4866.pdf
Media
- extracted text
-
Social and Economic Research Project Reports
No. 13
I
Mothers' Definition and Treatment of Childhood Malaria on the
Kenyan Coast
Halima Abdullah Mwenesi
Kenya Medical Research Institute
Medical Research Centre, Nairobi
P.O.Box 20752
Nairobi
Kenya
TDR
UNDPIWORLD BANK/WHO Special Programme for Research and Training in Tropical Diseases (TDR)
© WorldHealth Organization 1994
SER Project Reports appear as part ofa series of uneditedfinal reports resultingfrom
projects supported by the UNDP/WORLD BANK/WHO Special Programme for
Research and Training in Tropical Diseases (TDR). These reports are submitted to the
TDR Steering Committee on Social and Economic Researchfor review and evaluation
upon completion of a project. Project reports included in this series have not been
published in their entirety elsewhere.
The designations employed and the presentation of the material in SER Project
Reports do not imply the expression of any opinion whatsoever on the part of the
Secretariat of the World Health Organization concerning the legal status of any
country, territory, city or area or of its authorities, or concerning the delimitation of
its frontiers or boundaries.
Authors alone are responsible for the views expressed in SER Project Reports and
for the presentation of the material contained therein.
Community Health Cell
Library and Documentation Unit
367, "Srinivasa Nilaya"
Jakkasandra 1st Main,
1st Block, Koramangala,
BANGALORE-560 034.
Phone : 5531518
I
TDR/SERZPRS/13
Social andEconomic Research ProjectReports
Mothers' Definition and Treatment of Childhood Malaria on the
Kenyan Coast
Halima Abdullah Mwenesi
Kenya Medical Research Institute
Medical Research Centre, Nairobi
P.O.Box 20752
Nairobi
Kenya
(June 1993)
UNDP/WORLD BANK/WHO Special Programme for Research and Training in Tropical Diseases (TDR)
No. 13
Foreword
TheUNDP/WORLD B ANK/WHO Special Programme for Research and Training in Tropical Diseases
(TDR) is a globally coordinated effort to bring the resources ofmodem science to bear on the control of major
tropical diseases: malaria, schistosomiasis, filariasis (including onchocerciasis), the trypanosomiases (both
African sleeping sickness and the American form, Chagas’ disease), the leishmaniases and leprosy. The
Programme has two interdependent objectives:
• To develop new methods of preventing, diagnosing and treating selected tropical diseases, methods
that would be applicable, acceptable and affordable by developing countries, require minimal skills
or supervision and be readily integrated into the health services of these countries;
• To strengthen - through training in biomedical and social sciences and through support to institutions
- the capability of developing countries to undertake the research required to develop these new
disease control technologies.
The Social andEconomicResearchProjectReports series represents a communication venture undertaken
by TDR's Social and Economic Research (SER) component. This series was launched in 1987 to facilitate and
increase communication among social scientists and researchers in related disciplines carrying out research on
social and economic aspects of tropical diseases and to disseminate social and economic research results to
disease control personnel and government officials concerned with improving the effectiveness of tropical
disease control.
Research reports published in this series are final reports ofprojects funded by TDR and usually include more
material than ordinarily published in peer review journal articles. TDR considers this material to be valuable
both for investigators involved in the study of social and economic aspects of tropical diseases and for
professionals involved in training programmes in the social sciences, economics and public health. The series
should acquaint those working on similar problems with approaches undertaken by others, in order to test new
approaches in different settings, and should provide useful information to personnel in disease control
programmes and related agencies.
Although SER, as of 1 January 1994, was integrated into an Applied Field Research (AFR) component of
the TDR Programme, the AFR Steering Committee considers that the series deserves to be continued so that
results of the social and economic research in TDR will continue to be disseminated.
In the interests of rapid dissemination of social and economic research findings, much of the supporting
material, e.g., tabulated data, has not been included in the present report. This materialis, however, available
upon request to interested researchers. All requests for such material, citing in full the number, title and author(s)
of the SER Project Report, should be addressed to: Dr C. Vlassoff, Secretary, Manager, Gender and Tropical
Diseases Task Force, TDR, World Health Organization, 1211 Geneva27, Switzerland.
Tore Godal, Director
Special Programme for Research
and Training in Tropical Diseases
TDR
iii
PREFACE
This report is based on research undertaken for doctoral work by Dr. Halima Mwenesi,
supported by TDK’s Research Strengthening Group. This project is an example of a truly
worthwhile investment from TDK’s perspective, not only because it resulted in a successful
PhD but also because the research itself produced extremely interesting and useful results.
A major contribution of the report is that it demonstrates women’s clear knowledge and
understanding of malaria in the study area. The distinctions made by mothers between signs
of fever and signs of convulsions and the way in which they treat them are particularly
insightful and instructive. The study provides a closer view of people's perceptions,
knowledge and behaviour at the community level than is available to medical personnel in
health facility settings, or to researchers using traditional survey instruments. Through her
insights and recommendations, Dr. Mwenesi has provided an opening for the development of
future research and interventions.
Carol Vlassoff
Manager
Gender and Tropical Diseases Task Force
Special Programme for Research and Training in Tropical Diseases (TDK)
v
Mothers9 Definition and Treatment of Childhood Malaria on the Kenyan Coast
Final Report of a project supported by the TDR Social and
Economic Research Component
Halima Abdullah Mwenesi
Kenya Medical Research Institute
Medical Research Centre, Nairobi
Advisors
Trudy Harpham
London School of Hygiene and Tropical Medicine
University of London
Robert W. Snow
Kenya Medical Research Institute
Coastal Unit, Kilifi
June, 1993
vi
Acknowledgements
The research described in this report is a modest venture of an inexperienced social
science student who set out to learn more about social aspects of health and disease,
with a bias towards tropical diseases, specifically, malaria.
In my quest to understand human behaviour in relation to health and illness, I have
interacted with several people who have been a source of encouragement and from
whom I have learned a great deal. Some of these people deserve to be mentioned: Dr.
Patricia Rosenfield of the Carnegie Foundation of New York (formerly of TDR), and
Prof. S. Migot-Adholla of the World Bank (formally chairman, sociology department,
University of Nairobi) encouraged my interest in tropical diseases research, and my
studies in social science and medicine. I thank Dr. Carol Vlassoff and the entire Tropical
Disease Research and Social and Economic Research (TDR/SER) Committee for
supporting my PhD training programme as a Principal Investigator.
I am also grateful to the numerous colleagues and friends who have contributed to and
encouraged me through the course of my studies. Special and sincere appreciation goes
to my supervisor, Dr. Trudy Harpham, an untiring source of inspiration, who has
encouraged and guided me through the rigours of scientific enquiry throughout the entire
study period. Many thanks to Dr. P.R. Kenya and Dr. Kevin Marsh for giving me the
resolve to embark on the project, and Dr. R.W. Snow, who unwaveringly offered me
invaluable advice and time since the commencement of the study. The help obtained
from Dr. Dayo Forster on data management is sincerely acknowledged.
I am grateful for the financial support I received from the UNDP/World Bank/WHO
Special Programme for Research and Training in Tropical Diseases (TDR/SER).
Secondly, I thank Dr. D. Koech, Director, Kenya Medical Research Institute for allowing
me time off to further my studies; and all members and colleagues of the Health Policy
Unit, at the London School of Hygiene and Tropical Medicine. Others I wish to mention
are the Head of Unit, Dr. N. Peshu and all colleagues at the KEMRI Unit, Kilifi; the
field workers and all the support staff from Kilifi and the Medical Research Centre
(MRC), Nairobi.
My deepest appreciation goes to all the people in Kilifi, especially the mothers, without
whose cooperation this undertaking would not have been possible. Thanks are also
extended to the administrators of the Kilifi district hospital, Ngerenya dispensary and the
private clinics in Kilifi town.
After several trials and errors at producing a proposal, progress reports, financial reports,
managing a team of ten field workers, (and a young family) I have finally produced a
thesis for a doctoral degree; the last three years have not been in vain.
Halima Abdullah Mwenesi
Principal Investigator
June, 1993
vii
Table of Contents
Acknowledgements
3
Abbreviations
6
EXECUTIVE SUMMARY
7
INTRODUCTION......................
Statement of the problem
Justification for the project
Background to the study .
Project Objectives ...........
8
8
8
9
10
THE SETTING OF THE STUDY: KILIFI DISTRICT
Study sites..........................................................
Study population...............................................
10
11
12
METHODOLOGY .............................................................................................
Theoretical framework and research methods.........................................
Development of research tools.................................................................
Mapping and enumeration of household retail outlets and health
facilities ..................................................................
The census ....................................................................................
The mothers’ survey .....................................................................
The health facility users’ survey....................................................
The retail outlet users’ survey ......................................................
The retail outlet proprietors’ survey.............................................
The drugs in homes survey ..........................................................
The ethnomedical survey...............................................................
Non-participatory observation ......................................................
Fieldwork supervision and data analysis.......................................
12
12
13
13
14
14
16
17
18
18
18
19
19
RESULTS................................................................................................................ 20
Definition and management of childhood malaria in Kilifi .....................
20
Mothers’ beliefs and behaviour in relation to childhood malaria . . . 22
Management of childhood malaria at home..................................... 23
Convulsions (fits)................................................................................ 25
Anaemia............................................................................................. 26
Splenomegaly .................................................................................... 26
Mothers’ health seeking behaviour (HSB) for malaria........................
27
Mothers’ use of health facilities........................................................ 27
Mother’s use of retail outlets............................................................ 27
Retail outlet proprietors’ knowledge of malaria and its
medications . . .
28
Antimalarial drugs available in retail outlets and
homes.
28
ix
DECISION-MAKING DYNAMICS IN HEALTH SEEKING BEHAVIOUR ... 29
DISCUSSION ...................................................................
A biocultural definition of malaria........................
"Cultural blindness"? - The mosquito - malaria link
Does malaria have any complications?.................
Morbidity and mortality from childhood malaria . .
Who makes decisions for childrens’ health care? .
Mothers’ diagnostic abilities..................................
Treatment selection...............................................
Retail outlet proprietors .......................................
METHODOLOGICAL ASSESSMENT..........................
Integration of methodologies................................
Limitations: scheduled and unexpected.................
29
31
32
33
34
35
36
37
37
39
39
41
IMPLICATIONS AND FUTURE RESEARCH NEEDS
43
REFERENCES
49
List of tables
Table 1: Socio-demographic characteristics of mothers
20
Table 2: Mijikenda and Luo mothers’ knowledge, beliefs and
management of childhood malaria............
22
Table 3: A summary of mothers’ knowledge, beliefs and
management of convulsions in children ...........
25
Table 4: Summary of major findings
30
Table 5: Issues on malaria control that need to be addressed
or improved upon..................................... 44
x
EXECUTIVE SUMMARY
Presumptive treatment of clinical malaria by health personnel, village health workers
(VHWs) or mothers is currently the mainstay of malaria control in Africa. Surprisingly,
little is known about the management of malaria and its complications in the home. This
study examines mothers’ definitions of malaria, recognition of complications associated
with malaria, the use of proprietary drugs and decision-making dynamics in the health
seeking process.
The research was carried out in Kilifi district, situated on the Kenyan coast, an area
holoendemic for malaria. Complementary data collection methods (quantitative,
qualitative and non-participatory observation) were used to explore the four main
objectives of the study:
1. To determine whether the different communities under study perceived malaria as a
major health problem and how they translated this perception into action.
2. To determine whether overt complications of childhood malaria were recognised and
how they were managed.
3. To determine decision-making dynamics and the significance of family and other social
networks in the health seeking process.
4. To determine the extent and reasons for use of proprietary treatments for malaria in
the context of other traditional and modern therapeutic measures.
A total of 1408 respondents, drawn from three different settings; a) peri-urban, b) rural,
and c) a slum settlement, were included in the study.
The main findings of the study were:
1. Mothers have limited knowledge of childhood malaria as a biomedical disease, but
have a clear and precise biocultural definition of the disease.
2. The biocultural definition of childhood malaria does not recognise biomedical
definitions of complications of malaria, but provides separate explanatory models for
these conditions.
3. Decision-making for health care is the prerogative of males.
4. Use of proprietary treatments for malaria is partly a function of the biocultural
definition and the easy availability of over-the-counter (OTC) drugs. More importantly,
mothers are treating their children promptly, albeit not effectively.
«
This study contributes to the understanding of psychosocial processes that generate and
mediate viable management of childhood malaria at the household and community
levels. Recommendations are made which call for the re-assessment of the image of
malaria that is portrayed to populations in malarious areas and the promotion of health
education in malaria control. It is hoped that the findings can be linked to malaria
programmes in the study area and as pointers to control programmes elsewhere - thus
contributing to policies for malaria control and health promotion.
xi
INTRODUCTION
Statement of the problem
Presumptive treatment of clinical malaria is currently the mainstay of malaria control in
Africa. This has been in response to the failure of conventional malaria control efforts.
Recently, there have been trials of community-based presumptive treatment services
delivered by Village Health Workers (VHWs), in line with the implementation of
primary health care (PHC) (WHO/UNICEF, 1978). VHWs are a cadre of health
personnel perceived to be more accessible and acceptable to members of their
communities. They are expected to improve overall coverage and equity of services by
increasing service use especially by disadvantaged individuals and households (Berman
1984). Recent studies however, indicate that community-based chemotherapy strategies
in Africa are yet to help reduce morbidity and mortality from malaria (Walt, 1988;
Slutsker, Breman and Campbell 1988; Hoffman, Masbar, Hussein et al. 1984; Spencer,
Kaseje, Sempebwa et al. 1987; Greenwood and Bradley 1988; Menon 1991).
Little is known about the management of malaria and its complications in the home.
This study did not investigate the reasons for failure or success of PHC in malaria
control. Rather, it aimed at highlighting some issues that may mediate viable
management of childhood malaria at the household level. The position taken in this
study is that the general axiom about health and illness in a culture mediates upon
information on mothers’/carers’/guardians’ (hereinafter referred to as mothers)
definition of childhood malaria; their perceptions of the seriousness of the disease; their
response to its different presentations; the time-lag between onset of illness and remedial
action; the case for self-medication, available services and the decision-making dynamics
involved in the choice of care. This is important in formulating useful criteria for malaria
control programmes.
Justification for the project
The behaviour of humans has not been given an equal position with the other factors in
the malaria equation. Arguments such as:
Malaria control activities ... depend on the prevailing epidemiology of the disease,
on the efficacy of technological methods, on the structure of health services, on
the logistic and financial capability of each country and, last but not least, on the
national commitment (The Lancet 1983: 963);
»
or the contention by Ruebush, Breman, Kaiser et al, (1986), that the practical realities
of controlling malaria through PHC are the need for a thorough understanding of
malaria epidemiology as well as the recent call by Okelo (1990) that "malaria control
strategies of the future should focus on a wide variety of areas such as schizonticidal
drugs to replace chloroquine, insecticides, and vaccines..." (p: 293) illustrate the neglect
of human behaviour in current thought.
1
However, as Oaks, Mitchell, Pearson et al. (1991) put it:
human behaviour and social organisation - one side of malaria’s host-vector
parasite triangle- are clearly vital determinants for the success of control
programmes. Unfortunately, we do not know enough about how humans respond
to malaria to be able to build strong multidisciplinary control programmes (p:
258).
Bradley (1991) concludes, "no one method will on its own control malaria, but a
combination of skills and methods, because there is not an either/or choice". He says:
"... when the insights of anthropology and education fully permeate the way in which
environmental control is implemented and are not just added on as afterthoughts ... then
real progress in control can be made" (Bradley 1991:28).
These were the challenges of this social science study. It sought to fill the gap between
the epidemiology and the social aspects of malaria control, by trying to explain the
meaning and values of some socio-cultural and behavioural factors that may affect
household and community based efforts for malaria control. There is a dearth of such
studies which could help us understand not only why programmes fail or succeed, but
also how to sustain successful ones.
Background to the study
This project was carried out parallel to a hospital based case-control study on severe
malaria in children. Several issues emerged from the early phases of the case-control
study:
a) Severe disease presented from the study area frequently even though
entomological challenge was low (i.e few mosquitoes but a lot of severe
disease).
b) Mothers were frequently administering antimalarial (AM) drugs and other
over-the-counter drugs (OTC) to their children before presenting to hospital.
c) Convulsions were a common feature of severe malaria compared to other
serious febrile illnesses.
d) Severe malaria appeared to be geographically clustered with some areas
generating more cases than others.
e) Reported use of traditional healers (THs) for the treatment of malaria
was low.
This study was conducted to explain the specific behaviours highlighted above.
2
Project Objectives
The initial objectives of the study approved in May 1989 were:
i) To assess the degree to which the basic understanding of the illness
"malaria” and treatment practises vary between mothers of children taken
to health centres with malaria and mothers in the general community.
ii) To assess the extent to which the incidence of clinical episodes of
malaria in children aged 0-9 years treated at a health centre are
representative of the incidence of clinical episodes of malaria actively
detected in the community.
In August 1989, the objectives were revised after lengthy discussions, between the
investigator and the field advisor Dr. R. Snow, on the issues arising from the case-control
study. The first objective was found to be more involved than originally anticipated and
the second objective to require a more indepth prospective surveillance of fevers than
feasible in the available time. Subsequently, the second objective was dropped and the
first one amended and expanded.
The amended objectives of the study were:
i) To determine whether the different communities under study perceived malaria
as a major health problem and how they translated this perception into action.
ii) To determine whether overt complications of childhood malaria were
recognised and how they were managed.
iii) To determine decision-making dynamics and the significance of family and
other social networks in the health-seeking process.
iv) To determine the extent and reasons for use of proprietary treatments for
malaria in the context of other traditional and modern therapeutic measures.
THE SETTING OF THE STUDY: KILIFI DISTRICT
Kilifi is one of the six districts in the Coast Province of Kenya, an area holoendemic for
malaria. It measures about 12,523 km2 (KDDP, 1989) and is divided into five
administrative divisions with its headquarters in Kilifi town. The district’s population in
1991 was 711,838, with an average density of 57 persons per km2 (CBS, 1991). Kilifi
experiences two rainy seasons. Temperatures and humidity are generally high all year,
with daily means averaging a minimum of 22° Celsius and a maximum of 30-34° Celsius.
This type of climate, coupled with a large expanse of water, swamps and forests, favours
continuous breeding of mosquitoes and development of parasites.
3
Kilifi district has three hospitals: two government-run hospitals at Kilifi and Malindi
towns, a mission hospital at Kaloleni, six government-run health centres and thirty-four
dispensaries. The disease pattern in Kilifi shows a preponderance of preventable diseases
as the major causes of morbidity and mortality in both children and adults. Further, it
is estimated that one in three children in Kilifi is malnourished (Thompson, 1990) and
educational levels in the district, especially among women, are among the lowest in
Kenya. However, malaria is seen as the major factor contributing to the under
development of the area (Ewbank, Henin and Kekovole 1986). Blacker, Mukiza-Gapere,
Kibet et al. (1987) state the determinants of differential mortality in coastal Kenya thus:
...coastal districts show most of the correlates of high mortality particularly
Kilifi... These districts are highly malarious, have high rates of stunting and
wasting, are relatively poorly provided with HFs; their women are badly
educated and the lack of progress in female education correlates with lack
of progress in the decrease of child mortality. Only on the score of piped
water do they come out reasonably well (: 9)
Earlier work in Kenya correlating malaria endemicity, maternal education, malnutrition,
health facility (HF) accessibility and access to clean water, estimated child mortality rates
for Kilifi to be 212 /1000 live births (PSRI 1982). More recent statistics show that
malaria is by far the major single cause of death in hospitalised children and malaria
specific mortality in Kilifi are at approximately 6/1000 per year (Snow pers.comm. 1992).
The vector responsible for malaria in this area is the A.gambie s.l which accounts for
over 90% of infections (Mbogo, pers. comm. 1992).
Study sites
Three sites, all within the case-control study area, were chosen for the study. Kilifi town
is a peri-urban, relatively affluent setting with a population of 11,700 people at the 1989
census (CBS, 1991). It covers an area of six km2, with a population density of >800 per
km2. The town hosts the district’s headquarters and has a heterogenous population,
engaged in different occupations. The only major industry in Kilifi town is the cashew nut
factory. The town is reasonably well supplied with transport services. Kilifi district
hospital (KDH), at the town centre, provides out-patient and in-patient services. There
were three private clinics in the town.
Mtondia is a densely populated slum situated about 15 km from Kilifi town. It is
inhabited by a squatter community of the Luo, a Nilotic group which migrated to the
area from the Lake basin region of Western Kenya. They are mainly quarry diggers but
also engage in fishing for their subsistence. The population in 1989 was estimated at 1000
people. The area has one private HF and has access to clean piped water at a fee.
Sokoke is a remote rural settlement scheme about 40 km from Kilifi town. In the early
seventies, families were settled in the area by the government on twelve acres of land
each. Transport services are intermittent and almost non-existent during the rainy
season. The population was estimated to be 6,563 people in 1989. The people are
typically rural, subsistence farmers and predominantly Giriama. There is one dispensary
4
and people have access to clean piped water at a fee, although water points are few and
far between.
These areas were purposively, not randomly, selected. As alluded to earlier, there were
differences in the recruitment of cases from some parts of the study area designated for
the case-control study. Preliminary results (Snow, pers. comm. 1989) of the case-control
study gave the following indications:
Sokoke: The rural area 40 km away and with poor accessibility to Kilifi District Hospital
(KDH) had surprisingly large numbers of children recruited into the case-control study
during 1989. Kilifi town: The peri-urban area within easy reach of KDH had
proportionately fewer children recruited into the case-control study in 1989, despite
having an apparently large number of under five year olds. Mtondia: The slum
settlement community about 15 kms away, with fairly adequate accessibility to KDH, also
had few children recruited into the case-control study in 1989.
Study population
Seven of the nine Mijikenda peoples: the Giriama, Jibana, Chonyi, Rabai, Kambe,
Kauma and Ribe are in Kilifi district. The Giriama are the largest group making up 90%
of the total district population. The Mijikenda family system is patriarchal and mostly
polygamous. They believe in natural and supernatural causation of illness and practice
sorcery. As a corollary, different types of healers, diviners, sorcerers and “witch-finders’
who formally define the causes of personal and social misfortune and even death are
central to the lives of the people (Thompson, 1990; Parkin, 1991; Mwenesi, 1993). For
a detailed description of the population, see Mwenesi (1993).
The Luo belong to the Nilotic ethnic group that occupies the Lake Basin region of
Western Kenya, Nyanza Province. They have the highest rate of out-migration of ablebodied Luo men in the country (Alila, 1978), hence the Mtondia Luo settlement. The
family system is also patriarchal and mostly polygamous (Parkin, 1978) and they also
believe in natural and supernatural causation of illnesses and practice sorcery (Whisson,
1964; Mwenesi, 1993). Different healers deal with different diseases and misfortunes.
The region from which the Luo community migrated resembles Kilifi district in terms of
disease distribution, high child and infant mortality rates, low socio-economic levels,
education levels and the available infrastructure (GK/UNICEF, 1989; 1990).
METHODOLOGY
Theoretical framework and research methods
t
The cultural-ecology theoretical perspective, which views the interaction between people
and their micro-environments (ecological, economic and cultural) as determinants of
health problems in relation to local cultural definitions and expectations together with
and the social construction approach which emphasises the cognitive and interactional
processes involved in the management of disease were used as the overall conceptual
frameworks for data analysis (Stanton, Black, Engle et al 1992). Both these perspectives
5
acknowledge the differences between biomedical and biocultural knowledge and
practices, as opposed to approaches that explain behaviour in terms of biomedical
rationality.
Complementary methods of data collection were used to explore the four main objectives
of the study. Several workers (Jick, 1983; Rubinstein 1984; Glik, Parker, Mulingande et
at 1987; Heggenhougen and Clements 1987; Mechanic 1989; Steckler et al. 1992; Yach
1992) have argued that combining different methodological approaches is important for
social research intended to provide information upon which interventions can be
designed. This study required both qualitative and quantitative data to accomplish the
proposed objectives. The methods used were:
i) quantitative survey research methods where structured interview questionnaires were
used to collect quantitative data.
ii) qualitative methods where the ethnographic approach of using unstructured in-depth
interviews and non-participatory observation methods were used.
Development of research tools
All draft questionnaires were prepared in London. In Kilifi, ethnographic inquiries were
made and the drafts were discussed informally with colleagues and some residents. The
aim of this exercise was to identify and understand any traditional terms and concepts
used to describe disease and illness in the different study areas. Finally, six
questionnaires and an in-depth interview schedule were developed and pre-tested. The
pilot study assisted in refining the questions and establishing codes for most possible
responses. During the pre-tests, it was decided to use a tape-recorder and to develop a
drug collage. The questionnaires were translated into Kiswahili, Kigiriama and Dholuo;
then translated back into English to identify any idiosyncrasies in the interpretations.
They were, however, administered in the languages in which the respondents were most
conversant. The investigator could follow conversations in both Kigiriama and Dholuo.
A detailed field manual which doubled as a codebook, was developed. It explained to the
field worker why and how each question was to be asked and recorded.
Meetings with relevant administrative authorities were held, where they were briefed on
the purpose of the study. Ten local school-leavers were recruited. They spoke the local
Mijikenda languages and Kiswahili (the national language) fluently. No Dholuo-speaking
field workers were recruited - none presented themselves for recruitment. They were
suspicious of the motives of the research because of their squatter status. However, it
had been established during the pilot study that most Luo people could speak Kiswahili.
It was decided to use "on the spot" interpreters if the need arose. The field workers were
trained for two weeks on general aspects of field procedures.
Mapping and enumeration of household retail outlets and health facilities
Households were mapped and enumerated to facilitate easy tracing of the respondents
and subsequent follow ups and to devise an accurate sampling frame. Enumeration of
6
»
retail outlets (ROs) and health facilities (HFs) facilitated recruitment of the self-selected
group of mothers who resort to these facilities for the treatment of their children.
Enumeration of ROs was also necessary for the recruitment of proprietors who stocked
and sold antimalarial (AM) and other over-the-counter (OTC) drugs.
Smith and Morrow (1991) recommend that definitions of "households” be made explicit
for any studies using them as units of analysis. The diversity of household types
(Mwenesi, 1993) necessitated the development of an operational definition of the term.
A household was defined at two levels. The first level involved the physical and
geographical identification of a home and the second level was created to fit in with the
study design. Social-anthropological parameters were incorporated, to create a
"household unit" that included all members with influence upon one another’s decision
making processes, which would transcend the geographical identification. Using the
outlined definition of households, sketch-maps of the study areas were made since no
current maps of the areas were available. Also mapped were all ROs that sold OTC
drugs both within the study areas and within a one kilometre radius of their outskirts.
The census
The de jure system (Smith and Morrow, 1991) was used to conduct a census, which was
necessary in order to provide precise demographic profiles of each area and to provide
a comprehensive sampling frame for further surveys (Mwenesi, 1993).
A household roster was developed and used. Each individual was given an identity
number. Other information gathered through the roster was the name of the household
head, the respondent’s name, his/her relationship to the household head and others in
the household, gender, month or year of birth, ethnic or tribal group, whether present
or absent at the time of enumeration, and the "kid’’ code which was used to identify
women who had or were guardian to children under nine years of age. Age estimations
for the study population were made by use of a historical and local events calendar.
All mothers and female guardians of children under nine years of age were included in
the study. Similarly, all retail outlet proprietors (ROPs) who stocked OTC drugs from
the mapped areas were included. The sample was large because analysis involved
comparisons of people possessing multiple, interrelated characteristics so that meaningful
conclusions could be arrived at from the data collected from sub-groups of the sample
i.e. mothers re-interviewed at HFs and at ROs.
The mothers’ survey
!
The mothers’ survey gathered data on respondents’ knowledge, beliefs and behaviour
towards malaria and associated complications as well as their demographic and socio
economic characteristics. A questionnaire was administered to all available study area
mothers. The mothers’ survey questionnaire was developed to include only questions
relevant to the objectives of the study and to determine how questions were to be
formulated, especially the important question on the local definition of malaria. These
questions were tried out during the pilot study:
7
a) Tell me the types of fever you know.
b) What do you do when your child has fever?
c) What disease name would you give the following symptoms?
fever + vomiting + headache?
d) What is malaria?
Important points came to light. The four questions were inadequate for the study
objectives. The first question produced an array of febrile illnesses which were difficult
to interpret. Mothers recognized fever and had terms to describe several types including
malaria and other conditions that present with pyrexia. This question’s advantage over
the others was that it drew malaria as one type of fever.
Question (b) assumes that malaria is synonymous with fever and does not consider other
types of fevers. Although it has been used to elicit answers for malaria by several
workers, for example: Dabis, Breman, Roisin et al. (1989) and Glik, Parker, Mulingande
et al, (1989), in Guinea; and Deming, Gayibor, Murphy et al, (1989), in Togo, it was
unsuitable for this study. The position of this study is that the answers to this question
may be misleading if different types of fevers are perceived to have different aetiologies
by a given community.
Question (c) was not useful either. For most of the mothers, symptoms were perceived
as disease entities in their own right. Fever, vomiting and headache were three distinct
illnesses.
The fourth question (d), which is the straight forward approach used in Kenya by
Abdullah (1985) and Ongore, Kamunvi, Knight et al. (1989) was not understood by most
mothers, except those who had secondary school level of education and Luo mothers who
all distinguished other febrile illnesses from the condition they referred to as malaria,
even when they did not describe malaria correctly. Clearly, the direct question of ’what
is malaria’ can only be asked of respondents with some amount of schooling or where
malaria control efforts have been carried out over a long period of time.
The preceding discussion illustrates difficulties involved in questioning people about a
complex disease like malaria, which has no specific symptoms and mimics many other
conditions. Neither malaria nor fever can be investigated as individual disease categories,
especially in areas of low literacy, as fever is one of the many symptoms used to denote
the disease of malaria. Consequently, the study concentrated on mothers’ perceptions of
severe childhood malaria. It was hoped that identifying manifestations of severe malaria
perceived by mothers to be serious, would enhance the case for malaria drug
intervention.
Clinically, manifestations and complications of P.falciparum are numerous and varied.
They include cerebral malaria, anaemia, repeated generalised convulsions (Warrell,
Molyneux and Beales 1990) and splenomegaly (Greenwood 1987). Convulsions, anaemia
and splenomegaly were chosen and included in the study. These were chosen because (a)
all the three study communities had local names for them indicating that they were
8
recognised, (b) they have fairly overt signs (convulsions and anaemia) and (c) they were
of interest to the case-control study. Preliminary results of the case-control study showed
that more than 50% of the children diagnosed as having severe malaria had a history of
convulsions and were anaemic (Snow, in press). Cerebral malaria was not included in the
study because it was not mentioned in relation to children at all during the pilot study.
Modules of symptoms and conditions including malaria as a disease category were
included in the questionnaires, as was the generic term hpma which covers a range of
febrile conditions, convulsions, anaemia and splenomegaly. To be questioned about any
of the five conditions, a mother would have had to mention it as a childhood illness that
presents with high fever. The field workers were instructed to use the literal
transliteration of fever - "hot body" - which had the same meaning in the three languages
used for interviews. These criteria removed the preconceived assumption inherent in
many studies that communities living in malarious areas define malaria biomedically
(Abdullah 1985; Ongore, Kamunvi, Knight et al. 1989; Deming, Gayibor, Murphy et al.
1989; Dabis, Breman, Roisin et al, 1989; Glik, Parker, Mulingande et al. 1989) and
ensured that respondents knew or were aware of the illness about which they were
answering questions.
The final questionnaire had four sections. (1) background information of the respondent
such as education, marital status, religion, occupation of self or partner and parity. (2)
illness modules, the core of the questionnaire which addressed mothers’ definitions and
reactions to malaria and its complications. Sections (3) and (4) addressed health
consultation behaviour, decision-making dynamics and socio-economic variables. To
introduce some amount of flexibility in the interview, all the illness modules allowed for
interviewer prompting after the respondent had exhausted spontaneous information.
All questions were closed and pre-coded. A manual and an aide memoire, as well as a
collage of all the drugs in the ROs was prepared and used with the questionnaire. The
collage was used to ascertain whether mothers could identify any OTC drugs they
mentioned. All households that had a mother/s and a child or children under nine years
were included in the study. The interviews (15-20 minutes) were conducted in the
respondents’ homes.
The health facility users’ survey
The HF users’ survey, conducted after the mothers’ survey, aimed at validating data
relating to service use and monitoring patterns of health-seeking behaviour (HSB) during
a self-diagnosed episode of malaria. It had been proposed to interview only those
mothers who had brought a child suspected of malaria or any of its complications to a
HF. The pilot study revealed that very few mothers would be recruited at HFs if that
inclusion criterion was retained. The proposal was dropped and all mothers from the
study areas presenting at a HF during the survey were interviewed regardless of the
child’s illness. This did not compromise the objectives of the study and enhanced
information on general HSB for children.
9
The questionnaire covered matters on accessibility, decision-making for choice of service,
duration of illness, management at the household level and an assessment of the
concordance between mothers’ and health workers diagnosis of a childs’ illness.
Management of malaria at the HF was noted, and mothers’ understanding of prescription
descriptions was assessed. The last section of the questionnaire assessed mothers’ abilities
to follow treatment regimens, and the multiple use of services. This was administered to
the mothers three days later in their homes. They were not forewarned about the follow
ups.
Every Monday, Tuesday, Thursday and Friday between eight a.m. and six p.m., one field
worker was stationed at each of the six HFs in the area except for the KDH which was
manned by three field workers. Starting at the out-patients department or reception
areas, any adult female accompanying a child/children to the HF was approached. If it
was established that the adult woman was from the study area, had been interviewed
during the mothers’ survey, and had brought the child because she suspected that she/he
was ill, the questionnaire was administered. If on the other hand the female caregiver
had only brought the child to the "Well Baby Clinic” for routine procedures, the interview
included only part one of the questionnaire and was terminated with no follow-up.
The identified mother was then followed along the queue at the HF. After her child was
seen, the field worker recorded the treatment prescribed and followed her to the
pharmacy. Here, drugs given were recorded and duration and regimen of treatment
noted; by counting the number of tablets and measuring the amount of syrup given in
centimetres against the bottle used. This was used only as an estimation. The mother was
then asked to repeat the instructions given to her about the medication and the answer
was recorded.
Wednesdays and Saturdays were follow-up days. Mothers were re-interviewed in their
homes about how they had administered the medications to the child, including the
measure they had used for giving syrups, whether any other person had shared the child’s
drugs, whether they had given any other treatment since being seen at the HF and the
reasons for doing so. Tablets were re-counted and syrups re-measured. The interview
lasted about ten to fifteen minutes.
The retail outlet users’ survey
The RO users’ survey, like the HF survey, aimed to validate some of the data collected
in the mothers’ survey, and to gather information on actions taken for self-medication
during a current self-diagnosed episode of malaria or its complications. The survey also
aimed at gathering information on the extent and reasons for use of OTC drugs for
treatment.
The questionnaire was adapted from the HF questionnaire. It had two sections. The first
section sought information on health consultations, the current illness for which the drugs
were being purchased, history of previous treatment and the type and combination of
drugs that were purchased. The mother was asked about the mode of transport, time and
money spent for the journey to the RO. Section two sought information on any other
10
treatment the mother might have given to the child since purchasing the drugs, and the
rationale.
For four days in a week field workers were stationed at ROs from opening to closing
time. Any adult female who purchased any drugs, was from the study areas, had been
interviewed during the mothers’ survey and was purchasing the drugs for her
child/children was included in the survey. Section one of the RO user’ questionnaire was
administered to the mother at the RO. The mothers were re-interviewed at their homes
on the third day to find out whether they had replenished the drugs, whether they had
used other types of treatment, and the reasons for using other treatments. They were not
forewarned about the follow-up. The interview lasted five to ten minutes. It had been
intended that any persons purchasing drugs at ROs be questioned, just in case a mother
in the study had sent someone to the RO. It proved difficult because most of the proxies
(especially children) knew neither the names of the mothers nor the names of heads of
households (necessary for follow-up). Secondly, it was not practical to follow each of
these proxies back to their homes because the RO sentinels would be broken. Thus only
mothers who actually purchased drugs themselves were interviewed.
The retail outlet proprietors’ survey
The aims of the ROPs’ survey was to find out their knowledge, attitudes and practices
(KAP) on malaria and the extent of the use of OTC drugs on malaria in the
communities. The questionnaire for this survey had three sections. The first section dealt
with demographic information: age, gender, education, religion and ethnicity. This
information was included because biographic information collected during the census was
only for ROPs within the study area. Even for those from within the study areas,
available information was only on age, gender and ethnicity. The second section dealt
with the type of drugs stocked, and the quantities sold. The final section dealt with KAP
of ROPs in relation to malaria, and their knowledge of correct dosages of the different
AM drugs in stock.
This survey was conducted by the investigator. All 64 ROPs already identified as stockists
of OTC drugs were included in the study. The person behind the counter of the RO at
the time of the visit was interviewed only if they were the usual store-keepers. Stand-ins
were not interviewed.
The drugs in homes survey
A small survey was conducted to validate and verify some of the information collected
from ROPs, on available drugs in households. By use of random numbers, 390
households were picked from the three study areas. The adult person who was found in
the household at the time of the visit was interviewed. This small survey also aimed at
assessing the quantities of chloroquine and other AM drugs consumed by the use of
proxy indicators: quantities of AM drugs sold, dispensed or available in households
(Horgerzeil, 1985).
11
The ethnomedical survey
This approach was firstly used to gather preliminary data to help in development of
questionnaires and secondly, for collection of data on beliefs and perceptions surrounding
matters of health and disease. The questions focused on disease aetiology, treatment and
prevention. The reasons for questioning the people about disease in general (without
focusing on malaria) was to locate the place of malaria in the disease profile of the
people. This strategy also provided information on priority health problems in the area,
reasons for use and non-use of services and the people’s perception of the cause of the
very high infant and child mortality rates in the area. Suggestions for remedies for the
perceived problems were also obtained.
Key informants were chosen through purposive and snow-ball sampling methods. The
first few informants were picked because they were perceived by other people in the
community to know the information required. The subsequent informants were suggested
by those already chosen. In-depth informal interviews (Khan and Manderson, 1992) were
conducted. The data were gathered by the investigator and the interviews took place in
the homes of the key informants. Kiswahili was the language used and when necessary,
an interpreter was used. There was no fixed duration but the interviews took between
60-90 minutes. Only one person was interviewed at a session. The discussions were
recorded on tape and notes taken when necessary (Mwenesi, 1993).
The data were transcribed and indexed under several categories determined after sorting
the pilot study data. These were: priority health problems in the areas, diseases perceived
to be prevalent in the areas, their aetiology, symptomatology, treatment, prevention,
perceived causes of mortality in children, reasons for use of different health services, and
their ideas on the solutions for the perceived health problems.
Non-participatory observation
Structured and unstructured observations were made parallel to the survey and
ethnographic interviews. The mothers’ questionnaire included questions that required the
field worker to observe indicators for socio-economic status assessment, among others.
Observations on domestic, social and occupational activities as well as environmental
factors that expose people to the problems of malaria, were made by the investigator.
These observations sought to supplement, validate and verify data from the other
approaches. For instance, data were gathered from chance observations of certain
activities and factors that had not been addressed by either the surveys or in-depth
interviews. These included house types, sanitation and cultural activities that enhance
man-mosquito contact (Mwenesi, 1993).
Fieldwork supervision and data analysis
The fieldwork was conducted between April 1990 and December 1990, and was finalised
between December 1991 and January 1992. Field workers were briefed every day on
what was required of them for each phase. Personal checks were made at each point of
the fieldwork to ensure that field workers understood and followed instructions, and their
12
work was verified by re-interviewing five percent of households already interviewed by
each of them. Recall checks on reported refusals and on reported unsuccessful contacts
were also routinely made. No significant interviewer/interviewee problems were
encountered.
Editing and preparing data for analysis was done simultaneously with data collection.
Data processing facilities and senior scientists who were valuable resource persons "on
site” at the KEMRI unit Kilifi enabled this. Each questionnaire was checked for
consistency errors and completeness at the earliest and when necessary re-visits were
done. After coding, data was entered twice by independent data entry clerks using the
Dbase IV programme and was cleaned and analysed using the EPI-INFO programme.
Ethnomedical data were transcribed according to relevant categories. The information
was used as the background for the analysis of the quantitative data and the basis for
understanding the study population’s thinking on matters of health and illness and
specifically their definition of malaria.
13
RESULTS
Overall, 894/1,120 mothers were identified as eligible respondents were interviewed. Of
these, 883 mothers were successfully interviewed as 11 questionnaires were incomplete;
a response rate of 99%. Table 1 summarises findings on socio-demographic
characteristics of the mothers included in the mothers’ survey.
Table 1: Socio-demographic characteristics of mothers (N=883)
VARIABLE
ATTRIBUTE
FREQUENCY (%)
Residency
Kilifi town
Mtondia
Sokoke
352 (40%)
224 (25%)
307 (35%)
Ethnicity
Mijikenda
Luo
Other
608 (69%)
152 (17%)
123 (14%)
Age group
10-34 = Young
35-59 = Middle age
60+ = Old
633 (76%)
245 (28%)
5 (6%)
Religion
Christian
Traditional
Muslim
353 (40%)
280 (32%)
250 (28%)
Marital status
Married
Partnered
Un-partnered
716 (81%)
13 (4%)
154 (15%)
Education
None
Primary
Secondary/Tertiary
Other
445 (50%)
304 (34%)
97 (12%)
37 (4%)
None/Housewife
Service/Pros
Labourers/Farmers
’ercentages for attributes with > 1 answer are > 100.
776 (88%)
89 (10%)
18 (2%)
Occupation
Definition and management of childhood malaria in Kilifi
The Mijikenda and the Luo peoples in Kilifi district explain illness as part of misfortunes
of being human. Like any calamities, illnesses can happen naturally "when so God
wishes" or when there is disharmony in the elements which include, wind, heat and food.
They can also happen as a result of witchcraft or sorcery and the "evil eye", or one can
bring illness on oneself or one’s progeny by neglect of taboos and other ancestral or
religious obligations.
Some of the illnesses attributed to the (cold or ‘evil’) wind are all febrile illnesses
including malaria, homa, colds, aches and pains, convulsions and splenomegaly. Heat was
reportedly responsible for tsango or mshipa which could refer to conditions like
14
strangulated hernia, hydrocele and non-specific lower abdominal pains in men and
women. Food was implicated for diabetes and hydrocele. All the mentioned illnesses are
seen to be inevitable and unavoidable because they result from natural processes which
are beyond mans’ control. Illnesses that were attributed to negligence of taboos and
obligations included a childhood folk illness referred to as chirwa or kirwa in Mijikenda
and chira in Dholuo. Clinically, it presents with symptoms of marasmus or severe protein
energy malnutrition (PEM).
The childhood illnesses frequently mentioned either spontaneously or on prompting as
presenting with high fever in order of magnitude were: homa (90%), malaria (83%),
convulsions (56%), splenomegaly (24%) and anaemia (20%). Mention of fever was
associated with higher education levels and the younger age groups. It is important to
note that 150 (17%) mothers did not mention malaria as a childhood illness.
The Mijikenda had no vernacular name for malaria. They defined the illness as one of
the many manifestations of a condition known locally as homa. This concept refers to a
febrile condition that has symptoms ranging from mild to severe headache, influenza,
fever and general malaise. Four different types of homa were identified: non-specific
homa (feeling unwell), homa ya mafua (flu or common cold), homa ya matumbo (typhoid
fever) and homa ya ripeho (a febrile condition accompanied by chills). The concept of
homa was also used to refer to fever especially in the context of "fever rising" or joto
mwilini literally "hot body". This literal meaning of fever was used to establish if there
was a vernacular description or understanding of malaria.
The Luo referred to malaria as ’meleria’ and perceived it as a disease entity on its own,
distinct from other febrile conditions such as athung’a (flu or common cold). When asked
if they knew of a disease called malaria, most Luo informants answered in the
affirmative, while most Mijikenda key informants answered in the negative. The probable
explanation is that malaria control efforts have been going on almost continuously for
the last 50 years in Nyanza Province (GK/UNICEF 1989), the home area of the Luo,
thus raising their awareness.
In most instances the Mijikenda informants, talked about malaria in terms of what they
"hear" about it. Answers to questions about malaria were punctuated by "I don’t know"
or "I’m not sure about the answer" but "they say" or "you people say...”. The "they" and
"you" referred to health workers or the media. The disease was perceived to be among
the five top priority health problems in the area, but not the major one. Several
respondents alluded to an advertisement over the radio for one AM/OTC drug which
says: "malaria kills. Each year it kills millions of people". This advertisement does not
say, however, where exactly the "millions" are "killed". Thus most of the informants
thought that malaria must be a general problem in Kenya.
15
Mothers’ beliefs and behaviour in relation to childhood malaria
A summary of principle results pertaining to mothers’ knowledge, beliefs and
management of malaria are presented in table 2.
Table 2: Mijikenda and Luo mothers’ knowledge, beliefs and management of childhood
malaria (n = 733)
VARIABLES
QUANTITATIVE
RESULTS
SIGNIFICANT
ASSOCIATIONS
(P< =0.05)1
QUALITATIVE
COMMENTS
Aetiology
56% ‘mosquitoes’
35% ‘D.K’
9% ‘God/weather
changes’
Mention of mosquitoes
was associated with
residency in Kilifi
Town, age <35 years
and higher
education levels.
Weather changes,
especially sudden
changes from hot or
cold and being
rained on were
implicated.
Symptoms and age
specificity
75% ‘fever’
53% ‘vomiting’
42% ‘chills’
42% ‘general malaise’
N.S
Headache was
mentioned in
relation to adults.
Age-specificity was
not recognised.
Recognition of
complications
associated with
malaria
16% ‘convulsions’
4% ‘anaemia’
2% ‘splenomegaly’
N.S
Complications
viewed as illnesses in
their own right and
not associated with
malaria.
Communicability
and transmission
47% ‘communicable’:
of these:
10% ‘correct’
22% ‘droplets and sharing
bedding and utensils’
36% ‘cold entering the
bod/
and 32% ‘D.K’
Knowledge that
malaria is
communicable was
associated with
residency in Kilifi
Town and ’’higher"
education levels.
Other febrile
illnesses such as flu
may have given
credence to the
contact explanatory
model.
Preventability and
prevention
53% ‘preventable’:
of these:
23% ‘correct’
48% ‘keep warm/dry5
25% ‘D.K’
2% ‘prophylaxis’
Knowledge of
preventability and
prevention modes was
associated with
residence in Kilifi
Town, age <35 years
and "higher" education
levels.
As for other febrile
illnesses, no specific
modes of prevention
were mentioned for
malaria.
16
Cont- VARIABLES
QUANTITATIVE
RESULTS
SIGNIFICANT
ASSOCIATIONS
QUALITATIVE
COMMENTS
Anti-mosquito
measures used
65% ‘currently used’ Of
these:
53% ‘insecticides’
15% ‘bed-nets’
11% ‘local repellents’
21% ‘cover-up when
sleeping and close
windows earl/
Use of anti-mosquito
measures generally
and bed nets
specifically was
associated with
residency in Kilifi
Town and "higher"
education levels.
Anti-mosquito
measures are not
used for the purpose
of preventing malaria
but to decrease the
nuisance of bites.
Management of
childhood malaria
and malaria-like
illness within
households.
16% reported child
malaria currently or in
last 2 weeks. Of these:
59% gave OTC drugs. Of
these: 49% ‘AM drugs’
and 51% ‘other
medication’
Reporting malaria in a
child was associated
with "higher" education
levels, so was selfmedication with other
drugs especially
antibiotics.
Self-medication with
OTC analgesics and
AM drugs is
rampant. A local
febrifuge for all
febrile conditions is
available and emetics
and purgatives are
used.
Recognition of OTC
drugs
70% mentioned
OTC/AM drugs. Of
these: 56% could not
identify them on the drug
collage provided.
Recognition of OTC
drugs was strongly
associated with
"higher" education
levels.
Mothers rely on
ROPs when
purchasing OTC
drugs that they know
by brand names only.
Footnote: The independent variables tested for associations included residency, ethnicity, age,
religion, marital status and education. NA = Testing for association not applicable and N.S = no
significant association found. D.K = Don’t Know. OTC = over-the-counter. AM = Antimalarial.
OTC/AM = over-the-counter antimalarial. The sum of certain percentages are >100 because some
questions had > 1 answer.
Management of childhood malaria at home
Beliefs on the aetiology of illnesses invariably dictate the type of therapy and thus
healers to be consulted. The reported first-line treatment for childhood malaria was
OTC drugs from ROs. One hundred and seven (91%) out of 118 mothers who diagnosed
their child as having malaria currently or in the two weeks prior to the survey did
something about it. Twenty-nine percent reported to have given antimalarial drugs, 30%
gave antipyretics and other medications including antibiotics, 25% said they took child
to health facility, while 9% gave no treatment and 7% gave a home remedy. Residency,
age and education were associated with correct knowledge of malaria.
17
The following explanation from a mother who was an informant illustrates a typical
reaction to childhood malaria or any other febrile illness.
If I suspect my child is unwell; for example the child may be restless, have no
appetite and not happy, I observe him more closely. I wait and see for a day or
two if the restlessness will go away. If the child continues to be restless and cries
or has hot body, I look for any children’s tablets or medicine in the house and
give the child. If there is none and I have some cash then I get some aspirin or
cafenolR and give the child. The child will get better if the illness is malaria or flu
[homa la mafua]. If the child does not get better or develops cough, I have to find
a way of getting to the hospital. If on the other hand the child develops nyago, I
do what I can in the house and then rush the child to the traditional healer
(Mwenesi 1993: 157).
A typical and casual reaction where the mother portrays experiential wisdom in child
rearing (Mwenesi, 1993). She was observant and discerning in monitoring her child’s
illnesses, she is aware of and knows by name which OTC drugs to use and also portrays
some knowledge of differential diagnosis, based on the prognosis of each symptom or
initial diagnosis. The selected therapy and the lapse between onset of illness and
remedial action seems to depend on the perceived aetiology, prognosis of the illness and
its perceived seriousness. Note the informant’s casual observance on malaria, while
"...nyago..." (convulsions) is treated as an emergency. Note that the mother does not
allude to consulting anyone about the child’s condition.
The traditional home therapy commonly used for malaria and malaria-like illness in
Kilifi is a herbal febrifuge prepared from the "neem" tree (azaderachta indica), locally
referred to as muarobaini [literally, ‘the herb that can cure forty diseases’], mkilifi and
mzerekta. However, only one percent of mothers reported to use it for childhood
malaria.
Malaria transmission as shown on Table 2 was not understood. Although 15% of
households reportedly used bed-nets, people were generally indifferent towards
mosquitoes. Nevertheless, all the respondents conceded that mosquitoes were a major
nuisance in the area. For them, malaria control was not related to mosquito control but
the availability of adequate curative services. In general people were dissatisfied with the
health system, especially in relation to drug availability and "user fees" after 27 years of
free services.
18
Convulsions (fits)
A total of 498/883 (56%) mothers mentioned convulsions as a childhood illness that
presents with fever. Table 3 summarises their knowledge, beliefs and management of
convulsions at the household level.
Table 3: A summary of mothers’ knowledge, beliefs and management of convulsions in
children (n=498)
VARIABLES
QUANTITATIVE
RESULTS
SIGNIFICANT
ASSOCIATIONS
(P< =0.05)
QUALITATIVE
COMMENTS
Aetiology,
symptoms and
susceptibility
47% ‘D.K.’
28% ‘fever’
15% ‘malaria’
8% ‘spirits’
2% ‘worms in head’
Symptoms and age
specificity well
recognised.
Knowledge that
malaria could cause
convulsions
associated with age
<35 years and higher
education levels.
An inevitable
childhood illness.
Almost a ‘folk’
illness for the
Mijikenda people.
Transmission,
communicability
and prevention
20% ‘preventable’. Of
these: 43%
‘avoidance of
vectors’; 19%
‘charms/amulets’
38% ‘D.K.’
Knowledge of correct
prevention associated
with age >35 and
higher education
levels.
Not understood but
explained as
corollaries of
aetiological beliefs.
Management of
cases at the
household level
5/12 ‘OTC drugs’
5/12 ‘traditional
therapy’
2/12 ‘no treatment’
N.S
Management starts
at home, then
traditional therapy is
sought. Allopathic
therapy is
contraindicated and
no specific OTC
drugs for
convulsions.
Footnote: Independent variables tested for associations included residency, ethnicity, age,
religion, marital status and education. NA = Testing for association not applicable and N.S
= no significant association found. D.K= Don’t know.
Both the Mijikenda and the Luo have local names for childhood febrile convulsions.
They recognise them as a serious childhood illness. The Mijikenda refer to the condition
variously as nyago, dege, nyuni and nyama wa dzulu. They also call it ukongo wa kitoto,
literally meaning ‘the childhood illness’. The Luo call it oriere. The Mijikenda attributed
the cause of convulsions to a figurative ‘animal or bird’ which gets into a child by
frightening the victim thus inducing the fits. Luo informants attributed the cause of
convulsions to intestinal worms somehow finding their way into the head. No explanation
was elicited as to how the worms got into the head.
19
Most of informants could describe convulsions in a child. Twenty-eight percent (N=498)
of mothers gave fever as a symptom of convulsions. However, they could not clearly
discern the link between fever and convulsions. Fits were perceived to be neither
communicable nor preventable. The Mijikenda first-line treatment for fits was reported
to be sponging with the mother’s or any other close female relative’s urine. The child is
then taken to a traditional healer, always a man. Treatment consists of herbal
preparations for drinking and bathing. Smoke from different herbs or elephant dung
(kufukiza) and charms/amulets called hirizi or vuje may be used. They are supposed to
decrease the virulence of the nyama and not necessarily to prevent further attacks. Strict
rules are attached to the management of convulsions both at the healers and at the
household (Mwenesi 1993). The different options for the treatment of febrile convulsions
must be exhausted before a child is taken to a HF, where it is believed the chances of
dying as a result of mishandling are high. Especially dreaded is the prospect of a child
being injected, which they believe could kill the child instantaneously (Mwenesi 1993).
Luo treatment of fits also commences at the household. Certain roots are dried and
crushed and then the child is made to sniff the powder to induce sneezing, to get the
worms out of the head. Most mothers know these roots and they treat their children
themselves. A visit to a HF on account of a child fitting is a last resort. One thing to
note however is that many of the Luo informants thought that febrile convulsions were
more common on the coast than in their home area in Western Kenya.
Anaemia
Anaemia was mentioned by only 174 (20%) of all the mothers interviewed. It was not
perceived as a priority disease. Residency in Kilifi town and "higher” education levels
were associated with the answer. However, the condition was well known in the
communities. The local names for the condition are: safura which literally means ‘the
swollen or puffy’ in Mijikenda and rembe orumo which literally means ‘blood is finished’
in Dholuo. The local name in Dholuo suggests that the Luo did not recognise anaemia
in their cultural disease profile because they have literally transliterated the medical
explanation to lay people that one has anaemia when they have insufficient blood in their
bodies.
Although everyone could get anaemia, children under five years were perceived to be
more susceptible because they tend to eat soil and spoilt foods which are believed to be
the causes of the condition. Transmission and communicability was not understood but
mothers believed that anaemia in children could be prevented, unless it was caused by
sorcery. Prevention measures mentioned were corollaries of perceived causes. Anaemia
was not perceived to be a life threatening illness. Only seven mothers reported anaemia
in their children, currently or in the past two weeks prior to the survey and none of them
had presented her child at a HF or had given a home remedy. Anti-helminths were
mentioned by 8% of the mothers, no other OTC drugs were indicated for the treatment
of anaemia.
20
Splenomegaly
A total of 216/883 (24%) mothers mentioned splenomegaly as a childhood illness that
presents with fever. The response was associated with the younger age-group and
schooling. A large percentage of these mothers (72%) included fever as a major symptom
of splenomegaly. However, most of the mothers (71%) could not describe an enlarged
spleen. Both communities recognised splenomegaly and had local names for it: luwengu
in Mijikenda and ima in Dholuo. Ingested soil could solidify in the stomach causing
splenomegaly. Worms could also form a mass which presents as an enlarged spleen.
The first-line treatment for this condition is OTC anti-helminths and laxatives. Certain
herbs with laxative properties referred to as msahala are also used to induce diarrhoea
which is perceived to be a wash out for the worms and the soil. Traditional treatment
is sought if the condition persists. The initial diagnosis changes from ‘soil mass’ and
‘worms mass’ to ‘bad blood mass’. Bad blood refers to the swelling which is perceived to
be coagulated blood which has to be sucked out. The usual treatment is scarification of
the swelling followed by ‘letting’ out the blood by suction, followed by rubbing poultices
into the cuts (Mwenesi, 1993). A HF is a last resort. Three mothers who reported
splenomegaly in their children in the two weeks preceding the survey reported to have
given their children antipyretics and one had taken the child to a HF.
Mothers’ health seeking behaviour (HSB) for malaria
Mothers’ use of health facilities
Sixty-nine mothers were re-interviewed at health facilities. The socio-demographic profile
of these mothers was not different from mothers who were not re-interviewed. Fifty-three
of these, whose children were ill were interviewed in detail about their child’s current
illness and the whole process of care seeking. Cough was the illness most likely to
present at a health facility (50%) and only one mother diagnosed her child’s illness as
malaria prior to coming to the health facility. The average number of days between onset
of illness and health facility visit was three days. Some of the reasons given for the three
days lapse (62%) in order of magnitude were: perception that illness was mild, partner
being absent, other important matters to attend to and lack of someone to mind the ill
child’s siblings. These were categorised as "predisposing" factors. The remaining 38%
gave their reasons as lack of money for transport, having given the child OTC drugs and
having taken child to a traditional healer and were categorised as "enabling" factors. No
"health systems" factors were given. At the facilities, 55% of mothers who received
antimalarial drugs did not follow instructions on how to dispense the medication and
none asked for clarification for fear of the health workers.
Mother’s use of retail outlets
An inventory of medications available in 64/200 retail outlets in the study areas yielded
150 types of drugs: 19 (13%) types of antimalarial drugs, 27 (18%) types of analgesics
and antipyretics, the rest were gastrointestinal (23%) and respiratory drugs (13%). The
remaining 33% were a host of preparations including antibiotics (3%). Results from the
drugs in homes survey showed that 67% of households had purchased drugs in the past
21
f
ANO
ntation
)
two days before the survey and 44% had drugs at the time of the survey. Of these, 41%
had antimalarial drugs in varying quantities.
Eighteen mothers were re-interviewed at retail outlets. The type of retail outlet used
depended on distance and availability of required medication. Fourteen out of 18
mothers purchased drugs for homa, with symptoms ranging from general febrile
conditions, anorexia, crying and restlessness to cough. Most mothers purchased the drugs
within one day of noticing the symptoms or behaviour changes and thus acted promptly
to administer presumptive treatment. The most popular choice of medications was
combinations of analgesics (10/18 mothers), followed by analgesics together with
antimalarial drugs (4/14), while the remaining four purchased other drugs. Mothers
based the combination of drugs on what each was supposed to treat and on their brand
names. However, although 70% (N = 883) mentioned OTC drugs that they thought could
cure malaria, 56% of them could not identify the medications from a drug collage
provided.
The treatment seeking pattern that emerged from this data was that mothers took
remedial action for their child’s illness as soon as possible. They first gave any
medication available within the household. They repeated the treatment with medication
from the ROs, which was repeated if necessary. They only went to HFs when, in
consultation with a significant other, the mother thought it was necessary.
Retail outlet proprietors’ knowledge of malaria and its medications
Almost all (91%) ROPs thought malaria was a serious illness but they did not recognise
age specificity. Seventeen percent (N=58) of these thought malaria was not preventable
and all those (83%) who thought it was preventable mentioned prophylaxis as the best
preventive measure, in keeping with their vocation of selling drugs. They routinely
offered advice to mothers who purchased drugs for children. The best drugs for the
treatment of malaria were reported to be chloroquine-based AM drugs, of which one
brand was so popular that one ROP remarked that "it sells like aspirin". However, 24
(38%) of them did not know the correct dosages of the drugs, and only 2 (3%) of them
indicated that their customers bought correct dosages of AM drugs.
Antimalarial drugs available in retail outlets and homes
Of 19 types of AM drugs recorded from ROs, 14 were chloroquine phosphate-based
drugs, all being sold without control under different brand names. Most had "catchy"
names and were vigorously advertised in all media including large posters strategically
billed on RO walls and often doors. The remainder were aminoquinolines including
amodiaquine (CamoquineR) (recommended for chemotherapy only) and proguanil
hydrochloride (PaludrineR) (recommended for prophylaxis only), and second-line drugs,
intended for treatment failures due to parasite drug resistance to ingredients such as
quinine, examples of these are: sulphadoxine + pyrimethamine (FansidarR) and
sulfalene+pyrimethamine (MetakelfinR). The second-line AM drugs should be used
under the direction of a qualified person but they were readily available over the
counter.
22
One hundred and seventy (44%) out of 390 households had drugs at the time of the
interview. Of these, 70/170 (41%) had AM drugs. Four of the households had secondline drugs and 66 had different brand names of chloroquine phosphate-based AM drugs.
None of the households had a full course of any of the AM drugs in their possession.
The respondents reported that they buy AM drugs only when in need and what was
found in the house was a recent purchase. Twenty-two percent had purchased the drugs
in the last 48 hours before the survey. The majority of households that had any type of
drugs were in Kilifi town. One hundred and twenty households had no drugs at the time
of the visit but they reported to have purchased some type of drug in the last 48 hours.
Of these, 90 (75%) had evidence of purchase in form of drug wrappings. This implies
that a total of 260 (67%), households had purchased drugs between one and two days
prior to the survey. The analysis did not include medications recorded as having been
obtained from HFs. The data indicates a high level of self-medication in the community.
Qualitative information indicated that low utilisation of HFs and high use of ROs was
directly a result of the introduction of user fees at HFs.
DECISION-MAKING DYNAMICS IN HEALTH SEEKING BEHAVIOUR
The Mijikenda and Luo people have a well defined social structure in which everyone
knows their place. It was surprising that almost all mothers, 880/883, regardless of
marital status, age and education level reportedly sought advice before taking an ill child
to a health facility, while another 386 (42%) sought advice for retail outlet use,
(responses to a retrospective hypothetical situation). Results obtained from mothers at
both health facilities and retail outlets during a current illness episode on consulting
behaviour were in concordance with the results from a hypothetical situation. The
consultations involved family members. Husbands were consulted more frequently (63%).
Fifty percent of all mothers interviewed consulted because it was expected of them and
50% consulted only when they perceived an illness to be serious or they did not
understand its nature.
Other male members of households and in their absence senior females were also
consulted. The reasons advanced were that illnesses, whatever their nature, were a
matter of life and death and a second opinion on the likely cause of a child’s illness was
necessary. Further, having to seek advice for a child’s illness not only acts as a check for
‘bad behaviour’, especially on the part of women among the Mijikenda, and men among
the Luo, but, according to the informants, hastens the child’s chances of being treated
quickly and effectively without wasting time and sometimes money at HFs, especially for
illnesses involving sexual transgressions. Financial considerations were not cited as
reasons for consulting a significant other.
DISCUSSION
The discussion of the results includes qualitative and quantitative data. The findings are
discussed and their implications for malaria control spelled out.
23
Table 4 summaries the major findings of the study.
Table 4: Summary of m^jor findings
MAIN OBJECTIVES
SUB-STUDY
MAJOR FINDING(S)
1. Determine how the
communities under study
define and perceive childhood
malaria and how they translate
this perception into action.
- Mothers’ survey.
- Ethnomedical survey.
- HF and RO users’
surveys
- ROPs survey.
1. They have limited knowledge of
childhood malaria as a biomedical
pathology and react to it from a
biocultural framework.
2. The image of malaria presented by
health workers and the media to
populations living in malaria endemic
areas has been based on presentations
of the illness in non-immune or semiimmune populations and may be the
major barrier in malaria control in this
areas.
3. The time-tested simplistic
mosquito-malaria link presented to
people by malaria workers of all
cadres: no mosquitoes no malaria may
be another obstacle in malaria control.
2. Determine whether certain
complications of malaria in
children are recognised and
their management.
- Mothers’ survey.
- Ethnomedical survey.
1. Mothers in Kilifi do not recognise
convulsions, anaemia and
splenomegaly as complications of
malaria in children, but as illness
entities in themselves.
2. Antimalarials are not thought to be
appropriate for their management and
are therefore withheld or withdrawn
from children suffering from these
conditions.
3. Assess decision-making
dynamics and the significance
of family and other social
networks in the health seeking
process.
- Mothers’ survey.
- Ethnomedical survey.
- HF and RO users’
surveys.
1. Mothers rarely make decisions on
health matters related to their
children.
The male head of a household almost
exclusively decides what therapy is to
be used when a member of the
household is taken ill.
4. Determine the extent and
reasons for use of proprietary
drugs for malaria in children in
the context of multiple
therapeutic systems.
- Mothers’ survey.
- Ethnomedical survey.
- RO users’, ROPs and
drugs in homes surveys.
1. Mothers are promptly treating their
children for malaria like illness.
2. ROs are extensively used as sources
of health care; not as alternatives to
HFs but as the first tier of care for
‘mundane’ illnesses.
3. The range of types and
formulations of OTC drugs including
antimalarials is bewildering to both
users and proprietors. They know
little or nothing about them and use
or sell them inappropriately.
24
A biocultural definition of malaria
Most mothers in Kilifi, especially the Mijikenda, have limited knowledge of childhood
malaria as a biomedical disease, its cause, prevention and treatment. They define it as
a mild self-limiting illness likely to cause death. Their ethnoetiological explanations
leaned more on thermal and climatic theories, where sudden body temperature changes
from hot to cold or vice-versa, or sudden weather changes, compromised the body’s
ability to resist febrile illnesses like malaria. Ramakrishna and Brieger (1987) in Nigeria
mention that mothers also believed that malaria was a temporary illness caused by
excessive heat. The Kilifi mothers’ explanatory model did not extend to other imbalances
such as hot-cold foods and other humoral features. The thermal definition as used by
Mayall (1986) who reported similar results from mothers in England when studying
common colds is therefore preferred. Jackson (1985) in Liberia and Bledsoe and
Goubauld (1985) in Sierra Leone reported similar findings.
These ethnoetiological factors associated with malaria are natural processes, ‘acts of
God’, which individuals perceive to be beyond their control. The only precautionary
measures against malaria in Kilifi are common sense practices that are a corollary of the
ethnoetiologies such as avoiding extreme temperatures, excessive wetness and weather
changes. Except among a few mothers with some schooling who use sporadic
antimosquito measures in Kilifi town there are no conscious efforts made to avoid
mosquitoes as a malaria prevention effort; since they were not associated with malaria.
The only other studies known to the investigator that have systematically investigated
mothers’ definition of childhood malaria are by Jackson (1985) in Liberia and Glik,
Parker, Mulingande et al. 1987, Glik, Gordon, Warden, et al. 1988). In contrast with
findings in this study, Jackson (1985) reported that mothers in Liberia defined malaria
as a serious illness which they were apprehensive about and therefore regularly gave
their children some local prophylactic teas. In both studies however, malaria was cited
as the fifth major health problem in Liberia and on the Kenyan Coast. The contrast in
definition and perception may be explained by a probable difference in the amount of
exposure the communities have had to information about malaria, since ethnoetiologies
for malaria are similar in both countries. The difficulty of Mijikenda mothers to
recognise childhood malaria is largely a result of the explanatory model for febrile
illnesses which is imbued in their culture. The generic term homa, which covers a wide
range of febrile conditions may denote malaria, but the fact that mothers differentiate
between typhoid and influenza would make that assumption incorrect. Luo mothers on
the other hand differentiate malaria from other febrile illnesses because they do not
posses a generic term for febrile illnesses in their taxonomy of illnesses. Spencer, Kseje,
Sempebwa et al. (1987) came to the same conclusion while working in Nyanza, the home
are of the Luo. Most malaria control programmes in Kenya for the last 50 years have
been carried out in the area.
Kilifi mothers had fairly good knowledge of symptoms that present in mild malaria. This
knowledge may have been acquired through contact with health workers where exhibited
symptoms may be diagnosed as malaria, from media messages which describe symptoms
for conditions that their brand medications can treat or from learning at school, from
friends and relations. Similar results were reported by Jackson (1985), Spencer et al.
(1987), Ejizie, Ezedinachi, Usanga et al. (1991) and Rooth and Bjorkman (1992).
25
The findings illustrate that Kilifi mothers would clearly not react to childhood malaria
from a biomedical premise, but from a biocultural ’’explanatory" position. According to
Kleinman (1980) and Young (1987), "explanatory models" are generated by individuals
within their cultural background. The models contain knowledge about the kinds of
illnesses, what to expect about the onset of the symptoms, courses of the illness episodes
and the appropriate treatment. The explanatory models illuminate our understanding as
to why people react in certain ways to a given illness.
These findings imply that health information for malaria should include these divergent
explanatory models and also appreciate the context within which they are formulated.
That may be the only way to make them acceptable to the people. The most important
point emanating from the findings, however, is the indication that health promoters of
all persuasions have been using an incorrect premise in health education for malaria. It
is argued here that mothers in Kilifi may not have biomedical knowledge or even a local
name for malaria, but they have the correct perception of malaria, as a chronic disease,
gained through empirical observation and experience. It is not surprising that mothers
and key informants described malaria as the disease they "hear about" from "you people"
(health workers and the media). The disease that is presented to them is consistent with
the presentation observed in non-immune and semi-immune patients: acute, serious and
almost always fatal if not promptly treated at a HF. This illness is alien to them because
being in an endemic area they identify with a self-limiting illness, that is chronic and
debilitating. They do not experience the malaria that is portrayed to them. Any health
information strategy in an area like Kilifi that fails to appreciate peoples’ perception and
definition of the illness is bound to fail.
The explanatory model for malaria presents to Kilifi mothers an illness that is not serious
and which one need not be apprehensive about because they may have no control over
it, since it is firmly placed in the realms of natural processes - fate and God. Blaxter
(1990) argues that if ill-health is seen largely as "self-inflicted" then education, persuasion
and an emphasis on self-responsibility will work as policy. But, if it is perceived to be
principally outside the individual’s control, then social policy issues are paramount. In
the case of malaria in Kilifi, the issues would relate to availability of medication and the
persistence on raising awareness, especially the recognition of a malaria episode and the
link between malaria and mosquitoes.
"Cultural blindness"? - The mosquito - malaria link
Fonaroff, (1968) argued that peoples’ cultures are to blame for the perceptual difficulty
of linking mosquitoes to malaria by most populations living in malarious areas. While
working in Trinidad, he noted that East Indian ethnic groups who presumably knew
about mosquitoes and malaria, preferred to live near marshy land where they could
cultivate their highly valued food - rice. These cultural preferences reportedly put these
groups at a greater risk of malaria infection relative to other ethnic groups on the
Islands.
Three decades later, one wonders about whose "culture is blind", that of the people in
malarious areas or that of malaria workers? This study and most other social science
studies (for example: Abdullah 1985; Ongore et al. 1989; Jackson 1985; Hongvivatana,
26
Leerapan and Chaiteeranuwatsiri 1985; Silva 1991; Agyepong 1992; Lipowsky, Kroeger
and Vasquez 1992) attest to the fact that mosquitoes are not perceived to be linked to
malaria, especially by people who have had little or no schooling. The question is, why
has this perceptual difficulty persisted through almost the whole period that malaria has
been an issue?
It could be that the simplistic messages that were given to the people about how to
prevent malaria, which may have seemed reasonable then, have contributed to the
problem of the malaria-mosquito link. People in malarious areas were given the mistaken
impression that eradication of malaria would be brought about by the eradication of
mosquitoes. The belief was based on earlier ideas that: "...the disappearance of malaria
related directly to the disappearance of anophelines" (Fonaroff, 1968: 538).
People were (and are still) taught that malaria control would be achieved through
environmental manipulation, especially by clearing bushes and vegetation and draining
of stagnant waters to destroy mosquito breeding sites. The fact that there can be
mosquitoes without malaria (‘anophelism without malaria’) is not made clear to them.
The messages were simple for simple societies, but they were also ambiguous. It may be
that the malariologists of the 1920s did not know as much as is now known about the
bionomics of different types of mosquitoes and their breeding habits. However, there is
evidence to show that as early as 1940s, some of the procedures advocated such as
clearing vegetation, were found to be useless (Hackett, Russell, Scharff et al. 1938;
Ribbands 1946).
Empirically it is difficult for people to visualise any method that would have an impact
on mosquito densities and thus they perceive most ‘simple’ measures against mosquitoes
as a waste of time. While working in Kisumu, Western Kenya, Millman (1967) found that
villagers refused to clear vegetation for malaria control purposes, because they needed
fodder for their animals. Abdullah (1985) reports that 87% of her respondents also in
Kisumu, claimed to clear vegetation as a malaria control measure. However when asked
whether the measure was effective, only five percent of them answered in the affirmative.
Malaria workers of all cadres must clear the confusion through properly formulated
health education messages if the "cultural blindness" is to be rectified.
Does malaria have any complications?
The explanatory model for malaria in Kilifi does not provide for the illness becoming
severe and therefore causing other illnesses. As such, convulsions, anaemia and
splenomegaly were not perceived as complications of malaria but as specific illnesses
with different ethnoetiologies and in the case of convulsions, sophisticated management
procedures. The three conditions did not even fall into the same explanatory models.
Convulsions are perceived as a serious childhood condition, non-communicable and
unavoidable. The condition must have been encountered frequently in the area because
it has acquired a ‘folk illness’ status known by different names but commonly called
nyago. It has a supernatural ethnoetiology where spirits are implicated. Nevertheless,
convulsions are treated as emergencies. Similar results have been reported by Boerma,
(1989) for South Coast Mijikenda and Coreil (1983) for Haitians. In contrast, Jackson
27
(1985) reports that Liberian mothers linked convulsions to malaria, but it was found that
convulsions as a reported symptom overpredicted evidence of malaria and may have
reflected the high prevalence of neonatal tetanus in the study area.
Biocultural definitions of anaemia were not as entrenched or as elaborate as beliefs on
convulsions. Among the Mijikenda, anaemia could be caused by natural processes such
as eating soil (dirt) and spoilt food, which results in helminthiasis or by sorcery. Heredity,
where a mother passes on the pica syndrome to her child in utero was also implicated.
The Luo on the other hand had no biocultural explanatory model about the condition
but explained anaemia as lack of blood - the lay explanation of low haemoglobin levels.
It was not viewed as a life-threatening illness and could be prevented by avoiding the
implicated aetiological factors or protection against sorcery. The explanatory model for
splenomegaly, as for anaemia, was not elaborate. The condition was attributed to natural
processes which included eating soil and helminthiasis. It could also result from unknown
aetiologies which manifest as "bad blood” in the ’painful swelling’. It was not viewed as
a life-threatening illness.
Mothers did not mention cerebral malaria, which presents as delirium or coma. This
confirms the observation by Greenwood (1991) that in Kilifi, anaemia is responsible for
more malaria related deaths in children than cerebral malaria. Jackson (1985) reports
similar findings in Liberia. It has been reported that cerebral malaria and severe
anaemia are responsible for most deaths from malaria in tropical Africa (Greenwood
1991), but Rooth and Bjorkman (1992) concluded that perhaps cerebral malaria in
children is uncommon in malaria endemic areas.
It is important at this point to comment on the link between good prognosis of malaria
infection and bile in vomitus. Mothers believed that malaria was exacerbated by bile
excretions. If the bile was vomited, the patient was perceived to be recovering. This has
not been mentioned anywhere else in the social science literature on malaria or even in
the literature on the pathophysiology of malaria where not much on the bile factor was
found. Gastrointestinal dysfunction is discussed in relation to nausea and vomiting,
especially in hyperpyrexia (White and Ho, 1992) and only in cerebral malaria is the
colour of the vomitus described, especially where there is involvement of other organs
in the gut.
Morbidity and mortality from childhood malaria
The objective was not to measure the rates but to determine the extent to which
childhood morbidity and mortality was attributed to childhood malaria. Recall bias is
appreciated for this result. Of 733 mothers, only 16% reported malaria in their children
in the two weeks prior to the survey and only 1.3% reported to have ever lost a child due
to malaria. The figures for anaemia and splenomegaly were 4% (n=174) for morbidity
and no mortality, 1.3% (n=216) and no mortality, respectively. These conditions were
perceived as non life-threatening.
Findings on morbidity and mortality from convulsions may reflect the true situation
because the nature of the illness reduces recall bias significantly. Twelve (2.4%) out of
498 mothers reported convulsions in their children in the two weeks prior to the survey.
28
This would translate to about half of all the children having one attack of convulsions
annually; indicating substantial cases of severe illness in the community. Five percent of
the mothers reported to have ever lost a child due to convulsions. Convulsions, therefore,
were implicated for more deaths than all the other conditions put together, while malaria
was implicated for more morbidity than all the other conditions.
The findings on ethnoetiological explanatory models for complications associated with
malaria and the perception that these conditions are not life-threatening (except
convulsions) have several implications for malaria control. Firstly, any health intervention
that assumes their link with malaria but neglects these psychosocial factors would be
misplaced. Secondly, none of the conditions is revered enough to act as a proxy for
malaria during health intervention programmes.
Who makes decisions for childrens9 health care?
Health care especially for children begins at the household level. It is assumed that
mothers as their carers have the prerogative of making health care decisions for their
children. This assumption, which has directed past health interventions for malaria,
presupposes the existence of a situation of choices, with mothers and significant others
bidding for their preferences, within the context of equal say in households (Rogler,
1989). However, the Kilifi study suggests that this is not always the case. Decision-making
processes for health care in different types of households and settings depends generally
on the social structure and particularly on who "owns" the child: an important
consideration in patrilineal and matrilineal societies.
Clear social roles delineate who makes what decisions. Among the Mijikenda and Luo,
women may make decisions regarding what their family meal will comprise of but they
had no mandate to make decisions for their husband’s children or even themselves. This
applied to partnered or un-partnered mothers. The decisions for health care lay
exclusively with the male head of the household (especially the husband), who not only
decided the final diagnosis of an illness, but also the therapy to be used. Male in-laws
in the case of married and widowed women, fathers and brothers in the case of
unmarried women, also hold the mandate to make health decisions.
Similar findings were reported by Janzen (1978). Zairian mothers and even female heads
of households did not make decisions on health matters for either themselves or their
children. However, unlike in Kilifi, matriarchal kinsmen made health decisions in Zaire.
In Tanzania, McCauley, West and Lynch (1992) reported that although mothers were
responsible for the care of their ill children, they sought permission from their husbands
before utilising a health facility or traditional healer. In contrast, Glik et al. (1987) while
working in Rwanda, reported that mothers generally made decisions on child health
matters independently.
Generally in Kilifi, consultations for health care matters rarely went beyond the
immediate household. This was especially so for illnesses that had any moral
ethnoetiologies implicated or were suspected to result from sorcery. The element of
social control implied by ‘folk’ illnesses such as chirwa, which presents as PEM, and
which these communities believe can only be treated traditionally is a case in point.
29
Keeping matters within the household protects the household male head’s positions.
Finally, sorcery may be practised by kin against kin. Thus, matters of illness remained
within immediate households until other kin even in next door households had been
cleared by a muganga.
The implications of these findings on malaria control is that any interventions for malaria
that are targeted at women as child carers are bound to have little impact. This is best
summed up by McCauley et al. (1992) based upon their experience in Tanzania. They
report:
Although the mothers were responsible for both water use and the health
and cleanliness of their children,...the decision to change behaviour [to
control trachoma in children] had to be sanctioned by the husband in the
household and the community as a whole (: 817).
Mothers’ diagnostic abilities
The study findings demonstrated that mothers, regardless of their socio-economic and
socio-demographic characteristics make the first diagnosis of illness in their children by
defining and interpreting bodily and behavioural changes. It has been argued that health
and illness constitute a continuum but the point at which one becomes the other is often
vague (Lieban, 1992). The mechanisms by which Kilifi mothers conclude that certain
bodily and behavioural changes are illness are not explicitly laid out by the results of the
study. However, the findings suggest that a mothers’ decision to confer the sick-role to
a child and to subsequently do nothing or seek treatment is a function of various factors.
In the case of malaria, they include the mothers’ experiential knowledge (determined by
familiarity with the illness, as a result of an ecology conducive to malaria); based on its
definition (derived from their nosology and explanatory model of malaria and/or contact
with biomedical information on malaria), and the social structure which determines
household decision making dynamics. All these factors together constitute the functional
biocultural definition of malaria; what Sevilla-Casas (1992) refers to as subjective
perception and evaluation of risk. Other factors relate to the availability of other sources
of care and the adequacy of health facilities.
This study did not detect any belief-conflicts on the part of mothers or significant others
in choice of treatment for mild malaria. The type of therapy selected and the lapse
between onset of illness and remedial action taken seems to depend on the perceived
aetiology, prognosis of the illness and its perceived seriousness. As one mother reiterated
when referring to convulsions in children: ”...We know what we can treat ourselves and
what you can treat for us”. The explanatory model for malaria in Kilifi provided mothers
with precise courses of actions to take for malaria-like illness.
A conclusion by a mother who notices and decides something is the matter with her child
seems to culminate into two pathways (1) deals with behaviour or bodily changes
(‘symptoms’) she views as precursors to illness by observing for a few days or
administering self-medication, either OTC drugs or home therapies; and (2) deals with
30
the ‘symptoms’ as concomitant with an illness and may observe, administer selfmedication or, depending on the prognosis of the symptoms, seek care immediately,
based on the ethnoetiology of the disease.
All the factors discussed above are psychosocial processes which have not been seriously
investigated as literature on use and non-use of services shows (see Gilson, 1988). This
study demonstrated that they are important in health-seeking behaviour, but as Foster
(1976) concluded, economic and social costs are equally important.
Treatment selection
In Kilifi, mothers used ROs as the first tier in the hierarchy of care sources, for
symptoms perceived to be mild, mundane or those conditions that respond rapidly to
modern chemotherapy. OTC drugs were viewed to be efficacious and their use was very
high. The mothers were not only aware of, but also knew by name, some OTC drugs that
they believed "cured" childhood malaria. Fifty-nine percent reported to routinely treat
their children for malaria with OTC drugs from ROs. However there is cause for concern
because the study showed that 56% of these mothers could not visually identify the drugs
they mentioned. They depended on advice from ROPs, notwithstanding their low
knowledge of correct dosages. The most serious cause of concern however, is the
saturation of OTC drugs in Kilifi. Out of a total of 150 types of drugs and other
preparations, there were 19 types of antimalarial drugs, both first-line and second-line
drugs. Of these, 14 types were chloroquine based, all being sold under different brand
names and very vigorously promoted through the media. The findings of this study
reinforce results from previous studies on availability of excessive proprietary drugs in
the markets (Maitai et al. 1981; Raynal 1985; Van de Geest and Whyte 1989; Logan
1988; Kasilo, Nhachi and Mutangadura 1991); especially antimalarial drugs (Maitai et
al, 1981; Olatunde 1981; Van der Geest 1987; 1988; Foster 1991a; 1991b).
An important finding in this study however, was the fact that mothers were treating their
children promptly, within 24 hours of onset of illness; although the time lag between
onset of illness and taking a child to a health facility was 3 days. What is not happening
however, is the correct recognition of true episodes of malaria. Further, mothers not only
used inappropriate drugs for what they define as malaria, but they also gave incorrect
doses. Similar findings were reported by Dabis et al. (1989); Deming et a L (1989) and
Makubalo (1991). Whereas the prompt treatment at the household level may protect
children with mild malaria from advancing to severe forms, delay in seeking proper care
for malaria may have serious neurological consequences (Brewster, Kwiatkowski and
White (1990) and could lead to death in less than 24 hours (Greenwood, Greenwood,
Bradley et al. 1987; Taylor and Molyneaux 1988).
In areas where use of health facilities is high and medication can only be obtained from
health facilities, children are not treated promptly. Glik et al. (1987; 1988). The reasons
for this trend in Ivory Coast are not clear but in Rwanda it was due to a colonial statute
which had outlawed keeping drugs in households and therefore self-medication. Though
the laws were defunct at the time of the study, the habit was embedded in peoples’
treatment-seeking behaviour.
31
LIBRARY
f [ documentation
AND
1
)
I
\
UNIT
10
Retail outlet proprietors
ROPs cooperated in the research but they felt that they could not take it upon
themselves to give advice if not asked, or to question any one purchasing drugs from
their concerns, about what the drugs were for, or why they bought incorrect doses.
Firstly, it would not be in their interest to pry into peoples’ illnesses since it was not
expected of a ROP - they are not health workers. Secondly, OTC drugs in themselves
were not commodities with a good margin of profit, but their being available attracted
other business. ROPs were only fulfilling a need and thus it was, they felt, upon the
government to control the manufacture and pricing of OTC drugs, which are not on price
control. ROPs price them according to the procurement price from dealers. This
sentiment was also expressed by key informants who felt that the malaria problem in the
area could only be dealt with by the government "ensuring that HFs were well stocked
with antimalarial drugs" and "lowering and regulating the prices of proprietary drugs".
The Luo of Karateng in Kisumu district expressed the same sentiments, that only the
government could solve the malaria problem by "bringing the best medicines to hospitals
and shops" (Abdullah 1985: 83). This scenario presents a situation where ROPs may not
be keen to be involved in malaria control strategies.
The neem tree, Azaderachta indica, was commonly used as a home therapy for malaria,
but specifically for convulsions. Management of convulsions at hospitals kept mothers
away because it is not in concordance with the biocultural explanatory expectations of
how therapy should be administered. Injections and other invasive procedures are
perceived to be the cause of the childrens’ deaths and not the illness itself. This may
explain the observation that most paediatric deaths associated with malaria occur in
emergency wards of hospitals. In Zaire, Greenberg, Ntumbazondo, Ntula et al. (1989)
reported that 62% of 1323 paediatric deaths occurred in the emergency ward prior to
admission. The children are brought in as emergencies and few survive because the
illness is too advanced. Their deaths only help to reinforce the belief that children who
are taken to health facilities with convulsions do not survive - and the vicious cycle
continues.
The neem tree is used for the same purpose in same purpose in Ghana (Abbiw, 1990;
Agyepong, 1992). The medicinal and insecticidal properties of this plant have been and
are continuously being investigated with regard to its efficacy as an antimalarial and as
a vector control (Reuben and Rao 1991; Nkunya 1992). The conclusion so far is that its
antimalarially active limonoids are in very small amounts and thus the therapeutic effects
claimed by patients may be due to other factors, such as the recently established anti
inflammatory and immunomodulating activities (Nkunya 1992). Thus when introducing
health intervention programmes, this herbal remedy belief should not be discouraged,
neither should it be vigorously promoted. Its insecticidal properties are also encouraging
and its products are used widely in rice paddies in India and other countries of South
East Asia (Reuben and Rao 1991).
The use of HFs for the treatment of malaria was low. What is usually referred to as
‘delay’ in bringing a child to a hospital is in fact the period that a mother has to make
complex decisions, in real life situations where as Blaxter (1981) puts it, issues may not
be discrete, roles may conflict and priorities have to be assigned in accordance with an
32
individuals’ value systems. Thus although mothers did not present their ill children to
HFs on the first day of symptoms, at home they were not ignored either. Further, when
their differential diagnosis had to change either because the child did not respond to the
treatment being given at home or from a traditional healer; or the condition was
perceived as getting worse, then mothers did bring their children to a HF.
It has been suggested that purchasing OTC drugs incurs relatively large out-of-pocket
health expenditure especially on poor people (Melrose 1982; Cored 1983; Ugalde and
Homedes 1988). However, on the whole, self-medication for Kilifi mothers was cheaper
than a HF visit. A single visit to a HF would cost a mother between Ksh.20-30 (£0.300.55), depending on whether there were laboratory investigations or not. The hospital
journey may also include transport money which for the respondents ranged between
Ksh. 10-25 (£0.20-0.50) and may also include money for other incidentals such as lunch.
If no transport funds were available, a mother in Sokoke for example would have to be
well enough herself to make the 40 kilometre journey with a child strapped on her back,
while a mother in Mtondia would have a 15 kilometre journey to make. A correct course
of an antimalarial for a child (syrup or tablets) would cost the mother between Ksh.5-10
(£0.10-0.20). The user fee charge at HFs therefore increased the need for selfmedication.
It must be pointed out however, that mothers had their own preferences in choice of
what HF to use. For example, mothers from Sokoke preferred to make the 40 kilometre
journey to Kilifi district hospital, by-passing the government dispensary in the area, which
was more easily accessible to them. They did so in the hope that they would be
’’examined by more qualified doctors", especially the "muzungu (white) doctors in the
‘malaria clinic’". The issue of by-passing HFs on account of drug shortages, always cited
in literature on utilisation of services did not arise in Sokoke at the time of the survey
because the drugs ration had recently arrived and therefore there was no drug shortage.
Further, no user fees are charged at dispensaries and yet Sokoke mothers still went to
Kilifi district hospital where a fee was levied. Gilson (1992) observed in Tanzania that
mothers by-passed the nearest facility to go to one perceived as better. In the Gambia,
mothers complained they lacked money to pay for drugs from VHWs yet they
subsequently payed more at HFs (Menon 1991). This may explain why mothers do not
use services even when proximity is not an issue, and charges are minimal or non
existent.
It was suggested that mothers are using antimalarial drugs inappropriately and that they
do not get proper advice from ROPs. The study also found that it was not only OTC
drugs that were used in this manner, that even medication prescribed in HFs was used
irrationally. This is partly because no clear instructions were given to the mothers by the
health workers or because they (mothers) used the wrong measures when giving the
medication to the children at home. Fifty-five percent of mothers whose children were
given medication did not understand instructions and they did not ask for clarification.
They explained that they did not ask for clarification because health workers can be
harsh when asked too many questions, an observation also made by Nyazema,
Chavunduka, Dzimwasha et al. (1991) in Zimbabwe.
33
The positive finding made at HFs and in the study area in general was the fact that
injections were neither a preferred mode of treatment for malaria in HFs nor was there
a belief among the people that injections were the best treatment, hence the lack of the
culture of ‘quack injectionists’ in the area. This is in contrast to findings from Guinea by
Dabis et al. (1989), and Ejezie et al. (1991) in Nigeria, who reported excessive use of
injections for malaria in their studies. Dangers of this procedure are well documented
(Warrell et al. 1991).
METHODOLOGICAL ASSESSMENT
Integration of methodologies
This study seeks to contribute to the development of health communication information
for malaria control, specifically, childhood malaria, through appropriate and prompt
presumptive treatment of fever and malaria-like illness. While acknowledging the debate
about whether it is possible or even desirable to combine research methodologies
(Werner and Schoepfle 1987; Buchanan 1992) and the promotion of triangulation of
methods (Mechanic 1989; Steckler, McLeroy, Goodman et al. 1992; Yach 1992); the
latter position was adopted for this research. It views quantitative and qualitative
methodologies as complementary and assumes that weaknesses in either would be
compensated by the other. Integration of data collection methodologies provided a
broader understanding of both the context and factors that influence mothers’ definition
and treatment of childhood malaria.
The triangulation technique allowed for coverage of a large study population (1408) in
which 1348 respondents were interviewed through structured interviews and enabled the
in-depth interviewing of 60 respondents concurrently. Observations of any phenomena
of interest to the study were also made parallel to the other methodologies. Through
quantitative methods and their sampling procedures the study investigated knowledge,
beliefs and behavioural differentials as determined by socio-demographic and
socioeconomic variables. Qualitative methods were used to investigate intensely the basis
of the knowledge, beliefs and behaviour. Hence, it was possible to obtain the general
feelings and attitudes of the mothers towards childhood malaria, and to measure the
strength of the feelings and factors influencing them.
For example, quantitative results (table 2) at a glance showed that 56% of mothers knew
the cause of malaria to be mosquitoes, the majority knew the symptoms, 47% understood
communicability and 53% thought malaria was preventable. More impressive is the
finding that 65% of the mothers currently used anti-mosquito measures. The results also
indicated that education levels had a great influence on what mothers knew about
malaria. However, further data presented in the same table 2 demonstrate that the
figures do not auger well for childhood malaria control. Overall, only 10% of the
mothers had correct (biomedical) knowledge of malaria, of the remaining 90%, some of
the mothers had beliefs that were a mixture of biomedical and biocultural beliefs while
the majority had only biocultural beliefs.
It was through qualitative methods however, that one could clearly determine the basis
of the beliefs and behaviours. Malaria is viewed in the same light as other febrile
34
illnesses that are inevitable, not serious and self-limiting but with the possibility of being
severe and could perchance cause death. Like other febrile illnesses, childhood malaria
was attributed to weather or temperature changes, which Mayall (1986) refers to as
thermal theories. While working among mothers of different social classes in Britain (an
inner city area of North London), Mayall found that mothers believed that colds were
caused by effects of temperature changes, or of extremes in temperature. Like in Kilifi,
it was mothers in low socio-economic groups who invariably also had relatively low levels
of education who subscribed to these thermal theories. HSB for childhood malaria,
especially with regard to therapy choice would not have been clearly understood if only
quantitative data was used. The fact that the male head of household is almost
exclusively the sole decision-maker in the choice of therapy for children and mothers
would have been difficult to elucidate.
It was also through qualitative methods that beliefs on complications associated with
malaria were fully discerned. For example, the quantitative data showed that only 16%
of mothers recognise convulsions as a childhood illness that presents with fever. On
prompting (which is arguably more a qualitative technique than a quantitative one), the
figure rose to 56%. Further, quantitative data indicated that only 8% of mothers believed
convulsions had a spiritual origin and only 19% reported use of charms and amulets to
prevent the illness. The interpretation of the quantitative data would arrive at the
conclusion that mothers in Kilifi do not recognise convulsions as a childhood illness. This
would be far from the truth. Qualitative findings have demonstrated that almost all
mothers not only know the condition but know what to do in the event that their child
has an attack. The condition has the same status as other ‘folk’ illnesses like chirwa,
included into the larger cognitive classification of illnesses specific to children. Similar
results are illustrated for anaemia and splenomegaly. Further, ethnoetiological beliefs
and the very elaborate management procedures would have been completely missed if
integration of methodologies was not employed.
Another example of information that could have been missed if the data collection
methods were not integrated was the dissatisfaction with and non-use of HFs because of
user fees. Twenty-two out of 53 mothers explained the three day time lag between onset
of illness in their children and seeking care from HFs as being due to "predisposing"
factors such as the perception that illness was mild, partner being absent, other important
matters to attend to and lack of someone to mind the ill child’s siblings while others
(11/53) cited "enabling" factors such as lack of money for transport, having given the
child OTC drugs and having taken the child to a traditional healer. No "health systems"
factors such as accessibility, lack of user fees, waiting time at facilities or problems with
health workers were explicitly cited, although they may be implied in the factors
mentioned above. The mothers may have said nothing about user fees and accessibility
being a hindrance because they were already at HFs when they were interviewed, but it
came out very clearly during the in-depth interviews that most mothers chose not to go
to HFs, unless it was absolutely necessary, because it was beyond their means. It was
more unlikely for people to come to HFs especially when they thought the illness may
involve laboratory tests which would have to be paid for in addition to the consultation
charge and then be given "only chloroquine".
35
The complementarity of the methods thus allowed for clarification and wider
interpretation of the data. Results based on these data would be far more useful as a
basis for intervention strategies because all possible aspects of maternal reaction to
childhood malaria were explored. The triangulation approach has great potential for
areas where communities are much more at ease talking than responding to questions
in mono-syllables. However it is necessary to point out that while no undue disadvantages
of using triangulation were encountered, there were a few problems that were
experienced in the analysis of the data, especially with regard to deciding the cut off
point (by quantification) for categorisation of themes in content analysis. Because
methodological integration is still not fully entrenched and in essence is still "currently
being attempted and struggled with..." (Steckler et al. 1992: 6), there are few "how to"
documents for reference. Nevertheless, with more investigators using the methodology,
the tools should become better (see entire issue Hlth. Educ. Q. 19(1) 1992), and health
communication planning, implementation and evaluation might have more positive
results than hitherto.
Limitations: scheduled and unexpected
Firstly, although the study aimed at focusing on mothers’ definition and treatment of
childhood malaria, it was deemed necessary to confine mothers to answering questions
about childhood illnesses that present with fever. This was intentionally done to focus on
childhood malaria and some of its complications. However, key informants and ROPs
could discuss any childhood illnesses without restrictions. The questions put to the key
informants and ROPs were in effect a community diagnosis of childhood health
priorities, without overwhelming mothers with too many questions, while also verifying
their responses against the illness taxonomy obtained from the key informants and ROPs.
Secondly, although the HF users’ and RO users’ sub-studies elicited important
information about who uses these services, the data could not be used to determine
usage levels in the communities because only a self-selected group of mothers already
included in the mothers’ survey were re-interviewed in these sub-studies. The criterion
proposed for inclusion into the HF users’ survey was ‘mother accompanying an ill-child
to the facility, with a self-diagnosis of malaria’. However, it became apparent that few
mothers indeed would be included in the study because malaria was not often self
diagnosed. Only one out of 53 mothers re-interviewed at a HF (in the space of four
weeks) diagnosed her child as suffering from malaria before coming to the HF. Similar
results were reported by Rooth and Bjorkman (1992), who while working in Rufiji,
Tanzania, an area almost similar endemicity to Kilifi, found that only two out of 164
mothers (in the space of six months) mentioned malaria as a possible diagnosis before
bringing the child to the HF.
Thus the criterion was abandoned and all mothers bringing an ill child to the HF, who
had been previously interviewed in the mothers’ survey, were included in the study. This
ensured that those mothers whose children would be diagnosed as having malaria by a
health worker would also be re-interviewed. This did not compromise the results.
36
Similarly, the inclusion criterion for RO users’ survey had to be modified. It had been
decided that mothers and ‘others’ purchasing OTC drugs on behalf of mothers be
interviewed at ROs if they purchased drugs for children. However, the field-workers
proved to be too few for the survey to facilitate follow-up of those ‘others’. It was
imperative that the mother herself be interviewed especially because of the questions on
decision-making processes involved in the use of the services. This in effect made it
difficult to recruit all users; consequently, sub-samples were small and generalisation of
the findings difficult. However, the RO users’ results are intended for use in intervention
programmes in Kilifi, making the issue of the small sample and thus generalisability less
critical (Basch 1987; Ramakrishna and Brieger 1987). As Ramakrishna and Brieger
(1987) argue, "...in-depth research cannot be generalised to a larger population. But this
is not the aim of the research. The results are intended to be valid for the local situation,
an essential for PHC planning...[the need is to] generalise not the particular results...but
the process by which the results were obtained" (Ramakrishna and Brieger 1987: 173).
The methods used in this study can therefore be replicated in other areas.
Thirdly, it is recognised that SES of a household may influence maternal behaviour in
relation to malaria. However, the study did not adequately measure the SES position of
the households. This would require information on household structures, information on
direct and indirect income generators, land tenure and other indicators. It was hoped that
occupation of the head of household, household possessions and house quality would give
an estimation of SES, but these markers only reinforced the expected differences
between the three areas - peri-urban, rural and slum settlement. It was thus not fully
used in the data analysis.
Overall, no major contradictions came through the findings because of triangulation of
methods. Any differences noted may be explained by the nature of the research question
or the nature of questionnaire interviews. For example, many more key informants than
mothers linked fever to convulsions, and most mothers played down the need to consult
someone for health decision matters. As Gilson (1992) concludes with regard to
integration of methodology, " contradictions... could be traced back to the structured
nature of questionnaire interviews...particularly closed questions..."(:196).
IMPLICATIONS AND FUTURE RESEARCH NEEDS
The study has described the biocultural definition of childhood malaria and how
decisions are made as to whether, and how to react to the illness within the existing
health care sources. These psychosocial processes may hinder or facilitate appropriate
management of malaria at the household level. The study has also identified other
factors that may militate against malaria control, namely: (a) the image of malaria
depicted to populations in malarious areas, (b) the simplistic depiction of the malaria
mosquito link and (c) the excessive availability of over-the-counter antimalarials and
other drugs. The implications of these findings for malaria control are presented and
recommendations are made for consideration in future malaria control policies.
37
The key findings and recommendations emanating from this study are summarised in
table 5. The findings are discussed in detail in Mwenesi (1993).
Table 5: Issues on malaria control that need to be addressed or improved upon
ISSUE
SITUATION
RECOMMENDED ACTION
KAP on malaria, mosquitoes
and its complications
Limited knowledge
1. Health education messages should aim
at depicting malaria as it is experienced in
holoendemic areas - chronic and
debilitating. 2. Difficult as it may be the
mosquitoes and malaria link should be
made clear to people. 3.The fact that
malaria can progress to serious
complications should be carefully included
into malaria messages.
Decision-making roles
Clear and not
flexible
1. Whole populations including men must
be targeted in HE for malaria.
Availability and use of OTC
drugs
Excessive
1. Closer control of licensing procedures,
promotion, distribution and selling of OTC
drugs.
2. Reduce the number of brand names on
the market.
3. Control drugs at HFs to minimize
prescription drugs finding their way into
ROs.
4. Educate the public and especially ROPs
on the need to use/sell correct and
relevant medication and the dangers of
misuse of OTC drugs.
5. Enforce drug regulation laws.
Use of health facilities
Low
1. Explain to the public why ‘user fees’ are
an unavoidable part of future health
services and the hope that they will
improve quality of care.
2. Health workers should improve their
communication skills especially on the
need to complete courses of malaria
medication.
Traditional healers
Strong1
1. Identify, forge and promote links with
them.
Kenya is in the process of formulating a long over-due malaria control policy. Lessons
to be learned from the findings of this study are that the issue of malaria is not a straight
forward one. The policy should be informed by Kenyan and other experiences. For
Kenya, community orientation for CBHC exists, as indicated by the fact that such
programmes have covered more than half of the 41 administrative areas (districts) in the
1 Traditional healers are frequently used in Kilifi although not for malaria. Convulsions, anaemia & splenomegaly are treated by
traditional healers making them part of the general sources of care.
38
12 years since CBHC commenced in the country (Maneno and Mwanzia 1991), like the
oft cited Saradidi (Kaseje and Sempebwa 1989), Chogoria (Deboer and McNeil 1989)
and Kibwezi (Johnson, Kisubi, Mbugua et al. 1989) projects. Whatever programmes or
strategies are to be used to implement the malaria policy (PHC, selective PHC or the
Bamako Initiative), the implementation has to be thought out carefully. Lessons should
be learned from earlier malaria control policies which have not been successful (Bradley
1991).
While there are clouds over the prospects of ‘Health for All by the Year 2000’
particularly as a result of structural adjustments programmes (SAP) in the developing
world (Stone 1992; Woelk 1992; Rathwell 1992), CBHC is still perceived to be a viable
strategy for providing health care for the majority of the people (Garner and Walt,
forthcoming). However, certain health care strategies currently being promoted need re
assessment, both nationally and internationally. For example, why are findings of social
science studies not taken up seriously; and why is health education not being given the
due attention it deserves if informed change is to be achieved where necessary?
The fact that most CBHC projects do not base their programmes on results of needs
assessment was alluded to earlier. The reason for this state of affairs has been
expounded by Stone (1992), which she sums up as the contradiction inherent in CBHC
(PHC) as a policy:
it is supposed to foster community participation and wherever possible
assist local communities define their own health needs, and initiate ways
of meeting them. At the same time PHC has already set its own
parameters around both the needs of the people and the range of possible
means of meeting them (: 411).
The implication is that often, well-meaning programmes like the strategy of prompt
presumptive treatment of malaria through VHWs are rendered ineffective. This is
because even when a community diagnosis is done in an area, there is a bias towards
current strategies, already decided upon. Stone (1992) describes a hypothetical situation
where ‘well meaning’ health planners may assume that even if local communities are
more interested in curative care they will still be responsive to health education. For
example, mothers in Kilifi asked for subsidies on OTC drugs which to them are the
strategy that they feel can tackle the malaria problem. They were promptly treating their
children albeit inappropriately. They did not link mosquitoes to malaria or malaria-like
illness. Any intervention programme introduced must convince these mothers why
malaria is a major problem and not malnutrition for example. They may not be keen to
be taught by VHWs, what they already know and do - prompt treatment of feverish
symptoms - and if it also includes anti-mosquito measures they would have to be fully
convinced of their role in the aetiology of malaria.
The relevance of new technologies must be made clear to the people through whichever
CBHC medium is used (PHC, selective PHC, or BI). Recently, the call for use of bed
nets has gained momentum and because the BI strategy is seen to lend itself well to the
promotion of bed net use, it is the medium used. However, it is not clear how this ‘new’
technology will be implemented and sustained in areas of deprivation like Kilifi where
39
user fees for the preferred curative services are already a problem. This raises doubts as
to the viability of impregnated bed nets as a measure for malaria control in Kilifi.
Health education is not the panacea for the problems encountered in malaria control
efforts, but it could be a strong catalyst in bringing about informed change. It should
however be based on the premise that the perceptions that people hold are diverse and
may offer opportunities as well obstacles to change. As was demonstrated by the findings
of this study, peoples’ beliefs form part of a valued framework which helps them cope
with their day to day health problems. Such beliefs are not open to simple correction
(Williams and Wood 1986).
Health education, acquired through schooling or community education prepares an
individual’s base for behaviour change, that may facilitate ’healthy’ decision-making.
Maternal education especially has been shown to have a strong impact on child health
(Cleland and van Ginneken 1988; Rasmussen 1989). On average, each one-year
increment corresponds with a seven to nine percent decline in under five year olds’
mortality. Further, education is found to exercise the strongest influence in early
childhood (Cleland and van Ginneken 1988), which corresponds with the critical agegroup that is most at risk of malaria infection. Thus although they may not be the
decision-makers for health care matters, they would be better off informed.
However, factors that affect child health cannot be tackled only through individual
behaviour change. Those whose behaviours are targeted, be they mothers or whole
communities, must be empowered to enable them to change their behaviour by
addressing the other problems that put them in the situation they find themselves, such
as unfavourable micro-environments in ecological and socio-economic terms as well as
concomitant structural inequalities. In other words the cycle of deprivation should be
broken (Zaidi 1988), by realising that economic powerlessness of most people who suffer
from malaria is real.
For example in Kilifi, while tackling the problem of childhood malaria through health
education, there would also be need to address the causes of the rampant malnutrition,
issues of housing and sanitation and the low levels of education amongst women. Malaria
is but one of the many problems that face the community. The reality though, is that
resources required to address these issues are limited. Dialogue with the community
should allow for the most urgent to be tackled.
Health education in general should equip populations with awareness of the benefits and
dangers inherent in self-medication with OTC drugs. However, the Kenya government
which already has in its statutes laws like the Pharmacies and Poisons Act, which is
chapter 244 of the Laws of Kenya to regulate the production, licensing, distribution,
promotion, pricing and sale of this widely used source of care should enforce the law and
probably re-assess it.
The chain of distribution of OTC drugs which include antimalarial drugs and analgesics
is long - manufacturer -> wholesaler -> retailer. Pricing of OTC drugs works on a
supply and demand basis - it is not controlled. Consequently, the further in the interior
a retailer is, the more difficult it is to police the sale of expired drugs or prescription only
40
drugs, including antibiotics such as chloramphenicol and tetracycline, found in ROs in
Kilifi. But, regulations must be enforced.
In Kilifi the number of drugs and formulations available over the counter is bewildering.
Even more unacceptable, is the availability of the same generic drug in many different
brand names. There were 19 types of antimalarial drugs. Of these, 14 types were
chloroquine-based. Twenty-seven types of analgesics and antipyretics, based on aspirin,
paracetamol and codeine, singly or in combination, were also being marketed under
different brand names. Drugs sold in informal outlets account for 35% of all drugs
distributed in Kenya (Foster 1990). The Kenya government may not individually be able
to change the generic/patent names regulations, but most of the OTC drugs are
manufactured locally. Through strict licensing and advertising regulations, it may be
possible to reduce the number of drugs available and to ensure that promotional
materials are reasonable.
At present, it is estimated that only 20% of the population is effectively covered by
services (Mwabu and Mwangi, 1986) and approximately half the population has no access
to public health services (Dahlgren, 1991). It is obvious then that other sources of care
will be used. Self-medication will continue to be practised, but unacceptable risks that
go with it should be checked and minimized. The wrongly held belief that ‘a pill exists
for every ill’ (Laing, 1990) should be a focus for health education.
In the recent past, the call has been for the serious consideration of social aspects in
tropical disease control. There is enough information that has been generated in this
field to justify a serious focus on health education, one of the most important aspects of
CBHC. While certain social sciences have recently gained momentum in the field of
health and medicine, especially in improving the methodologies employed in gathering
and disseminating data, there has been no corresponding acceleration in the field of
health education, which would enable the input of anthropological and sociological
efforts to be implemented and evaluated.
The greatest need is to develop convincing and persuasive communication materials to
reverse the effect of old teachings of past decades on malaria and the mosquito to
populations that are mostly illiterate and posses strongly held alternative explanations a tabula rasa situation does not exist. Commensurate with the promotion of health
education, will probably be the training of a new cadre of personnel who will receive
proper training in health education - not the three to four days training handed out to
VHWs to teach about a complex disease like malaria. Health educators should rise to
the occasion.
The essential drugs programme policy (WHO 1988a; 1988b), that addresses the issue of
drug selection, procurement, tendering and distribution to health facilities could
profitably be extended to retail outlets. This informal sector will continue to provide a
health service so urgent positive action is imperative. Generic name use for drugs is a
contentious issue for commercial drug companies, but there is evidence that their use is
possible (Mamdani and Walker 1985). The World Health Organisation (WHO) and the
newly formed International Network for the Rational Use of Drugs (INRUD) (Ross
1992), should seriously consider the possibility of an essential drugs list for ROs, akin to
41
the essential drugs list for HFs. It could be a useful tool for guiding countries about what
should be sold over-the-counter. This appears to fit well into the primary objective of
INRUD, which is to identify through a coordinated set of country studies, a set of
effective interventions to recommend as policy options for the promotion of rational drug
use. An essential drugs list for retail outlets is long over-due. This could be prescribed
under existing drug regulation statutes.
Malariologists, entomologists, social scientists and health educators must urgently come
together to formulate health education messages and strategies that will get the message
of malaria across to populations in malarious areas. Instead of workers in different
countries working in isolation, trials of the different types of strategies could be
conducted in comparable areas of these countries to find out the most feasible ones. For
example, the feasibility of different channels such as the radio, television, street theatre
or public meetings could be examined.
Research issues arising from this study are:
a) There is evidence that severe malaria when not fatal results in some serious
consequences. For example, repeated generalised convulsions have been implicated in
the aetiology of secondary epilepsy (Bahemuka, 1981; Matuja, 1989). In their study of
severely ill children, Brewster et al. (1990) reported that 25% of those presenting with
cerebral malaria ended up with major neurological handicaps. It would be important to
determine the prevalence of these neurological problems in malarious communities. If
their rates are high and people perceive them as priority health problems, and if their
link to malaria can be established, they could be used as proxies for malaria in HE
interventions for malaria. This would be an indirect way of tackling the malaria menace.
b) The magnitude of OTC drug-related iatrogenesis such as aspirin and paracetamol
poisoning should be investigated. This would require a multi-disciplinary approach where
social science methods would be used to elicit drug use histories of suspected cases,
epidemiological methods to map out incidence and prevalence rates in the community
and biomedical and clinical methods. Hard data would be needed to convince authorities
of the need to be strict about licensing, manufacture, distribution and sale of the drugs.
c) Since traditional healers and retail outlet proprietors are part of the health service
sources in Kilifi, possibilities of incorporating them into malaria intervention efforts
should be explored.
The battle against malaria will continue. The serious effort and enthusiasm being put in
the search for new insecticides, new drugs and the eagerly awaited vaccine, must also be
put into health education and communication sciences. This suggests that all those
working in malaria are a necessary part of the equation. Ignorance on the part of health
care providers and recipients must be fought through continuing education for the former
and health education for the latter. ‘All for Health’ should be the guiding motto for
‘Health for AH’.
42
REFERENCES
Abbiw, D. 1990. Useful plants of Ghana. Intermediate Technology Publications, London.
Abdullah HR. 1985. Social aspects of malaria control: A knowledge, attitudes and practices study among the Luo of
Karateng, Kisumu district. Unpublished MA Thesis. University of Nairobi, Nairobi.
Agyepong IK. 1992. Malaria: Ethnomedical perceptions and practices in an Adangbe farming community and implications
for its control. Soc. Sci.& Med. 35(2): 131-137
Alila PO. 1978. The role of public bureaucracy in agriculture in Western Kenya. Unpublished PhD. Thesis, Indiana
University.
Alonzo AE. 1979. Everyday illness behaviour: a situational approach to health status deviations. Soc. Sci. & Med. 13(A):
397-404.
Bahemuka M. 1981. Management of epilepsy: Therapeutic aspects. East Afri. Med. Jnl. 58(6), 389-400.
Basch CE. 1987. Focus group interview: An under-utilized research technique for improving theory and practice in health
education. Hlth. Educ. Q. 14(4): 411-448.
Berman DA. 1984. Village Health Workers in Java, Indonesia: Coverage and equity. Soc. Sci. & Med. 19(4): 411-422.
Blacker J, Mukiza-Gapere J, Kibet M. et al. 1987. Mortality differentials in Kenya. Paper presented at IUSSP seminar
on Mortality and Society in Sub-Saharan Africa, Yaounde
Blaxter M. 1981. The health of children: A review of research on the place of health in cycles of poverty. Heinemann
Educational Books, London.
Blaxter M and Paterson E. 1982. Consulting behaviour in a group of young families. Jnl. Roy. Coll. GPs. 32: 657-662.
Blaxter M. 1983. The causes of disease: Women talking. Soc. Sci. & Med. 17(2): 59-69.
Blaxter M. 1990. Health and life-styles. Tavistock/Routledge, London.
Bledsoe CH and Goubaud MF. 1985. The reinterpretation of Western pharmaceuticals among the Mende of Sierra Leone.
Soc. Sci. & Med. 21: 275-282.
Boerma JT. 1989. Maternal and child health in an ethnomedical perspective: Traditional and modern medicine in
Kwale. Preliminary final report. Kwale district child survival and development programme. Monitoring and evaluation
project 1987-1989. UNICEF/WHO.
Brabin L and Brabin BJ. 1992. Parasitic infections in women and their consequences. Advances in Parasit. 31: 1-81.
Bradley D. 1991. Malaria - whence and whither. In Targett GAT (ed). Malaria: Waiting for the vaccine. John Wiley and
Sons, London.
Brewster DR, Kwiatkowski D and White NJ. 1990. Neurological sequelae of cerebral malaria in children. The Lancet. 336:
1039-1043.
Buchanan DR. 1992. An uneasy alliance: Combining qualitative and quantitative research methods. Hlth. Educ. Q. 19(1):
117-135.
CBS. 1991. Economic survey. Central Bureau of Statistics. Ministry of Planning and National Development, Nairobi,
Kenya.
Cleland JG and van Ginneken JK. 1988. Maternal education and child survival in developing countries: The search for
pathways of influence. Soc. Sci. & Med. 27(12): 1357-1368.
Corel! J. 1983. "Allocation of family resources for health care in rural Haiti". Soc. Sci. & Med. 17(11): 709-719.
Cunningham-Burley S. 1990. Mothers’ beliefs about and perceptions of their children’s illness. In Cunningham-Burley S
and McKeganey N (eds); Readings in Medical sociology. Tavistock/Routledge, London.
Dabis F, Breman JG, Roisin AJ, et al. 1989. Monitoring selective components of primary health care: Methodology and
community assessment of vaccination, diarrhoea, and malaria practices in Conakry, Guinea. Bull. WHO 67(6): 675-684.
Dahlgren G. 1991. Strategies for health financing in Kenya - the difficult birth of a new policy. Scand. J. Soc. Med. Supp.
46:67-81.
43
Davis A. 1982. Children in clinics. Unwin Hyman, London.
De Boer CN and McNeil M. 1989. Hospital outreach community-based health care: The case of Chogoria, Kenya. Soc.
Sci. & Med. 28(10): 1007-1017.
Deming MS, Gayibor A, Murphy K. et al. 1989. Home treatment of febrile children with antimalarial drugs in Togo. Bull.
WHO 67(6): 695-700.
Ejezie GC, Ezedinachi ENU, Usanga EA. et al. 1991. Malaria and its treatment in rural villages of Aboh Mbaise, Imo
State, Nigeria. Acta Tropica 48:17-24.
Ewbank D, Henin R, and Kekovole J. 1986. An integration of demographic and epidemiological research on mortality in
Kenya. In Determinants of mortality change differentials in developing countries: The five-country case study project.
United Nations Department of International Economics and Social Affairs, New York.
Fonaroff LS. 1968. Man and malaria in Trinidad: Ecological perspectives of a changing health hazard. Annals Assoc.
Amer. Geogr. 58: 526-556.
Foster GM. 1976. Disease aetiologies in non-Westem medical systems. American Anthrop. 78: 77-82.
Foster S. 1990. Improving the supply and use of essential drugs in Africa. World Bank Working Paper. WP5 456. Pop.
& Human Resources Dept. The World Bank, Washington.
Foster SD. 1991a. The distribution and use of antimalarial drugs-not a pretty picture. In Targett GAT. (ed). Malaria:
Waiting for the vaccine. John Wiley and Sons, London.
Foster SD. 1991b. Pricing, distribution, and use of antimalarial drugs. Bull. WHO 69(3): 349-363.
Fosu GB. 1989. Access to health care in urban areas in developing countries. J. Hlth. Soc. Beh. 30(4): 398-411.
GK/UNICEF 1989. Situation analysis of children and women in Kenya. Government of Kenya/UNICEF, Nairobi.
GK/UNICEF 1990. Socio-economic profiles: Kwale, Kitui, Embu, Baringo, Kisumu, South Nyanza Districts: Nairobi
City and Mombasa municipality. Government of Kenya/Ministry of Planning and National Development/UNICEF.
Gilson L. 1992. Value for money? The efficiency of primary health units in Tanzania. Phd Thesis. LSHTM, University
of London.
Glik DC, Parker K, Mulingande G and Hategikamana B. 1987. Integrating qualitative survey techniques. Int. Q.
Community Hlth. Educ. 7(3): 181-200.
Glik D, Gordon A, Ward W. et al. 1988. Focus group methods for formative research in child survival: An Ivorian
example. Int. Q. Community Hlth. Educ. 8: 291-316.
Glik DC, Ward WB, Gordon A and Haba F. 1989. Malaria treatment practices among mothers in Guinea. J. Hlth. & Soc.
Beh. 30: 421-435.
Greenberg AE, Ntumbazondo M, Ntula N et al. 1989. Hospital-based surveillance of malaria-related paediatric morbidity
and mortality in Kinshasa, Zaire. Bull. WHO 67: 189-196.
Greenwood BM. 1987. Asymptomatic malaria infections - do they matter? Parasitology Today 3(7): 206-214.
Greenwood BM, Greenwood AM, Bradley AK et al. 1987. Mortality and morbidity from malaria among children a rural
area of the Gambia, West Africa. Trans. Roy. Soc. Trop. Med. & Hyg. 81: 478-486.
Greenwood BM, Greenwood AM, Bradley AK et al. 1988. Comparison of two strategies for control of malaria within a
primary health care programme in the Gambia. The Lancet (i): ,4112-1127.
Greenwood BM. 1991. An analysis of malaria parasites in infants: 40 years after Macdonald - A response. Trop. Dis. Bull.
88(2): R1-R3.
Hackett LW, Russell PF, Scharff JW and Senior White R. 1938. The present use of naturalistic measures in the control
of malaria. Bull. Hlth. Org. League of Nations 7: 1016-1064.
Heggenhougen HK and Clements J. 1987. Acceptability of childhood immunization: Social science perspectives. EPC
Publication No. 14 LSHTM, London.
Helman C. 1990. Culture, health and illness. (2nd edition), Butterworth-Heinemann, Oxford. Hibbard JH & Pope CR.
1983. Gender roles, illness orientetion and the use of medical care. Soc. Sci. & Med. 17: 129-137.
44
Hoffman SL, Masbar F, Hussein PR et al. 1984. Absence of malaria mortality in villagers with chloroquine-resistant R
falciparum treated with chloroquine. Trans. Roy. Soc. Trop. Med. & Hyg. 78: 175-178.
Hogerzeil VH. 1985. Standardized supply of essential drugs in Ghana. Drukkerij Elinkwijk bv-Utrecht.
HongvivatanaT. 1986. Human behaviour and malaria. South East Asian Jnl. Trop. Med. & Publ. Hlth. 17(3): 353-359.
HongvivatanaT, LeerapanP and ChaiteeranuwatsiriM. 1985. Knowledge, perceptions and behaviour of malaria. Center
for Health and Policy Studies. Mahidol University, Thailand.
Jackson LC. 1985. Malaria in Liberian children and mothers: Biocultural perceptions of illness vs clinical evidence of
disease. Soc. Sci. & Med. 20(12): 1281-1287.
Janzen JM. 1978. The quest for therapy - Medical pluralism in Lower Zaire: Comparative studies of health systems
and medical care. University of California Press, Berkeley.
Janzen JM. 1987. Therapy management: Concept, reality and process. Med. Anthrop. Q. 1(1): 68-84.
Jick TD. 1983. Mixing qualitative and quantitative methods: Triangulation in action. In van Maanen J (ed). Qualitative
methodology. Sage Publications, Beverly Hills, Cal.
Johnson KE, Kisubi WK, Mbugua JK. et al. 1989. Community-based health care in Kibwezi, Kenya: 10 years in retrospect.
Soc. Sci. & Med. 28(10): 1039-1051.
Kaseje DCO and Spencer H. 1987. The Saradidi, Kenya rural health development programme. Ann. Trop. Med. &
Parasit. 81(l)Supl.: 1-6.
Kaseje DCO and Sempebwa EKN. 1989. An integrated rural health project in Saradidi Kenya. Soc. Sci. & Med. 28:
1063-1071.
Kasilo OJ, Nhachi CBF and Mutangadura EF. 1991. Epidemiology of household medications in urban Gweru and Harare.
Central Afri. J. Med. 37(6): 167-171.
Khan ME and Manderson L. 1992. Focus groups in tropical diseases research. Hlth. Policy Plan. 7(1): 56-66.
Kleinman A. 1980. Patients and healers in the context of culture. University of California Press, Berkerly.
Kleinman A. 1981. On illness meanings and clinical interpretations: not "rational man", but a rational approach to man the
sufferer - man the healer. Cult. Med. Psychiat. 5: 378-377.
KDDP. 1989. Kilifi district development plan - 1989-93: Final Draft. GK/Kilifi District Office, Kilifi.
Laws of Kenya. Chapter 244: The Pharmacies and Poisons Act. Government Printers, Nairobi.
Laing RO. 1990. Rational drug use: An unsolved problem. Tropical Doctor. 20: 101-103.
Lieban RW. 1992. From illness to symbol and symbol to illness. Soc. Sci. & Med. 35(2): 183-188.
Lipowsky R, Kroeger A and Vasquez M. 1992. Sociomedical aspects of malaria control in Colombia. Soc. Sci. & Med.
34(6): 625-637.
Logan K. 1988. ‘Casi como doctor’: Pharmacists and their clients in a mexican urban context. In Van der Geest S and
Whyte SR (eds). The context of medicines in developing countries. Kluwer Academic Publishers, Dordrecht.
Maitai CK, Guantai A and Mwangi JM. 1981. Self medication in management of minor health problems in Kenya. East
Afri. Med. J. 58(8): 593-600.
Makubalo EL. 1991. Malaria and chloroquine use in Northern Zambia. Phd Thesis. LSHTM. University of London,
London.
Mamdani M and Walker G. 1985. Essential drugs and developing countries: A review and selected annotated
bibliography. EPC Publication No.8, LSHTM, London.
Maneno J and Mwanzia J. 1991. 12 years of primary health care in Kenya. GK/MOH, WHO/UNICEF. Nairobi.
Massard J. 1988. Doctoring by go-between: Aspects of health care for Malay children. Soc. Sci. & Med. 27(8): 789-797.
Matuja WBP. 1989. Aetiological factors in Tanzanian epileptics. East Afri. Med. J. 66(5): 343-347.
Mayall B. 1986. Keeping children healthy. Allen & Unwin, London.
45
Mayall B and Foster MC. 1989. Child health care: Living with children, working for children. Heinemann (Nursing),
London.
McCauley AP, West S and Lynch M. 1992. Household decisions among the Gogo people of Tanzania: Determining the
roles of men, women and the community in implementing a trachoma prevention programme. Soc. Sci. & Med. 34(7): 817824.
Mechanic D. 1989. Medical sociology: Some tensions about theory method and substance. J. Hlth. Soc. Beh. 30: 147-160.
Melrose D. 1982. Bitter pills. Medicine and the third world poor. Oxfam, Oxford.
Menon A. 1991. Utilization of village health workers within a primary health care programme in the Gambia. J. Trop.
Med. & Hyg. 94: 268-271.
Menon A, Joof D, Rowan KM et al. 1988. Maternal administration of chloroquine: an unexplored aspect of malaria
control. J. Trop. Med. & Hyg. 91: 49-54.
Merlin M et al. 1990. Aspects epidemiologiques du paludisme au Gabon. Medicine Tropicale 50: 39-46.
Millman RN. 1967. Settlement, change, and challenge on the Kano plains of Western Kenya. Dept, of Geography,
University of Nairobi (Mimeo).
Mulumba MP et al. 1990. Le paludisme de 1’enfants a Kinshasa (Zaire): influence des saisons, de Page, de I’environment
et de standing familial. Medicine Tropicale 50: 53-64.
Mwabu GH and Mwangi WM. 1986. Health care financing in Kenya. A simulation of welfare effects of user fees. Soc.
Sci. & Med. 22(7): 763-767.
Mwenesi, HRA. 1993. Mothers’ definition and treatment of childhood malaria on the Kenyan Coast. Ph.D Thesis,
University of London, London School of Hygiene and Tropical Medicine.
Nkunya MHH. 1992. Progress in the search for antimalarials. NAPRECA Monograph Series No.4 NAPRECA, Addis
Ababa University, Addis Ababa.
Nyazema NZ, Chavunduka D, Dzimwasha M et al. 1991). Drug information for the community: Type and source. Centr.
Afri. J. Med. 37(7): 203-206.
Oaks SC, Mitchell VS, Pearson GW and Carpenter CCJ. (eds). 1991. Malaria: Obstacles and opportunities - A report
of the committee for the study on malaria prevention and control: Status review and alternative strategies. Division
of International Health, Institute of medicine, Washington DC.
Okelo GBA. 1990. Prospects in the control of malaria. East Afr. Med. J. May, 293.
Olatunde A. 1981. Use and misuse of 4-aminoquinoline antimalarials in tropical Africa and re-examination of itch reaction
to these drugs. Tropical Doctor 11: 97-101.
Ongore D, Kamunvi F, Knight R and Minawa A. 1989. A study of knowledge, attitudes and practices (KAP) of a rural
community on malaria and the mosquito vector. East Afr. Med. J. 66(2): 79-89.
Parkin D. 1978. The cultural definition of political response: Lineal destiny among the Luo. Academic Press, London.
Parkin D. 1991. Sacred void: Spatial images of work and ritual among the Giriama of Kenya. Cambridge University
Press, Cambridge.
PSRI (1982). Population Studies Research Institute. Working paper on malaria. University of Nairobi.
Ramakrishna J and Brieger WR. 1987. The value of qualitative research: Health education in Nigeria. Hlth. Policy. Plan.
2: 171-175.
Rasmussen F. 1989. Mothers’ benefit of a self-care booklet and a self-care educational session at child health centres. Soc.
Sci. & Med. 29(2): 205-212.
Rathwell T. 1992. Realities of health for all by the year 2000. Soc. Sci. & Med. 35(4): 541-547.
Raynal L. 1985. Use of over-the-counter medicines in rural Matabeleland, Zimbabwe: The case of upgrading the dispensing
skill for storekeepers. Central Afr. J. Med. 31: 92-97.
Raynal L. 1985. Use of over-the-counter medicines in rural research: Health education in Nigeria. Hlth. Policy. Plan. 2:
171-175.
46
Reuben R and Rao R. 1991. Biological control of mosquitoes in paddy fields. In Curtis CF; Control of disease vectors
in the community. Wolf & Publishing LTD-CCR Press, London
Ribbands CR 1946. Effects of bush clearance on flighting of West African Anophelines. Bull. Ent. Res. 37: 33-41.
Rogler LH. 1989. The meaning of culturally sensitive research in mental health. Cited in Yach D. 1992. The use and value
of qualitative methods in health research in developing countries. Soc. Sci. & Med. 35(4): 603-612.
Rooth I & Bjorkman A. 1992. Fever episodes in a holoendemic malaria area of Tanzania: Parasitological and clinical
findings and diagnostic aspects related to malaria. Trans. Roy. Soc. Trop. Med. & Hyg. 86: 479-482.
Ross DM. 1992. Improving drug utilisation at the local level - an opinion. Tropical Doctor 22: 80-81.
Rubinstein RA. 1984. Epidemiology and anthropology: Notes on science and scientism. Communication and Cognition
17: 163-185.
Ruebush TK, Breman JG, Kaiser RL and Warren M. 1986. Malaria. Rev. Infectious Dis. 8(3): 47-59.
Rutabanzigwa-Ngaiza J, HeggenhougenK and Walt G. 1985. Women and health in Africa. EPC Publication No.6 London
School of Hygiene and Tropical Medicine, London.
Sevilla-Casas E. 1992. Commentary: Old themes and new directions in malaria studies. In Chen L, Kleinman A and Ware
N (eds); Advancing Health in Developing Countries: The role of Social Research. The Health Transition Project, Center
for Population Studies, Harvard University. Auburn House, New York.
Silva KT. 1991. Ayurveda, malaria and the indigenous herbal tradition in Sri Lanka. Soc Sci. & Med. 33(2): 153-160.
Slutsker L, Breman JG and Campbell CC. 1988. Strategies for control of malaria in Africa. The Lancet, (i): 283.
Smith PG and Morrow RH (eds). 1991. Methods for field trials of interventions against tropical diseases: "A tool box".
Oxford University Press, Oxford.
Snow RW, Peshu N, Forster D, Mwenesi H and Marsh K. 1992. The role of shops in the treatment and prevention of
childhood malaria on the Kenyan Coast. Trans. Roy. Soc. Trop. Med & Hyg. 86: 237-239.
Snow RW et al. in Press; Transactions.
Spencer NJ. 1984. Parents’ recognition of the ill child. In MacFarlane J (ed), Progress in child health. Churchill
Livingstone, London.
Spencer H, Kaseje DCO, Sempebwa EKN et al. 1987. Impact on mortality and fertility of a community-based malaria
programme in Saradidi, Kenya. Ann. Trop. Med. & Parasit. 81(1) Supl: 36-45.
Stanton B, Black R, Engle P and Pelto G. 1992. Theory-driven behavioural intervention research for the control of
diarrhoeal diseases. Soc. sci. & Med. 35(11): 1405-1420.
Steckler A, McLeroy KR, Goodman RM et al. 1992. Towards integrating qualitative and quantitative methods: An
introduction. Hlth. Educ. Q. 19(1): 1-8.
Stone L. 1992. Cultural influences in community participation in health. Soc. Sci. & Med. 35(4): 409-417.
Subedi J. 1989. Modem health services and health behaviour: A survey in Kathmandu, Nepal. J. Hlth. Soc. Beh. 30: 412420.
Taylor TE and Molyneaux ME. 1988. Cerebral malaria in children: Presenting features and prognosis. Malawi Med. Q.
5:3-11.
The Lancet. 1983. Malaria control and primary health care. The Lancet April 30: 963-964.
Thompson SG. 1990. Speaking "truth" power: Divination as a paradigm for facilitating change among Giriama in the
Kenyan hinterland. PhD. Thesis. School of Oriental and African Studies, London.
Trape JF, Zoulani A and Quinet MC. 1987. Assessment of the incidence and prevalence of clinical malaria in semi-immune
children exposed to intense and perennial transmission. Amer. J. Epid. 126(2): 193-201.
Ugalde A and Homedes N 1988. Medicines and rural health services: An experiment in the Dominican Republic. In Van
der Geest S and Whyte SR (eds); The context of medicines in developing countries: Studies in pharmaceutical
anthropology. Kluwer Academic Publishers, Dordrecht.
Van der Geest S. 1987. Pharmaceuticals in the third world: The local perspective. Soc. Sci. & Med. 25(3): 273-276.
47
Van der Geest S. 1988. The articulation of formal and informal medicine distribution in South Cameroon. In Van der Geest
and Whyte SR (eds) The context of medicines in developing countries: Studies in pharmaceutical anthropology. Kluwer
Academic Publishers, Dordrecht.
Van der Geest S. and Whyte S. 1989. The charm of medicines: metaphors and metonyms. Med. Anthrop. Q. 3: 345-368.
Velema JP, Alihonou EM, Chipaux J et al. 1991. Malaria morbidity and mortality in children under three years of age on
the Coast of Benin. Trans. Roy. Soc. Trop. Med & Hyg. 85: 430-435.
Walt G. 1988. Community health workers: Policy and practice in national programmes: A review and selected
annotations. EPC Publication No. 16. London School of Hygiene and Tropical Medicine, London.
Warrell DA, Molyneux ME and Beales PR. 1990. Severe and complicated malaria. Trans. Roy. Soc. Med. & Hyg. 84(2)
suppl: 1-65.
Watson E. 1984. Health of infants and use of health services by mothers of different ethnic groups in East London.
Community Med. 6(2): 127-135.
Werner O and Schoepfle GM. 1987. Systematic Fieldwork (Vols. 1 & 11). Beverly, CA, Sage Publications.
Whisson MG. 1964. Some aspects of functional disorders among the Kenya Luo. In Kiev A (ed) Magic, faith, and healing:
Studies in primitive psychiatry today. The free Press of Glencoe, New York.
White NJ and Ho M. 1992. The pathophysiology of malaria. Advances in Parasitology 31: 82-173.
WHO 1988a. The use of essential drugs. Third report of the WHO Expert Committee. Technical Report Series No.
770. WHO, Geneva.
WHO 1988b. The world drug situation. WHO, Geneva.
WHO/UNICEF. 1978. Primary Health Care. Report of the International Conference on Primary Health Care, Alma-Ata,
USSR. WHO, Geneva.
Williams GH and Wood PHN. 1986. Common-sense beliefs about illness: A mediating role for the doctor. The Lancet
December 20/27: 1435-1437
Woelk GB. 1992. Cultural and structural influences in the creation of and participation in community health programmes.
Soc. Sci. & Med. 35(4): 419-424.
Wyke S, Hewison J and Russell I. 1991. Children with cough: who consults the doctor? In Wyke S and Hewison J (eds)
Child health matters. Open University Press, London.
Yach D. 1992. The use and value of qualitative methods in health research in developing countries. Soc. Sci. & Med.
35(4): 603-612.
Young A. 1987. Peace-time and past-time in the clinical construction of combat-related post-traumatic stress disorder. Cited
by Rogers Stainton W. 1991 - Explaining health and illness: An exploration of diversity. Harvester-wheatsheaf,
Hertfordshire.
Zaidi S.A. 1988. Poverty and disease: Need for structural change. Soc. Sci. & Med. 27(2): 119-127.
48
T
Abbreviations
AM
Antimalarial Drugs
BI
Bamako Initiative
CBHC
Community-based Health Care
CDC
Centres for Disease Control, Atlanta, USA
CHW
Community Health Worker
HE
Health Education
HF
Health Facility
HP
Health Promotion
HSB
Health Seeking Behaviour
INRUD
International Network for the Rational Use of Drugs
KAP
Knowledge, Attitudes and Practices
KDH
Kilifi District Hospital
KEMRI
Kenya Medical Research Institute
OTC
Over-the-counter (drugs)
PEM
Protein Energy Malnutrition
PHC
Primary Health Care
RO
Retail Outlet
ROP
Retail Outlet Proprietor
SES
Social Economic Status
SAP
Structural Adjustment Programme
TH
Traditional Healer
VHW
Village Health Worker
WHO
World Health Organisation
Position: 477 (8 views)