Recogni
tion
Non
discrimin
ation
2
3
Health information
5
4
6
7
8
lOt
96
National health
plan
11
12
13
14
N
N
N
N
N
N
N
N
N
N
15
16
18
17
21
Underlying determinants
22
24t
rural
24t
urban
N
93-8
98- 3
31395
N
53-6
931
8-7549
91-6
93-9
25*
26t
rural
26t
urban
Acc
ess
Medicines
27t
28
29
30
Y
Y
(Continued from previous page)
Turkey
6/6
N
5
Turkmenistan 4/5
Y
7
1
N
Y
2/5
N
4
1
N
Y
Uganda
7/8
N
6
Ukraine
5/6
N
8
Tuvalu
UAt
4/5
N
UK
6/6
N
USA
1/7
N
Uruguay
7/7
N
4/5
N
3/5
N
Uzbekistan
Vanuatu
1
1
Y
1
Y
1
Y
1
Y
1
N
N
N
N
N
N
Y
Y
Y
Y
Y
Y
Y
N
N
N
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
N
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
N
N
N
N
N
N
Y
Y
Y
N
N
N
Y
N
N
N
Y
N
N
N
N
N
N
N
8
6
6
1
5/7
Y
4
1
N
Vietnam
5/5
Y
1
1
N
Yemen
Zimbabwe
5/5
N
7/8
N
6/8
N
100
N
Venezuela
Zambia
4
Y
92
Y
Y
Y
87
N
5
6
1
N
7
1
Y
Y
10
N
N
N
42
Y
Y
N
N
57
84
00651
89-3
991
6-9802
58 7
N
100
99- 6
377966
331
N
N
N
100
100
9-7934
N
Y
Y
N
N
Y
N
100
100
20-3792
100
100
1-6479
79 9
97-4
5- 2619
N
Y
Y
94
63
04192
N
N
Y
N
Y
Y
N
N
Y
Y
N
N
N
76-7
70-5
84-6
6-5735
87 3
96-2
1-1768
65 5
71
1-0311
Y
44-5
90-4
0-203
28-9
N
Y
Y
Y
69-2
96-8
0-8106
40-6
N
Y
Y
Y
N
9-5
Data are numbers or Y=yes or N«no, unless otherwise stated. NA-not applicable, -rnot’ available. IAC-=international assistance and cooperation. Indicators 9.19, 20, 23, 36, 39-42, 51, 52, 60-64, 69. and 71
Emirates. ‘Number of treaties actually ratified/number of treaties the state is eligible to ralitify. tProportion (%), ICO, emissions per capita. SUSS value per capita. IIMedian availability (%). |(Probability per are
1000
Table 1: Globally processed data for indicators of health systems and right to health
Nondiscrimin
ation
3
4
5
Health information
6
7
8
9
10*
National health plan
11
12
13
14
15
16
17
18
19
20
21
22
Participation
Underlying determinants
Acces$
Medicines
Health
promotion
23
24*
rural
27*
34*
35’
36
100
102
Y
90-6
80-2
Y
97
97
N
96
99
Y
24*
urban
Ecuador
8
2
1
Y
Y
Y
3
Y
Y
Y
Y
Y
Y
Y
N
N
Y
Y
N
39 3
78 3
Mozambique
7
0
0
Y
Y
Y
2
N
N
N
Y
Y
Y
Y
Y
N
N
Y
Y
N
48-5
40
Peru
7
1
1
Y
Y
Y
3
866
Y
Y
Y
N
N
N
N
Y
N
N
N
N
N
62
84
Romania
11
0
1
Y
Y
Y
4
99-9
Y
Y
Y
Y
Y
Y
Y
N
N
N
Y
Y
N
34
92
Sweden
6
0
0
Y
Y
Y
2
100
Y
Y
Y
Y
Y
Y
Y
N
Y
N
Y
N
100
100
25t
26*
54-2
17-8
37
f*
3
6
N
3-87
6-25
37
m*
76
are rubbers or Y.yes o. N-no, unless otherwise stated. NA.not applicable, -no. available lAC-international assistance and cooperation. Data not collected for indicators 1. 2. 28-33. and 72 at the national
Table 2: National data for indicators of health systems and right to health
We also asked whether the country’s national health
plan explicitly recognises the right to the highest
attainable standard of health {indicator 21). Explicit
human-rights language can be useful for policy makers
and empower disadvantaged individuals, communities,
and populations. On the basis of global data and our
approach, two national health plans secured a yes, four
secured a no, and data were not available for the
remaining 188 countries. From the right-to-healpi
perspective, global data collection is seriously deficient.
Participation and its preconditions
Despite the importance of participation to both health
systems and the right to health, no global data, with our
2<p70
approach, were available for any country for indicator 23.
At the national level, of the five countries, none legally
required participation of marginalised groups in the
development of their national health plan. This finding
suggests that participation is not receiving the attention
it demands: although some countries have made
provision for the participation of citizens, without
specifying marginalised groups. WHO building blocks of
a health system give insufficient attention to the role of
participation.’"
Active and informed participation depends on several
factors. Preconditions for meaningful participation include
having access to information (eg, access tU the health
budget), being free to speak openly without intimidation
www.thelancet.com Vol 372 December 13, 2008
1/02
8002 '£t jaqwaoaa 2ZE PA uiOTjaaueiaqjMMM
■asiuioid Xjduia ue aq ued uopsdppjEd
pijSuTUEaui ‘uotjeuijojut qjjEoq {ejia qans jnoqji^\
(£1—II sJOjEaipui) sqjEapjEUjajEiujojaquinuaqjajqEjTEAE
Ajaqqnd aqEtu puE ‘asijsjjuaa UaqjES jou op saujunoa gg
jsqj paA\oqs EjEp {Eqojf) -uotjeuijojut oj ssaaas
oj uoijuajjE {Epads qjiM ‘suoijjpuoaajd jo uoisuaiuip
{ESapuou '{EoijaEjd-ajoiu siqj passajpps a/x\ ‘ji asn
oq/A asoqj jo XjiAijEaja aqj Suipnpui ‘saauEjsuinajia uo
spuadap ssaujnjasn sjt ajojaiaqj pus ‘jooj e si axej aqj jsqj
sasiuSoaaj ji ‘jaqjEtf -anjEA jnoqjiAx si a;\e{ aqj jsqj usaui
jou saop silll ■uoijEjuauja{duii pus a\e{ uaaAAjaq sjsixa
dsS apiAA e saujunoa aiuos ui ‘XjquiassE pus ‘uoijEpossE
‘uoissajdxa ‘uoijeuijojui jaajoid jqSiui me{ aqj qSnoqjjv
•jsaSSns sjsp asaqj usqj asiOAx si Xjqsaj 'jaAaA\o{-{
isixa jou op uopoiuojd q]|Boq aAipojjo
io uopedprued jtySuTUBam joj suopipuooaxd ‘sauiunoo
asoxp xq •(££ jojeoipux) uoiiETDOsse jo iqSu aqi paiojd
iou soop mej axp saujunoo auxu tn pus 1(95 lOjEDipux)
uoissajdxa jo xuopaajj pajoid iou saop saixjunoo [p
ui mej aqj 1(9 JOiEOipux) uoueluiojut 01 iq§u aqj pajojd
jou saop saujunoD i?i ut mej aqj ‘ajduiExa joq •siojEDipui
jo uorpaps jno ui juasaid 9je suoippuooa-id asaqj jo smog
•uoTjotuojd qjjEaq se qons 'sajruEaj qjjEaq-oj-iqSu .isqiojo
uouEjuatuajdiui jnjoumEauj .ioj sp-iEnSajss XjESsaoau osje
oje ssaqx •(Supaaui oxjqnd e pjoq ‘§a) juaxuipadun jnoqjTM
Supaaui pus ‘(dnoj§ sjuapsd jo uopetdosse jEoipauj
juapuadapui ue qsqqsjsa ‘Sa) uoipppsaj ou qjiM asiusSjo
oj aajj Sujaq ‘({punoo qqsaq jedoj e aspijua Xpijqnd ‘Sa)
(sjeaX) Xjuejoadxa ajpi^ ■squiqaAii OOOOOT Jadopey^ sq^iiqaAH 0001 jadXjinqeqojdt ejidea jad suoissima '03! (%) uoipodoja. |aAa|
6ZZ
VN
•
ZV
S
x-E
A
A
6-XZ
VZ
V-E
6 EX
N
N
N
ZXZ
S8X
S2
8X
N
N
N
V-ZV 80V
8ZX
VZX
N
N
A
A
N
N
VN /V28
2
N
N
N
A
A
A
A
A
A
A
A
A
A
A
A
A
A
8S zs
gs
N
VN
VN
N
A
ZVZ
S8
8V2
ESX
N
N
N
N
N
A
A
XZ OZ
69
lk89 SZ9
*99
tS9
V9
£9
29
X9
09
6S
ssajpaj puE
'Xii|iqEiuno33E 'uopEnjEAa 'Buuoiiuoyy
A
A
Buisiej ssauajEMy
A
A
A
A
A
A
VN
A
A
N
VN
VN
VN
N
N
VN
VN
VN
A
A
VN
VN
VN
EXO
ss
vs
,ES
2S
IS
•OS
spjen6ajes (Euopippy
A
6E
I
VX
V-8
L
SZ
X-9
2
SV
60
A
N
A
A
A
80
l-S
,6V ,8V
N
,ZV
,9V
SV
A
A
A
A
A
A
N
N
N
N
N
N
A
A
A
A
A
A
A
A
N
ZV XV OV 6E
8£
A
A
A
A
N
N
N
N
VV EV
sj3>|.iom qi|EaH
Supueuy |euopeN
)VI
(sjeaX) Xjuejaadxa ajmi sq^jiqaAn 000 001 Jadoyey,. sqpiqaAii
qt-jy paj!un=3vn eauinc) mon endec|=9Nd ’Bisauoniw jo sajej$ pa4ejapaj=y\isj 06003 aqj jo ajiqnday Dijenotuao-Qaa 2!iqndaH ueaujy |ejjua3=8V3 Xjjunoa Xue joj a|qe|ieAe ajaM ejep ou asneaaq uwoqs jou
A
A
A
A
A
A
A
A
VN
EV
088
88
SS
VN
EV
0E8
Z8X
ZOX
VN
OEV
oox
sz
OSX
zx
SX
ZS
XZ
8X
9£
OE
VZ
VV
8E
02
SX
EX
XX
8
Z
N
X9
ZZ
VZ
69
89
SZ
8Z
6Z
8
9
S
VN
8Z
ZE
8
8
VN
Z9
8X
VZ
02
VN
OS
OSS
VEX
8Z
VN
89
8£
XE
VN
£9
OEx
XS
SV
N
EZ
VV
92
VZ
A
ZZ OZ
89
..Z9 1199 1189
6S
VN
VN
2
VN
VN
VN
VN
A
A
N
VN
VN
Z-9
E2
X-8
VN
VN
EE
8X
9'8
622
N
VX
x-S
XS
99X
N
9 SVOX
N
EX
Z-v
VN
8X
VN
N
N
EV
A
A
SO
S
N
vx
X-8
9
V
z-Sx
SSO
OX
zz
Z-8
V00
6X
92
8-Z
Z
Z
A
A
A
A
A
VN
VN
V
A
VN
VN
zo
N
VN
S
98
A
VN
8
EEX
N
9ZZ
A
ZEZ
A
VN
VN
XZ
N
VN
X
2
VZ
VN
VN
VN
VN
8E
62
8-V
VN
9-xx
82
ZS
A
A
A
*0S
I-6V
*8V
+ZV
+9^ SV
N
N
A
N
A
A
A
A
A
A
A
A
A
A
A
88
ZS
98
A
N
8'8
Z8E
N
A
xvs
sov
ZV
E 98
X8
6Z
X-9V
08
V8
OV
EV
OS
Vl 2
06
S
VEO
ZZ
£8
6S
ZS
96
Z6
V6
x6
88
E8
6EZ
XX9
9ZZ
SS
ES
08
02
9X V
28
V9
9Z
6Z
*SE
XVE
1KEE
IkzE
SxE
V00
N
N
8
Z
N
6S-6
SZ8
x6x
*
++
ssajpaj pus 'Xji|iqE;uno3DE
‘luaiussassE 'Buuoiiuoyj
I
A
A
A
A
A
o;iq6iy
VS
*ES
spjenGajES (Euoiiippy
DVI
SS
Bupueuy ieuoiien
VV
EV
8£
SJa>pOM qi|EdH
UZE
!
Right to Health
these data because it shows where the original data come
from,
as opposed to some other sources that do not or
1.92
| §
only use estimates. We identified the method and year of
172
| % 150158
data collection in the extended data tables (webtables 4
| |
and 5) to emphasise that data were obtained from various
sources, including national surveys and the joint
, I .00monitoring programme estimates, and that some data
i
were from 1990.'2' Comparison of data between countries
50is therefore difficult, which is why we did not rank
countries. Romania has less than 20% of its rural
population with access to water, and the difference in
o-|-------- -------CRC
iCERD
access to water between rural and urban populations is
ICESCR
striking (16% vs 91%).121 Data available at the national level
Figure2: Number of countries that have ratified treaties that include the
were
more recent and showed that, by 2005, the gap had
right to health
CRC=Convention on the Rights of the Child. ICERD=lnternational Convention Jn
narrowed a little but was still extreme (34% ps 92%).122
the Elimination of Racial Discrimination. ICESCR=lnternational Covenant on
We aimed to provide a basis to monitor, over time, the
Economic, Social, and Cultural Rights
progressive realisation of the right to the highest attainable
standard of health. The indicator for access to clean water
120 -n
I is especially useful for monitoring a country's progressive
□ Treaty-based grounds (indicator 3)
□ Non-treaty-based grounds (indicator 4.1
realisation. Romania's data for access to clean water in
100 rural and urban areas should be revisited in a few years. If
there is an acceptable measure of improvement, the
8c> government will be able to argue that, in accordance with
its international human-rights obligations, it is progres
sively realising this aspect of the right to health. But if
3 60access to clean water remains the same or becomes worse,
the government will have the burden of proving that all
40 has been done to try to improve access to clean water. If the
government cannot show that all that is possible has been
20 done, it will be in breach of its international human-rights
obligations. This example illustrates the importance of
- H H u 1,1 i,l H 1,1 1,1 H H H II m.rn
n
independent, transparent, and accessible accountability
mechanisms that can decide whether any improvement
that might have occurred is acceptable in the circumstances.
If the government has fallen short of its responsibilities,
accountability mechanisms should consider appropriate
redress, which ranges from guarantees of noh-repetition
I iqure 3: Number of countries protecting grounds of discrimination
to compensation.8’*
Definitional issues restricted data availability for violence
Underlying determinants of health
against women (indicator 26), the indicator of which has
We included indicators on fthe underlying determinants suitable data only for eight countries (Bangladesh, Brazil,
of health (eg, access to clean water, CO? emissions per Ethiopia, Japan, Namibia, Peru, Tanzania, and Thailand
person, and violence against women [[indicators
‘ ”
‘24-26]). from the WHO multicountry study on women’s health
The relation between diarrhoea (and other health and domestic violence against women).'21 Although a lot of
conditions) and access to clean water is well known.y data exist for this subject, definitions vary for violence,
Thus, we asked what percentage of the population has domestic violence, violence against women, and even
access to clean water—defined as at least 20 L per person women (as defined by age), making it challenging to find
per day from a source within 1 km of the user’s dwelling comparable data,
(indicator 24).ll,,'12(’ This indicator raised important defini
tional issues that were briefly considered in the Access to health services
explanatory notes (webappendix 1).
We addressed different dimensions of access, such as
Only 54 countries had more than 90% of their rural those of antenatal care (indicator 27), access to ^.lean water
population with access to clean water and only 115 countries (indicator 24), and catastrophic health expenditure
had moie than 90% of their urban population with access (indicator 46), but more work is needed to identify approto clean water (panel 8). Rural dwellers are disadvantaged priate indicators that measure access. We asked what is the
in most countries for clean water access (figure 4). We proportion of women with a livebirth in the last 5 years
used the joint monitoring programme as the source for who, for their last pregnancy, were seen at least three times
200
ii
■
I
I
nn
y
20/2
y
www.thelancet.com Vol 372 December 13, 2008
Right to Health
by a health-care professional, had their blood pressure
checked, had a blood sample taken, and were informed of
signs of complications (indicator 27). This indicator uses
antenatal care as a proxy for primary care and coverage to
represent access. We obtained data for this composite
indicator from the world health survey with information
available for 51 countries, 19 of which are categorised as
low-income and 22 as middle-income countries. 124 '25
lhe world health survey includes women aged 18 years
and older, and therefore data exclude those under 18 years
who also need antenatal care and might be a group with
reduced access to services. Those women whose pregnancy
did not result in a livebirth are not included and they may
be less likely to have received adequate antenatal care.
Furthermore, women could be reluctant to acknowledge a
livebirth when the child only survived a few minutes, and
could falsely refer to this as a stillbirth. These are a few
examples of limitations associated with this
indicator showing an overestimation of the percentage of
women who received care. Even with these overestima
tions, less than 50% of women had comprehensive
antenatal care in 33 of 51 countries (figure 5).
Similar to the indicator on access to clean water, the
antenatal-care indicator can be used to measure the
progressive realisation of an important aspect of the right
to health. For example, according to global data and our
a pproach, only 12% ofwomen in I ndia have comprehensive
care. India’s data should be revisited in the future to
assess whether the government is progressively realising
this important aspect of the right to health.
We asked whether donors’ international development
policies explicitly include specific provisions to promote
and protect the right to health in recipient countries
(indicator 51), and whether these policies explicitly include
provisions to support the strengthening of health systems
(indicator 52). No global data based on our approach were
available for any donor in relation to either indicator,
Nationally, however, some data were available. National
data in Sweden confirmed that the country’s international
development policies explicitly include specific provisions
on the right to health, whereas national data in Romania
led to the opposite conclusion. Also, national data in
Sweden and Romania confirmed that both countries'
international development policies explicitly include
provisions to support the strengthening of health systems.
We also asked about the percentage of net official
development assistance directed to health sectors
(indicator 53): Italy (17%), Ireland (15%), Netherlands (14%),
and Denmark (11%) seemed to be far ahead, whereas
Japan, Greece, and Luxembourg (all 1%) seemed to be a
long way behind (panel 10).
However, in recent years some donors have moved away
from the provision of funds for specific sectors towards
general budget support. Because of the Paris Declaration
Panels: Civil registration in Sweden
The law on national civil registration in Sweden provides registration of births to
registered parents. When a child is born in Sweden to non-registered parents, such as
undocumented immigrants, the child will not be registered. Information about number
of births to non-registered parents is unreliable. The picture is complicated by the fear of
many non-registered people of Swedish authorities.
Health workers
One of the WHO building blocks of a health system,
health workers have a key role in the implementation
of the right to the highest attainable standard of
health (panel 9). Because of their importance,
11 indicators address issues directly related to them
(indicators 38-44 and 61—64). On the basis of global data
and our approach, we found that 21 countries have a
national health-workforce strategy and 12. do not, whereas
for 161 countries data were not available. Using doctors
and nurses as proxies, we looked at remuneration,
national self-sufficiency, incentives to promote stationing
in rural areas, awareness raising, and human-rights
training. Of the 11 indicators that relate to health
workforce, eight do not have global data available
(including four in the awareness-raising group).
Panel 6: National health situational analysis in Mozambique
Before the development of its national health sector strategic plan, Mozambique undertook
a comprehensive national health situational analysis. This analysis identified health
problems of disaggregated population groups, such as children, and the feminisation of
HIV. It also showed the effect of gender in the fight against HIV, and the need to improve
human resources for health at all levels. It identified possible interventions, such as high
vaccination coverage, to control diseases in children younger than 5 years and the
improvement of campaigns on behaviour change. The analysis emphasised the need to
reinforce support systems and focused attention on the importance of monitoring and
assessing health programmes and services. However, close examination also showed issues
with the quality of the information gathered and data analysis.
Panel 7: Participation in Peru's health councils
International assistance and cooperation
Human-rights responsibility has several components,
including the duty of high-income countries to provide,
and
low-income
countries
to
seek,
international
assistance and cooperation. We set five indicators for
human-rights responsibility in health (indicators 51- 53,71,
and 72). All five indicators focus on the responsibility of
donors rather than on the responsibility of recipient
countries.
www.thelancel.com Vol 3/2 December 13, 2008
Representatives of Peru's ForoSalud (a nationwide civil-society network) obtained
support from the minister of health and the national health council for their proposal to
change the composition of national, regional, and provincial health councils. The existing
composition included nine representatives of health providers and only one of health
service users. The new proposal promotes a more bottom-up approach to participation,
including discussion of new health policies, with a plan to repeat the participatory process
in 2 years and 6 months. The aim is to promote accountability of government officers for
both the achievements and shortcomings of health policy.
20/3
Right to Health
100-1
O Romania
O
Mongolia
80-
DR C ongo Q
(Zaire)
f
(
O
o•
(0
o °oo
o
O 0>
0
i
0
o O
o
Our research suggests that, for indicators 51 and 52, the
relevant global data, based on our approach, are globally
unavailable. We also noted that the relevant national data
were available for Sweden and Romania. Data gathered
for indicator 53 provided a basis to monitor, over time,
donors’ progressive realisation of their human-rights
responsibility of international assistance and cooperation
in health.
(9o
Republic of
Ethiopia
The Congo 0 ©
.
o o
o
O
O n
O 8
o
0
Marshall Islands
O Palau
o
o°
Mozambique
Guinea
©
8 oo
O°
Cape vercfe®
QpAncjola
^Cambodia
0
Vanuatu
0 Afghanistan
Mauritius 0
Honduras
Equatorial Guinea O
;o
40
60
80
100
Rural population (%)
Figure 4: Percentage of urban and rural populations with access to clean water
ISI=157 countries.
i
20-i
i
I
15-
i O 10-
5-
0-20
>20-40
>40-60
>60
Women (%)
Figure $: Percentage of women by number of countries who had
comprehensive antenatal examinations during their last pregnancy
N=51 countries.
Panel 8: Access to clean water in Mozambique
In Mozambique, poor access to clean water is responsible for serious illnesses and leads to
frequent outbreaks of cholera. Access to clean water has improved in rural communities
through close coordination with the ministry of public works and with support from various
donors. Recent data confirmed this improvement and showed a higher rate of access to clean
water in rural (48-5%) than in urban (40%) areas. However, urbanisation has taken place
across Mozambique in recent years, leading to a large influx of people into cities and,
therefore, in provision of the necessary infrastructure to ensure access to water in urban areas.
on Aid Effectiveness, this trend is likely to accelerate
For this reason, a ’brief narrative should accompany this
indicator to explain each country’s approach to overseas
development assistance and health, for example, to signal
if the donor is moving away from sectoral support towards
general budget support. Donors’ support should be in
line with the growing international recognition that, if the
health-related Millennium Development Goals are to be
achieved, health systems must be strengthened.
20/4
Monitoring, assessment, accountability, and redress
Eight indicators were grouped together under monitoring
and accountability (indicators 65-72); however, several
others might properly be regarded as monitoring and
accountability indicators, such as those on impact
assessments (indicators 19 and 20).sx
Monitoring and accountability depend on the availability
of reliable and relevant data. Without indicators,
benchmarks, and data, it is not possible to monitor the
progressive realisation of the right to health. Several
traditional health outcome measures have a key role. A
worsening health outcome, such as maternal mortality,
does not necessarily mean that a country is failing its
right-to-health responsibilities. However, it obliges a
country to explain to an appropriate accountability body
why the situation is deteriorating.
We took into account three mortality measures: infant
mortality (indicator 65), mortality’ of children younger
than 5 years (indicator 66), and maternal mortality ratio
(indicator 67), together with life expectancy (indicator 68).
Of these, the one with the least global data available was
maternal mortality ratio. Data exist for 169 countries with
a range from 1 (Ireland) to 2100 (Sierra Leone)
per 100000 livebirths (mean 331).I?h This range is
astounding because most deaths are preventable and a
high ratio shows that the health system is failing.
As with other indicators already discussed, such as
antenatal care and access to water, indicators 65-69 can
be used as a basis to monitor, over time, aspects of the
progressive realisation of the right to health.
Ill countries have national human-rights institutions,
many of which make a substantial contribution to the
promotion and protection of human rights. Although
independent, these institutions are non-judicial and
designed to be more accessible, flexible, and informal
than courts.127 One of their functions is to monitor and
hold governments accountable. We asked whether a
country has a national human-rights institution with a
mandate that includes the right to the highest attainable
standard of health (indicator 69). Is the institution
empowered to monitor public and private health and
hold accountable those with right-to-health responsi
bilities? No global data for this indicator were available
with our approach for any country. Data were available,
however, nationally. Of the five countries, three (Ecuador,
Peru, and Romania) have national human-rights
institutions, and the mandate of each extends to the right
to health. With leadership and resources, these
www.thelancet.com Vol 372 December 13, 2008
I
J
Right to Health
institutions could contribute greatly to constructive
accountability for the right to health (panel 11)."
We also asked whether national human-rights institu
tions have a mandate to monitor donors’ human-rights
responsibility of international assistance and cooperation
(indicator 71). However, no global data based on our
approach were available for any donor for this indicator.
National data showed that, although three of the five
countries have national human-rights institutions, none of
these institutions has a mandate to monitor international
assistance and cooperation in health.
Countries that have ratified international human-rights
treaties have an obligation to report on their activities
related to that treaty usually within 2 years.5 A committee
of independent human-rights experts publicly considers
the report and may ask country’s representatives
challenging questions about the government’s record,
publishing its concerns and recommendations. A few
years later the process is repeated and the experts’
committee asks the country to explain what has been
done in relation to the earlier recommendations. Under
the International Covenant on Economic, Social, and
Cultural Rights, donor countries should report on their
human-rights responsibilities of international assistance
and cooperation in health. Therefore, we took into account
country's reporting on international assistance and
cooperation in health to the UN Committee on Economic,
Social, and Cultural Rights (indicator 72). We showed that
only three donor countries (Belgium, Canada, and Japan)
reported adequately on this crucial issue.
Low-income countries have the perception that highincome countries escape accountability when failing to
fulfil their international pledges and commitments that
are important to developing countries.*”1 Data seem to
confirm that this perception is right. Our national data
suggest that the mandates of national human-right
institutions do not extend to holding donors to account for
their human-rights responsibility of international assist
ance and cooperation. Also, most donor countries are not
being held accountable by a key UN human-rights treaty
body for their responsibility of international assistance and
cooperation in health. We conclude that donor countries
are subject to only feeble independent, institutional
scrutiny for their international responsibilities.
In 63 countries, the constitution, bill of rights, or other
statute recognise the right to health. Legal recognition
serves many purposes. In several countries, for example,
it made the authorities accountable before the courts,
leading to tangible improvements in health services.55-56
We, therefore, enquired about the number of judicial
decisions that took into account the right to health
in 2000-05 (indicator 70). However, it is possible that
the decision may not promote and protect this
fundamental human right. Nonetheless, even in this
case, at least the country is held accountable for the
right to health, having to explain itself before an
independent accountability body.
www.thelancet.com Vol 372 December 13, 2008
Global data disclosed, on the basis of our approach,
only five countries (Canada, Nigeria, Peru, South Africa,
and Venezuela) with judicial decisions conforming to
this indicator. The combined number of judicial decisions
is less than ten. This is striking because in recent years
numerous national courts have decided right-to-health
cases, and yet our data show that they are not globally
available in accordance with our approach.55-56
Here, we looked at some of the right-to-health features
and their data, and conclude that health systems in
numerous countries do not have the features required by
the right to health. Also, there are insufficient data
currently available, especially at the global level, to assess
these indicators in relation to many countries. Figure 6
Panel 9’ Health workers in Romania
Romania has one of the lowest densities of health personnel in Europe in relation to
doctors, nurses, dentists, and pharmacists. Furthermore, there is an imbalance between
regions of the country. In rural areas, there are 98 communities (villages) without any
doctor, and the situation is comparable for nurses and other health workers. In a third of
Romania, more than 30% of medical specialties are not available. The accession of
Romania to the EU aggravated the situation with 10% of doctors seeking work outside
Romania, according to a recent survey of the Romanian College of Physicians. This
situation increases already existing inequities between rural and urban populations.
Payment types currently compensate rural better than urban clinicians for the same type
of services. However, the existing additional benefit payment programme has not met its
goal of providing a sufficient health workforce in the underserved areas. These data
suggest that a comprehensive approach is needed to tackle more than the financial
dimension of the issue. The ministry of health has proposed additional incentives to try to
increase and stabilise the number of health workers in rural areas.
Panel 10: Official development assistance for G8 member
countries (global level data)
ltaly-17-48%
USA-7-2 6%
Canada—6-43%
France—3-74%
Germany—2-93%
UK-237%
Japan—113%
Russia Federation—Not available
Panel II: Monitoring and accountability in Peru
Over the past year, Cooperative for Assistance and Relief Everywhere (CARE)-Peru and
Physicians for Human Rights have supported the development of citizen and civil-society
accountability mechanisms at both district and local levels. An example is in the Piura and
Puno regions, where Quechua and Aymara women community leaders have been linked
to regional offices of the human-rights ombudsman to monitor women's health rights,
particularly their right to good quality, appropriate maternal health services. Rural
women's leaders have also been empowered by a joint agreement between ForoSalud
and the human-rights ombudsman office in Puno. Partnerships have been mutually
enriched, with women leaders feeling better positioned to demand information and
changes in health services.
20/5
i
ii Rightto Health
shows the unavailability of global data in relation to a
selection of 25 indicators for which it is clear that, in
many cases, international bodies are not collecting the
appropriate right-to-health data. An overarching
conclusion is that those at the international and national
levels with responsibilities for health systems seem to be
giving inadequate attention to the right-to-health analysis
and some of the features required by the right to highest
attainable standard of health.
Opportunities and challenges
On a country-by-country basis, table 1 summarises the
degree to which health systems of countries include
some features that arise from the right to health in
relation to 72 indicators. Fable 2 summarises natiorial
2: Constitution or other statute
recognises right to hearth
data for the same indicators in relation to five countries.
When considering the performance of art individual
country, the country’s stage of economic development
(what human-rights treaties refer to as the countries
resource availability) is important.
Some of our findings are positive: for example, we
record high rates of vaccination with measles-containing
vaccine (MCV) and diphtheria, tetanus, pertussis (DTP3)
vaccine (indicators 34 and 35). General comment 14 places
a high priority on immunisation programmes. '6 Although
such programmes can occur as vertical interventions,
whenever possible they should strengthen health
systems. Weak health systems impede high immunisation
coverage1" and the GAVI Alliance and Fund Boards
recently increased funding for health systeM strength-
6.:<
EZ
6: State law protects the right
73
tc- information
7 State law requires registration
of births and deaths
121
TT
188
8: State has civil registration
I 9 I
63
11 Stale caliects data for
69
maternal deaths
12: State centralises data for
maternal deaths
69
13: State makes available data
for maternal deaths
69
14. State collects data for
neonatal deaths
88
37
88
3Z
88
3/
S3
17
15: State centralist's data for
neonatal deaths
S3
17.1
17
53
124
17
16 State makes available data
for neonatal deaths
17 State has compichensive
national health plan
13
T
18 State undertook health
57
sit»ationa= analysis
2- Nationa health p;.in lectigmses
nght totif-a4.il
137
4-f
188
22. National health plan includes
[.... :~t~ "T
t ommitment to universal access
28: Access to medicines in
l6S
ZE
constitution or olhei statute
20
188
29: Published national medicines
99
policy
43
30: Published list of essential
medicines
128
5?
.4?’
18
38: State has national health
21
workforce strategy
161
45: Health expenditure greater than
minimum required
122
54: State law protects
confidentiality of persona! data
26
20
70
45
123
55: State law requires informed
18
cons<?nt
164
56: Constitution protects freedom
of expression
132
41
<-institutiori protects freedom
of association
T~7~T
58: State has patients' rights
charter
12
39.1 atii.-iiis rights char ter
available m official languages
3211'
Yes
No
n/av
0
I
23
Z0
__J
182
184
20
40
60
80
100
120
140
160
180
f itjtjre 6 Availability of global data for a selection of 25 indicators
n/av-not available.
20/6
www.thelancet.com Vol 372 December 13, 2008
Right to Health
ening to US$800 million.IW We recorded good practices
for the implementation of right-to-health features of
health systems, such as Mozambique’s effective coordina
tion between different stakeholders to improve access to
safe water in rural communities (panel 8). Also, Sweden
has recently introduced the legal requirement for
inclusion of human rights in the training curriculum of
health workers, reflective of indicators 63 and 64.I2‘J
Countries that have not put in place some of the key
features of a health system—eg, a comprehensive national
health plan (indicator 17), a published national list of
essential medicines (indicator 30), a national health
workforce strategy (indicator 38), or government expen
diture on health per person above the minimum required
for a basic effective public health system (indicator 45)—
are in breach of their right-to-health responsibilities
whatever their stage of economic development.
We provide a basis on which to monitor health systems
and the progressive realisation of the right to health. A
suitably improved version of this project repeated in a few
years’ time will give an indication of whether countries
have progressively realised the right to health. Take, for
example, a high-income country in which prevalence of
violence against women has increased (indicator 26);
access to health services (indicator 27) has worsened;
immunisation of 1-year-old children against measles
(indicator 34) has decreased; total government spending
on health as a percentage of GDP (indicator 47) has
lessened; and life expectancy has fallen. In such a case,
progressive realisation of the right to health has not been
achieved. Unless the government has a rational, objective
explanation for the worsening situation (eg, a natural
disaster), this country would be in breach of its right-tohealth responsibilities.
We have emphasised the limitations of indicators
generally and some of our indicators specifically.
Moreover, as we have seen, there are some inconsistencies
between global data (table 1) and national data (table 2).
So the data must be used with caution.
In this Report, we concentrate on recognition of the right
to health; non-discrimination; health information; national
health plan; participation; underlying determinants of
health; access to health services; health workers; inter
national assistance and cooperation; and monitoring,
assessment, accountability, and redress. We do not discuss
some important right-to-health features and their data,
such as finance and medicines, because space restrictions
compelled us to be selective.
Limitations to indicator selection and data collection do
not contradict the profile of indicators or the findings, but
should be considered when analysing the results. We hope
data collection was not included in the methods, except
when checking for an unexpected result. The number of
indicators is large, but we did not want to compromise
too much. When the project is repeated, we suggest an
assessment of concordance of the data. Although one of
the objectives of this project was to assist with monitoring
progressive realisation of the right to health, some of the
indicators, such as maternal mortality ratio, are not
sensitive to change over short periods. Additionally,
so-called yes or no indicators do not lend themselves to
measurement of gradual change over time, although
many are complemented by a commentary (webtables 4
and 5) to explain the result that could indicate
improvement with time.
For this project, restrictions on collection ofworldwide
data were needed. However, the one-click rule
introduced elements of inconsistency in the data
collection. Some regions tended to have more
information available within the limits of the one-click
confines than others—for example, lists of WHO
member states each link to some documents about that
country. For some regions, detailed information was
available on global websites, whereas, for others, global
websites were structured in such a way that the one-click
rule allowed only an index page for the region, not the
actual information, to be reached. Also, there was a risk
of further discrepancy in data available for developed
countries and those available for developing countries.
This discrepancy might arise because international
organisations assist developing more than developed
countries in data collection or analysis, and these data
are subsequently more readily available worldwide.
Our research shows that insufficient data are
available, especially at the global level, in relation to
right-to-health features. From the perspective of
health systems and the right to health, UN bodies and
other international stakeholders are not collecting
appropriate data. International and national
institutions with responsibilities for health systems
seem to be giving inadequate attention to the
right-to-health analysis and some of the features
needed for the right to the highest attainable standard
of health. Here we focus on other areas of concern
and make recommendations. We do not attempt to
discuss all our concerns and recommendations but
focus on those arising from some key areas. Taking
into account both the right-to-health analysis outlined
in this report and the data collected, we make
additional recommendations in panel 12.
that the profile of indicators will prompt discussion and
Recognition of the right to health in international
treaties, national constitutions, and other statutes gives
rise to a legal obligation for countries to ensure that
their health systems have certain features, as discussed,
and also that the performance and quality of health
systems do not regress or stagnate but improve over
that subsequent revisions will make indicators more
robust.
This project relies on secondaiy data published by
others, and any limitations of the primary data affect our
dataset. Because of resource constraints, triangulation of
www.thelancet.com Vol 372 December 13, 2008
I
Recognition of the right to health
2077
Right to Health
Panel 12: Recommendations
We recommend WHO and the Office of the High Commissioner for Human Rights
•
adopt a stewardship role in the collection and collation of data for right-to-health
features of a health system
•
lead the process to establish universal definitions for commonly used terms and standard
•
ised units of measurement regarding the right-to-health features of a health system
maintain and regularly update a global data repository on the right-to-health features
of a health system
•
lead the process to establish, where appropriate, international benchmarks to assess
country performance regarding the right-to-health features of health systems
•
ensure that the Global Health Workforce Alliance gathers data relevant to the right to
health, such as for human rights training for health workers
We recommend other UN specialised agencies and bodies
•
coordinate with WHO, national governments, civil-society organisations, and other
international, regional, and national stakeholders to ensure coherence in global
monitoring with respect to the right-to-health features of a health system
•
provide technical assistance to national governments to facilitate data collection on
the right-to-health featuresofa health system
■
record descriptive and numerical data
•
cooperate with WHO, national governments, civil-society organisations, and other
•
standardised units of measurement of right-to-health features of health systems
ensure that the activities of the UN specialised agency or other body are aligned with
the comprehensive national health plan
relevant stakeholders to establish universal definitions for common terms and
We recommend national governments
•
explicitly recognise the right to health, and right-to-health features, such as access to
essential medicines, in the national constitution or statute
•
ensure explicit recognition of the right to health in the comprehensive national health
plan
•
ensure sufficient expenditure on medicines to provide, as a minimum, equitable
access to essential medicines
•
collect data on marginalised groups to inform the planning and development of the
health system
•
do health and human-rights impact assessments before finalising the comprehensive
national health plan
■
in partnership with WHO, UN specialised agencies, civil society, and others, collect and
■
disaggregate data on at least the five priority prohibited grounds of discrimination
regularly update information on right to-heal th features of health systems
sex, age, ethnicity, socioeconomic status, and rural or urban residence
•
regularly submit information updates about the right-to-health features of health
systems to the global data repository maintained by WHO/OHCHR
•
cooperate with the WHO, UN bodies, and others in establishing national and
•
•
ensure registration of births and deaths within a civil registration system
establish national human-rights institutions with a mandate that includes the
international benchmarks to monitor the right-to-health features of health systems
right-to-health and budgeted programme of activities for raising awareness about the
right to health
•
•
ensure the mandate of the national human-rights institution includes monitoring and
accountability with respect to international assistance and cooperation in health
include compulsory human-rights and right-to-health training for health workers,
judges, and lawyers
•
submit timely, full reports to the UN Committee on Economic, Social, and Cultural
Rig fits and other relevant UN treaty bodies
(Continues on next page)
20/8
time. However, most countries (121 of the 184 for which
data are available) do not recognise the right to health in
their national constitutions or other statute, although
every country has ratified at least one international treaty
that recognises the right to health. Recognition of the
right to health has not only generated judicial decisions
that have improved the delivery of health-related services, ’
it has also led to non-j udicial mechanisms ofaccountability,
such as the Right to Health Unit established by the
Uganda Human Rights Commission,w and led to
enhanced health policy and practice. For example, the
Department of Health in the United Kingdom recently
commissioned an assessment of the effectiveness of
implementing a human-rights based approach in health
and social care."" Focusing on five pilot projects, the
assessment concluded that such an approach had a
noticeable effect on the treatment and care ofhealth-service
users, and that it is one way of achieving good practice.
'Ifius, we recommend that countries ratify treaties that
encompass the right to health, explicitly recognise this
human right in their national constitutions or other
statute, and integrate the right to health into their national
health plans?6
Health information
Health information is the life-blood of effective, accessible
health systems and the right to health. Information
enables individuals and communities to promote their
own health and allows governments to formulate
evidence-based health plans. Monitoring, accountability,
and participation depend upon access to information.
Without reliable disaggregated data, whether health
systems are delivering access to services and facilities
without discrimination is impossible to know. However,
our research suggested that health information systems
in many countries are seriously deficient in several
ways.
Health information systems include a range of data
sources, such as censuses, household surveys, vital
registration systems, and other health-facility data
sources.'" Ideally, systems should not duplicate but
complement each other, providing accurate information
making best use of limited resources."'1 Data for maternal
and neonatal deaths, for example, should be included
within the vital registration system, but the data then
need to be specifically extracted, a position supported by
the WHO Maternal Mortality Report in 2005.126 Of
194 countries, 122 have a civil registration system;
however, many of these are incomplete, with fewer than
90% of events registered. Accurate recording of cause of
death according to international standards is important,1"
and the statistics generated should then form a part of
the situational analysis that contributes to health-system
planning. We recommend that all countries should
legally require registration of births, deaths, and cause of
death according to international standards (using the
international classification of disease)."4
www.thelancet.com Vol 372 December 13, 2008
Right to Health
The USAID-supported Demographic Health Surveys,
the Multiple Indicator Cluster Survey programme
developed by UNICEF, and other surveys provide much
information. Data provided by these surveys were used as
sources for many indicators in this profile, and for some
indicators, such as those considering access to water and
vaccination coverage (see extended data in webtables 4
and 5) the survey type and year were recorded along with
the score. Although surveys fill an important role, they are
not the best long-term solution to collection of routine
data if not done regularly and they should be included in a
civil registration system that is permanent, continuous,
compulsory, and universal.'l? At the very least, a system is
needed to record reliable information on birth and death
registration in all regions. Such a system would have
implications for several human rights. Without birth
registration, many entitlements may be denied throughout
life, such as access to health care, education, international
travel, and the right to own property. However, data for the
percentage of births registered in rural and urban areas
are only available for 78 countries. Furthermore, only
69 countries regularly collect, centralise, and make
publicly available data throughout the territory for
numbers of maternal deaths. We recommend, at a
minimum, continuous registration ofbirths and deaths in
all areas within a vital registration system.
Descriptive information is needed to understand the
issues behind quantitative data; however, we often found
a shortage of this information. In this project, all
indicators have been reduced to a number or to yes or no
answers, but the accompanying data and commentaries
(webtables 4 and 5 and webappendix 1) are crucial to
understanding our findings. Some issues are difficult to
accurately capture with purely quantitative data, and we
encourage greater emphasis on a complementary brief
narrative in some cases."’
As part of their human-rights responsibility of
international assistance and cooperation in health,
donors (webpanel) should accelerate their coordinated
efforts to provide training and technical assistance for
sustainable data collection and processing and to make
data available worldwide. Additionally, donors should
facilitate the establishment of national health information
systems, including a comprehensive civil registration
system in all countries, with clear mechanisms for
relaying this information to a globally accessible data
repository. WHO and the UN Office of the High
Commissioner for Human Rights should have a
leadership role in establishing and maintaining a global
system for collection and collation of up-to-date
information from different countries and UN bodies on
right-to-health features of health systems.
Disaggregation
Policy makers and health practitioners need accurate
information about marginalised groups as many are at
risk of worse health,1’6 because, in many cases, of
www.thelancet.com Vol 372 December 13, 2008
(Continued from previous page)
We recommend national and international civil society
•
participates in health system planning and monitoring
•
advocates that right to health is properly incorporated in health system planning
•
advocates the inclusion of marginalised groups in health decision making
•
advocates that the mandates of national human-rights institutions include the
right-to-health and budgeted programme of activities for raising awareness of
•
disseminate information about key judicial decisions about the right to health
•
ensure that the activities of the civil society organisation are in alignment with the
comprehensive national health plans
right-to-health
We recommend research institutions
•
provide assistance to national governments to do health and human-rights impact
assessments
•
•
do or commission research on the right-to-health features of health systems;
actively promote knowledge sharing among academics on the right-to-health and
•
collaborate with national governments, WHO, UN bodies, civil-society organisations,
right-to-health features of health systems
and others to promote a greater understanding of the right to health and
right-to-health features of a health system
We recommend donors
•
recognise the importance of strengthening health systems in international assistance
strategies
•
allocate greater funding for health in low-income and middle-income countries
•
ensure donor accountability for international assistance and cooperation in health in
•
both donor and recipient countries
align international assistance and cooperation strategies with the comprehensive
national health plans of recipient countries
problems in accessing health-related services. From the
human-rights perspective, the goal is to disaggregate data
in relation to as many of the internationally prohibited
grounds of discrimination as possible (indicators 3 and 4),
although some data cannot be disaggregated. The
collection of disaggregated data remains an enormous
challenge for many countries and, because of limited
capacity, reliable disaggregated data are often unavailable.
The contextual natures of vulnerability and discrimination
further hamper collection of relevant data: a group might
be especially vulnerable in one context but not in another.
We had difficulty identifying indicators that captured
vulnerability and marginalisation, probably because of
their contextual nature. We recommend five priority
grounds—sex, age, ethnic origin, rural or urban residence,
and socioeconomic group—for disaggregation as a
minimum.''8 These ground are similar to those identified
by the Health Metrics Network.’1’
Some health issues demand disaggregation on
particular grounds; for example, in the context of sexual
and reproductive health, disaggregation on the basis of
sex and age are crucial. Of the 21 countries with
information about the proportion of men and women
with comprehensive knowledge of HIV/AIDS, men fared
better than did women in 16, suggesting that different
20/9
Right to Health
I
|
I
!
strategies for education of women are required. The
Guidelines on Construction of Core Indicators in relation
to monitoring the Declaration of Commitment on
HIV/AIDS note that an important lesson from the
UNAIDS Global Progress Report of 2003 was that without
disaggregated data monitoring of access, equity, and
change over time is difficult.'” The guidelines
acknowledge that this disaggregation requires effort, but
point out that such data are commonly collected at the
subnational level but subsequently lost when passed to
the national level. Another recommendation therefore is
that coordinated efforts are made to collate and present
reliable data at the national and global levels broken
down on the basis of the five priority grounds.
Coordination of data
Because the right to health involves policies and practices
that lie beyond the health sector, effective coordination is
needed between different sectors (including health,
transport, environment, and education) and different
health services. General comment 14 recognised coordina
tion as a right-to-health feature of health systems.1'' Data
collection is one example of where coordination is
needed—as happens with global maternal mortality data.
UNICEF. WHO, UNFPA, and the World Bank periodically
assess maternal-mortality data.1"1 As a result of joint
assessment the data on maternal mortality are the same at
WF10 and UNICEF. However, life expectancy data differ
between the two sources; for example, WHO reports a
life expectancy of 61 years in Namibia, UNICEF reports
52 years. The cause of this large discrepancy is unclear.
We recommend coordination and collaboration between
countries, regional stakeholders, UN bodies, and others to
establish a global repository for health data with up-to-date
and consistent reporting.
For some indicators more information is available
worldwide for low-income than for high-income
countries, such as the data on DTP3 and MCV
vaccination.11'' However, high-income countries have the
resources to collect and process the relevant data, which
should be logged in a global repository.
Standardisation
When developing the indicators and collecting data, brie
major dilliculty was the lack of universal definitions for
many commonly used and important terms, such as
clear definitions for rural and urban. Disaggregation of
data on these grounds was therefore difficult; in practice,
individual data collectors decide whether a location is
rural or urban.'4" Similarly, violence against women does
not have a standard definition.14' Specific criteria and
definitions should be created, although they might not
capture every nuance of the relevant issue, to allow
consistency of data collection and comparison over time.
Standard definitions are needed to create national and
international benchmarks against which to measure a
country's progress. General comment 14 anticipates that
2080
countries and the UN Committee on Economic, Social,
and Cultural Rights will agree key indicators and then
identify appropriate benchmarks or targets to be achieved
by the country over the next few years in relation to the
selected indicators.1'’ Agreed indicators and country-specific
benchmarks are needed to measure progressive realisation
and hold governments to account. Sometimes benchmarks
are agreed internationally for all countries, or a group of
countries—for example, all donors should deyote 0-7% of
their gross national income to overseas development
assistance, and all African leaders have pledged to allocate
15% of their annual budgets to healtht'42 Whether
benchmarks are set nationally or internationally, standard
definitions are important. For example, CO? emissions are
recorded in various ways, which makes benchmarking
difficult.141 Although indicators and benchmarks are vital if
we wish to measure progressive realisation and hold
countries to account, their usefulness depends on widely
agreed definitions.
Standard formatting for data collection would be
especially helpful in low-income and middle-income
countries that have to collect similar information in
different formats to fulfil the demands of different donors
and international bodies.
We warmly welcome the Health Metrics Network, a
global partnership of UN bodies, donors and others aiming
to improve health information at country, regional, and
global levels.5,1 The network hopes that by 2011 its detailed
framework and standards for country health information
systems will be the universally accepted standard for all
developing countries and global agencies. The network
comments on the inappropriate use of data collection
methods (eg, surveys used to record adult mortality) and
advocates the disaggregation of health-status data. We
recommend that the network encourages the collection of
data outlined in this report, which is needed to measure
the right-to-health features of a health system.
Access to information is part of the right to health.
I lowever, in addition, information is protected by national
and international codes of civil and political rights."4
Many of these codes provide stronger accountability
mechanisms than are available to the right to health. We
recommend that human-rights workers in the domain of
civil and political rights use their expertise to improve
access to health information, following the example of
the London-based civil society organisation Article 19.
Comprehensive national health plans
Within the confines of our methods, no data currently are
available at the global level to show that any country has a
comprehensive national health plan, whereas 13 countries
have data available on the WHO website indicating that
they do not have a comprehensive national health plan.
We recommend adoption of a universal definition of a
comprehensive national health plan. Countries should
develop comprehensive national health plans consistent
with defined criteria, including budget allocation for all
www.thelancet.coni Vol 372 December 13, 2008
Right to Health
proposed activities. Information about all such plans
should be available both nationally and internationally.
Monitoring and assessment of these plans both nationally
and at the global level are also needed to ensure com
pliance with agreed criteria. Any gaps identified should
be systematically addressed; for example, if a national
health plan does not encompass the private sector then
this shortcoming needs to be identified and remedial
action taken.
Appropriate national and international human-rights
bodies- should monitor whether or not a country has a
comprehensive national health plan conforming to the
agreed criteria. For example, a national human-rights
institution should check whether or not the government
has an appropriate national plan. Most national human
rights institutions report annually to the legislature and
the status of a national plan could be publicly reported this
way. Also, appropriate international committees ofhuman
rights experts, such as the UN Committee on Economic,
Social, and Cultural Rights, should routinely ask countries
appearing before the committee about their comprehensive
national health plans. The UN Human Rights Council has
recently established a new procedure that all countries
must follow. Known as the universal periodic review, a
comprehensive national health plan is so important that
the council should routinely ask all countries about the
status of their plan.145
Monitoring, assessment, accountability, and redress
We recommend that much closer attention be devoted to
establishing accessible, transparent, and effective mech
anisms for monitoring and accountability of health
systems and the right to health. Analysis of the data
collected in this project reveals weak mechanisms at
international and national levels. Without indicators and
reliable data, neither the condition of health systems nor
the progressive realisation of the right to health is
possible. Accountability, however, is much more than
monitoring. Organisations and individuals with right-tohealth responsibilities must be held to account in relation
to the fulfilment of their duties, with a view to identifying
successes and difficulties—what freedman" calls con
structive accountability. In this way, accountability
strengthens health systems.
As explored by Potts,814 there are many different mech
anisms of monitoring and accountability for the right to
health—social, political, administrative, quasijudicial, and
judicial—each with a crucial role. Our indicator 69 tried to
address quasijudicial accountability by asking if countries
have national human-rights institutions with mandates
that include the right to health. However, there was no
globally available data for this indicator revealing either a
shortfall in the data available or in the mandates
themselves. We recommend that national human-rights
institutions include the right to health in their mandates
and budgets for programmes.146 In close collaboration with
the health sector, these institutions could provide
www.thelancet.coni Vol 372 December 13, 2008
human-rights training for health workers, raise public
awareness of right-to-health entitlements and processes,
work with public officials to integrate the right to health
into policies, help to prepare right-to-health protocols and
guidelines for health workers, monitor right-to-health
features (eg, comprehensive national health plans and
international assistance and cooperation in health), and
undertake independent public inquiries into particular
right-to-health issues, hold those responsible to account,
and make recommendations. Australia's Human Rights
and Equal Opportunity Commission has made health
inquiries on several occasions, for example, in its Social
Justice Report 2005 that applies human rights to
indigenous health policy.147 With the recent change of
government in Canberra, this important report is now
shaping health policy and practice. National human-rights
institutions should forge strategic partnerships with the
media, health workers, patients’ groups, judges, lawyers,
academics, and others.
Mechanisms of social accountability include public
hearings and social audits. The People's Health Movement
in India has set up the People's Rural Health Watch to
conduct independent health monitoring in seven states in
northern India. This initiative supplements the community
monitoring that is part of the National Rural Health Mission
launched by the government in 2005.88 In some situations,
Medecins Sans Frontieres, after listening to patients to
understand why a situation is occurring, will draw attention
to the issues, problems, and responsibilities.148
One example of administrative accountability is
addressed in the indicators for impact assessment (a
process through which the potential effects of a policy,
programme, or plan on the health of the population is
assessed). However, data for the national level show that
none of the five countries had made a health impact
assessment or any impact assessment including the right
to health before the implementation of their national
health plan. We recommend that countries make such
assessments before adopting their national health plan;
there should at least be an impact assessment in relation
to key elements of a plan. Such assessments can be
crucial to progressive realisation of the right to health.75
Effective monitoring and accountability depends on
numerous factors, including the recognition of the right to
health as a legally enforceable right. Where the rule of law
is respected, it helps to be able to say to the relevant
minister, local health council, or hospital director that a
particular health initiative is not only ethically appropriate
and good practice but is also required under binding
national and international human-rights law. Account
ability is not judicial accountability. As discussed, judicial
■h1I1
—i
accountability is one narrow, limited form; moreover, it is
accountability of last resort. Judges and lawyers must be
willing to learn about, and apply in a balanced manner,
such right-to-health concepts as progressive realisation,
resource availability, core obligations, and disadvantage;
they must be willing to listen to health experts and those
2081
I
Rightto Health
j
Ij
I
using health-related services. Information about key
judicial decisions for realisation of the rightt to health
should be widely accessible. Also, countries must report
regularly on regional and international treaties that they
have ratified.
Redress is another important component of accountability.1*” It comes in many forms, such as full and public
disclosure of the truth, apology, acknowledgment of
responsibility, a change in policy, law reform, rehabilitation
(eg, the provision of health-related services), and compensation.
Finally, careful attention must be given to the
human-rights accountability of international bodies, as
well as the private sector. The Human Rights Guidelines
of Pharmaceutical Companies in relation to Access to
Medicines can deepen the accountability of the
pharmaceutical sector.149
Additional research
This project highlights the need for more research on the
right to health. For example, what are the core obligations
signalled in paragraph 43 of general comment 14?
Research is needed on the application of the right to health
to the six WHO building blocks of a health system and
within both public and private sectors. More attention
should be devoted to right-to-health features of health
systems: for example, what are appropriate mechanisms
of monitoring and accountability? More research is needed
on the most appropriate indicators for assessing the
degree to which health systems include these
right-to-health features. We had particular difficulty
identifying appropriate indicators in relation to access
(including access for marginal groups), respect for cultural
difference, quality, participation, referral systems,
standards (ie, provisions that elaborate in more detail
upon the general right-to-health formulations found in
treaties, constitutions, and statutes), coordination (ie, the
need for effective coordination across a range of public
and private stakeholders, at the national and international
levels, both within and between health-related sectors),
and monitoring and accountability. We recommend that
particular thought be given to identifying appropriate
indicators for these issues. Echoing Gruskin and
coUeagues,19 we also recognise the need to build evidence
of the effects of the application of the right to health on
health systems.
Over 18 months of research, our interdisciplinary project
has depended upon the insights of experts in both health
and human rights. UN bodies, non-governmental
Contributors
As project director, GB contributed to the drafting of all sections and
explanatory notes and collected global data. PH contributed to the
drafting of all sections. GB and PH prepared figure 1. RK contributed to
the drafting of the introductory, key findings, and opportunities and
challenges sections and explanatory notes and collected global data. C) S
and BMP contributed to the drafting of the methodology section and the
explanatory notes and collected global data. CR contributed to the
drafting of the methodology, key findings, the opportunities and
challenges sections and explanatory notes and collected global data. DP
contributed by advising on statistical analysis and developing several and
reviewing all of the figures in the key findings section. DA and MAP
organisations, policy makers, academics, and others
collected national data on Ecuador. AF and DT collected national data on
have made indispensable contributions. We strongly
recommend that all those sharing the common ground
between health and human rights deepen their dialogue,
cooperation, and collaboration. Our findings have impli
cations for professions and institutions at all levels and
Conflict of interest statement
We declare that we have no conflict of interest.
Conclusion
2082
in both public and private sectors. For example, health
ministries and national human-rights institutions need
to meet and talk, and UN organisations must routinely
discuss health and human-rights issues. For example,
WHO, UNFPA, and the World Bank must engage with
the UN Human Rights Council, Office of the High
Commissioner for Human Rights, and human-rightstreaty bodies. All these organisations—and many
others—have the common aim of strengthening health
systems.
Countries have a legal obligation to progressively realise
the right to the highest attainable standard of health and
therefore to improve their health systems progressively.
Indicators and benchmarks are needed to measure
present conditions of health systems and to monitor them
over time. Indicators selected in this profile and methods
of data collection have limitations, but the findings have
generated several recommendations. We are drawn to the
conclusion that those with responsibilities for health
systems are giving inadequate attention to the
right-to-health analysis. Our main, overarching recommendation is that all those with health-related responsibilities explicitly consider the right-to-health analysis and
integrate this human right into their policies and
practices, with a view to strengthening health systems.
We hope that this project will be repeated periodically so
that the progress of individual countries will be monitored.
No doubt improvements will be introduced to the
methodology and profile of indicators before the exercise
is repeated in a few years’ time.
This project rests on the conviction that an equitable
health
health system
system isis aa core social institution, no less than a
fair court system or democratic political system. Because
of its importance, a health system is reinforced and
protected by the right to the highest attainable standard
of health and other human rights. Health systems should
have certain right-to-health features identified in this
report. These features are legally binding requirements,
not optional extras. Governments must be held to account
to ensure that health systems have, in practice, the
features required by international human-rights law.
Peru and drafted some panels. MM collected national data on
Mozambique and drafted some panels. CV and DP collected national data
on Romania and drafted some panels. DAP. C|-S. AP. and DT. write in a
personal capacity and the views expressed are not those of their affiliations.
www.thelancet.com Vol 372 December 13, 2008
Right to Health
Acknowledgments
We are extremely grateful to the following individuals who have provided
us with assistance: Idris A Bawa. Alexandra Cameron,
Edelisa D Carandang. Judith Bueno de Mesquita, Aditi Das. Louise Finer,
Maiebakeng Forere, Jane Godsland, Sofia Gruskin, Hans Hogerzeil,
Dragana Korljan, Richard Laing. Rhona MacDonald,
Gillian MacNaughton. Maxwell Madzikanga. Paulo Marchi,
Sophie Mathewson. Martin McKee. Ellen Nolte, Uju A Okereke,
Lisa Oldring, Ching-Ling Pang, Genevieve Paul. Justina Pinkeviciute,
Helen Potts, Anna Protano-Briggs, Mike Rowson, Julian Sheather,
Christine Tapp. Daniel Tarantola, Michael Wilks. This report was
researched with support from the Ford Foundation and Joseph Rowntree
Charitable Trust. However, the content of the report is the sole
responsibility of the authors and cannot be taken to reflect the views of
either the foundation or trust.
References
1
UN. Universal Declaration of Human Rights. G. A. Res. 217A (111),
2
3
4
5
6
8
9
10
11
12
13
14
15
16
17
18
19
UN GAOR. Res. 71. UN Doc. A/810. New York: United Nations,
1948.
Brownlie 1. Basic documents of international law. Oxford: Oxford
University Press, 2006.
Marks S. Health and human rights: Basic international
documents. 2nd edn. Cambridge: Francois-Xavier Bagnoud
Center for Health and Human Rights, Harvard School of Public
Health, 2006.
UN. International Covenant on Economic, Social and Cultural
Rights (ICESCR). New York: United Nations. 1966.
Sepulveda M. The nature of the obligations under die International
Covenant on Economic. Social and Cultural Rights. Utrecht:
Intersentia, 2003.
UN. Convention on the Rights of the Child (CRC). New York:
United Nations. 1989.
UN. International Convention on the Elimination of all Forms of
Racial Discrimination. New York, United Nations 1966.
Kinney E, Clark B. Provisions for health and health care in the
constitutions of the countries of the world. Cornell Int Law J 2004:
37: 285-355.
WHO. Constitution of the World Health Organisation.
http://www.who.int/gb/bd/PDF/bd46/ebd46_
p2.pdf (accessed Sept 21. 2008).
WHO. Declaration of Alma-Ata, international conference on
primary health care, Alma-Ata, USSR, 6-12 September, 1978
http://www.searo.who.int/LinkFiles/Healtli_Systems_declaration_
almaata.pdf (accessed Sept 21, 2008).
WHO. Ottawa charter for health promotion, first international
conference on health promotion Ottawa, 1986. http://www.who.int/
hpr/NPH/docs/ottawa_charter_hp.pdf (accessed Sept 21. 2008).
WHO. The Bangkok charter for health promotion in a globalized
world, 2005 http://www.who.int/healthpromotion/conferences/
6gchp/hpr_050829_%20BCHP.pdf (accessed Sept 21. 2008).
Ministry of Health, Government of Uganda. Review of the HSSP1I
in relation to human rights and gender as part of third-term review
for the health sector, second draft June 2008. Kampala: Ministry of
Health. 2008.
WHO. Human rights, health and poverty reduction strategies.
Geneva. World Health Organization, 2005.
Constitutional Court. Tutela decision T-760/2008, Judge: Manuel
Jose Cepeda, Second Reviewing Chamber, 31 July 2008. http://www.
corteconstitucional.gov.co/ (accessed Sept 21. 2008).
Gauri V, Brinks DM. eds. Courting social justice: judicial
enforcement of social and economic rights in the developing world.
Cambridge: Cambridge University Press. 2008.
Chai Y. Cottrell J. eds. Economic, social and cultural rights in
practice: the role of judges in implementing economic, social and
cultural rights. London: Interights. 2004.
Inter-American Commission on Human Rights. Access to justice as
a guarantee of economic, social and cultural rights: a review of the
standards adopted by the Inter-American Commission on Human
Rights. Washington: Inter-American Commission on Human
Rights. 2007
Baderin MA. McCorquodale R, ed. Economic, social and cultural
rights in action. Oxford: Oxford University Press, 2006.
www.thelancet.com Vol 372 December 13, 2008
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
Ely Yamin A, Parra Vera O. The role of courts in defining health
policy: the case of the Colombian constitutional court. Harvard Law
School Human Rights Program Working Paper, October 2008.
http://www.law.harvard.edu/programs/hrp/documents/Yamin_
Parra_working_paper.pdf (accessed Sept 21. 2008).
Asher J. The right to health: a resource manual for NGOs. London:
Commonwealth Medical Trust, 2004.
Asher J. Hamm D, Sheather J. The right to health: a toolkit for
health professionals. London: British Medical Association/
Commonwealth Medical Trust, 2007.
Physicians for Human Rights. The right to health and health
workforce planning: a guide for government officials,
non-governmental organisations, health workers and development
partners. Cambridge: Physicians for Human Rights. 2008.
People's Health Movement. The assessment of the right to health
and health care at the country level, http://www.phmovement.org/
files/RTH_assmt_tool.pdf.
Hunt P. Report of the UN Special Rapporteur on the right to the
highest attainable standard of health to the Commission on Human
Rights, UN Doc E/CN.4/2004/49. Geneva: United Nations. 2004.
Hunt P. Report of the UN Special Rapporteur on the right to the
highest attainable standard of health to the United Nations General
Assembly UN Doc A/61/338. New York: United Nations, 2006.
Hunt P. Report of the UN Special Rapporteur on the right to the
highest attainable standard of health to the Commission on Human
Rights. UN Doc. E/CN.4/2005/51. Geneva: United Nations. 2005.
Hunt P. Report of the UN Special Rapporteur on the right to the
highest attainable standard of health to the United Nations General
Assembly UN Doc A/59/422. New York: United Nations. 2004.
Hunt P. Report of the UN Special Rapporteur on the right to the
highest attainable standard of health to the United Nations General
Assembly UN Doc A/62/214. New York: United Nations. 2007
WHO. Everybody's business: strengthening health systems to
improve health outcomes: WHO’s framework for action. Geneva:
World Health Organization. 2007.
Freedman L. Achieving the MDGs: health systems as core social
institutions. Development 2005: 48: 19-24.
Weissbrodt D. The right to a fair trial under the Universal
Declaration of Human Rights and the International Covenant on
Civil and Political Rights. The Hague: Marlinus Nijhoff, 2001.
International Covenant on Civil and Political Rights (ICCPR).
New York, United Nations. 1966.
Zimmerman and Steiner v Switzerland. European Court of Human
Rights. A.66 (1983) 6 EHRR 17.
Hunt P. Report of the UN Special Rapporteur on the right to the
highest attainable standard of health to the Human Rights Council.
UN Doc. A/HRC/4/28. Geneva: United Nations, 2007.
Committee on Economic, Social and Cultural Rights. The right to
the highest attainable standard of health: 11/08/2000.
E/C.12/2000/4, CESCR General Comment 14. Twenty-second
session Geneva, 25 April-12 May 2000 Agenda item 3.
http://www.unhchr.ch/tbs/doc. nsf/(symbol)/E.C. 12.2000.4. En
(accessed Sept 21, 2008).
Hunt P. Backman G. Health systems and the right to the highest
attainable standard of health. Health Hum Rights: 2008:10: 81-92.
Hunt P. The human right to the highest attainable standard of
health: new opportunities and challenges. Trans R Soc Trap Med Hyg
2006:100: 603-607
Gruskin S, Mills E. Tarantola D. History, principles, and practice of
health and human rights. Lancet 2007; 370: 449-55.
Mann J. Gruskin S. Grodin M, Annas G, eds. Health and human
rights: a reader. New York: Routledge, 1999.
Gruskin S. Grodin M, Annas C, Marks S, eds. Perspectives on
health and human rights. New York: Routledge. 2005.
Yamin A. Journeys towards the splendid city. Hum Rights Q2004:
28: 519-38.
43
Hunt P. The health and human rights movement: progress and
obstacles. J Law Med 2008; 15: 714-24.
44
Hunt P. Mesquita J. Mental disabilities and the human right to the
highest attainable standard of health. Hum Rights Q 2006;
28: 332-56.
WHO. WHO Report 2001: Mental health: new understanding, new
hope. Geneva: World Health Organization, 2001.
45
2083
pight to Health
46
47
48
49
50
51
52
53
54
55
56
57
58
59
WHO. People at the centre of health care: harmonising mind and
body, peoples and systems. Geneva: World Health Organization. 2007
Tarimo E. Towards a healthy district. Organizing and managing
district health systems based on primary health care. Geneva: World
Health Organization, 1991.
Singh AJ, Govender M, Mills E. Do human rights matter to health?
Lancet 2007; 370: 521-27.
Hogerzeil HV, Samson M. Vidal Casanovas J, Rahmani L. Access to
essential medicines as part of the fulfilment of the right to health—
is it enforceable through the courts? Lancet 2006; 368: 305-11.
WHO. Who guidelines for drinking water quality, http://www.who.
int/water_sanitation_health/dwq/guidelines/en/index.
html (accessed Sept 21, 2008).
WHO. International health regulations (2005), 2nd edn, Geneva:
WHO. 2008.
WHO. Framework and standards for country health information
systems. 2nd edn. Geneva: WHO, 2008.
60
Sphere Project. Humanitarian charter and minimum standards in
disaster response. Geneva: Sphere Project, 2004.
61
Potts H. Participation and the right to the highest attainable
standard of health. Essex: Human Rights Centre. University of
Essex, forthcoming.
Manandhar DS, Osrin D. Shrestha BP. et al. Effect of a participatory
intervention with women’s groups on birth outcomes in Nepal:
cluster randomised controlled trial. Lancet 2004; 364: 970-979.
62
63
64
Medicines Transparency Alliance. Website. http://www.
medicinestransparency.org (accessed Nov 4, 2008).
Open Society Institute. Mediating Romani health: policy and
programme opportunities. New York: Open Society Institute, 2005.
65
Braveman P. Gruskin S. Defining equito in health.
J Epidemiol Community Health 2003; 57: 254-58.
66
Hoffmann v South African Airways, Constitutional Court of South
Africa. Case CCT 17/00 2000. decision of 29 September 2000. http://
www.saflii.org/za/cases/ZACC/2000/17.html (accessed Sept 21. 2008).
Eldridge v British Columbia. [1997] 3 S.C.R. 624. http://csc.lexum.
umontreal.ca/en/1997/1997rcs3-624/1997rcs3-624.html (accessed
Sept 21, 2008).
UNFPA. Culture matters—working with communities and
faith-based organisations: case studies from country programs.
New York: UNFPA. 2004.
67
68
69
70
71
72
71
2084
Horton R. What does a national health service mean in the
21st century? Lancet 2008; 371: 2213-18.
International Harm Reduction Association. Global harm reduction
report. London: International Harm Reduction Association. 2008.
Hunt P. Report of the Special Rapporteur on the right to the highest
attainable standard of health to the Human Rights Council, UN
Doc A/HRC/4/28/Add.2. Geneva: United Nations, 2007.
Hunt P, Khosla R. The human right to medicines. SUR 2008;
5: 99-116.
Amnesty International. Peru: poor and excluded women—denial of
right to maternal and child health. London: Amnesty
International, 2006.
Ministerio de Salud. Norma tecnica para la atencion del parto
vertical con adecuacion intercultural. Lima, Peru: Ministerio de
Salud, 2005.
WHO. World Health Report 2000. health systems: improving
performance. Geneva: World Health Organization, 2000.
Hunt P. Report of the Special Rapporteur on the right to the
highest attainable standard of health to the Commission on
Human Rights, UN Doc E/CN.4/2005/51/Add.3. Geneva: United
Nations, 2005.
Akunyili D. The fight against counterfeit drugs in Nigeria. In:
Transparency International—global corruption report 2006:
corruption and health. Berlin: Transparency International,
2006: 96-100.
Green A. An introduction to health planning for developing health
systems. Oxford: OUP, 2007.
Hunt P. Stewart R, Mesquita J. Oldring L. Neglected diseases:
a human rights analysis. Geneva: Special Programme for Research
and Training in Tropical Diseases (TDR). 2007
Alloliy P. Reidpath D. Ghalib 11. Pagnorn F, Skelly I. FfTicacious,
effective, and embedded interventions: implementation research in
infectious disease control. BMC Public Health 2008: 8: 343.
74
75
76
77
78
79
80
81
Chapman A, Russell S, eds. Core obligations: building a framework
for economic, social and cultural rights. Antwerp: Intersentia. 2002.
Hunt P, MacNaughton G. Impact assessments, poverty and human
rights: a case study using the right to the highest attainable
standard of health, http://www2.essex.ac.uk/human_rights_centre/
rth/projects.shtm (accessed Aug 15, 2008).
Minister of Health v Treatment Action Campaign, Constitutional
Court of South Africa, Case CCT 8/02, decision of 5 July 2002.
http://www.saflii.org.za/za/cases/ZACC/2002/16.html (accessed
Sept 21. 2008).
Paschim Banga Khet Mazdoor Samity v State of West Bengal,
1996 SCJ 25: 29. http://www.commonlii.Org//cgi-bin/disp.pl/in/
cases/INSC/1996/659.html?query=paschim (accessed Sept 21.
2008).
WHO. Report of the global partners' meeting on neglected tropical
diseases: a turning point. Geneva: World Health Organization, 2007.
UNAIDS. Enhancing results by applying the Paris Declaration at
sector level: progress update and lessons learnt from aid
effectiveness in AIDS responses. Geneva: UNAIDS. 2008.
WHO, Framework Convention on Tobacco Control, adopted by the
56th World Health Assembly, 31 May 2003. http://www.who.int/
fctc/en/ (accessed Sept 21. 2008).
Skogly S. Beyond national borders: States’ human rights obligations
in international cooperation. London: Intersentia. 2006.
82
High Level Forum on Aid Effectiveness. Paris declaration on aid
effectiveness, http://www.oecd.0rg/document/l8/0.3343,en_2649_
3236398_354O1554_l_l_l_l,OO.html (accessed Sept 21. 2008).
83
International Council on Human Rights Policy. Duties sans
frontiers: human rights and global social justice. Versoix:
International Council on Human Rights Policy, 2003.
Straight Talk Foundation, http://www.straight-talk.or.ug/
(accessed Oct 14. 2008).
Hunt P. Report of the UN Special Rapporteur on the right to the
highest attainable standard of health to the Human Rights Council.
UN Doc A/HRC/7/ll/Add,2. Geneva: United Nations, 2008.
84
85
86
87
88
89
90
91
92
93
94
Darrow M. Between light and shadow: the World Bank, the
International Monetary Fund and international human rights law.
Portland: Hart, 2003.
Van Genugten W. Hunt P, Matthews S. eds. World Bank. IMF and
human rights. Nijmegen: Wolf. 2003.
Potts H. Accountability and the right to the highest attainable
standard of health. Colchester: Human Rights Centre, University of
Essex. 2008.
Clapham A. Human rights obligations of non-state actors. Oxford:
Oxford University Press, 2006.
New Vision. Health rights unit launched. Kampala: New Vision,
Oct 1, 2008.
Nolte E. McKee M, Wait S. Describing and evaluating health systems.
In Bowling A. Ebrahim S. eds. Handbook of health research
methods. New York: Open University Press, 2006: 12-43.
Neuman E. Do international human rights treaties improve respect
for human rights? J Conflict Resolut 2005; 49: 6.
Nolte E. McKee M. Measuring the health of nations: analysis of
mortality amenable to health care. BMJ 2003; 327: 1129-32.
International conference on primary health care. Declaration of
Alma-Ata. Alma-Ata, USSR: World Health Organization, 1978.
95
Committee on the Rights of the Child. HIV/A1DS and the rights of
the child. March 17. 2003. CRC/GC/2003/3, General Comment 3.
96
Committee on the Rights of the Child. Adolescent health and
development in the context of the Convention on the Rights of the
Child. July 1. 2003. CRC/GC/2003/4, General Comment 4.
Committee on the Elimination of Discrimination against Women.
Women and health. A/54/38/Rev.l, chapter 1. General
Recommendation 24.
98 Hunt P. Report of the Special Rapporteur on the right to the highest
attainable standard of health to the Commissioner on Human
Rights, UN Doc E/CN.4/2006/48. Geneva: United Nations, 2006.
99 UNDP. Human development reports, http://hdr.undp.org/en/
statistics/ (accessed Aug 15. 2008).
100 UN. Report on indicators for promoting and monitoring the
implementation of human rights. HRI/MC/2008/3. 6 June 2008.
http://www2.ohchr.org/english/bodies/icm-mc/documents.htm
(accessed Dec 1. 2008).
97
www.thelancet.com Vol 372 December 13, 2008
Right to Health
101 UN. United Nations Millennium Development Goals Indicators.
http://mdgs.un.org/unsd/mdg/Default.aspx (accessed Aug 15, 2008).
102 UN Office of the High Commissioner for Human Rights. Geneva.
Principles and guidelines for a human rights approach to poverty
reduction strategies. Geneva: UN, 2007.
103 WHO. Medicines strategy 2004-2007. Geneva: World Health
Organisation, 2004.
104 UNICEF. Monitoring and statistics: core indicators in depth.
http://www.unicef.org/statistics/index_24296.html (accessed
Aug 15. 2008).
105 UNAIDS. 2006 Report on the global epidemic. Geneva:
UNAIDS. 2006.
106 The World Bank. World Development Indicators 2007 http://web.
worldbank.org/WBSITE/EXTERNAL/DATASTATISTICS/
0,.contentMDK:21298138~pagePK:64133150~piPK:64133175~theSite
PK:239419.00.html (accessed Aug 15. 2008).
107 The World Bank. Governance Matters VI: aggregate and individual
governance indicators 1996-2006. http://www-wds.worldbank.org/
external/default/WDSContentServer/WDSP/IB/2007/07/10/
000016406_20070710125923/Rendered/PDF/wps4280.pdf
(accessed Aug 15, 2008).
108 WHO. Indigenous peoples' health. WHO/SDE/HSD/99. Geneva:
World Health Organization. 2006.
109 CRR. Broken promises, human rights, accountability, and maternal
death in Nigeria. Centre for Reproductive Rights, 2008.
110 PHR. Deadly delays: maternal mortality in Peru—a rights-based
approach to safe motherhood. Physicians for Human Rights, 2008.
111 Hunt P. Mesquita ). Reducing maternal mortality: the contribution
of the right to the highest attainable standard of health. University
of Essex, 2006.
112 UNDP. Beyond scarcity: power, poverty and the global water crisis.
New York: Palgrave Macmillan, 2006.
113 United States National Library of Medicine National Institutes of
Health International Committee of Medical Journal Editors
Uniform Requirements for Manuscripts Submitted to Biomedical
Journals: sample references. 2005. http://www.nlm.nih.gov/bsd/
uniform_requirements.html (accessed Sept 21, 2008).
114 OECD. Ratification of the Convention on the OECD. http://www.
oecd.org/document/58/0,3343,en_2649_ 3448 3_1889402_l_1_1_l,00.
html (accessed Sept 21, 2008).
115 The World Bank. Cumulative IDA Subscriptions and Contributions.
http://siteresources.worldbank.org/IDA/Resources/CUMlDA2008.
pdf (accessed May 20. 2008).
116 European Parliament and Council. Directive 95/46/EC of the
European Parliament and of the Council, http://www.cdt.org/
privacy/eudirective/EU_Directive_.html#H D_NM_10
(accessed Aug 08. 2008).
117 Hunt P. MacNaughton G. A human rights-based approach to health
indicators. In: Baderin M. Mccorquodale R, eds. Economic, social
and cultural rights in action. Oxford: Oxford University Press.
2007: 303-30.
118 WHO. Countries, http://www.who.int/countries/en/
(accessed Sept 09, 2008).
119 WHO. UNICEF. Global water assessment and sanitation report
2000. http://www.wssinfo.org/pdf/GWSSA_2000_report.pdf
(accessed May 23. 2008).
120 Howard G, Bartram J. Domestic water quantity, service level and
health, http://www.who.int/water_sanitation_health/diseases/en/
WSHO3O2.pdf (accessed Dec 20. 2007).
121 WHO. UNICEF. Joint monitoring programme for water supply and
sanitation. 2008. http://www.wssinfo.org/en/welcome.html
(accessed Aug 8. 2008).
122 National Institute of Statistics. Annual report of National Institute
of Statistics, chapter Housing and Public Utilities, 2006. http://
www.insse.ro/cms/files/pdf/ro/cap5.pdf (accessed July 23. 2008).
123 WHO. WHO Multi-country study on women's health and domestic
violence against women. Geneva: World Health Organization, 2005.
124 WHO. World Health survey results, http://www.who.int/
healthinfo/survey/en/ (accessed July 23. 2008).
125 The World Bank. Data and statistics 2008 http://go.worldbank.org/
D7SN0B8YU0 (accessed Sept 21. 2008).
www.thelaricet.com Vol 3/2 December 13, 2008
126 WHO, UNICEF, UNFPA, and the World Bank. Maternal mortality
in 2005. Geneva: World Health Organization, 2007.
127 UN-OHCHR. Fact sheet no 19 national institutions for the promotion
and protection of human rights. Geneva: United Nations, 1993.
128 GAV1. Health system strengthening, http://www.gavialliance.org/
vision/policies/hss/index.php (accessed Sept 21, 2008).
129 Hbgskolefdrordning. SFS 1993:100 Amendment SFS 2007: 638.
http://www.notisum.se/index2.aspsTemplate=/template/index.asp
&iMenulD=314&iMiddleID=285&iParentMenuID=236&iLanguagel
D=1 (accessed Sept 21. 2008).
130 Department of Health. Human rights in healthcare evaluation.
31 July 2008. http://www.dh.gov.uk/en/Publicationsandstatistics/
Publications/DH_088979 (accessed Nov 30. 2008).
131 WHO. Country Health information systems, http://www.who.int/
healthinfo/systems/en/ (accessed Aug 8, 2008).
132 WHO. Department of Economic and Social Affairs' principles and
recommendations for a vital statistics system revision, http://
unstats.un.org/unsd/publication/SeriesM/SeriesM_19rev2E.pdf
(accessed June 15. 2008).
133 WHO. Beyond the numbers: reviewing maternal deaths and
complications to make pregnancy safer. Geneva: World Health
Organization. 2004.
134 WHO. International statistical classification of diseases and related
health problems 10th revision version for 2007. http://www.who.int/
classifications/apps/icd/icdlOonline/ (accessed Sept 21, 2008).
135 Robertson D. Bedell R, Lavery J, Upshur R. What kind of evidence
do we need to justify humanitarian medical aid? Lancet
360: 330-33.
136 WHO. Commission on social determinants of health-final report.
Geneva: World Health Organization, 2008.
137 United Nations General Assembly Special Session on H1V/AIDS.
Guidelines on construction of core indicators in relation to
monitoring the declaration of commitment on HIV/AIDS. http://
www.measuredhs.com/hivdata/guides/ungassmeguidejuly2005.pdf
(accessed Aug 8. 2008).
138 WHO. World Health Report 2005: Make every mother and child
count. Geneva: World Health Organization, 2005.
139 WHO, UNICEF. WHO vaccine-preventable diseases monitoring
system. http://whqlibdoc.who.int/hq/2007/who_IVB_2007_eng.pdf
(accessed May 20, 2008).
140 Demographic and Health Surveys. Model "A" questionnaire with
commentary for high contraceptive prevalence countries. 2001.
http://www.measuredhs.com/pubs/pdf/DHSQ4/
DHS-IV-Model A.pdf.pdf(accessed Sept 15. 2008).
141 WHO. Summary report: WHO multi-country study on women’s
health and domestic violence against women. Geneva: World
Health Organization, 2005.
142 Abuja Declaration. AIDS is a state of emergency in the continent.
Adopted 24-27 April 2001. http://www.un.org/ga/aids/pdf/abuja_
declaration.pdf (accessed Sept 21, 2008).
143 UNDP. MDG Series metadata. http://millenniumindicators.
un.org/unsd/mdg/Metadata.aspx?lndicatorld=0&Seriesld=776
(accessed Sept 21, 2008).
144 International Covenant on Civil and Political Rights (1CCPR). New
York, United Nations, 1966. United Nations G.A. Res. 2200A (XXI).
UN GAOR, 21st Session. Supp. No. 16, at 52, UN Doc. A/6316,
entered into force 23 March 1976. http://www.unhchr.ch/html/
menu3/b/a_ccpr.htm (accessed Sept 21. 2008).
145 UN Human Rights Council Res 5/1, Institution-building of the
United Nations Human Rights Council. 18 June 2007. http://www2.
ohchr.org/english/bodies/hrcouncil/ (accessed Sept 21. 2008).
146 OHCHR. Economic, social and cultural rights: handbook for
national human rights institutions. Geneva: OHCHR. 2005.
147 Human Rights and Equal Opportunity Commission. Social justice
report 2005. Canberra: Human Rights Equality Commission. 2005.
148 DuBois M. Civilian protection and humanitarian advocacy: strategies
and (false?) dilemmas. Humanitarian Exch 2008; 39: 12-15.
149 Hunt P. Report of the UN Special Rapporteur on the right to the
highest attainable standard of health to the United Nations General
Assembly UN Doc A/63/263, 2008. New York: United
Nations, 2008.
2085
)
BMC Pregnancy and Childbirth
BiaMed Central
Research article
Open Access
Women's health groups to improve perinatal care in rural Nepal
Joanna Morrison*1, Suresh Tamang2, Natasha Mesko1, David Osrin1,
Bhim Shrestha2, Madan Manandhar3, Dharma Manandhar2, Hilary Standing4
and Anthony Costello*1
Address: ’International Perinatal Care Unit, Institute of Child Health, University College, London, 30 Guilford Street London, WC1N 1EH, UK,
2Mother and Infant Research Activities (MIRA), GPO Box 921, Kathmandu, Nepal, 3Nepal Administrative Staff College, Kathmandu, Nepal and
•institute of Development Studies, Palmer, Brighton, Sussex, BN1 9RH, UK
Email: Joanna Morrison* - j.morrison@ich.ucl.ac.uk; Suresh Tamang - mira@vianet.com.np; Natasha Mesko - n_mesko@hotmail.com;
David Osrin - d.osrin@ich.ucl.ac.uk; Bhim Shrestha - mira@vianet.com.np; Madan Manandhar - rn.manandhar@info.com.np;
Dharma Manandhar - dsm@healthnet.org.np; Hilary Standing - h.standing@ids.ac.uk; Anthony Costello* - a.costello@ich.ucl.ac.uk
* Corresponding authors
Published: 16 March 2005
BMC Pregnancy and Childbirth 2005, 5:6
doi: 10.1 186/1471 -2393-5-6
Received: 14 June 2004
Accepted: 16 March 2005
This article is available from: http://www.biomedcentral.eom/l47l-2393/5/6
© 2005 Morrison et al; licensee BioMed Central Ltd.
This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.Org/licenses/by/2.0).
which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Abstract
Background: Neonatal mortality rates are high in rural Nepal where more than 90% of deliveries
are in the home. Evidence suggests that death rates can be reduced by interventions at community
level. We describe an intervention which aimed to harness the power of community planning and
decision making to improve maternal and newborn care in rural Nepal.
Methods: The development of I I I women's groups in a population of 86 704 in Makwanpur
district, Nepal is described. The groups, facilitated by local women, were the intervention
component of a randomized controlled trial to reduce perinatal and neonatal mortality rates.
Through participant observation and analysis of reports, we describe the implementation of this
intervention: the community entry process, the facilitation of monthly meetings through a
participatory action cycle of problem identification, community planning, and implementation and
evaluation of strategies to tackle the identified problems.
Results: In response to the needs of the group, participatory health education was added to the
intervention and the women's groups developed varied strategies to tackle problems of maternal
and newborn care: establishing mother and child health funds, producing clean home delivery kits
and operating stretcher schemes. Close linkages with community leaders and community health
workers improved strategy implementation. There were also indications of positive effects on
group members and health services, and most groups remained active after 30 months.
Conclusion: A large scale and potentially sustainable participatory intervention with women's
groups, which focused on pregnancy, childbirth and the newborn period, resulted in innovative
strategies identified by local communities to tackle perinatal care problems.
Page 1 of 12
(page number not for citation purposes)
BMC Pregnancy and Childbirth 2005, 5:6
Background
Participatory approaches to health have been advocated
since the 1978 Alma Ata declaration in which the World
I lealth Organisation emphasised the need for citizen par
ticipation in primary health care 11 (.This paper details the
development and implementation of a participatory
project to improve perinatal care at the community level
in ntral Nepal.
Community participation in health care
I'he vision of Alma Ata was that increasing community
participation in planning and implementation would
lead to more cost-effective delivery of health care and
increases in service utilisation. As communities took
greater ownership of services they would become more
culturally acceptable and responsive to local needs. Com
munity participation also aimed to increase self-reliance
and social awareness, which would lead to better health
outcomes |2-4]. Opinions differ about the extent to which
participation can achieve these results, and to what degree
governments and agencies have facilitated participation,
but the appeal of participatory approaches remains
strong. Participation may be considered as a continuum
151. In fully participatory approaches, needs are identified
by the community themselves, who then may seek exter
nal support. At the other end of the continuum, superfi
cial participation of community representatives is sought
to validate the aims of programme planners, usually
already decided.
Harnessing the strengths of participation in community
based interventions
Reproductive health is an area where participatory
approaches have been attempted. A structured literature
search for community-based interventions focusing on
perinatal health revealed no randomized, controlled tri
als, but two studies in developing countries which had
evaluated impact on Perinatal health outcomes. The
Warmi project in Bolivia, initiated by a collaboration of
Save the Children Federation, USA and USAID MotherCare project 16|, worked with women's groups to reduce
maternal and neonatal mortality and morbidity. They
used a participatory approach involving community diag
nosis, planning together, implementation of plans, and
participatory evaluation. The Warmi project, though nei
ther randomized nor controlled, and based on a beforeand-after analysis of 639 and 708 births, did report a
reduction in the perinatal mortality rate from 117 to 44
per thousand births. The activities initiated by women's
groups included literacy programmes, savings and credit
schemes, and programmes to increase access to family
planning.
Studies based in the community that are towards the low
end of the participation continuum also appear to have
http://www. biomedcentral.com/1471 -2393/5/6
been successful in enabling improvements in pregnancy
outcomes. A study in Maharashtra state, India, tested the
effectiveness of early detection of warning signs of illness
and village level management of neonatal sepsis (a cause
of many neonatal deaths in developing countries)!?|. Vil
lage health workers were trained to visit newborn infants
in their homes and identify and treat neonatal sepsis. This
intervention appeared highly successful as a drop in neo
natal mortality of 62% occurred. Village health workers
were intensively managed and supported by the research
team, and therefore large-scale implementation may be
difficult. The study did, however, provide evidence that
community level interventions to prevent or treat prob
lems of the perinatal period in developing countries could
be cost-effective.
The Nepal MIRA Makwanpur trial
The MIRA Makwanpur trial was designed to test the
impact on neonatal mortality of a participatory interven
tion with women's groups, based on the Warmi [Bolivia
model, but on a much larger scale and using a rand
omized and controlled design. In south Asia infant mor
tality rates fell steadily from 1970 to 1990, but the decline
has subsequently plateaued. In order to reduce infant
mortality rates further, a focus on the neonatal period, in
which most infant deaths occur, is necessary (8,91. Pri
mary and secondary care are deficient in rural areas of
Nepal and where services exist, the reasons for their
underuse are complex. The topographical barriers com
bined with limited expenditure on public health, poor
quality of care, a high turnover of service providers, a lack
of drugs and supplies and a lack of ownership of health
programmes by communities all contribute to issues of
demand and supply.
I’he trial was implemented by a Nepali non-governmental
organization, MIRA (Mother & Infant Research Activities).
MIRA has been working in Nepal since 1992, conducting
research specifically about newborn care, and is headed by
a senior pediatrician (DM). The trial involved 24 Village
Development Committees in rural Makwanpur district.
Ethical approval was sought from the Nepal Health
Research Council, and local meetings were held with the
District Development Office and Chief District Officer to
discuss the aims and objectives of the study. The chairper
sons for each Village Development Committee agreed to
take part in the study and provided signed consent, and
links were made with community leaders, district health
services and non governmental organisations. Each Vil
lage Development Committee has an average population
of 7000 (range 1576 to 23 429) divided between nine
wards. In twelve of the Village Development Committees
a trained, locally based facilitator was employed to mobi
lize women's groups. All pregnancies and births to mar
ried women of reproductive age were monitored in the
Page 2 of 12
(page number not for citation purposes)
BMC Pregnancy and Childbirth 2005, 5:6
community. Details of the monitoring and the design of
the trial have been described elsewhere 1101 and the effect
of this intervention on birth outcomes was reported in a
recent publication 1111. Astonishingly, there was a reduc
tion in neonatal mortality by 30% in intervention clus
ters, and an even larger and statistically significant effect
on maternal mortality rates (78% reduction), although
caution is required in interpretation given the relatively
few maternal deaths. This paper describes and analyses
the implementation of the first stages of the participatory
intervention over a 30 month period.
Methods
Setting of intervention
Nepal has a population of 23 million and a per capita
gross national product of 240 US dollars. Literacy rates
have improved steadily, particularly for females (currently
43%), but there remain gender disparities in literacy,
school enrolment, and school dropout rates 112|.
Life expectancy is now 61 years| 13). The total fertility rate
is 4.1, the under-five mortality rate 91, the infant mortal
ity rate 64, the neonatal mortality rate 39 per thousand
live births, and the perinatal mortality rate 47 per thou
sand births 191. The maternal mortality ratio is estimated
at 539 per 100 000 live births 114|. Access to health care is
limited as a result of geography, limited expenditure on
public health, variable quality of care, high turnover of
service providers, a lack of drugs and supplies, and lack of
ownership of health programmes by communities.
Makwanpur district, south west of Kathmandu, has a pop
ulation of 376 000 1151 and a Human Development Index
of 0.31, close to the national median. Makwanpur com
prises hill and plain areas, with 15 different ethnic groups,
the largest being Tamangs, a Tibeto-Burman group. Data
from our baseline survey showed that more than 90%
births take place at home and only five percent are
attended by a trained birth attendant. The first health care
provider in times of maternal or neonatal illness is the
shaman (dhami jhankri) or traditional healer [ 16|.
The intervention process
The first ten meetings of the women's group participatory
intervention were based on the design of the Warmi
project in Bolivia 117|. in order to enter the communities
successfully we gathered detailed information on local
social networks and organizations, as well as attitudes and
practice around the time of pregnancy and birth. Social
mapping and qualitative research were conducted and
served as a training exercise in facilitating focus group dis
cussions and building rapport in the community [18]
http://www.biomedcentral.eom/1471-2393/5/6
Establishing facilitated women's groups
Meetings were facilitated by a paid, locally based woman,
who was selected on merit and trained in facilitation tech
niques. The position of facilitator was locally advertised
and suitable candidates were interviewed by senior MIRA
employees. Each facilitator is paid a salary slightly higher
than the government equivalent (5330 Nepalese rupees,
or 71 US dollars). Her full-time responsibility was to plan
and facilitate monthly women's group meetings, each
facilitator leading nine groups per month, covering an
average population of 7000. Meetings were organized in
co-ordination with the local Female Community Health
Volunteer, an unpaid community based health worker. In
profiling our study area, we found that nongovernmental
organisations or community based organisations did not
routinely work in all 24 of our study Village Development
Committees and had different agendas. The female com
munity health volunteer works at ward level, and as part
of her job description she runs women's groups to con
duct health promotion activities.
The facilitator used a meeting manual, adapted from the
Warmi project, to guide the women's groups through
problem identification and community planning using
participatory iterative methods (see Figure 1 and Fable 1).
Facilitators were trained in the use of this manual and
were allowed scope for their own input. Facilitation super
visors were also appointed after national advertisement
and formal interview, and two men and three women
were selected. One supervisor was provided for every three
facilitators, providing support through community visits
and regular meetings.
First meetings and problem identification
Facilitators and supervisors were responsible for creating
awareness and interest in their communities about the
meetings, and in most wards at least 20 women attended
the first few meetings (see Figure 2). Time was taken to
introduce the study agenda to the groups, especially
important in areas where many non governmental organ
isations work and expectations were often high. The first
three meetings facilitated discussion of the reasons why
mothers and newborn infants die in the community. The
reasons for death were discussed in terms of social as well
as medical factors with the aid of a story 119|. Women
were introduced to the concept of ‘learning together'
through another story, and were encouraged to discuss
perinatal problems within the group and with their neigh
bours and friends. In this way the facilitator and the
women learned which perinatal problems affected their
community. Each group prioritised three problems of
newborn infants and/or pregnancy which were recorded
with justification for their inclusion. Most group members
were illiterate and therefore facilitators used pictures and
voting with stones to prioritise problems.
Page 3 of 12
(page number not for citation purposes)
BMC Pregnancy and Childbirth 2005, 5:6
http://www.biomedcentral.com/1471-2: 93/5/6
4 X
•
j,: ■ ■ k
a f?
II
Problem identification
Planning together
___________________ 5____
Evaluation and problem identification
.... -... T
.. " *
I. s.
Implementation
3
3
Figure I
The community action cycle
Table I: Content of first ten women's group meetings
Stage in the cycle
Meeting
Content
Introduction
I
2
To discuss why mothers and newborn infants die
To introduce the group to MIRA Makwanpur’s work
To introduce how MIRA will work in the community
Problem
3
To find out how women understand maternal and neonatal problems
Identification
4
5
To find out what kind of maternal and neonatal problems are in the community
6
To share the information collected from other women in the community
To decide what are the three most important maternal and neonatal health problems
Problem Prioritisation
Planning together
To discuss whether the maternal and neonatal problems are common
To identify strategies to collect information from the community
7
To discuss possible strategies for addressing the prioritised problems
8
To discuss which other community members should be involved in developing strategies
9
To discuss how to prepare for the community members meeting
IO
The community members will learn about what the women have been doing
The community members will learn about the three problems identified by the women
The community members will learn about the possible strategies
To reach a consensus of the strategies
Page 4 of 12
(page number not for citation purposes)
BMC Pregnancy and Childbirth 2005, 5:6
http://www.biomedcentral.eom/1471-2393/5/6
.w jiW.
■ti
A.
i
■
■»
\
X. ■ ^4;
Figure 2
Women's group and facilitation manager
Planning together
The objective of these meetings was to encourage women
to identify local and low cost ways of tackling the priori
tised problems using local resources. Many examples were
listed in the manual and the supervisor was encouraged to
support the facilitator during these meetings. The idea
behind these meetings was to enable the women to pre
pare a plan to tackle the problems they had found, which
would then be presented to their community.
The groups also discussed the way they would present
their findings and practised to develop their confidence.
The supervisor supported the local facilitator and the
group, and played a key role in facilitating the meeting.
After introducing MIRA and its role in the community, the
women's group presented their prioritised problems and
suggested strategies to tackle them.
Methods of data collection and analysis
The community meeting
A community meeting was planned and organized by the
groups to enable increased community participation and
to legitimize the work of the group. The community was
invited to hear what the women's group had been doing,
and io participate in planning together strategies. Most
groups decided that community leaders would be invited
by letter, and other households would be verbally invited.
Data were collected through a variety of qualitative tech
niques. Participant observation was carried out by the
technical advisors to MIRA (NM and JM). 'They helped
design and implement the intervention, and lived in the
vicinity of the head office in 1 letauda, Nepal, throughout
the study period (JM succeeded NM). They visited the
field many times and attended facilitation team meetings
regularly. The advisors are of British nationality but have
Page 5 of 12
(page number not for citation purposes)
BMC Pregnancy and Childbirth 2005, 5:6
http://www.biomedcentral.eom/1471-2393/5/6
an excellent spoken command of Nepali language and
have a background in anthropology and sociology.
Although they did not keep a diary as is usual in partici
pant observation, their reflections and observations were
noted in monthly reports and topic reports which have
been used in this analysis. These reports were also
contributed to and discussed with the facilitation team
(facilitators, supervisors, and senior facilitation manager)
who also added their reflections and observations. The
technical advisor (JM) and senior facilitation manager
(ST) analysed monthly reports and meeting minutes and
reached consensus on themes emerging from the data,
and issues of interest. Although there are limitations to
this method of data analysis (the cultural background of
the technical advisors and the fact that a diary was not
kept), we wished to present the results of operational
research that makes use of less formally recognized quali
tative data collection techniques. Analysis of the data by
more than one person strengthens the analysis and using
different methods of data collection enables triangulation
of the data.
Table 2: Problems prioritised by women's groups
Results
Ante partum haemorrhage
Neonatal Problems
Number of groups
Pneumonia in the newborn infant
31
16
12
7
6
4
Low birth weight
Jaundice in the newborn infant
Neonatal death
Breathing problem in the newborn infant
Infant not feeding
Green stool in the newborn infant
3
Wounds in the newborn infant
2
Tetanus in the newborn infant
I
Eye and ear infection in the newborn infant
I
Maternal problems
Missing data
58
42
30
22
20
IS
I3
II
II
6
6
5
2
2
I
3
Total
330
Retained placenta
Vaginal Discharge
Malpresentation
Headache in the mother
Post partum haemorrhage
Fever (unspecified if in mother or baby)
Breast problems
Out of 111 women's groups, 77 moved on to develop and
implement strategies and 1 00 groups continue to meet to
discuss perinatal health. Particular reasons for which the
remaining groups did not meet include the unstable secu
rity situation, lack of support from local leaders, husbands
or health workers, and general lack of interest.
Miscarriage
Abdominal pain in the mother
Oedema of hands and legs in the mother
Prolonged labour
Maternal death (delivery complications)
Anaemia in the mother
Vomiting in the mother
What makes an active women's group?
Hie continuing activity of most groups suggests that usu
ally group members found the experience useful and
enjoyable. Not surprisingly, the activity of the groups var
ied. We found no specific formula for an active women's
group. Previous studies suggested that homogeneity of
members was conducive to a successful group (20,211, but
our groups showed much ethnic and social diversity.
Issues of ethnicity, geography and distance from a market
area did not uniformly affect the activity of groups. For
example, there were two particularly active groups near
market areas, but in other areas factors concurrent with
living near a market - such as higher socioeconomic status
and less cohesion between households - did not facilitate
enthusiastic women's groups. Some groups were domi
nated by women from higher castes, but in others these
higher castes served as a stimulant to more traditionally
subservient or timid ethnic groups. Issues of local support
from political groups, local health staff and men also
seemed important. Other studies have also found that
supportive husbands make it easier for women to partici
pate in groups 1221. In most communities supervisors and
facilitators had been successful in establishing good com
munity rapport, and strategies were agreed to help main
tain community support, such as facilitators attending
antenatal clinics and supervisors presenting reports to Vil
lage Development Committee chairpersons.
Problem prioritization
Women actively participated in learning together and
gathered much information from their communities. The
prioritized problems reflected local perceptions of the
seriousness and frequency of specific problems and hence
were different in each community (See Table 2).
Planning together
During the community planning meeting groups nomi
nated members to present their findings and eight groups
performed small socio-dramas. When local health person
nel and area chairmen attended, discussions wer^ livelier
and planning more productive. Issues of health care
underutilization by the community or issues of poor serv
ice delivery were often raised. In nine places, communities
appeared apathetic towards the group and were not pre
pared to commit or participate in planning. In these
Page 6 of 12
(page number not for citation purposes)
http://www.biomedcentral.eom/1471-2393/5/6
BMC Pregnancy and Childbirth 2005, 5:6
instances, they were happy for the group to plan and
implement strategies and little discussion took place. In
four places the group met with hostility from community
leaders or health personnel or exceptionally low attend
ance from the community, usually due to local grievances
with staff selection procedures or to the unstable security
situation.
explored different selling points: local shops, Female
Community Health Volunteers and Traditional Birth
Attendants, and group members sold kits to their friends
and neighbours. Recently, groups from one Village Devel
opment Committee have used free distribution of clean
home delivery kits as an incentive to attend for antenatal
care; the kits are free for women who attend at least four
times.
Strategy development and implementation
The strategies that were discussed during planning together
and have been most successfully implemented were the
mother and child health fund, locally produced clean
home delivery kits, management and production of
stretchers, and awareness raising through video shows.
Mother and child health fund schemes
69 groups favoured mother and child health funds as a
way of overcoming the cost barriers to seeking and obtain
ing care. The cost of consultation, medicine and transport
is a real reason that families do not gain access to services
in Nepal |23|. MIRA provided training to fund manage
ment committee members elected from each group. These
committees sometimes included literate community
members not attending the group. Each group developed
their own policy with regard to how money would be col
lected, who would be able to access it, how often it would
be collected, and who would be responsible for managing
it. 23 months after the first mother and child health fund
was established, groups had generated between 731
rupees (10.5 US dollars) and 9635 rupees (133.8 US
dollars).
Clean home delivery kits
I’he clean home delivery kit is advocated by the World
Health Organisation as an effective way to promote clean
liness during home delivery and to reduce the risk of
maternal and neonatal infection [24,251. In Nepal, a local
private company (MCH Products Pvt Ltd) has produced a
clean home delivery kit, approved by the Ministry of
Health, which contains a blade, a bar of soap, three cord
ties, a plastic coin for cord cutting, a plastic sheet, and a set
of pictorial instructions. There are problems with distribu
tion to remote rural areas and other difficulties regarding
local acceptability and price [261.
A few groups were keen to develop their own locally pro
duced clean home delivery kit, and facilitators
disseminated this idea to motivate other groups by
example. 19 groups have made clean home delivery kits
and four groups have reproduced subsequent batches.
Groups have decided the price, but all groups sell at a
lower price than the MCI I products kit, with profits going
into the mother and child health fund. 'The pictorial
instruction leaflet was developed with a local artist and
was piloted in the community. The groups have also
Stretchers
In the study area, most births take place in the home 1271
and transportation of women who encounter problems is
difficult. Women's groups therefore identified the need
for stretchers. 19 groups decided to raise money for
stretchers themselves and the other 23 groups utilized
local resources such as forest user groups or Village Devel
opment Committee offices. Women's groups investigated
if there were any existing stretchers in their area, or if these
needed repair. One group felt that the modern style of
stretcher was not suitable for carrying women across diffi
cult terrain, and therefore made a bamboo basket (dhoko)
which is traditionally used to carry fodder or crops using
a head strap (tump line). Some women's groups assumed
management of unused stretchers, ensuring their accessi
bility and promoting their use, with 35 groups levying a
fee for usage
Awareness raising through video shows
During the community meetings, many groups felt that
there was a lack of awareness about perinatal health
problems and how to deal with them. MIRA had previ
ously researched and produced a 20 minute film about
newborn care in Makwanpur and the groups were keen to
use it in their communities. Group members approached
those households in the community with electricity and a
television, and the video was shown in homes or public
buildings. Although not all of the study area has electric
ity, the video was shown in 10 out of 12 Village Develop
ment Committee areas, attracting an audience of more
than 2100.
Participatory health education
During the identification of strategies to address prob
lems, there was a tendency to mismatch prioritized prob
lems and strategies. For example, one group suggested
tackling the problem of post-partum hemorrhage by
attending antenatal care. Another group considered that
the problem of vaginal discharge during pregnancy could
be addressed by training new Traditional Birth Attend
ants. During the first ten meetings, and from previous data
analyses, the team found an overwhelming preference for
care within the community, in terms of place of birth and
seeking solutions to health problems [ 16,27|. Home prac
tices with unequivocal allopathic clinical benefit were
rarely mentioned. There was also little knowledge about
Page 7 of 12
(page number not for citation purposes)
http://www.biomedcentral.eom/1471-2393/5/6
BMC Pregnancy and Childbirth 2005, 5:6
what kind of problems could be managed at different
health service institutions, and it appeared that communi
ties define the "seriousness" of a problem in a different
manner to the allopathic model.
Therefore, the team concluded that perinatal health edu
cation would be useful during the development of the
strategies. It was felt important to avoid turning the facili
tators into educators, and therefore a participatory form of
health education was developed, based around a picture
card game.
The picture card game
A packet of small hand held cards of different shapes was
developed in order to address the mismatch between
problems and strategies and to promote participatory
learning. Each shape of card represents either a problem
(circle shape), a prevention activity (triangle shape), a
home-care activity (house shape), or a health institution
(square shape). The cards are pictorial and were devel
oped with the MIRA health team and a local artist (see Fig
ures 3 and 4). They were extensively piloted with women's
groups and adapted accordingly. A manual for facilitators
was also developed to accompany the cards. The card
game is played in the group by passing round the cards
and discussing the pictures. The group members match
problem cards to their corresponding prevention activi
ties, home care activities or type of health institution that
could treat that problem. The card game worked well in
facilitating discussion, and women and facilitators both
enjoyed the learning experience. The team completed a
participatory evaluation of the game with a sample of
groups which indicated that the game also facilitated
learning about danger signs, home care and prevention
activities. Group members are presently taking the picture
cards on visits to pregnant women in the community who
are not group members.
Service quality spin-offs
Community health volunteers
The facilitator has worked to involve and support female
community health volunteers with their work in the com
munity. 70 group meetings have regular attendance and
active involvement of the local female community health
volunteer and traditional birth attendant. The female
community health volunteer is the lowest cadre of govern
ment appointed health staff and is responsible for one
ward. She is unpaid and has a broad job description
mainly focused around health education. In theory, she
should run monthly women's group meetings to facilitate
health education and discuss issues of maternal and new
born health. Although she receives initial and refresher
training, she is often left to work unsupervised and unsup
ported, and in practice female community health volun
teers find it difficult to run women's groups. By seeking
the participation of the female community health volun
teer in the groups, we gave her a forum to conduct her
work and increase her contact with her user group.
In twelve wards the women's group was invited by the
local health institution to play an active role in selection
of new female community health volunteers and tradi
tional birth attendants. The group had created good links
with the health institution, which could be exploited for
future service quality improvements. Clearly, the health
institution and the community consider many groups as
legitimate entities with a role to play in the health of
women and their children. It also appears that members
of women's groups have become more involved witih their
local health services.
In one area women's group members responded to the
needs of women visiting an outreach clinic for antenatal
care and family planning. Women were complaining of a
lack of privacy and there was no furniture in the clinic. The
group contacted the local forest user group to supply
materials for rudimentary furniture and collected money
for the purchase of cloth for curtains. In several capes, the
women's group was a medium for brokered links between
health service providers and users.
Proxies for empowerment effects?
One group put a sign on the door of their meeting place
indicating a sense of ownership. Ten out of 12 facilitators
have been invited to participate in other meetings and
community activities as their role as key actors in the
community is being recognized and developed. Women's
groups have sung the song from the video film at the
annual women's festival, and a supervisor initiated discus
sion about newborn health in a local bus using a cassette
of the song from the film. Another women's group organ
ised a perinatal care quiz that was carried out with nearby
women's groups and community members.
Discussion
We have described the development and implementation
of a large scale participatory intervention with women's
groups. This was adapted from a smaller scale project
developed in Bolivia, and implemented in a poor rural
population of 86 000 in Nepal. The effects on health out
comes, reported elsewhere, were dramatic.
During the process of developing and implementing the
intervention we had to be flexible and respond to the
needs of the group. Croup members and the wider com
munity clearly faced difficulties when thinking about
ways to tackle perinatal problems. These difficulties raised
issues around culture and our facilitation role.
Page 8 of 12
(page number not for citation purposes)
http://www.biomedcentral.eom/1471-2393/5/6
BMC Pregnancy and Childbirth 2005, 5:6
*
fH
p
F
■-
ki■
I
A
■
p
‘
I fe
/*'
‘f.
.
)p
Ji
w
Figure 3
Women's group using picture cards
Striking a balance between support and directiveness
It is highly likely that the facilitation team's attempts to
adequately support the facilitators may have led to less
participatory processes taking place, especially in the case
of strategy development. The facilitation manual was con
sidered by the facilitation team as an essential resource.
Examples were often given to enable facilitators to grasp
key concepts before conducting a meeting. The manual
was designed as a reference guide but evidently became
more like an instruction booklet, as the strategies most
commonly adopted during community meetings were
those given as examples in the manual. The reasons
behind this usage of the manual illustrate some of the key
issues in implementing a participatory project. To truly
facilitate, and not be directive, is a difficult technique to
learn, especially in a hierarchical society where the facili-
tator's education has emphasised rote learning rather than
independence of thought [28]. Our facilitators were also
keen to educate and provide the answers, and it may have
appeared easier (in the short term) to suggest things for
groups to do than to facilitate open discussion. The self
confidence or ability of the facilitator to manage the chaos
and unpredictability resulting from a truly participatory
process was often lacking, although their facilitation skills
developed with time.
Power and culture
Difficulties in linking problems to strategies may also be
explained in part by the cultural phenomenon of fatalism,
'ke i’arne' (what to do7.). Bista described ‘kegame' as a belief
in fatalism which leads to the feeling that "one has no per
sonal control over one's life circumstances, which are deter-
Page 9 of 12
(page number not for citation purposes)
http://www.biomedcentral.eom/1471-2393/5/6
BMC Pregnancy and Childbirth 2005, 5:6
111
■7
o
A
/
7\
C’'
r-""
7
ft
.x-..
*
I;
|
e
i
i
»
Uh I eV'
Tr
I
I
Figure 4
Picture card game and manual
.
mined through ti dipine or powerful external agency" 129 ]. He
argued that this fatalism and dependency affects the work
ethic and achievement motivation in Nepal. Concepts of
planning, orientation to the future, and sense of causality,
are all affected. Our study experience was that fatalism
affected both the way people viewed themselves in rela
tion to a problem, and also the power and capacity they
believed themselves to have in overcoming it.
What has been the impact of the women's groups and their
strategies?
In cases where there was a mismatch between problem
and strategy, or when groups developed strategies sug
gested by MIRA, we hope that these groups will benefit
from the implementation process alone. The strategies in
the manual are not necessarily evidence-based, and it may
be that the process of implementation is more beneficial
than the strategy itself. Through implementation, interac
tion between the wider community and the group may be
increased, knowledge about the group may spread, and
more people may become interested and involved in
issues of perinatal health. To enable a better understand
ing of the intervention process, evaluations using both
qualitative and quantitative methodologies are underway.
The impact of the women's group intervention was evalu
ated in a cluster randomized controlled trial which
showed a 30% reduction in neonatal mortality rates and
a reduction in maternal mortality rates in the first 30
months of the trial|ll|. Qualitative analysis will explore
perceptions and process indicators to assess how the inter
vention affected the study area and community stakehold
ers. Cost analysis of the intervention will enable estimates
Page 10 of 12
(page number not for citation purposes)
http://www. biomedcentral.com/1471 -2393/5/6
BMC Pregnancy and Childbirth 2005, 5:6
24.
25.
26.
27.
28.
29.
WHO: Guidelines for introducing simple delivery kits at the
community level. Geneva, Maternal and Child Health Unit, Division
of Family Health, World Health Organisation; 1987.
WHO: Essential Newborn Care: Report of a Technical Work
ing Group. Geneva, World Health Organisation; 2000.
Beun M: Use of the clean home delivery kits in Nepal: a qual
itative study. Seattle, USA, Program for Appropriate Technology
(PATH); 2002:1-34.
Osrin D, Tumbahangphe K, Shrestha D, Mesko N, Shrestha B, Manandhar M, Standing H, Manandhar D, Costello A: Cross sectional,
community based study of care of newborn infants in Nepal.
British Medical Journal 2002, 325:1063-1066.
Dixit S: Education, deception, state and society. In The state of
Nepal Edited by: Dixit KM and Ramachandaran S., Himal Books; 2002.
Bista DB: Fatalism and Development: Nepal’s Struggle for
Modernisation. India, Orient Longman Ltd; 1991.
Pre-publication history
The pre-publication history for this paper can be accessed
here:
http://www.biomedcentral.eom/1471-2393/5/6/prepub
Publish with BioMect Central and every
scientist can read your work free of charge
"BioMed Central will be the most significant development for
disseminating the results of biomedical research in our lifetime."
Sir Paul Nurse, Cancer Research UK
Your research papers will be:
• available free of charge to the entire biomedical community
• peer reviewed and published immediately upon acceptance
• cited in PubMed and archived on PubMed Central
• yours
you keep the copyright
Submit your manuscript here:
http://www.biomedcentral.com/info/publishing adv.asp
BioMedcentral
Page 12 of 12
(page number not for citation purposes)
http://www. biomedcentral. com/1471 -2393/5/6
BMC Pregnancy and Childbirth 2005, 5:6
of cost-effectiveness and sustainability to be made. A com
parison of the socio-economic status of women's group
members with non-group members will allow an estimate
of the equitability of the intervention.
The Millennium Development Goals for reductions in
maternal and neonatal mortality in developing countries
are unlikely to be met by 2015. In populations where
maternal and newborn mortality rates are highest, most
deliveries occur at home. It is essential that Safer Mother
hood and Newborn Care Programmes design interven
tions which reach out to the poorest groups in order to
change care practices at home, and care seeking for illness
or complications of childbirth. Our participatory work
with women's groups provides a model for an interven
tion that can be scaled rapidly in even the poorest and
most remote communities.
Conclusion
A large scale participatory intervention to improve preg
nancy outcomes in rural Nepal through 111 women's
groups has been described. Although we have faced con
textual, cultural and security problems, we believe that the
participatory approach can be a powerful tool in unleash
ing the creative potential to solve perinatal health prob
lems in communities. Such an approach may have lasting
benefits, affecting behaviour in subsequent pregnancies.
Competing interests
The author(s) declare that they have no competing
interests.
References
I.
2.
3.
4.
5.
6.
7.
8.
9.
IO.
II.
12.
13.
14.
Contribution by authors
|M wrote the first draft of the paper and contributed to the
study design and collection of field data. ST, and BS con
tributed to the study design, collection of field data and
analysis, and criticised later drafts of the paper. NM and
DO contributed to the study design and analysis, and crit
icised later drafts of the paper. MM and 1 IS contributed to
the design of the study and criticised drafts of the paper.
DM and AC contributed to the design of the study and
supervision of the field programme, and criticised drafts
of the paper. JM and AC will act as guarantors for the
paper.
Acknowledgements
15.
16.
17.
18.
19.
20.
The authors wish to thank the many individuals in Makwanpur District who
gave their time generously and without complaint, and the field staff of the
MIRA Makwanpur team, without whom the study would have been impos
2I.
sible. We also thank the Makwanpur District Development Committee and
the Village Development Committee members for their active and contin
22.
uing support; and the MIRA executive committee in Kathmandu. The study
was funded by the British Government Department for International
Development, UNICEF Nepal, and the Division of Child and Adolescent
Health. World Health Organisation, Geneva, with additional financial assist
ance from UNICEF, Nepal and UNFPA, Nepal.
23.
WHO: Declaration of Alma-Ata: 6-1 2 September; Alma-Ata,
USSR. World Health Organisation; 1978.
Rifkin S: Community participation in maternal and child
health/family planning programmes. Geneva, World Health
Organisation; 1990.
Rifkin S: Paradigms Lost: toward a new understanding of com
munity participation in health programmes. Acta Tropica 1996,
61:79-92.
Morgan L: Community participation in health: perpetual
allure, persistent challenge. Health Policy and Planning 2001,
16:221-230.
Arnstein SR: A ladder of citizen participation. Journal of the Amer
ican Institute of Planners 1969, 35:216-224.
Howard Grabman L, Seoane G, Daven port C: The Warmi
project: a participatory approach to improve maternal and
neonatal health - an implementor’s manual. Arlington, Vir
ginia, Mothercare/USAID; 1992.
Bang A, Bang R, Baitule S, Reddy MH, Deshmukh M: Effect of home
based neonatal care and management of sepsis on neonatal
mortality: field trial in rural India. The Lancet 1999,
354:1955-1961.
Paul V: Newborn care in India: a promising beginning but a
long way to go. Seminars in Neonatology 1999, 4:141-149.
Nepal Ministry of Health: Nepal Demographic and Health Sur
vey 200 I. Kathmandu, Family Health Division, Ministry of Health and
Calverton: New Era; ORC Macro; 2002.
Osrin D, Mesko N, Shrestha B, Shrcstha D, Tamang S, Thapa S, Tumbahangphe K, Shrestha J, Manandhar M, Manandhar D, Standing H,
Costello A: Implementing a community-based participatory
intervention to improve essential newborn care in rural
Nepal. Trans R Soc Med Hyg 2003, 97:18-21.
Manandhar DS, Osrin D, Shrestha BP, Mesko N, Morrison J, Tumbahangphe KM, Tamang S, Thapa S, Shrestha D, Thapa B, Shrestha JR,
Wade A, Borghi J, Standing H, Manandhar M, Costello AML, team
MIRAM: Effect of a participatory intervention with women's
groups on birth outcomes in Nepal: a cluster randomised
controlled trial. Lancet 2004, 364:970-979.
Central Bureau of Statistics: Statistical year book of Nepal 200 I.
Kathmandu, His Majesty's Government National Planning Commis
sion Secretariat; 2001.
Central Bureau of Statistics: Gender disaggregated indicators.
Kathmandu, Ministry of Women, Children and Social Welfare; 2002.
Pradhan A, Aryal R, Regmi G, Ban B, Govindasamy P; Nepal Family
Health Survey 1996. Kathmandu, Ministry of Health, Nepal; New
ERA; Macro International Inc.; 1997.
Nepal Research Associates: Makwanpur District Profile. Hetauda,
District Development Committee; 1999.
Mesko N, Osrin D, Tamang S, Shrestha B, Manandhar D. Manandhar
M, Standing H, Costello A: Care for perinatal illness in rural
Nepal: a descriptive study with cross-sectional and qualita
tive components. BMC International health and human rights 2003,
3:.
O’Rourke K, Howard Grabman L, Seoane G: Impact of commu
nity organization of women on perinatal outcomes in rural
Bolivia. Rev Panam Salud Publica/Pan Am /Public Health 1998, 3:9-14.
Tamang S, Mesko N, Shrestha B, Osrin D, Manandhar M, Standing H,
Manandhar D, Shrestha J, Costello A: A qualitative description of
perinatal care practices in Makwanpur District, Nepal. Contri
butions to Nepalese Studies 2001,28:10-19.
Werner D, Bower B: Helping Health Workers learn. California.
USA, The Hesperian Foundation; 1982.
Green CP: Mother Support Groups: A Review of Experience
in Developing Countries. Arlington, Va., U.S. Agency for Interna
tional Development, by the Basic Support for Institutionalizing Child
Survival (BASICS) Project; 1998.
Asthana S: Women’s health and women’s empowerment: a
locality perspective. Health and Place 1996, 2:1 -1 3.
Howard Grabman L: Reasons some women participate in local
organizations and others don't. Washington, D.C., Save the
Children; 1998.
JHPIEGO: Pre Analytical report and recommendations in
Financial Planning for Safe Motherhood in Nepal: a market
research and tools development report. Kathmandu, Nepal,
Johns Hopkins Program for International Education in Gynecology
and Obstetrics; 2001.
Page 11 of 12
(page number not for citation purposes)
■
st
Articles
Effect of a participatory intervention with women's groups
on birth outcomes in Nepal: cluster-randomised controlled
trial
l.uncet -2004; 364: 970-79
See Comment page 914
Mother and Infant Research
Dharma S Manandt^ar, David Osrin, Bhirn Prasad Shrestha, Natasha Meskojoonnu Morrison, Kirti Man Tumbohangphe, Suresh Tamang,
Sushma Thapa, Dej Shrestha, Bidur Tliapa, Jyoti Raj Shrestha, Angie Wade, Josephine Borghi, Hilary Standing, Madan Manandhar,
Anthony M de I. Costello, and members of the MIRA Makwanpur trial team
Activities (MIRA). PO Box 921,
Kathmandu. Nepal
(Prof D S Manandhar F RC.P,
B P Shrestha MSc,
K M I ijriibahangplw MSc.
S 1 aniniiy BSt. s ’>hapa MS<,
Summary
Background Neonatal deaths in developing countries make the largest contribution to global mortality in children
younger than 5 years. 90% of deliveries in the poorest quintile of households happen at home. We postulated that a
community-based participatory intervention could significantly reduce neonatal mortality rates.
D Shieslha BSi, G Tlwpa BSi,
I R Shrestha MPH); International
Perinatal Care Unit
(I) Osrin MRCP, N Mesko MSt.
J Morrison MSi, j Borghi MSc,
Pro! A M de I < osrello l;R( P- and
Centre for Paediatric
Epidemiology and Biostatistics
IA W.ide PhD 1, Institute of Child
Health. University College
London, JO Guilford Street
London WC1N 1EH, UK;
Institute of Development
Methods We pair-matched 42 geopolitical dusters in Makwanpur district, Nepal, selected 12 pairs randomly, and
randomly assigned one of each pair to intervention or control. In each intervention duster (average population
7000), a female facilitator convened nine women’s group meetings every month. The fiicilitator supported groups
through an action-learning cycle in which they identified local perinatal problems and formulated strategics to
address them. We monitored birth outcomes in a cohort of 28 931 women, of whom 8% joined the groups. Ihc
primary outcome was neonatal mortality rate. Other outcomes included stillbirths and maternal deaths, uptake of
antenatal and delivery services, home care practices, infant morbidity', and health-care seeking. Analysis was by
intention to treat. The study is registered as an International Standard Randomised Controlled Trial, number
ISRCTN31137309.
Studies, Palmer, Brighton BN1
9RE,UK(H Standing PhD); and
Nepal Administrative Staff
College, Kathmandu, Nepal
(M Manandhar PhD)
( orrespondcncc to:
Prol Anthony Costello
ipu@ich.ucl.ac.uk
Findings From 2001 to 2003, the neonatal mortality rate was 26-2 per 1000 (76 deaths per 2899 livebirths) in
intervention clusters compared with 36 9 per 1000 (119 deaths per 3226 livebirths) in controls (adjusted odds ratio
0 70 [95% CI 0 53-0-94]). Stillbirth rates were similar in both groups. The maternal mortality ratio was 69 per
100000 (two deaths per 2899 livebirths) in intervention clusters compared with 341 per 100 000 (11 deaths per
3226 livebirths) in control clusters (0-22 [0 05-0-90]). Women in intervention clusters were more likely to have
antenatal care, institutional delivery, trained birth attendance, and hygienic care than were controls.
Interpretation Birth outcomes in a poor rural population improved greatly through a low cost, potentially sustainable
and scalable, participatory intervention with women’s groups.
Introduction
Of the world’s 4 million annual neonatal deaths. 98%
occur in developing countries. Infant and child
mortality rates have declined, notably through better
control of diarrhoea, pneumonia, and vaccinepreventable disease, and the importance of the newborn
period has increased. In India, neonatal mortality now'
accounts for up to 70% of infant mortality/ Most
perinatal and neonatal deaths happen at home, and
many could be avoided w'ith changes in antenatal,
delivery, and newborn care practices.’ However, primary
and secondary health-care systems have difficulties in
reaching poor rural residents, and a potentially effective
perinatal health strategy must recognise this reality. In
Makwanpur district, Nepal, for example. 90% of women
give birth al home, and trained attendance at delivery is
uncommon.4
We are unaware of any randomised controlled trial of
community-based strategies to reduce neonatal mortality,
a shortfall that indicates the absence of information on
demand-side interventions. Two sludics have made
important contributions in this area. Bolivia’s Warmi
970
project—an uncontrolled before-and-after study—was
implemented in a poor rural population of 15 000 people
with little health-system infrastructure. The project
worked with women's groups to encourage participatory
planning for mother and infant care.' “ and showed a fall
in perinatal mortality' rate from 117 to 44 per 1000 births
over 3 years, hi India, the SEARCH group reported a non
randomised controlled study from a rural population of
80000 in Gadchiroli, Maharashtra.1' The intervention
entailed training of traditional birth attendants, health
education, and a new cadre of supervised village health
workers who visited newborn infants at home, identified
warning signs, and managed sepsis with antibiotics. After
3 years the neonatal mortality rate had fallen by 62%.
Replication and scaling up of this exciting community
based model presents policy makers with some
challenges, particularly because of the need for a new
cadre of community health worker to deliver injectable
antibiotics al home.
Community participation has long been advocated to
build links between primary services and their users.''
and to improve service quality.' * However, the evidence
www thelamei com Vol 364 September 11, 2004
Articles
base for the effectiveness of participatory models is
scarce. '" Previously, we showed no effect of direct
education by health workers on infant care practices and
care-seeking behaviour after delivery.” In view of the
Bolivian model, we thought that a participatory approach
might have more effect on perinatal care practices and
might increase consultation for difficulties in pregnancy
and the newborn period. Although external facilitators of
user groups have proven valuable in agriculture and
forestry',S i l to our knowledge no study has rigorously
assessed such a potentially scalable approach to
improving reproductive health outcomes.
We postulated that a community-based participatory'
intervention could reduce the neonatal mortality rate
from 60 to 40 per 1000 livebirths. The MIRA
Makwanpur trial was a cluster-randomised controlled
trial of such an inlcrvcnlion in a rural mountainous area
ot Nepal. I he trial tested a large-scale intervention, using
facilitators to work with women s groups in a population
of 170000 covering 1600 knT. A cluster design was
chosen because the intervention was structured around
communities rather than individuals.
NEPAL
• Kathmandu
INDIA
A
4 Makwanpur
\ 7 21\
•w-
J / 18
02,
5
713
$ Hetauda rriuriKipallty
□ Intervention area
Methods
TIBET
20
24
□ Control area
Study location and population
With a population of more than 23 million and a gross
national income of US$240 per person/0 Nepal is a p-''
”- Figure 1: Makwanpur District, Nepal, and distribution of study village
poor
country whose development challenges are exacerbated| development committees
by its geography and unstable political situation. Life
expectancy is 61 years. The total fertility rate is geopolitical unit, committee representatives were key
4-4 children per woman in rural areas/1 and the points ot liaison, and discussions with local people
estimated maternal mortality ratio is 539 per suggested that randomisation of smaller units would
100000 livebirths.'' 57% of women cannot read.-1 The increase the risk of contamination. All 43 village
estimated infant mortality rate is 64 per 1000 livebirths, development committees in Makwanpur district were
the neonatal mortality rate 39 per 1000 livebirths, and eligible for randomisation, of which one was excluded at
the perinatal mortality' rate 47 per 1000 births/' In rural baseline for security reasons.
areas, 94% of babies are born at home/- and only 13% of
We enrolled a closed cohort of married women of
births are attended by trained health workers/
reproductive age. Inclusion criteria were: consent given
Makwanpur district lies in Nepal's central region for involvement age 15-49 years inclusive on
where the middle hills join the plains. The population of June 15. 2000; married; and potential to become
about 400000 subsists mainly on agriculture and the pregnant. Exclusion criteria included long-term
largest ethnic groups are Tanlang and Brahmin-Chhetri. separation from spouse and widowhood. Women who
The district hospital in the muniicipality
' ’’ of Hetauda has chose to participate in the study gave verbal consent and
facilities for antenatal care and delivery', althoughi were free to decline to be interviewed at any time.
operative delivery was not available during the study
period. There are 7852 people per hospital bed/4 The Procedures
district health system makes perinatal care available We matched 42 village development committees into
through a network of primary' health centres, health 21 pairs. Because we did not have disaggregated neonatal
posts, subhealth posts, and outreach clinics. Traditional mortality figured, pairing was based on a process
birth attendants are available throughout the district, but of topographic
stratification,
. w
. grouping of village
their attendance al births is less common than in some
development- committees with similar ethnic group
other parts of south Asia.4
Nepal is administratively divided in descending order
ol size—into development regions, zones, districts,
village development committees, and wards. We chose
the village development committee as the cluster unit of
randomisation lor the following reasons: it is a standard
distributions, and matching of pairs with similar
population densities. We used a list of random numbers
wv.v? •hcl.nucl iioiii Vol 364 September 11, 2004
to select 12 pairs. These 24 village development com
mittees formed the study clusters. We randomly allocated
one cluster in ea'ch pair to either intervention or control
on the basis of a coin toss (figure 1). Because of the
971
Articles
Figure 2 Typical women's group meeting
Picture cou'tesv of Thomas Kelly and Save the Children, USA.
nature of the intervention the trial allocation was not
masked, but analysis of primary and secondary- outcomes
was not done until just before the data monitoring
committee meeting at 30 months. We generated the
cluster allocation sequence in Kathmandu before
enrolment of participants.
Enrolment activities were done from September, 1999,
to November, 2000. A team oflocal enumerators mapped
the 24 village development committees on foot, identified
and allocated a unique identification number to every
household (defined as a group of individuals sharing one
kitchen), did a baseline census of demographic and
socioeconomic indicators, and generated a list of female
household members according to predefined written
protocols. This document was scrutinised by a data
auditor, a surveillance manager, and a group of local
supervises and converted into a list of women meeting
the inclusion criteria for the cohort. From March, 2001,
to July. 2001, a team of 44 field interviewers visited every
potential member of the cohort, reassessed her for
inclusion, explained the study, asked for her consent,
allocated her a unique, identification number, and
completed an individual questionnaire, which included
questions on demography, education, maternity history,
details of any preceding pregnancy, home-care practices,
and use of health services for perinatal illness.'
Surveillance began in February', 2001, and involved
28931 participants in 28 376 households. The strategy we
used was adapted from one used by the Nepal Nutrition
Intervention Project, Sariahi,-” and has been described in
detail elsewhere/1' It entailed 255 ward enumerators,
25 field interviewers, and nine field coordinators. The
local female enumerator visited all cohort members in
die ward she was responsible for every' month over the
study period to record menstrual status. She recorded
data on individualised printed forms. The nine ward
972
enumerators of every cluster met with a cluster
interviewer once a week. In the absence of other
explanatory circumstances, pregnancy was registered
when a cohort member ceased menstruation, for
3 months. The cluster interviewer did two interviews for
every pregnancy: the first, at 7 months of gestation, as
near as possible to the transition between our definitions
of miscarriage and stillbirth: and the second at 1 month
postpartum, as near as possible to the transition between
neonatal and infancy periods. In the event of an
unfortunate outcome (miscarriage, maternal death,
stillbirth, or neonatal death), the interview was done by a
senior field coordinator. The interview was developed in
Nepali and piloted and repiloted by the local team. It was
modular to deal with different outcomes, covering
antenatal, delivery and postpartum care, home care
practices, maternal morbidity, neonatal morbidity, health
service usage, and cause of death in the event
of mortality.
In the event of neonatal death, we used an approach
relined and locally adapted from existing questionnaires
to establish cause of death. An open question about the
cause of death was followed by a modular series ol
closed questions. The answers to these questions were
designed to produce a classification of 14 causes of
neonatal death based on those used by SEARCH in
India/' and were classified by a paediatrician (DO) on
Phase
Meeting Aim
Introduction
1
To introduce the study to the gioup
To discuss why mothers and newborn infants die and
how the itilervenlion will work in the lonirnttnily
Problem
identification
To as< ertain how women imdersland inalernal and
neonatal problems
I o find out about maternal and neonatal problems in
the cciinmiHitly
To understand the frequency of maternal and
neonatal problems and to identify strategies to obtain
information in the community
Problem
prioritisation
6
I g share mformatic’ii from other women in the
coniniunity and to priori Use three important rnaiemal
and neonatal ht'alih problems
Planning
together
To discuss possible strategies for addressing the
priority problems
8
I o discuss involvement of other community members
in developing strategies
9
To discuss preparation for a meeting of community
members
10
To hold a meeting involving other community
members to discuss the activities of the
women's yroups, the piiority problems identified
by the groups, and possible strategies, and reach
consensus
Table 1: First ten women's group meetings
www ihelance’. COm Vol 364 September 11, 2004
Articles
the basis of open text responses, modular closed
questions, and a computer algorithm.
The average population per cluster was about 7000,
spread over an area of 60 kmf For every intervention
cluster we recruited one local female facilitator. Shortlists
for this role were derived from nomination by community
leaders, advertisement, and word of mouth, after which all
potential candidates were interviewed. A cluster consisted
of nine wards. The facilitator—a literate locally resident
woman—convened one women’s group meeting a month
iti every ward (figure 2). Some groups set up by local
female community health volunteers already existed but
their activity was sporadic. The role of the facilitator was to
activate and strengthen groups and support them through
an action research cycle.
The intervention needed a facilitator rather than a
teacher, with abilities and training in participator}'
communication techniques. She needed to have a grasp
of perinatal health issues and some knowledge of
potential interventions so she could act as a broker of
information and a catalyst for change. Although it was
important that none of the facilitators had a health
background, we gave them brief training in perinatal
health issues. Supervision, and a manual based on the
Warmi project methodology,” was integral to facilitator
training and support. One supervisor provided support
for every' three facilitators by attending group meetings
and making regular community visits.
The first step of the intervention was to discuss
issues around childbirth and care behaviours in the
community, which allowed facilitators to develop
participatory learning skills and generated information on
pregnancy and childbirth, covering beliefs and practices
in both uncomplicated and complicated pregnancies."':s
file facilitators then supported the women's groups
through monthly meetings (table 1). This phase of ten
meetings lasted almost a year. In the next steps of the
intervention, the women's groups implemented and
assessed their strategies. One result of the process was
that women sought more information about perinatal
health. This information was provided through the
iterative design and playing of a picture card game that
addressed prevention, treatment, and consultation for
typical problems in mothers and babies.
The form and content of discussions within women’s
groups varied, as did levels of involvement and potential
strategies. Some typical strategies were community
generated funds for maternal or infant care, stretcher
schemes, production and distribution of clean delivery'
kits, home visits by group members to newly pregnant
mothers, and awareness raising with a locally made film
to create a forum for discussion. Throughout the
process the groups were involved in other health-related
activities in their communities.
A baseline service audit identified weaknesses in the
provision of antenatal, delivciy. and newborn care in
Makwanpur district. Because we aimed to lost solely the
www thelancel coni Vol 364 September 11, 2004
effects of the women’s group intervention, health-service
strengthening activities were undertaken in both
intervention and control areas. We decided to do this
process on ethical grounds because w’e hoped that it
would benefit control areas"’ and on theoretical grounds
because we thopght that a degree of improvement in
sei'vices would be necessary' for the success of the trial
intervention. We therefore ensured that primary health
centres in the study area were equipped with locally made
resuscitaires (open incubators that allow' access to
newborn babies while keeping them warm), phototherapy
units, warm cots, and neonatal resuscitation equipment.
We remedied some shortfalls in essential neonatal drugs
once only and discussed strategies for resupply with local
heal th-service managers. In partnership with the District
Public Health Office, we organised training in essential
newborn care foi[ all cadres of government health staff and
for female community' health volunteers and traditional
birth attendants. Community-based workers received a
basic newborn care kit containing a rubber bulb for
suction, tube-and-mask for assisted respiration, iodine,
gauze, a baby wrapping cloth, and a pictorial manual.
We postulated that the women’s group activities might
lead to reductions in neonatal mortality rates in
intervention clusters compared with control clusters. At
the outset, we did not think perinatal mortality' rates
would be affected much, since we did not envisage that
changes in honie-care practices would lead to reductions
in stillbirth ratejs.
The primaiy outcome was neonatal mortality rate
(deaths in the first 28 days per 1000 livebirths).
Prospective interviews undertaken through the
surveillance system provided information on several
other outcomes, including stillbirths and maternal
deaths, uptake of antenatal and delivery services, home
care practices at delivery' and postpartum, infant
morbidity, aiijd health-care seeking. We obtained
background
demographic
and
socioeconomic
information to investigate cluster comparability.
Surveillance! coordinators observed
10% of
interviews and reviewed all questionnaires at nodal
points in the field before transmitting them for review
by data auditors. After audit and correction, sometimes
needing transfer back to the site of collection, data
were doublefentered into a relational database
management system in Microsoft SQL Server 7.0
(Microsoft Corporation. Redmond, WA, USA). The
system further addressed data quality through
predefined acceptability constraints.
We defined miscarriage as cessation of a
presumptive pregnancy before 28 weeks of gestation
and stillbirth las fetal death after 28 weeks of gestation
but before delivery of the baby’s head, which was a
modification of the 22-week definition to meet local
practicalities. We classified neonatal death as death of a
liveborn infant within 28 completed days of birth. Early
neonatal deaths refer to deaths within 7 completed
973
Articles
team were recruited locally and undertook their activities
in their home areas. When the study surveillance team
noted minor illness in mothers or infants, they
encouraged attendance at an appropriate health facility. In
1 duster excluded for security reasons
the event of severe illness, team members had an ethical
responsibility to assist with rapid and appropriate
I 21 pairs of clusters matched
transport and treatment, irrespective of allocation. All
I 12 pairs of clusters randomly selected
information provided by participants remained
confidential. Access to information was restricted to
4
interviewers, supervisors, data auditors, and officers, and
Clusters randomly allocated within pairs
research staff at the analytical level. No analyses or outputs
I
included the names of participants.
4_______
____________ 4’
43 dusters assessed for eligibility
Enrolment
Allocation
12 clusters allocated intervention
Median households 1133 (range 433—2838)
12 clusters allocated control
Median households 733 (range ?3f>—3814)
14 884 participants
12 clusters received intervention
14 047 participants
12 dusters received control
4
4
Follow-up
3036 participants became pregnant
288/ had one pregnancy
144 had two pregnancies
S had three pregnancies
3190 pregnancies
J
Lost to follow-up:
0 dusters
37 participants moved out of study area
73 participants miscarried before 7 months
1 participant declined to participate
134 participants with incomplete data
4
Analysis
12 dusters analysed
Participants analysed:
2945 deliveries
2972 infants born (2/ pairs of twins)
73 stillbirths
2899 livebirths
76 neonatal deaths
2823 infants alive at 1 month
3344 participants became pregnant
3166 had one pregnancy
176 had two pregnancies
2 had three pregnancies
3524 pregnancies
I
Lost to follow-up:
0 dusters
1 participant died in first 7 months of
pregnancy
52 participants moved out of study area
77 participants miscarried before 7 months
1 participant declined to participate
123 participants with incomplete data
4
12 dusters analysed
Participants analysed:
3270 deliveries
3303 infants born (33 pairs of twins)
77 stillbirths
3226 livebirths
119 neonatal deaths
3107 infants alive at 1 month
Figure 3: Trial profile
days and late neonatal deaths from 7 to 28 completed
days of birth. Perinatal death describes either a
stillbirth or an early neonatal death.
The study was approved by the Nepal Health
Research Council and the ethics committee of the
Institute of Child Health and Great Ormond Street
Hospital for Children, and was done in collaboration
with His Majesty’s Government Ministry' of Health,
Nepal. We discussed the aims and design of the trial at
a national meeting in 1998. After this time, we held a
series of meetings with members of the Makwanpur
District Development Committee, the Chief District
Officer, and local stakeholders. In early 2000, the
chairpersons of the 24 village development committees
involved in the study gave signed consent on behalf of
their communities.
Benefits to the control clusters were improvements in
equipment and training provided at all levels of the health
care system. All community-based members of the study
974
Statistical analysis
To determine the number of cluster pairs to be enrolled,
we had to estimate the coefficient of variation in
outcome between clusters within matched pairs
and
the expected number of births per cluster over the
timescale of the study. Based on national and district
estimates, we assumed a neonatal mortality rate of
60 per 1000 livebirths, an average of 480 births per
cluster, and a k.„ value in the range 0-15-0-3. We
estimated that inclusion of 12 pairs of clusters would
allow us to detect a reduction in neonatal mortality of
between 27% and 38% (37-44 per 1000 livebirths) with
80% power at a 5% significance level.1' The
corresponding estimates of intradass correlation are
between 0-0055 and 0-0061. Because we did not
envisage any adverse effects of the intervention at either
cluster or participant level we did not use any stopping
rules. After the first year of surveillance, we saw that
birth rates were lower than expected on the basis of
estimates. The trial steering group decided not to assess
neonatal mortality rates until we had obtained data for
2 complete years of births from introduction of the
inteivention. We therefore undertook a preliminary7
analysis in November. 2003, and presented the findings
to an independent data monitoring committee. The
committee considered issues of quality, confidentiality,
and analysis and recommended definitive analysis and
publication of the 2-year findings.
The analysis was undertaken as intention to treat at
both cluster and participant levels. Participants who had
begun the trial as residents of a given cluster were
retained as residents even if they had moved to another
cluster during the trial period.
Within the prospective cohort, we compared neonatal
mortality rates, stillbirth rates, and maternal mortality7
ratios between control and intervention groups, taking
account of clustering and the paired nature of the data,
with hierarchical logistic models (Mlwin version 1.1).
We estimated intraclass correlation coefficients from
retrospective neonatal mortality and stillbirth data by
analysis of variance within Stata version 8. Secondary7
outcomes and process indicators were compared with
adjustment for clustering. All estimates are presented
with 95% Cis. This study is registered as an
www.lhelancet.com Vol 364 September 11, 2004
Articles
International Standard Randomised Controlled Trial,
number ISRCTN31137309.
Role of the funding source
Representatives of the UK Department for International
Development (DFID) suggested that no health-care
activities should be carried out in parallel with existing
government services and that—for sustainability
reasons—no funding should be available for women’s
group activities. Apart from these issues, the sponsors of
the study had no role in study design, data collection,
data analysis, data interpretation, or writing of the
report. The corresponding author had full access to all
the data in the study and had final responsibility for the
decision to submit for publication.
Results
Figure 3 shows the trial profile. All 24 clusters selected
lor inclusion received their allocated intervention.
Between Nov 1,2001. and Oct 31.2003. 3190 pregnancies
happened in inteivention clusters and 3524 in controls.
Presumptive miscarriage rates were 2-3% (73/3190) in
intervention clusters and 2-2% (77/3524) in control
clusters. Loss to pregnancy follow-up as a result of
migration, withdrawal of consent, or incompleteness of
surveillance data was 5-4% (172/3190) in intervention
clusters and 5-0% (176/3524) in control clusters.
2972 births (including 54 twins) were available for
analysis in intervention clusters and 3303 (including
66 twins) in control dusters.
Table 2 presents baseline characteristics of
intervention and. control clusters. Although the median
number of households per cluster was lower in control
clusters, the total numbers of households and
participants who became pregnant were similar. Some
evidence exists of less poverty in intervention than
control clusters: household asset scores and participant
schooling—but not recalled annual food sufficiency—
seem to slightly favour the pooled intei-vention clusters.
The age breakdown does not suggest differences
between intervention and control clusters, either for
population structure or for participants who became
pregnant.
For estimated baseline mortality' rates, participants in
intervention clusters recalled 11415 livebirths and
290 neonatal deaths in the 5 years preceding our census
(neonatal mortality rate 25-4 per 1000 livebirths).
Participants in control clusters recalled 12 132 livebirths
and 304 neonatal deaths (neonatal mortality rate 25-1
per 1000 livebirths). Cluster-specific breakdown of these
pooled data showed that neonatal mortality rates were
higher in intervention than control clusters in four pairs.
similar in four pairs, and higher in control than
iiiiervcnlion clusters in lour pairs. Although valid for
cluster comparison, the prospective findings in the same
population suggest that maternity histories substantially
underestimated actual neonatal mortality rales.
VAvw thelanceUom Vol 364 September 11, 2004
Number of households
Median per i.luster (range)
Number of participants
Median per duster (range)
Household asset score
None <’( (he assets on (he list
Clock, radio, iron, or bicycle
More costly appliances
Household food sufficiency
Intervention
clusters
Control
clusters
148/9
1133(433-2838)
14884
1110 (487-2824)
13532
6122 (45%)
4094 (30%)
<316(2^%)
13532
13497
2923
/33 (236-3814) J‘> / (1S8-451)
1404/
3036
77/ (219-4069) 214(164-463)
121/0
3036
1545(51%)
6233(51%)
Pregnancies
in intervention
clusters
Pregnancies in
coihtrol
dujsters
31p9
.48(i9-/<i6)
3 44
U4 (61-8<5j
3:44
I £ i>(> (56%)
4476(37%)
954 (31%)
1461 (12%)
53/(18%)
3036
3:44
’>7J. (32%)
3036
1( 02 (30%)
Less than 8 months annually
4090(30%)
12170
3372(28%)
Participant age
13532
121/0
Younger (han 20 years
1130(8%)
973 (8%)
20-29 years
5192 (3«%)
II /I (35%)
0/ (')%)
3144
711 (22%)
4758(39%)
/19 (24%)
KOlf' (jt>%)
Dirp’ 154%)
704 (21%)
30-39 years
4265 (32%)
3/82(31%)
556(18%)
40 years oi older
2945(22%)
2657(22%)
65 (2%;
Participant schooling
13532
12170
2893
None
11031(82%)
I.0741 (88%)
Primary
957(8%)
Secondary nr higher
1562 (12%)
939 (7%)
472 (4%)
2122(73%)
503(1/%)
268 (9%)
Participant could not read
8981 (66%)
9664 (79%)
1.734 (60%)
....... .... ... j----
rl...... :.
105 (3%)
3141
2f>« 1 (85%)
807 (10%)
153(5%)
?1|18(77%)
Table 2: Baseline characteristics of intervention and control clusters and pregnancies in intervention and
control areas
Figure 4 shows within-cluster neonatal mortality rates
for each of the lu cluster pairs. The line of equality has
been superimposed on this graph. In 11 cluster pairs,
neonatal mortality rates were lower in the intervention
group. The pooled rate in the intervention group was
nearly 30% lower than in the control group (table 3).
Hierarchical modelling—taking clustering into account—
yielded an oddsl ratio of 0-70 (95% Cl 0-53-0-94) for
neonatal mortality in the intervention clusters compared
with the control clusters. The intraclass correlation
coefficient estimated from retrospective data was 0 ■ 00644
(95% CI 0-00004-0-0128).
Stillbirth rates did not differ between intervention and
control clustery (table 3). The intraclass correlation
coefficient estimated from the retrospective data was
90
80-
I
.
1
i
'
!
1
70-
I
60 —
I
40-
I
I
50-
3020100
10
I
I
30
I
40
I
50
I
6o
T
70
SO
90
( itiiliol NMFt (><-r 1000 livfbi'th>
fFigure4: Neonatal mortality rates in intervention and control areas
rNMR-neonalal mortality tale.
975
Articles
Intervention
clusters
Control
dusters
Documented births
2972
livebirths
2399
3303
3226
Stillbirths
73
77
Neonatal deaths
76
119
l arly (0-6 days)
50
26
l ate (7-28 days)
Adjusted odds
ratio (95% Cl)
70
49
11
Maternal deaths
Stillbirth rate per 1000 births
24 6
23-3
106(076-l-47)
Neonatal mortality rate per
26 2
369
070(0-53-094)
69
^4 1
0 72 (0 05-0 90)
1000 livebirths
Maternal mortality ratio per
100 000 livebirths
Table 3-’ Mortality rate comparisons between intervention and control
clusters
0-00438 (95% CI 0-0 00948) for the stillbirth rate. With
limited sample size, maternal mortality was not a
predefined outcome of the study. The maternal mortality
ratio was about 80% lower with intervention than with
control clusters (adjusted odds ratio 0-22 [95% CI
0-05-0-90]).
Table 4 presents process indicator outcomes. In
general, they suggest so-called healthier behaviours in
intervention clusters: women in these clusters were
more likely than those in the control clusters to have had
antenatal care, to have taken haematinic supplements, to
have given birth in a health facility, with a trained
attendant or a government health worker, to have used a
clean home delivery kit or a boiled blade to ent the
umbilical cord, and for the birth attendant to have
washed her hands. No differences were noted in delayed
Intervention clusters
Control clusters
Pregnancies
3190
3574
Any antenatal tare
1747 (55%)
1051(30%)
Any non and folic acid supplenn?nts
1574(49%)
1152(27%)
Any pen eiveil maternal illness from 7 months
668(21%)
926(26%)
2 82 (1 41-5-62)
1-99(1-14-3-46)
0-74 (0-43-1 28)
3-37(l 78 637)
Adjusted odds
ratio (95% Cl)
gestation io 1 month postpartum
Visited health facility in event of illness
333(50%)
20/(22%)
Deliveries
2945
3’270
Institutional deliveries
201 (7%)
66(2%)
Birtfi attended by goveininerit health provider
2/2(9%)
102 (3%)
|12 (1-62-6 03)
Spei ilii ally, doctor, nurse, or midwife
207 (7%)
69 (2%)
3 53(1 54-8-10)
Birth attended by traditional birth attendant
199(7%)
129(4%)
1-70 (0-93-3-11)
Used a dean home delivery kit
550(19%)
4- 59 (2 83-7-45)
5 5 (2 40-12-6)
Used a boiled blade to cut the cord
1580(54%)
154(5%)
827(26%)
Attendant washed her hands
1,988 (68%)
1064(33%)
Livebirths
2899
3226
Cou! ijidressi'd 01 dressed with antiscptii
2356(81%)
2349(73%)
1 62(0 58-4 55)
Baby wrapped within 30 min
1.975 (68%)
2257(76%)
0-92 (037-2 31)
Raby bathed within 1 h
2880(99%)
3?O7(99%)
1 11 (0-46- 2-71)
Infants alive at 1 month
282?
310/
Any >>1 three infant illnesses (roiH-|h. fi'ver. diarrhor-a)
919(33%)
1 320 (42%)
()-65(0361 >0)
iaken to healtli farilily in event o( illness
219(24%)
1 31 (10%)
2 8-1 (1 65-4 88)
Breastfed infants
2864
3181
h litiulud wit hili 1 h
1/80 (62%)
1/18(54%)
1 40(0•‘>•'-3 79)
Discarded colostrum
820(29%)
1 344 (42%)
O SS (0-27 -1 10)
Table 4: Process indicator outcomes
976
3- 55(1-56-8-05)
3-4/(l-39-8-69)
wrapping of newborn infants, early bathing, or
breastfeeding. Rates of maternal morbidity were similar,
but women in intervention clusters were more likely
than those in control clusters to have visited a health
facility in the event of illness. Likewise, infant illness was
more likely to have led to a visit to a health facility.
The most usual causes of neonatal death were
complications of preterm birth, presumptive birth
asphyxia, and infection. The pattern of causes did not
differ between groups, but we rioted that infectionrelated deaths were less frequent in intervention
clusters.
Discussion
We have shown that an intervention in rural Nepal,
entailing women's groups convened by a local woman
facilitator, reduced neonatal mortality by 30%. Maternal
mortality, although not a primary outcome of the trial, was
also significantly lower in intervention areas. The
intervention seemed to bring about changes in home-care
practices and health-care seeking for both neonatal and
maternal morbidity. The activities of one facilitator in a
population of 7000 rapidly reached a high proportion of
pregnant women, even in poor and remote communities.
Only 8% of married women of reproductive age ever
attended a group, but the groups attracted 37% of newly
pregnant women, and members raised awareness of
perinatal issues outside the groups themselves.
Cluster-randomised trials are susceptible to bias. The
intervention and control areas had similar retrospective
neonatal mortality rates, but some differences were
noted in literacy and poverty indicators. We do not think
these factors could account for the noted differences in
mortality rates, but they do merit further investigation.
Suiveillance methods could have affected outcomes,
although this activity would have taken place in both
intervention and control areas.
Two potential effect modifiers were the convening of
women's groups in collaboration with governmenttrained female community health volunteers and health
system strengthening activities across intervention and
control areas. Would work with women's groups have
the same degree of effect in areas where no community
health volunteer was present or no training of health
workers in essential newborn care took place?
Security problems in the district escalated during the
third year of the study. Supervisory’ activities were
intermittently compromised in four clusters (two
intervention and two control), and although no
women’s group was disbanded, four groups had to
postpone their meetings several times.
The intervention seemed to be acceptable: 95% of
groups remained active at the end of the trial despite no
financial incentives and the opportunity costs incurred by
women spending time away from other tasks. With
appropriate investment and political commitment, we
think the intervention could be scaled up rapidly. Scaling
www.thelancet.com Vol 364 September 11, 2004
Articles
up could be achieved through both government and non
government organisations and would not necessarily need
to be managed by health-sector personnel, although
coordination would be essential. Local rather than central
government might be preferable to lead the process for
reasons of participation, accountability, and sustain ability.
A cost-effectiveness analysis was done alongside the
study. The cost per newborn life saved was US$3442
($4397 including health-service strengthening costs) and
per life year saved $111 ($142 including health-service
strengthening costs). This value compares favourably with
the World Rank's recommendations that interventions
less than US$127 per disability-adjusted life year saved are
some of the most cost effective.'1 Our estimates probably
underestimate the programme's cost-effectiveness. They
do not. include benefits to infants born outside the closed
cohort surveillance; they ignore longer-term benefits of
the intervention to subsequent pregnancies; they exclude
benefits to infants of reduced morbidity and to mothers
from reduced morbidity and mortality: and they omit
potential savings in set-up and supervision costs if the
activities were replicated elsewhere.
Two key elements distinguished our approach from
conventional health education. First, women’s groups
looked at demand-side and supply-side issues. Second,
the approach emphasised participatory learning rather
than instruction. The women's group strategies—the
picture card game, health funds, stretcher schemes,
production and distribution of clean delivery kits, and
home visits—also entailed interaction outside the
groups, which increased awareness of perinatal issues.
The renewed interest in community participation in
health care'2 is attributable partly to the scarcity of
resources committed to primary care and partly to the
perceived failure of conventional health education and
primary health care to deliver substantial health
benefits.
A major challenge has been to engage
users and enable them to adopt positive health care
behaviours. In many countries, local-health committees
have had little accountability to their communities, and
the level of representation of beneficiaries such as
women is low.'1" Beneficiaries themselves can be
passive in the face of service bureaucracies*''’’” because of
an absence of local ownership, different perceptions of
priorities, and capture of resources by powerful groups.
If participation is a key element of primary health care
then few controlled studies have been done of its effect
on health outcomes. Participation is typically seen as an
adjunct to implementation rather than as a primary
intervention, and the distinction between a didactic
approach to health education at community level and a
participatory approach to developing strategies is
blurred. For example, community-based health
promoters have increased exclusive breastfeeding rates
in Mexico ‘ and India." where diarrhoeal morbidity was
also diminished. In Ethiopia, a randomised controlled
trial of mother coordinators trained to teach other local
www.tlwlana’lxoni Vol 364 September 11, 2004
mothers to recognise symptoms of malaria in their
children and to promptly give chloroquine achieved a
40% reduction in under-five mortality *l
The procedure used to establish cause of death
suggested that infection accounted for fewer deaths in
intervention than control clusters. This finding lends
support to the noted rises in antenatal care, trained birth
attendance, clean delivery- kit use. hand washing by birth
attendants, and care seeking in the event of neonatal
morbidity. These data complement the work of
SEARCH,"' whosp intervention consisted of a package of
activities. Scaling up the use of injectable antibiotics by
community health workers presents difficulties for
policy makers, and our less intensive intervention
achieved half the SEARCI I mortality reduction.
The effect of the intervention on maternal mortality
was surprising in view of the size and power of the study
and obviously needs replication. If validated, the finding
would be noteworthy for the potential of this approach to
achieve Millennium Development Goals." The partici
patory strategy could benefit other health outcomes such
as stillbirths, infant and childhood mortality, and
malaria and 11IV infection in pregnancy. The absence of
effect on stillbirth rates shown in this trial does not rule
out future success if issues such as nutrition received
greater emphasis in women’s groups."
The trial findings raise several issues that we intend to
address in subsequent work: differential changes in care
practices between group members and non-members,
the process of d iffusion of behaviour changes within the
population, an examination of potential confounding
within the cluster-randomised design, further analysis
and refinement of the verbal autopsies, and a detailed
discussion of cost-effectiveness.
Progress towards the Millennium Development Goals
for maternal and child mortality reduction has faltered.
Our findings suggest that a demand-side intervention
can achieve great reductions in neonatal and maternal
mortality’ in poor and remote communities. The
approach—a local woman facilitating women’s
groups—is potentially acceptable, scalable, sustainable,
and cost effective as a public-health intervention.
Assessment of demand-side interventions needs greater
attention in primary care/’ Studies are needed to assess
how we can replicate the approach in different settings,
as are large trials to examine effects on maternal
morbidity and mortality.
Contributors
All authors contributed to the design of the study and criticised drafts of
the paper. D S Manandhar and AMdeL Costello were responsible for
the conception and overall supervision of the trial. B 1’ Shrestha
managed the project under initial guidance from 1 u Shrc-siha
l< M Tumbahangphe. S Tarnang. S Thapa, D Shrestha. and B Thapa
managed data collection, field intervention, health service activities, data
entry and adimrnslralion. respectively. N Me.sko and | Morri.son were
technical advisors on intervention and qualitative aspects of the study,
and D Osrin on quantitative aspects. J Borghi conducted the economic
analysis with help from B Thapa. M Manandhar and I I Standing
advised on the facilitation process. O Osrin and A Wade carried out the
977
Articles
quantitative analysis. A M de I. Costello and D Osrin wrote the first draft
of the paper and were responsible for subsequent collation of inputs
and redrafting.
Conflict of interest statement
We declare that we have no conflict of interest.
Acknowledgments
We thank the many individuals in Makwanpur District who gave their
time generously and without complaint, and the field staff of the MIRA
Makwanpur team. Field coordinators were Laxmi Ghirnire, Rajita
Shrestha, Rita Shrestha, and Prarnod Thapa for the intervention, and
Dhruba Adhikari. Biswas Aryal. Bishnu Bhandari. Surendra Bhatia.
Bharat Budathoki. Bhim Khadka. Sukra Raj Lama. Mukunda Neupane.
and Rishi Neupane for die surveillance. We thank the Makwanpur
District Development Committee and its chairman, Rameshwar Rana,
and the village development committee member? for their active and
continuing support; the District Health Officerand District Public
I ft ahh Oflicei lor tin ir help; u v MIRA executive committee in
Kathmandu; the staff of the Nepal Nutrition Intervention Project.
Sariahi for their help and advice; the Department for International
Development. London for their creative and critical input to the project:
Ashish Shrestha and Anama Manandhar for designing the database and
Kamala Badan for auditing the data;
Lisa Howard Grabrnan of Johns Hopkins School of Public Health for
her advice on the participatory intervention based on her experience in
Bolivia; members of the data monitoring committee for their time and
expertise (Patricia Hamilton, Royal College of Paediatrics and Child
Health. London; ProfChitra Gurung. Institute of Medicine,
Kathmandu; Bishnu Prasad Pandit, Policy Planning and International
Cooperation. Ministry of Health, Kathmandu; Stephen Wall, Saving
Newborn Lives Initiative, Washington: Bharat Arnatya and
Sushan Man Shrestha, Nepal Health Research Council;
Dibya Shree Malla, Nepal National Academy of Medical Science); and
Julia Fox-Rushby, Abhay Bang. Bill Musoke., Sonia Lewycka,
Prof Andrew Tomkins. Sally Hartley, Jose Marlines, Hugh Moffatt,
Claire Puddephall, Sarah Ball. Jovan I lie, and Susie Vickery for useful
discussions during the implemenlalion of the project. The study was
funded by DFID, with important support from the Division of Child and
Adolescent Health. WHO, the United Nations Children’s Fund, and the
United Nations Fund for Population Activities.
10
11
12
13
14
15
16
17
18
19
20
21
2.2
23
References
Saving iicwborn lives Slate of the world ' newborns Washington.
DC Save the Children. 2001
I’aul V. Newborn care in India' a promising beginning, but a long
way to go. Semin Neonaloi 1999: 4: 141—19.
3
WH O. Essential newborn care: report of a technical working
group (Trieste, April 25-29, 1994). Geneva: World Health
Organization. Division of Reproductive Health (Technical
Support), 1996.
4
Osrin D, Turnbahangphe K. Shrestha D. et al. Cross sectional,
community based study of care of newborn infants in Nepal. BM)
2002;325: 1063-66.
5
Ensor T, Cooper S. Overcoming barriers to health service access:
influencing the demand side. Health Policy Plan 2004: 19: 69-79.
6
Howard-Grabman I... The ■'autodiagnosis’’: a methodology to
facilitate the maternal and neonatal health problem identification
and prioritization in women’s groups in rural Bolivia, working
paper 16A. Virginia: MotherCare, John Snow International. 1993.
7
Howard-Grabman I... “Planning together": a methodology to
facilitate the development of strategies and actions to address
priority maternal and neonatal health problems in rural Bolivian
communities, working paper 16B. Virginia. MotherCare. John
Snow International, 1993.
8
I loward-Grabman I.. Seoane G. Davenport C. MotherCare, and
Save the Children. Tlie Warmi Project: a participatory approach to
improve maternal and neonatal health, an implementor’s manual.
Westport; John Snow International, Mothercare Project. Save the
Children, 2002..
9
Bang AT, Bang RA. Baitule SB. Reddy MH, Deshmukh MD. Effect
of home-based neonatal care and management of sepsis on
neonatal mortality: field Inal in rural India. Lancet 1999; 354:
1955-61.
978
24
Stone 1... Primary health care for whom? Village perspectives from
Nepal. Soc Sci Med 1986; 22: 293-302.
Carlough M. LeMaster J. Flow can motherhood be made safer in
Nepal? More than hospitals are needed. J Inst Med (Kathmandu)
1998; 20: 20-28.
Justice J. Neglect of cultural knowledge in health planning: Nepal’s
Assistant Nurse-Midwife program. In: Hahn R, ed. Anthropology
and public health: bridging differences in culture arid society.
New York: Oxford University Press. 1999.
Dhungel B. The qualitative community judgement: the role of the
intermediate health practitioners in Nepal’s family health services.
In: Allen M, ed. Anthropology of Nepal: peoples, problems and
processes. Kathmandu: Mandala Book Point. 1992.
Kutzin J. Obstacles to women's access: issues and options for more
effective interventions to improve women's health: Human
Resources Development and Operations Policy (HRO) working
paper no 1.3. Washington, DC: World Bank. 1993.
Tipping G, Segall M. Health care seeking behaviour m developing
countries: an annotated bibliography and literature review,
development, bibliography 12. Sussex: Institute of Development
Studies. 1995.
Morgan L. Community participation in health: perpetual allure,
persistent challenge. Health Policy Plan 2001; 16: 221-30.
Rolarn A. Manandhar D. Shrestha P. Malla K. Ellis M. Costello A.
The effects of postnatal health education for mothers on infant care
and family planning practices in Nepal: a randomised, controlled
trial. BMJ 1998; 134: 805-10.
Hobley M. From passive to active participatory forestry. Nepal. In:
Oakley P, ed. Projects with people: the practice of participation in
development. Geneva: International Labour Organisation. 1990.
Hobley M. Participatory forest management in South Asia; the
process of change in India and Nepal (Rural Development Forestry
Study Guide). London: Overseas Development Institute (ODI), 1996
Ministry of Finance. Economic survey, fiscal year 2001/2002.
Kathmandu: His Majesty’s Government. Ministiy of Finance. 2002.
Ministry of Health. Nepal demographic and health survey 2001
Kathmandu. His Majesty's Government. Ministry of Health, Family
Health Division. 2002
Pradhan A. Aryal R. Regrni G. Ban B, Govindasamy P. Nepal
Family Health Survey 1996. Kathmandu: His Majesty's
Governrnenl, Ministry of Health. 1997.
Central Bureau of Statistics. Statistical year hook of Nepal 2001.
Kathmandu: His Majesty’- Government, National I'larming
Commission Secretariat. 2001
Sharma I I, Goutam R. Vaidya S District demographic profile of
Nepal. Kathmandu: Informal Sector Research and Study Center,
2002.
Pradhan E, Katz J, LeClerq S, West K. Data management for large
community trials in Nepal. Controlled Clin Triala 1994. 15: 220-34
26 Osrin D, Manandhar A. Shrestha A, el al. Design of a surveillance
system for pregnancy and its outcomes in rural Nepal.
I Nepal Med Assoc 2003; 41: 411-22.
27 Tarnang S. Mesko N, Shrestha B, et al. A qualitative description
of perinatal care practices in Makwanpur district, Nepal.
C'onlrib Nepalese Stud 2001; 28: 10-19.
28 Mesko N, Osrin D, Tarnang S, el al. Care for perinatal illness in
rural Nepal: a descriptive study with cross-sectional and qualitative
components. BMC In! Health Hum Rights 2003; 3: 3
29 Edwards S. Braunholtz D. Lilford R. Stevens A. Ethical issues in the
design and conduct of cluster randomised controlled trials. BMJ
1999;318: 1407-09.
30 I layes RJ. Bermell S. Simple sample size calculation for duslerrandonnzed trials. Int J Epidemiol 1999; 28: 319-26.
11 World Bank. World Development Report. Washington. DC: World
Bank. 1993.
32 WHO. Community participation: advocacy and action.
http://www.who.int/whr/2004/chapler3/en/. World Health Report
Geneva- World Health Organization. 2004
1< Bryce J. el Anfeen S. Pariyo G, Lanala C, Gwalkiu D, I labichl J.
Reducing child mortality: can public health deliver? Lancet 2003:
362: 159-64.
14 Victor? CG. Wagstaff A, Armstrong Schellenberg J. Gwatkin D.
Claeson M, I labichl J. Applying an equity lens to child health and
25
www.lhelancet.com Vol 364 September 11, 2004
Articles
mortality: more of the same is not enough. Lancet 2003; 362:
23.3-41.
35
.36
37
IS
39
40
Manandhar M. An exploratory study of local health committees in
Nepal London- Centre for International Child Health, Institute of
Child Health, 1996.
Sepehri A. Pettigrew |. Primary health care, community
participation and community-financing: experiences of two middle
hill villages in Nepal. Health Policy Plan 1996:11: 93-100.
Rifkin S. l essons from community participation in health
programmes. Health Policy Plan 1986; 1: 240-49.
Rilkin S, Muller E, Bichrnann W. Primary health care: on
measuring participation. Soc Sci Med 1988; 26: 931-40.
Rilkin S. Community participation in maternal and diild
health/family planning programmes. Geneva: World Health
Organization, 1990.
41
42
43
44
Bhandari N, Bahl R, Mazuindar S. et al. Effect of conirnunity-based
promotion of exclusive breastfeeding on diarrhoeal illness and
growth; a cluster randomised controlled trial, l ancet 200.3; 361:
1418-23.
Kidane G. Morrow R. Teaching mothers to provide home
treatment of rnalpria in Tigray. Ethiopia: a randomised trial. Lancet
2000; 356: 550-55.
United Nations. Millennium Development Goals.
http://www.un.org/millenniurngoals (accessed July 10, 2004).
Ceesay SM. Prentice AM, Cole TJ, et al. Effects on birth weight and
perinatal mortali y of maternal dietary supplei
iimenls in rural
Gambia'
Gambia: 53 vest
year r.inrlnmiwvl
randomised mntrnllpd
controlled h-iai
trial. RBM] 1997; 31S:
786-90.
Morrow AL, Guerrero ML, Shults J, el al. Efficacy of home-based
peer counselling to promote exclusive breastfeeding: a randomised
controlled trial, lancet 1999; 353: 1226-31.
I
i
I
I
I
I
I
www.thelancet.coni Vol 364 September 11, 2004
979
—--------- - ,i.
Impact of community organization of women
on perinatal outcomes in rural Bolivia
Kathleen OrRourke^ Lisa Howard-Grabman,2 and Guillermo Seoane2
ABSTRACT
INTRODUCTION
infant mortality has been attacked in
developing countries with varying
degrees of success. Costa Rica, with
one of the most successful programs,
lowered infant mortality from 68 per
1 000 live births in 1970 to 20 per 1 000
in 1980 (1). Costa Rica's success has
been attributed largely to the develop
ment of primary and secondary health
care (2). Increasing access to health
services and to primary care practi
tioners, predominantly nurses, also
dramatically reduced infant mortality
in Nicaragua (3).
1
University of Texas-Houston, School of Public
Health. Mailing address: UT-Houston School of
Public Health, 1100 N. Stanton, Suite 110A, El
Paso, TX 79902. E-mail: Kathleen@mail.UTEP.edu.
MotherCare, Arlington, Virginia.
An intervention to improve maternal and child health was conducted in a remote Bolivian
province with limited access to modern medical facilities. The intervention focused on initiat
ing and strengthening women's organizations, developing women's skills in problem identifi
cation and prioritization, and training community members in safe birthing techniques. Its
impact was evaluated by comparing perinatal mortality rates and obstetric behavior among
409 women before and after the intervention. Perinatal mortality decreased from 117 deaths
per 1 000 births before the intervention to 43.8 deaths per 1 000 births after. There was a sig
nificant increase in the number of women participating in women's organizations following
the intervention, as well as in the number of organizations. The proportion of women receiv
ing prenatal care and initiating breast-feeding on thefirst day after birth was also significantly
larger. The number of infants attended to immediately after delivery likewise increased, but the
change was not statistically significant. This study demonstrates that community organiza
tion can improve maternal and child health in remote areas.
Bolivia has one of the highest rates
of infant mortality in Latin America.
Estimates indicate that in 1994, be
tween 75 and 105 of every 1 000 Boli
vian infants died in the first year of life
(4). Within Bolivia, infant mortality
varies regionally and rates are approx
imately 50% higher in rural areas than
in urban areas. Rates are higher among
the country's Aymara and Quechua
indigenous groups than in the Ladino
population (persons of Spanish or
mixed heritage) (4). Because health
care resources are limited in many
rural areas, one recommended ap
proach for reducing infant mortality is
through training relatively inexpen
sive and culturally appropriate pro
viders, such as nurses and community
health workers (5).
An evaluation of local community
health workers in Bolivia concluded
that they did not fit easily into leader
Rev Pancnn Snlud Publico/Pan Am I Public Health 3(1), 1998
ship roles in either modern or tradi
tional medicine (6). The medical and
local communities were more accept
ing of community health workers
whose roles and responsibilities they
helped define (6). Training women to
evaluate the health status of their chil
dren and family members has been
shown to be a cost-effective means of
improving the health of communities
in which there is limited access to
modern health care (7).
The Warmi project
The Warmi project, conducted from
July 1990 to June 1993, attempted to
improve maternal and child health
through involving communities in
health care. The World Health Organi
zation (WHO) endorsed this strategy
at the Alma-Ata conference (1978),
9
with the recommendation that health
programs strive for "community par
ticipation and ultimately self-reliance,
with individuals, families, and com
munities assuming more responsibil
ity for their [own] health" (8).
The province of Inquivisi was
selected as the Warmi project site
based on its remoteness, rural charac
ter, and limited access to modern
medical care. In addition, this area
had previously been the site of Save
the Children/Bolivia projects. Conse
quently, population census results and
maternal mortality data were available.
Key project components included
(a) organizing women's groups, (b)
developing an approach to identify
ing problems, (c) implementing a "for
mal action plan" for the problems
identified, and (d) training birth atten
dants and husbands in safe birthing
techniques.
The primary purpose of this study
was to evaluate the potential effect of
organizing women's groups on peri
natal mortality in a remote, rural area
of a developing country. A secondary
goal was to identify the impact of
specific components of the project,
such as increasing prenatal care,
improving immediate newborn care
and breast-feeding, and increasing the
number of deliveries attended by
trained personnel.
but also because people do not con
sider the facilities adequate to deal
with medically complicated situations.
Consequently, women who develop
obstetric complications during preg
nancy and labor are usually referred to
hospitals in La Paz or Oruro, a journey
of four to six hours by road.
In rural Bolivia, only 23% of deliver
ies are attended by either a physician
or trained nurse. The majority of births
(55%) are attended by relatives, usu
ally the fathers; 14% are attended by
midwives or traditional birth atten
dants; 6% are unattended (i.e., the
woman gives birth without a helper);
and the circumstances are unknown
for the remaining 2% (9). In Bolivia the
term "midwife" refers to trained med
ical personnel who deliver babies,
while traditional birth attendants
(TBAs) are individuals within the
community who deliver babies but are
not trained as part of the formal med
ical system.
MATERIALS AND METHODS
nique called "autodiagnosis" was em
ployed to address community prob
lems. Autodiagnosis consists of the
following four steps: (a) identification
and prioritization of problems, (b)
group development of a formal action
plan, (c) implementation of the plan,
and (d) evaluation.
Each community identified a differ
ent set of problems and approaches,
and, accordingly, specific interven
tions varied by community. However,
certain objectives were addressed by
all the women's groups: to (a) increase
knowledge of reproduction, contra
ceptive use, danger signs of complica
tions, and self-care, (b) improve imme
diate newborn care, and (c) increase
the percentage of women who receive
delivery care from trained birth atten
dants. Further details of the study pro
tocol can be found in the project implementers' manual (10).
Prior to the implementation of the
Warmi project interventions, a study
of perinatal mortality (infant deaths
occurring from the 28th week of preg
nancy through 28 days of life) was
conducted. This study identified all
births and perinatal deaths in the com
munity area for the prior two years
(November 1988-October 1990). For
each infant who died, two community
controls were randomly selected
among children who were born in the
same year, lived in similar communi
ties, and survived for at least 28 days.
Following the intervention, which
took place from January 1991 through
June 1993, a second case-control study
was conducted. Again, all births and
perinatal deaths that occurred in the
two years prior to the survey (April
1991-March 1993) were identified, and
each deceased infant was matched
with four to five controls.
During the first case-control study, a
questionnaire was administered to 237
mothers in their homes (or to other
family members when the mothers
were not available). Demographic
characteristics, obstetric history, and
details of the most recent childbirth
were elicited. In the second study, 172
questionnaires were administered.
Outcome variables included perina
tal mortality; the numbers of women's
Inquivisi Province encompasses
three geographic zones—Inquivisi,
Licoma, and Circuata—each of which
has one health post. The posts in
Inquivisi and Licoma are staffed by one
physician and an auxiliary nurse, while
the one in Circuata has only an auxil
iary nurse. All three posts possess only
basic medicines and equipment. Two
referral hospitals are located outside
the project area at a distance of one to
two hours' travel. These hospitals fall
short of minimum WHO standards,
lacking sterilization and anesthesia
equipment and properly trained staff.
Local health facilities are underuti
lized by the population. This is partly
due to economic and cultural factors,
Fifty communities in Inquivisi
Province participated in the Warmi
project. They varied greatly in tradi
tions and demographic characteristics
depending on the geographic zone in
which they were located. Settlements
in the zone of Inquivisi consist of longestablished, stable communities with
widely dispersed houses. The villages
in Circuata, in contrast, are newer and
contain large numbers of recent immi
grants. Housing is more densely con
centrated than in Inquivisi, and the
women are more apt to be bilingual.
Licoma is a mixed zone that combines
characteristics found in both of the
others (10). Villages comprise 40 to 300
families. The total population in the
demonstration area is 15 000.
Study personnel included five to six
teams, each consisting of two auxiliary
nurses from the Save the Children
staff. Monthly or more frequently,
each team met individually with all of
the zone's women's organizations,
which numbered approximately 50. At
these meetings, attended by approxi
mately 10-30 group members, a tech-
10
O'Rourke et al. • Impact of community organization of women on perinatal outcomes in rural Bolivia
Health care infrastructure
organizations and the extent of
women's participation in them; the
prevalence of specific obstetric prac
tices, including prenatal care and uti
lization of trained birth attendants;
and the timing of newborn care and
initiation of breast-feeding. Mortality
data were based upon total numbers
of births and perinatal deaths recorded
in the community registry and thus
were not limited to the study popula
tion. In the analysis of assistance at
delivery, trained attendants were con
sidered to include TBAs, health pro
moters,3 physicians, and nurses.
The central program strategy was
to increase participation in women's
groups. As described elsewhere (10),
there were a variety of types of
women's groups in the Inquivisi area,
including women's organizations,
cooperatives, mothers' clubs, and
agrarian unions. The groups' functions
varied together with their degrees of
effectiveness. The project staff consid
ered women's organizations best able
to organize women around health
issues. Consequently, the staff focused
on these groups, initiating or strength
ening 50 women's organizations.
Statistical analyses were performed
using SPSS statistical software (11).
Chi-square tests were performed on
categorical outcomes as defined
above. The Breslow-Day test for
homogeneity of odds ratios was used
to compare the degrees of change seen
in cases and controls (12).
RESULTS
By this criterion, the majority of
women in all communities were
Aymara, with Inquivisi having a
higher percentage of Aymaras and
fewer Quechuas than the other zones.
The presence of a dirt floor was used
as an indicator of lower socioeconomic
status. The majority of women in
every community had dirt floors, but
the highest percentage was found in
Inquivisi. Overall, prior to the study
intervention, women in Circuata par
ticipated in women's organizations at
a higher rate than women in the other
zones.
vention periods (Table 2). During the
first study period, 639 births were
identified. Of these, 36 infants were
classified as either fetal deaths or
stillbirths, and 38 others were born
alive but died within 28 days of birth
(data not shown). One perinatal death
could not be classified as prenatal or
postnatal. During the second period,
708 births were identified. Of these,
21 were classified as fetal deaths or
stillbirths, and 10 as postnatal deaths.
Therefore, the perinatal mortality rate
decreased from 117 per 1 000 before
the intervention to 43.8 per 1 000 births
after.
Mortality
Intermediate outcome data
Perinatal and neonatal mortality
decreased significantly between the
pre-intervention and the post-inter-
Table 3 presents the numbers of
women who were aware of women's
TABLE 1. Comparison of baseline sociodemographic characteristics of women living in
three zones of Inquivisi Province, Bolivia
Zone
Characteristic
Married6
Language spoken0
Spanish
Aymara
Quechua
Literate
Dirt floor
Participate in women's
organizations
Inquivisi (n = 72)
No.
(%)
Licoma (n = 68)
No.
(%)
Circuata (n = 97)
No.
(%)
Pa
67
(93.1)
57
(83.8)
83
(85.6)
0.206
66
65
15
(95-7)
(94.2)
56
53
(21-7)
16
58
66
(80.6)
(91.7)
53
49
(87.5)
(82.8)
(25.0)
(77.9)
(72.1)
85
66
26
80
82
(95.5)
(74.2)
(29.2)
(82.5)
(84.5)
0.094
0.004
0.560
0.769
0.007
27
(37.5)
20
(29.4)
53
(54.6)
0.003
a Chi-square tesl.
b Includes women living with a partner but not legally married.
c Information not available lor three women from Inquivisi. four from Licoma, and eight from Circuata.
Demographic comparisons of
communities
Table 1 compares relevant socio
demographic characteristics in the
three study zones al baseline. Women
in Inquivisi were more likely to be
married than women in Circuata and
Licoma. The language spoken at home
was used as a surrogate for ethnicity.
TABLE 2. Outcome of births during two study periods (n = 1 347), show
ing number of infants surviving at least 28 days (“living”) and number
dying in the perinatal period (“not living”), Inquivisi Province, Bolivia
Period
Living
No. (%)
Not living3
No. (%)b
Total
Pre-intervention
(1988-1990)
564 (88.3)
75 (11.7)
639
677 (95.6)
31
708
Post-intervention
1
Health promoters are health workers, both male
and female, selected by the community to provide
care. They receive training in primary health care
from the Ministry of Health.
Kez’ Pmwm Saliui Publica/Pan Am I Public Health 3(1), 1998
(1991-1993)
(4.4)
a Includes deaths from 28 weeks gestation through 28 days after birth.
bX2: P< 0.001, 1 df.
11
TABLE 3. Awareness of and participation of women in women’s organizations in pre- and
post-intervention periods (n = 409)
Participation in group
Awareness of groups
Aware
n (%)
Group
Member
n (%)
P-value
<0.001
Mother's clubs
Pre-intervention
Post-intervention
41 (17.3)
1 (0-6)
Women’s organization
Pre-intervention
Post-intervention
75 (31.6)
119 (69.2)
Any group3
Pre-intervention
Post-intervention
231 (97.5)
162 (97.0)
P-value
0.072
26 (11.0)
10 (5.8)
<0.001
<0.001
18 (7.6)
93 (54.4)
<0.001
0.779
for controls. However, the increases in
use of prenatal care were not statisti
cally different in cases and controls.
Presence of traditional birth atten
dants. Results were mixed for this
objective. While the percentage of
childbearing women attended by
TBAs increased for cases following the
intervention, it decreased for controls.
These differences, however, were not
statistically significant.
100 (42.2)
147 (86.5)
a Includes mother's clubs, women's organizations, agrarian unions, co-operatives, neighborhood committees, and credit
programs.
groups in their communities together
with the numbers of women par
ticipating in groups, pre- and post
intervention. Mothers' clubs had dis
tributed food prior to this study, but
had recently ceased doing so be
cause the international organization
that supplied the food left the area.
Their numbers declined sharply from
pre- to post-intervention. Meanwhile,
women's organizations proliferated
with growing membership. Overall,
more women were participating in
groups of one kind or another at the
end of the intervention.
Table 4 presents changes in the use
of prenatal care, the presence of a
trained attendant at birth, and the tim
ing of newborn care between the two
study periods. Separate results are pre
sented for cases and controls, together
with the results of Breslow-Day tests
for significant differences in the magni
tudes of change in the two groups.
Prenatal care. Both cases and con
trols were more likely to receive pre
natal care following the intervention,
with a statistically significant change
Timing of newborn care. The per
centage of newborn controls who were
attended immediately after delivery
rose between studies, while the op
posite occurred among cases. The
Breslow-Day test for homogeneity
approached statistical significance (P 0.058), suggesting a difference in the
amount of change registered between
the two groups.
Timing of breast-feeding. This objec
tive was evaluated for control infants
only, as most of the cases did not sur
vive childbirth. A significantly greater
percentage of control infants were
breast-fed on the first day of life fol
lowing the intervention—50.3% as
compared with 25.3% prior to the
intervention (x2 = 18.77, P < 0.001, data
not shown).
DISCUSSION
TABLE 4. Comparison of obstetrical practices in pre- and post-intervention periods for
cases and controls, Inquivisi Province, Bolivia
Controls
Cases
Variable
n
Yes
(%)
Yes
n (%)
0.175
Received prenatal care
Pre-intervention
Post-intervention
34 (45.3)
18 (60.0)
Trained attendant al birth
Pre-intervention
Post-intervention
21 (28.8)
11 (35.5)
Immediate newborn care
Pre-intervention
Post-intervention
15 (24.6)
2 (8.7)
P-value
Breslow-Day
P- value3
0.009
0.952
0.169
0.206
0.276
0.058
74 (49.0)
86 (64.2)
0.497
56 (37.1)
40 (29.4)
0.106
a Breslow-Day test lor homogeneity ol the odds ratios
12
P-value
50 (34.2)
52 (40.6)
Mortality rates
Infant mortality rates provide an
important measure of community
health status and are often used as an
indicator of overall socioeconomic
development. Perinatal mortality rates,
on the other hand, are a measure of
women's health and the quality of
health care provided during preg
nancy and the intrapartum period (13).
Prior to the intervention described
here, rates of perinatal mortality were
extremely high in the Inquivisi area,
with approximately 117 deaths per
1 000 births. Following the interven-
O'Rourke et al. • Impact of community organization of women on perinatal outcomes in rural Bolivia
tion, perinatal mortality in the project
area decreased by 65%.
Because there was no control com
munity, it is difficult to say with cer
tainty whether the intervention or
other factors caused the decrease in
mortality. A possible alternative expla
nation credits the effect of changes in
the survey populations between the
two studies. Such changes could occur
in more than one way: (a) if more
women from a given zone were sur
veyed following the intervention than
before, or (b) if demographic changes
occurred within the area due to migra
tion. However, neither of these expla
nations is likely. The pre-intervention
and post-intervention surveys both
sampled the same percentages of
women from each zone. Furthermore,
between the pre-intervention and post
intervention periods, there was no sta
tistically significant change in socio
economic variables such as maternal
literacy, housing characteristics, or lan
guage spoken (data not shown).
One possible cause of the decrease
in perinatal mortality may have been a
regression towards the mean owing to
the initially high rate during the first
period. This is not likely to be the full
explanation, however, because of the
magnitude of the decrease. Thus,
while it is not possible to say with cer
tainty that the decrease in mortality
resulted from the intervention pro
gram, it seems reasonable that the pro
gram was responsible for at least some
of the impact.
Women's organizations
One of the greatest changes ob
served in this study was a doubling of
participation in women's organiza
tions. Women in these groups were
encouraged to become more actively
involved in identifying health needs
and designing programs to address
them. They were encouraged to dis
cuss obstetric complications, and in
the process they learned that many
features of reproduction which they
had supposed to be matters of course
could in fact be altered. For example,
women were taught to care for their
babies immediately after delivery
rather than wait for the delivery of the
placenta. Other specific activities var
ied by community and included train
ing in literacy, fostering use of credit
programs, presenting educational pro
grams about safe pregnancy, and
implementing family planning.
The increase in participation in
women's organizations was accompa
nied by a concurrent decrease in the
number of mothers' clubs. As a result,
there was not an overall increase in the
numbers of groups in the community,
but rather a restructuring of the types
of groups women attended.
Improvement in pre- and postnatal
practices
The number of women who re
ceived prenatal care during their preg
nancies was significantly greater after
the intervention. Previous studies
have identified decreased risk of still
birth and neonatal mortality for
women who receive prenatal care (14).
In Mexico, expanded use of prenatal
care was identified as a major determi
nant of lower perinatal mortality (15).
While the exact mechanism by which
prenatal care reduces mortality is not
known, such care provides health pro
fessionals with a means of identifying
potential problems and educating
women about health in pregnancy,
labor, delivery, and postpartum.
Although changes in the timing of
infant care were not statistically signif
icant, there was an increase in the
percentage of control infants who
received care immediately after deliv
ery. Fewer cases received immediate
care, but this may be due in part to the
lack of attention given to stillborn
infants.
There was also a significant increase
in the number of women who breast
fed their infants on the first day of life.
While changes in breast-feeding prac
tices would not affect fetal deaths or
stillbirths, early breast-feeding can
increase the likelihood that breast
feeding is successful and could pre
vent mortality related to unsanitary
infant feeding practices (16).
Rev Panam Stilud Publica/Pan Am / Public Health 3(1), 1998
Study limitations
Reporting bias could potentially
influence results. This is unlikely,
however, since the same community
registration methodology was utilized
for the studies before and after the
intervention. In fact, the project staff
identified more births along with
fewer deaths during the second set of
interviews.
Classification of the time of death
was based upon maternal recall and
should be considered approximate. It
is not clear how well women were able
to differentiate between a fetal death,
which occurred prior to the onset of
labor, and death that occurred dur
ing labor and delivery. Furthermore,
infants were not always attended to
immediately after delivery. It is there
fore possible that some infants who
were born alive but died shortly after
delivery were incorrectly classified as
stillborn. Consequently, for purposes
of this evaluation, all deaths were
grouped together and no subanalyses
were done by time of death.
One project goal that could not be
fully evaluated was the training of
birth attendants and husbands in safe
birthing techniques. In Inquivisi, few
individuals function in the role of tra
ditional birth attendant—that is, a per
son within the community who deliv
ers babies and has acquired her skills
by working with other TBAs. Thus, in
the project, the objective of increasing
the pool of trained birth attendants led
to a focus on teaching basic skills to
individuals who were identified as
potential birth attendants rather than
on improving the skills of existing tra
ditional birth attendants. Husbands
were trained in safe birthing skills, but
it was not possible to identify the
training status of the husband from
the women's interview responses.
Thus, the analysis of care provided by
trained birth attendants did not in
clude trained husbands.
CONCLUSIONS
Indirect causes of infant mortality
are not well understood. Infant mor-
13
tality has been consistently associated
with low socioeconomic status (13),
decreased maternal educational level
(17), and low social status of women
(IS). These associations remain even
when analyses control for access to
care (19). It is not known precisely
how improving women's decision
making ability could affect the out
come of their pregnancies.
In general, the results of this study
support the use of community training
and organization of women as a means
of improving pregnancy outcomes for
women residing in remote areas with
limited access to modern health care.
However, there is limited evidence
defining the impact of individual
components of the program, such as
increasing prenatal care, improving
immediate newborn care and breast
feeding, and providing trained birth
attendants. Further studies should
focus on specific aspects of this type of
program to identify which compo
nents are most effective, as well as the
efficacy of this approach in urbanized
areas and in different populations.
Acknowledgments. The authors
are grateful to Dr. Marge Koblinsky,
Colleen Conroy, and Dr. Alfred Bart
lett for their assistance on this project.
Publication of this work was sup
ported by the United States Agency for
International Development (USAID)
under contract DPE-5966-Z-8083-00.
The contents of this paper do not nec
essarily reflect the views or policies of
USAID or MotherCare.
REFERENCES
1. Rosero-Bixby L. Infant mortality in Costa
Rica: explaining the recent decline. Stud Fam
Plann 1986;17:57-65.
2. Bahr J, Wehrhahn R. Life expectancy and
infant mortality in Latin America. Soc Sci Med
1993;36:1373-1382.
3. Sandiford P, Morales P, Gorter A, Coyle E,
Smithy G. Why do child mortality rates fall?
An analysis of the Nicaraguan experience. Am
I Public Health 1991;81:30-37.
4. Gutierrez M, Ochoa L, Raggers H. Encuesta
national de demografia i/ salud. La Paz, Bolivia:
Institute Nacional de Estadfstica; 1994.
5. Vargas-Lagos V. How should resources be
reallocated between physicians and nurses in
Africa and Latin America? Soc Sci Med 1991;
33:723-727.
6. Bastien J. Community health workers in Bolivia:
adapting to traditional roles in the Andean
community. Soc Sci Med 1990;30:281-287.
7. Bender D. Los sistemas cosmopolitas y tradicionales de salud: la mujer como nexo. Educ
Med Salud 1984;! 8:393-101.
RESUMEN
Impacto de la organizacion
de las mujeres en la
comunidad sobre los
resultados perinatales en
zonas rurales de Bolivia
14
8. Green L. The theory of participation: a
qualitative analysis of its expression in
national and international health policies.
Adz> Health Educ Promot 1986;!(part A):
211-236.
9. Bolivia 1989: Results from the demographic
and health survey. Stud Fam Plann 1991;22:
272-276.
10. Howard-Grabman L, Seoane G, Davenport C.
The Warmi project: a participatory approach to
improve maternal and neonatal health—an imple
mentor's manual. Arlington, Virginia: MotherCare/USAID; 1992.
11. SPSS release 6.0 (Vol. I). Chicago: SPSS Inc;
1994.
12. Rothman K. Modern epidemiology. Boston: Lit
tle, Brown; 1986.
13. Edouard L. The epidemiology of perinatal
mortality. World Health Stat Q 1985;38:
289-301.
14. Mistra P, Bajpal P, Tripathi T, Gupta R, Kutty
D. Perinatal mortality: a hospital study. Indian
Pediatr 1973;10:545-550.
15. Holian J. Infant mortality and health care in
Mexican communities. Soc Sci Med 1989;29:
677-679.
16. Delgado H, Valverde V, Martorell R, Klein R.
Relationship of maternal and infant nutrition
to infant growth. Early Hum Dev 1982;6:
273-286.
17. Caldwell J. Education as a factor in mortality
decline: an examination of Nigerian data.
Popul Stud 1979;33:395-413.
18. Jayachandran J, Jarvis G. Socioeconomic
development, medical care, and nutrition as
determinants of infant mortality in lessdeveloped countries. Soc Biol 1986;33:301-315.
19. Kim K, Moody P. More resources, better
health? A cross-national perspective. Soc Sci
Med 1992;34:837-842.
Manuscript received on 1 October 1996. Revised version
accepted (or publication on 9 June 1997.
Se llevo a cabo una intervencion destinada a mejorar la salud materna e infantil en una
provincia aislada de Bolivia con acceso limitado a instalaciones de salud modernas. La
intervencion se centre en la creacion y el fortalecimiento de organizaciones para muje
res, en el desarrollo de habilidades entre las mujeres, en la identificacion de problcmas
y la determinacion de prioridades y en el adiestramiento de habitantes de la comuni
dad en la aplicacion de tecnicas seguras para la atencion del parto. Para evaluar su
impacto se compararon las tasas de mortalidad perinatal y las practicas obstetricas de
409 mujeres antes y despues de la intervencion. La mortalidad perinatal bajo de 117
defunciones por 1 000 nacimientos antes de la intervencion a 43,8 defunciones por
1 000 nacimientos despues de ella. Se produjo un aumento significative del numero de
mujeres que participaron en organizaciones femeninas despues de la intervencion, asi
como del numero de dichas organizaciones. Asimismo, hubo un aumento significative
de la proporcion de mujeres que recibieron atencion prenatal y que iniciaron la lactancia materna desde cl primer dia despues del parto. El numero de neonates atendidos
inmediatamente despues del alumbramiento tambien aumento, pero el cambio no fuc
estadfsticamente significative. Esto estudio demuostra quo la organizacion comunitaria puede mejorar la salud materna e infantil en lugares aislados.
O'Rourke et al. • Impact of community organization of women on perinatal outcomes in rural Bolivia
07/06/2010
Community Health Cell Mail - The Lancet ar...
Deepak Kumaraswamy <deepak@sochara.org>
The Lancet articles
7 messages
Deepak <deepak@sochara.org>
To: Thelma Narayan <thelma@sochara.org>
31 May 2010 21:08
Dear Thelma
Attached are the articles from The Lancet. The references are done
with the help of reference manager
Rosato, M., Lax^rack, G., Grabman, L. H., Tripathy, P., Nair, N.,
Mwansambo, C., Azad, K., Morrison, J., Bhutta, Z., Perry, H., Rifkin,
S., & Costello, A. 2008, "Community participation: lessons for
maternal, newborn, and child health", Lancet, vol. 372, no. 9642, pp.
962-971.
Tnpathy, P., Nair, N., Barnett, S., Mahapatra, R., Borghi, J., Rath,
S Rath, S., Gope, R., Mahto, D., Sinha, R., Lakshminarayana, R.,
Patel, V., Pagel, C., Prost, A., & Costello, A. 2010, "Effect of a
participatory intervention with women's groups on birth outcomes and
maternal depression in Jharkhand and Orissa, India: a
cluster-randomised controlled trial", Lancet, vol. 375, no. 9721, pp.
1182-1192.
Also located this interesting video by Ekjut available from
http://www.youtube.com/watch?v=en1ubgj69wg
Some other articles that might be of interest
Azad, K., Barnett, S., Banerjee, B., Shaha, S., Khan, K., Rego, A. R.,
Barua, S., Flatman, D., Pagel, C., Prost, A., Ellis, M., & Costello,
A. 2010, "Effect of scaling up women's groups on birth outcomes in
three rural districts in Bangladesh: a cluster-randomised controlled
trial", Lancet, vol. 375, no. 9721, pp. 1193-1202
Backman, G., Hunt, P., Khosla, R., Jaramillo-Strouss, C., Fikre, B.
M., Rumble, C., Pevalin, D., Paez, D. A., Pineda, M. A., Frisancho,
A., Tarco, D., Motlagh, M., Farcasanu, D., & Vladescu, C. 2008,
"Health systems and the right to health: an assessment of 194
countries", Lancet, vol. 372, no. 9655, pp. 2047-2085.
Hope this is Helpful
Thanks
Deepak
Deepak Kumaraswamy
Research and Training Assistant
' Centre for Public health and Equity
27.1st floor, 6th cross. 1st main,
http://mail. google. com/a/sochara.org/?ui...
i/s
07/06/2010
Community Health Cell Mail - The Lancet ar...
1st block. Koramangala. Bangalore -34
email: deepak@sochara.org
PH: 91-80-41280009
4 attachments
lancet 2010; 375 1182-92.pdf
914K
Lancet 2008; 372; 962-71.pdf
161K
Lancet 2010; 375; 1193-202.pdf
277K
Lancet 2008; 372; 2047-85.pdf
890K
Deepak <deepak@sochara.org>
To: Lavanya Devdas <lavanya.devdas@gmail.com >
31 May 2010 21:11
Dear Lavanya
Some articles which might be of interest to you
[Quoted text hidden]
4 attachments
lancet 2010; 375 1182-92.pdf
914K
Lancet 2008; 372; 962-71.pdf
161K
Lancet 2010; 375; 1193-202.pdf
277K
—-J
Lancet 2008; 372; 2047-85.pdf
890K
1 June 2010 09:48
Thelma Narayan <thelma@sochara.org>
To: Deepak <deepak@sochara.org>
Cc: Prasanna Saligram <prasanna.saligram@sochara.org>, Rakhal Gaitonde <rakhal@sochara.org>, "E. Premdas"
<premdas@sochara.org>, Ruth Vivak <ruth@sochara.org>, Ravi Narayan <chcravi@gmail.com>
Dear Deepak,
Thank you for tracing the Lancet articles that I had requested you for and the additional articles from the Lancet
series. I am marking this to SOCHARA team members. I have met the Director of Ekjut the organisation in
Jharkand where some of this work has been done. They hava developed their thinking based on their previous
experience, but it is a very good collaboration between several institutions, development projects and academics.
It is worhtwhile for us in SOCHARA to consider whether we should engage in such collaborative research work.
I would suggest that all our team members go through the articles. Circulation to fellows and interns can be done
after each team has discussed the artice themselves as part of the weekly journal club sessions.
with best wishes,
http://mail.google.com/a/sochara.org/?ui.
2/s
07-/06/2010
Community Health Cell Mail - The Lancet ar...
Thelma
[Quoted text hidden]
Dr. Thelma Narayan, MBBS, M.Sc (Epidemiology) Ph.D ( London)
Public Health Consultant,
Centre for Public Health And Equity
Community Health Cell,
Society for Community Health, Awareness, Research and Action (SOCHARA)
#27, 6th Cross, ( 1st Floor)
1st Main, 1st Block, Koramangala,
Bangalore - 560 034, Karnataka, India.
Ph: 0091-80-41280009; Fax: 0091-80-25525372
E.mail: thelma@sochara.org; thelma.narayan@gmail.com
Website: www.sochara.org
Prasanna Saligram <prasanna.saligram@sochara.org>
To: Thelma Narayan <thelma@sochara.org>
Cc: Deepak <deepak@sochara.org>, Rakhal Gaitonde <rakhal@sochara.org >, "E. Premdas"
<premdas@sochara.org>, Ruth Vivek <ruth@sochara.org>, Ravi Narayan <chcravi@gmail.com>
1 June 2010 10:07
Dear Deepak,
Did you attach the articles? If so, Thelma’s mail has left it out because of it being a reply mail rather than a
forwarded mail. In a reply mail, the attachments get left out. Let me know. Otherwise if you have not got the
articles I can extract them and send it to the group
Best Wishes,
Prasanna
[Quoted text hidden]
Project Manager,
Center for Public Health & Equity (CPHE)
Society for Community Health Awareness Research and Action (SOCHARA)
E - 8/74, Basant Kunj, Arera Colony, Bhopal - 39.
Ph. - 0755-2561511
1 June 2010 10:16
Deepak <deepak@sochara.org>
To: Prasanna Saligram <prasanna.saligram@sochara.org>
Cc: Thelma Narayan <thelma@sochara.org>, Rakhal Gaitonde <rakhal@sochara.org>, "E. Premdas"
<premdas@sochara.org>, Ruth Viuek <ruth@sochara.org>, Ravi Narayan <chcravi@gmail.com>
Dear Prasanna
I just dropped into office, I had attached the articles in the mail to
Thelma, those attachments might have got discarded when the reply was
sent, am attaching the articles for the teams perusal, Very sorry for
the delay
[Quoted text hidden]
Thanks
Deepak
[Quoted text hidden]
http://mail. google. com/a/sochara.org/?ui...
3/5
07*/06/2010
Community Health Cell Mail - The Lancet ar...
[Quoted text hidden)
4 attachments
lancet 2010; 375 1182-92.pdf
914K
. Lancet 2008; 372; 962-71.pdf
■J 161K
Lancet 2010; 375; 1193-202.pdf
277K
Lancet 2008; 372; 2047-85.pdf
890K
4 June 2010 10:01
Ruth Vivek <ruth@sochara.org>
To: Thelma Narayan <thelma@sochara.org>
Cc: Deepak <deepak@sochara.org>, Prasanna Saligram <prasanna.saligram@sochara.org>, Rakhal Gaitonde
<rakhal@sochara.org>, "E. Premdas" <premdas@sochara.org>, Ravi Narayan <chcravi@gmail.com>
Dear Thelma
I think it will be good to be invalid in such collaborative research
initiatives to strengthen our networks as well as to hone our
technical skills.
I too had met Dr. Tripathy, his wife and the team once when I visited
their organisation during one of the commmunity monitoring visits.
They seemed to have a good team working there. Even then they were
involved in some sort of study. I was actually surprised to see such
kind of work happening in a remote part of Jharkhand.
So we should certainly start thinking on these lines
Regards
Ruth
On Tue, Jun 1, 2010 at 9:48 AM, Thelma Narayan <thelma@sochara.org> wrote:
[Quoted text hidden]
Ruth Vivek V.
Training & Research Assistant
Community Health Cell
85/2, 1st Main, Maruthi Nagara,
Madiwala,
Bengaluru - 560068
Tele/Fax:+91-80-25531518/25525372
Website: www.sochara.org
4 June 2010 13:08
Rakhal Gaitonde <rakhal@sochara.org>
To: Deepak <deepak@sochara.org >, Prasanna Saligram <prasanna.saligram@sochara.org>
Cc: Thelma Narayan <thelma@sochara.org >, "E. Premdas" <premdas@sochara.org>, Ruth Vivek
<ruth@sochara.org>, Ravi Narayan <chcravi@gmail.com>, shabanaameer2006@gmail.com, suresh@sochara.org
Dear Friends,
http://mail.google.eom/a/sochara. org/?ui...
4/5
07^06/2010
Community Health Cell Mail - The Lancet ar...
The following articles on the studies in Nepal and from Bolivia were the pioneers in studying community based
interventions using the cluster randomised trials. The bolivia study was probably the first study of this kind
(though it did not use the cluster randomised design).
There is one more study in this genre that I know of but have not been able to lay my hands on - and that is the
IMAGE study in south africa - this showed a decrease in domestic violence through community action - this was
just one of the outcomes - the main outcomes focusing on HIV / AIDS - this study was done by LSHTM and John
Porter is an author - so maybe we can get the references from him.
this makes a very impressive collection
in solidarity
rakhal
— Original Message — From: "Deepak" <deepak@sochara.org >
To: "Prasanna Saligram" <prasanna.saligram@sochara.org>
Cc: "Thelma Narayan" <thelma@sochara.org>; "Rakhal Gaitonde" <rakhal@sochara.org>; "E. Premdas"
<premdas@sochara.org>; "Ruth Vivek" <ruth@sochara.org >; "Ravi Narayan" <chcravi@gmail.com>
Sent: Tuesday, June 01, 2010 10:16 AM
Subject: Re: The Lancet articles
[Quoted text hidden]
3 attachments
bolivia warmi.pdf
325K
• .-i
LancetManandhar.pdf
495K
nepal womens group manandhar.pdf
1311K
http://mail.google.eom/a/sochara.org/2ui...
5/5
Reference
r
1.
O'Rourke K, Howard-Grabman L, Seoane G. 1998, "Impact of community organization of
women on perinatal outcomes in rural Bolivia". Pan American Journal of Public Health.
3(1):9-14.
2.
Manandhar, D. S., Osrin, D., Shrestha, B. P., Mesko, N., Morrison, J., Tumbahangphe, K. M.,
Tamang, S., Thapa, S., Shrestha, D., Thapa, B., Shrestha, J. R., Wade, A., Borghi, J., Standing,
H., Manandhar, M., & Costello, A. M. 2004, "Effept of a participatory intervention with
women's groups on birth outcomes in Nepalj-^tister-randomised controlled trial", Lancet,
vol. 364, no. 9438, pp. 970-979.
3.
Morrison, J., Tamang, S., Mesko, N., Osrin, D., Shrestha, B., Manandhar, M., Manandhar, D.,
Standing, H., & Costello, A. 2005, "Women's health groups to improve perinatal care in rural
Nepal", BMC Pregnancy and Childbirth, vol. 5, no. 1, p. 6.
4.
Backman, G., Hunt, P., Khosla, R., Jaramillo-Strouss, C., Fikre, B. M., Rumble, C., Pevalin, D.,
Paez, D. A., Pineda, M. A., Frisancho, A., Tarco, D., Motlagh, M., Farcasanu, D., & Vladescu, C.
2008, "Health systems and the right to health: an assessment of 194 countries", Lancet, vol.
372, no. 9655, pp. 2047-2085.
5.
Rosato, M., Laverack, G., Grabman, L. H., Tripathy, P., Nair, N.,Mwansambo, C., Azad, K.,
Morrison, J., Bhutta, Z., Perry, H., Rifkin, S., & Costello, A. 2008, "Community-participation:
lessons for maternal, newborn, and child health", Lancet, vol. 372, no. 9642, pp. 962-971.
6.
Tripathy, P., Nair, N., Barnett, S., Mahapatra, R., Borghi, J., Rath, S., Rath, S., Gope, R.,
Mahto, D., Sinha, R., Lakshminarayana, R.,Patel, V., Pagel, C., Prost, A., & Costello, A. 2010,
"Effect of a participatory intervention with women's groups on birth outcomes and
maternal depression in Jharkhand and Orissa, India: a cluster-randomised controlled trial",
Lancet, vol. 375, no. 9721, pp. 1182-1192.
7.
Azad, K., Barnett, S., Banerjee, B., Shaha, S., Khan, K., Rego, A. R., Barua, S., Flatman, D.,
Pagel, C., Prost, A., Ellis, M., & Costello, A. 2010, "Effect of scaling up women's groups on
birth outcomes in three rural districts in Bangladesh: a cluster-randomised controlled
trial", Lancet, vol. 375, no. 9721, pp. 1193-1202
- Media
RF_L_3_B_SUDHA.pdf
Position: 2229 (3 views)