RF_WH_18_SUDHA.pdf

Media

extracted text
RF_WH_18_SUDHA

UJH I
A
AIDS Education and Prevention, 7(1), 74—1995
© 1995 The Guilford Press

r-.<<?:;>:•

DETERMINANTS OF SAFE AND
RISK-TAKING SEXUAL BEHAVIOR
AMONG GAY MEN: A REVIEW
Harm J. Hospers and Gerjo Kok

......I

i

Since the outbreak of the AIDS epidemic, gay men have made impressive
changes in their sexual behavior. Notwithstanding these changes, there
are several studies that show considerable within-subject variability in
risk categories, indicating that for a substantial number of gay men it is
difficult to be consistently safe. A large number of studies have examined
the relationship between sexual behavior and determinants of that be­
havior. This review reports and discusses the findings with regard to
demographic, situational, behavioral, and psychosocial determinants.

Since the outbreak of the AIDS epidemic in the beginning of the 1980s. an estimated
12.9 million people around the world have been infected with the Human Im­
munodeficiency Virus (HIV) by early 1992 (Mann, Tarantola, & Netter. 1992). In
several parts of the world, gay men comprise the group that has been hit hardest by
the disease. Estimated proportions of gay men of the total number of HIV-infected
people per region arc 47% for Western Europe, 54% for I^tin America, 56% for
North America, 80% for Eastern Europe, and 87% for Oceania (Mann et al.. 1992).
At this moment primary prevention is the only possible way to limit the spread
of HIV. In their effort to influence the course of the disease, social scientific
researchers have devoted considerable attention to the identification of determi­
nants that distinguish between safe and risk-taking sexual behavior of gay men,
because insight into these determinants is vital for the development of effective
AIDS prevention interventions. In this review we present the knowledge that is
currently available. Before giving an overview of determinants we briefly recapitu­
late the behavioral changes among gay men that have occurred as a result of the
epidemic.

The authors are with the Department of Health Education. University of Limburg at Maastricht, the
Netherlands.
This work was supported in part by Praeventiefonds grant 28-2356 and the Ministry of Welfare. Health,
and Cultural Affairs of the Netherlands.
The authors wish to thank the anonymous reviewers for their valuable comments on earlier drafts of
the article.
Address reprint requests to Harm Hospers, Department of Health Education. University of Umburg,
P.O. Box 616, 6200 MD Maastricht. The Netherlands.

74

j

BEHAVIORAL CHANGES AS A RESULT OF AIDS

1.8

Once the routes of viral transmission were known, two major recommendations ior
gay men to prevent infection were given: 1) reduce the number of partners, 2)
refrain from insertive and receptive anal intercourse, or use condoms when having
insertive or receptive anal intercourse (McKusick, Conant, & Coates, 1985; Solomon
& Dejong, 1986).
Many studies have documented drastic reductions in the number of partners
(Centers for Disease Control. 1985; Joseph et al., 1987a; McKusick, Conant, &
Coates, 1985; Martin. 1987; Schechter et al., 1988; van Griensvcn et al., 1987;
Winkelstcin ct al., 1987). This reduction has resulted in a considerable decrease of
the probability of infection at an aggregate level. However, compared with con­
sistent condom use, reducing the number of sexual partners is far less effective
(Reiss & Leik, 1989). Especially in high-prevalence areas, reducing the number of
partners only leads to a marginal reduction of risk (Handsfield, 1985). Thus, the
‘reduce-your-numbcT-of-partners* approach may have the serious drawback that
individuals might use their compliance with this recommendation to justify the
high-risk behavior in which they still engage.
The majority of studies report impressive changes with respect to unprotected
anal intercourse (see Becker & Joseph [1988]; Stall, Coates, & Hoff [1988] for
reviews). Siegel. Bauman, Christ, and Krown (1988) showed that 22% of their
sample shifted to safer-sex behavior over a period of six months. Winkelstcin ct al.
(1987) reported a decrease of 60% in the prevalence of risk-taking sex over a
two-year period and Martin (1987) found a reduction of 70% in risk-taking sexual
episodes since respondents heard about AIDS. These changes in risk-taking behavior
were accomplished by either refraining from anal intercourse or using condoms
with anal intercourse. Although condom use with anal intercourse is now generally
accepted as an adequate means of reducing risk, failure rates of condoms, (breakage,
improper use) van' around 10% (Ross, 1987; Valdiserri et al., 1988; van Griensvcn &
de Vroomc, 1987) and thus do not totally eliminate risk of HIV infection.

-4 >

z

3§3

y

= x -m e = > “r o x
— w r* x !N w
r- ? r-

J P $'

•5 >

I2

5

o “

o

g
tc O

'g ■o

si

2

SO
u o

i

h

5 §

5i

S

□1
w a

£ 7. ? z £ r

•go

z5r

n o

E 5
P E

1

= 5

hi —nil
hi
—. - - — —



o •o

B- C
O *

X X = -J 5 5 3 2 I §

h

r^xcvxr-e^xx
xxxxxxxxxx

Hi

« E

BEHAVIORAL CHANGES ACROSS TLME

eS

Sa §

A Jess promising picture emerges when the available data are inspected on an
individual level rather than on an aggregate level. A considerable number of people
continue to engage in risk-taking sexual behavior. It also seems that people may shift
from one risk category to another over time. Table 1 provides an overview of
longitudinal studies that examined these kinds of changes. All studies compared
individual sexual behavior at two points in time, with varying intervals. Respondents
were categorized as either low or high risk at both time 1 and time 2. Cross­
tabulation of time 1 with time 2 behavior results in four categories: low risk at both
time 1 and time 2 (stable low risk), high risk at time 1 and low risk at time 2 (change
to low risk), low risk at time 1 and high risk at time 2 (change to high risk), and high
risk at both time 1 and time 2 (stable high risk).
Despite the differences in study characteristics, some general conclusions may
be drawn from these studies. First, it is clear that a great number of gay men have
either maintained safe sex behavior or have changed to safer sexual behavior. In fact,
many of them have successfully maintained this behavior for long periods of time.
Second, a substantial number of men reported one or more instances of risk-taking

— v

2 E N
.2 P ~U
2 •= £
u (g H
a _ o
•= « ~

s
- = pE

£ S 8
v
£ & >

!

s

1

II
lit

L IL - tii
•m

“5?

> - 5
u

I

~

x
?
5 £ = =?

pl c -= s 5 =
■=

Z •= z z

I z. •= = z i

.■ m

tr. “

I •





82 5-"

lih 2 Si

O

Hh

n

II
& 2 -3 H
2 g G n
“32a

iihi

3 X
3

2o

k
£
Q v)

>

2

sh
I =i £
• g §• §
13

th
n O n

a >3
K ? *5*
&3.§
a K
n r» »•
S. a- S

■S § 2-

In

■St S 3

?|

tn



S- tn

c r»
p 3
IT. w

g □

n

XT 3-

o> S

•u n

S <5
° n
Q.

r> m
<z> c.

-1

•-I

11n

Table 2.

n

A ° M

Hii

I

o
►rt

I

- §* 3 o S R §. 3

HH

Ia
c/>

5H<I

«• x " -5 <

I 2
Q

n 3. b> 2.

iVsh

«£:go-»n

I !hs

3 S3
5> c
3^

II

IIn
3=

n 2 7T

I & n IHI

aH
2 ° S
It'S 3
2 Er ?

~ -o 2 =• =:


ah
n

2 3
•5

a
I

3 q
•< P c
£ §• *' >^3
3
is- o' -5 “ °
tn
r n

n n

p

O

^50 □<§ £ S

8 2.“

3

<z>

o

a = q X' 5 " o
^3^35^35^0
&85S9§g8
□ 5*0 a gr 2 □
tn
- • “ 3 2 S"2
2
-S == gB gs III-£ ==■ os 2 -& E eg
tn
| & 5 | S'
g-ss-s-^ofza
X
’o § o 5 t?2 § S = g a g 5 S'
P.

!-HhH

hH

£> §■ 3 £ I feg §
» B 2= Sa * §■«
s s p o\§ £ * §-3. ^"2


-------| 5 P - 3 “
2.; * *■
>
S P <2
£
tn a “ c
S J | o- ■§ '’> = l § s 3 a
5 a 3- § H. g § -. - ”
H' 111 i-8■2. 3§ S p
<• o

o
73

f? ?

2-• 5
<P ^ 3 o
P <F> D d , 3-

V

<

•"*

1

S n 3
n 5
a. p p
r

° □

- r? n

- "> tn

■a
i | s 4POt
r

l^r
"Hrgs
I ?aKi £ n
tn y

O 3 c

5?2.oe-5’

O. 2 3

?IS

8 ?jIa
|l h
F-3 ? a a 2,

2

_ EL v)

&Z.M'
5- 2/,g3aq;2
Z I s g-

thiP



Ihls

3 9 p

z

tt

5
c

a jp g|a

O' 3

n

| ?
g.&5
? 5 s § i3 iI sa S- B

ih

ppp

S’ K- 1

I M-h-*8
P § i. ? s 15- a | 8 & 3 a.

2 □ rt

e Q

o g. ? g g- 2.

x IHh?
X
p.

xi.

Q 3 ‘g ” &

C

I
I nil >?5ir
P >
c
I5H n?
Pl S? X5 2 2 I o. - MIH-a H I" - I
S x

I £
i °u
£= a •? i gi §■8 3 S? I o8 wm
!
8? 3-iB.ga
|h p
S3?P .p 3 3 2 2.
o

!

O’

C. C

2*
O

£

3 ■

a” 5-a o ?

’ ; 1 3 9
. “5 £1 8 a 3 s

ssx.
5Ht<Hi|;
g.
SifsIlhiHl!
o ^o
2 ‘oiocg
n
o o □’p52=
o. <n p □' £“

&n 2 a o JP ”□&*</» O 3 -2

Studies on Determinants of Safe and Risk-taking Sexual Behavior among Gay Men

Dcinograplilc

Alcohol/drug use

Hclalion-

C/l

<v

Age

Incuinc

Race

Education

Sllip

Risk
Alcohol

Drug

noth

Nitrites

Knowledge

I’erception

Altitude

Norms Self Efl

Sliuly and behavioral tom-

parison

Abid ct al. (1991)

(change to URAI versus

402

0

0

0

0

402

0

0

0

0

0

0

0

0

maintenance of SB)
(change to UIAI versus

0

maintenance of SB)
Bauman & Siegel ( 1987)
(from SB to RTB, three

153

t)

levels)
Bochow (1990)
(UAI versus Al with con­

c

333

dom )
Connell ct al. ( 1990)

(UAI versus no UAI)

394

o

o

O

I)

Dallas (1990)
(Al versus no Al)

70

Ekslrand ft Coates (1990)
(UAI versus no UAI)

(Maintenance of URAI or

6H6
L

592

<l

UIAI versus maintenance of

SB or change to SB)
Fitapatrick cl al. (1989)

225

(Al versus no Al)
Fitapatrick ct al. (1990)

(RAI versus no RAI)

<:

356

c

IHI

c

99

0

o

(•old ct al. ( 1991)

(UAI versus no UAI)

0

0

0

Hays. Kcgcles & Coates
(1990)

(UAI versus no UAI)

0

0

o

0

4/0

o

Ul



(xapui 03)

(2661 ) smrpw V s«>M
(1U.H J” aaueuai

ottz

-u|ru> snu3A US °> aHurqi)

(0661) p: P mojiso
(lYD oi aSurqo

o

0

0

o

0

0

0
0

0

SnSJ3A ns JO 33UrU3>U|tUl)

i

(AS oi aSuxqa

0

snsiOA |Vf) JO 33UCU31Up:U>)

PM
HOf

(IVO ou snsjOA ivn)

3

(0661) Tt P >PI*n>pW

0

0

6ZI

(03 snsasA 03 ou)

3

(6861) Tt P

'Pipnaow
(flS o» aSutq}

0

0

0

0

66

1

fntjSA' IVin JO 33UtU3>U|CUl)



T

•J3A IVUn jo asinruaiurrui)

(OS o> dSueqo sns
0

0

(6861)
|t >3 pjtppoij 'J3>t*n>>W
(S|3A3|
Z18

0

0

Jnoj •fjj.H o> ns moq)

3

(6861) |t P “un

szt

(ivn ou snsjoA ivn)
( 1661) le P -*113)1

(ivun ou SOSJ3A IVHH)

ZS«
ZSf

(IVia ou sn<j3A IVin)

3

(0661) I® P uo«u3A3is
ppijnJO •33u3Jmn >s •Xiph

0
0

0

Z6Z
IHJ-

(xspuj ns)

3
3

(3JO3S X3pu|->|S|J)

(0661) 1« p 3>|U131
'ppu-nug *33U3JMtn 'IS '^IPX
(tlS JO 33ueu.siu|cui

tnsJSA ivn o> 38ueq3)

TO

0

0

0

(1661)
ppiJSTJIl >g 3DU31MI-] IS ‘AlPX

^>»3U/3-fl’S su*JON

spnipiv

UO|ld33J3J

oSp.ii.wouji

ssipiiN

l||Ofl

8iuO

uoponpii

diip

|0l|<O|V

33CH

3tU(>3U|

3«v

A/

V3

-uoiirpM

Wil

3!1|cIcj8oiu3q

3sn Snjp/iotio.iiv

p;)rup?0? ■?
Table 2.

Continued
Demographic

Alcohot/drug use
Kclation-

C7L

N

C

481

L

100

c

46}

C

955

c
c

611
611

Age

Income

Race

lidncation

Ship

Risk
Alcohol

Drug

Uoih

Nitrile*

Knowledge

Percept km

Atliludc

Norms Sctf Efficac-

Scagc ct al. (1992)

(URAI versus no URAI)

0

Siegel ct al. (1989)
(consistent SB versus con­

0

o

+/0"

0

0

sistent RTD)
Stall et al. (1986)

(low versus medium versus

high RTD score)
Valdiscrrl cl al. ( 1988)

(consistent CU versus never

0

0

0

0

CU)

de Wit. Sandfori ct al. ( 1991)

(Al versus no Al)
(consistent CU versus in­

0

consistent CU)

For each study is listed which behavioral comparison was made, whether this comparison svxs cross-sectional «>r longitudinal (Czl.). the number of subjects for the comparison
(N). ami the relationship with determinant*. For each determinant that was studied, is listed svhcihcr there is a positive relationship ( + ). no relationship (0). or a negative­
relationship (-) with safe sexual behavior (sec coding legend below for detailed information).
Abbret’lallmii- Sil - safe beliavior, RTD » risk taking Iwhavior. Cl' ■* condont use. Al - anal imcrciHirse. DAI » mipr.Hecu-d anal liitcrc.nirse. UltAI - unprotected receptive
anal ititcrctHtrsc. UIAI - unprolcctcd inscrlivc anal intercourse.
A-o/ex- ‘conimunication skills < , sclf-cllicacy• scale o
••emotional support ♦. network aflili;
ialion 0
CODINC. l.l-GliND
DemojjnilMc tur/uhles:
age: + means older age related Io SD
income: + means higher income related to SD
race: + means white more SB than other ethnical groups
education: f means higher educational level related to SB
RrlatiMUbip ittilut + means sex with regular partner safer
Mcoholhlntg use:
alcohol: 4 means (more) alcohol use related to SD
drugs: + means (more) drug use related to SD
both: + means (more) alcohol and/or drug use related to SB
nitrites: ♦ means (more) nitrite use related to SB

A'»n.«-/<v/ge i means higher knowledge related to Stl
Kish l’frcvl>litin: + means men who behave safe perceive less risk to getting
infected with HIV or perceive more risk of unsafe sexual practices
Atlilmle ♦ means cither positive attitude towards SB related to SB or posi­
tive attitude towards RTII related to RTII
SsM'itil minus: ♦ means either positive social mtrms towards Sil related to Sil
or positive social norms toward Rill related to Rill
Sclf e/ficac)- ♦ means higher self-efficacy related to SB

DETERMINANTS OF SAFE AND RISK-TAKING BEHAVIOR

frequency as
activity for different timeframes. The test-retest procedure for ’
s: 0.47-0.62)
well as number of partners for several practices were moderate (k.
(Saltzman et al., 1987). Kauth, St. Lawrence, and Kelly (1991) obtained self-reports
of the frequency of several sexual practices for three overlapping timeframes in the
past: two weeks, three months, and twelve months. They found that the frequency of
infrequent sexual practices showed the greatest consistency across timeframes.
They further found that the average correlation over all practices decreased signifi­
cantly as the period of recall increased: the average correlation between two weeks
and three months was 0.73 while the correlation between two weeks and twelve
months was 0.31. They conclude that self-report measures of sexual behavior should
use relatively short retrospective periods. Based on a review of studies of response
bias in sex research, Catania, Gibson. Martin, Coates and Greenblatt ( 1990) advise
the use of 1- or 2-month recall periods. Another study compared agreement be­
tween 1- and 12-month recall periods with several cognitive strategics used to recall
sexual practices and partners. They found that the degree of consistency was lowest
for non-person-oriented strategics (such as thinking of places or locations where
partners were met), and highest for person-oriented strategics (such as names or
faces of partners) (Blake, Sharp, Temoshok, & Rundcll, 1992).
To summarize, the few studies which were conducted on validity and reliability
of self-reported sexual behavior highlight the limitations of these measures. More
methodological research is needed to optimize our measures of AIDS-related be­
haviors. Meanwhile, researchers in this field should tty to increase recall accuracy by
using relatively short recall periods. One way to do this is to use person-oriented
strategies to estimate the frequencies of sexual practices and the number of partners.
To reduce self-presentation bias, Catania. Gibson. Chitwood et al. ( 1990) suggest the
construction of a nonjudgmental context, the use of open-response formats, the use
of a sexual terminology that the group under study is comfortable with, and
maximizing respondent privacy.

is complicated. Although homosexual behavior is not limited to the segment of the
population that labels itself as homosexual (Kinsey, Pomeroy, & Martin, 1948), it is
usually this segment that is studied. As Doll et al. (1992) have demonstrated, there
are important differences between homosexually and non-homosexually identified
men who have sex with men. In their study, homosexually identified men were
more likely to belong to gay organizations. Furthermore, significantly more heterosexually identified men reported engaging in unprotected anal intercourse.
Second, the majority of studies included in this review reported on samples that
consisted of self-selected individuals. As Davies and colleagues (1993) have argued,
those samples may overrepresent self-confident open homosexuals, as homosexual
behavior continues to be stigmatized. In addition, studies that looked at differences
between volunteers and non-volunteers found substantial differences in sexual
behavior and attitudes toward sex (see Catania, Gibson, Chitwood, & Coates [1990]
for an overview). Furthermore, within studies included in this review, samples are
predominantly white and overrepresent individuals with higher education and
income. For example, according to the 1990 ILS. census, whites make up 80% of the
United States population (Bureau of the Census, 1992). By contrast, in all U.S. studies
shown in Table 2 (except Hays. Kegeles, & Coates 11990], and Linn et al. [ 1989]),
the proportion of whites exceeds 90% (range 91-96%). Finally, the majority of
studies were conducted in geographical areas with high AIDS prevalence rates
among gay men. These biases may constitute drawbacks for the generalizability of
results. The sampling strategy employed in some studies to reduce the effects of
selection bias, is to target as many segments of the gay community as possible, such
as members of gay organizations, patrons of gay bars, saunas and discos, and health
and STD clinics with a high proportion of gay clientele.

VALIDITY AND RELIABILITY OF SELF-REPORTED BEHAVIOR
Without exception, the studies that were selected for this review had to rely on
self-reports from respondents concerning their sexual behavior. As Catania et al.
(1990) have noted, a solid validity index for individual sexual behavior does not
exist. One could use other data to corroborate self-reported sexual behavior, such as
incidence of sexually transmitted diseases (STDs) or number of condoms sold.
However, the authors conclude that these indices are also biased, since STDs do not
have a one-to-one relationship with frequency of sexual behavior and condom sales
do not have a one-to-one relationship with use.
Several methods have been used to get an impression of the validity of data on
self-reported sexual behavior. Firstly, studies used sexual behavior data from partn­
ers of respondents. One study found moderate to high respondent-partner agree­
ment with respect to frequency of receptive and insertive anal and oral sex (Spear­
man’s r. 0.56-0.81) (Scage et al., 1992). A similar study investigated the agreement
on frequencies of sexual activities between men with either AIDS or AIDS-related
condition and their sexual contacts (Coates et al., 1988). Agreement on the
frequencies of 17 sexual activities was moderate to high (Spearman's r. 0.49-0.91).
However, agreement significantly decreased as the lapsed time between the in­
terview and the date of the last sexual encounter increased. The authors of both
studies conclude that the results indicated that self-reported sexual behavior data
were reasonably valid for use in epidemiological analyses.
A few studies have examined the reliability of self-reported sexual behavior by
either using a test—retest procedure or comparing retrospective reports of sexual

I/

HOSPERS AND KOK

82

81

DESIGNS
Research on determinants of safe and risk-taking sexual behavior among gay­
men usually used cross-sectional designs, longitudinal designs, or a combination of
both. Cross-sectional studies generally report comparisons between respondents
who practice safe sex and respondents who engage in high-risk sexual acts (with
varying definitions of these behaviors). Longitudinal studies usually use data from
cohorts that were established in the AIDS epicenters (e.g.. San Francisco, New York.
Amsterdam). These studies generally report correlates that arc associated with
changes in these sexual behaviors over time.
BEHAVIORAL MEASURES

While all the studies in Table 2 examined self-reported safe and risk-taking
sexual behavior, there is considerable variation in the operational definitions of
these behaviors. The majority of cross-sectional studies compared men who had
engaged in unprotected anal intercourse with men who had not engaged in anal
intercourse or who had consistently used condoms. Some of these studies di­
fferentiated between receptive and insertive anal intercourse. Another group of
studies examined the differences between men who consistently used condoms and
men who did not consistently use condoms. Finally, one study used a risk-index

I

'I

1
1

DETERMINANTS OF SAFE AND RISK-TAKING BEHAVIOR

respondents showed higher levels of unprotected anal intercourse ...an older re­
spondents in both 1985 and 1988. De Wit and colleagues (1991) found that the
younger men in their sample had more anal sex and used condoms less consistently
than older men. Similar results were reported by Fitzpatrick et al. (1990), Kelly et al.
(1990), Kelly, St. Lawrence, and Brasfield (1991), Kelly et al. (1991). McKusick ct
al., (1990), and Ostrow et al. (1990).
Few authors have tried to explain why younger men have this tendency. As a
possible answer, Ekstrand and Coates ( 1990) suggest that younger men might have
less social support for practicing safe sex, they might perceive themselves to be
invulnerable, or they might have poorer skills for buying condoms or negotiating
safe sex.
The abovementioned studies reported an association between age and risk­
taking behavior. It is important to note that mtn who engaged in risk-taking sex
were younger, but not necessarily young. There are two studies that specifically
looked at sexual behavior of young gay males. Rothcram-Borus and Koopman
(1991) studied sexual behavior of 59 black and Hispanic gay adolescents (ages
ranging from 14 to 18 years old). Fourteen adolescents (24%) had engaged in
unprotected anal intercourse in the previous three months. Hays and associates
(1990) compared low versus high risk-taking among gay men aged 18-25. They also
found high levels of risk-taking behavior: 43% of their sample had engaged in
unprotected anal intercourse in the previous sLx months. A comparison of low risk
takers with high risk takers showed that the latter group reported more enjoyment
of anal intercourse without condoms, perceived less risk with regard to unprotected
anal intercourse, and reported poorer communication skills with sexual partners.
Hays and colleagues (1990) conclude that special efforts should be made to
reach younger gay males. They also add that if interventions succeed in reaching
young gay males, the likelihood of low-risk behavioral maintenance might be higher
than usual since young gay men with relatively little sexual experience have had less
opportunity to develop anal sex as a habitual behavior.

(frequency of unprotected anal intercourse x number of intercourse partners)
(Kelly et al., 1990), and two studies used a condom use index (proportion of anal
intercourse occasions when condoms were used (Kelly, St. Lawrence, Brasfield,
Lemke et al., 1990): sum of frequency of condom use during receptive and insertive
oral and anal intercourse (Ross & McLaws, 1992)).
Most longitudinal studies compared men who maintained high-risk behavior
over a period of time with men who maintained low-risk behavior or changed to
low-risk behavior. It is apparent that this variety in operational definitions provides
us with detailed information on the sexual.behavior of gay men and the changes in
this behavior as a result of AIDS. However, it also limits the comparability of results
across studies.

DEMOGRAPHIC VARIABLES
The value of sociodemographic variables for behavior change interventions is lim­
ited, because these factors themselves are, in general, not easily influenced. De­
mographical characteristics of target groups are primarily used to guide the develop­
ment and implementation of interventions. Relatively few sociodemographic factors
have been investigated and even fewer seem to peruin to risk behavior. For
example, studies in general find no relationship between risk-taking or preventive
behaviors and income (Adib et al., 1991; Connell et al., 1990; Fitzpatrick et al.,
1990), or educational level (Adib et al., 1991; Connell et al., 1990; Hays et al., 1990;
Kelly, St. Lawrence. & Brasfield, 1991; Kelly, St Lawrence, & Brasfield, Lemke et al.,
1990; McCusker, Zapka, Stoddard, & Mayer, 1989; Valdiserri et al., 1988).
With respect to the relationship between risk-taking behavior and ethnic origin
of respondents, it must be concluded that little is known. With respect to the studies
in Table 2, all studies conducted outside the United States and a minority of studies
conducted in the United States do not report the racial or ethnic origin of their
respondents. Of those studies that do provide such information (12 studies), four
included it as a variable in their analyses. These analyses showed no relationship
between ethnic origin and risk-taking behavior in two of the four studies (Adib et al.,
1991; Valdiserri et al., 1988). In the study by Hays et al. (1990) a higher percentage
of non-whites engaged in high-risk behavior compared with whites (30% versus
14%, respectively). This difference approaches significance (OR = 1.61, 95% CI =
0.98-2.64, p < .06). Finally, the study by Linn et al. (1989) showed a highly
significant relationship between race and level of risk-taking behavior. Their unsafe
category contained 59% of whites, 69% of African Americans, and 81% of Latin
Americans. It is possible that these dissimilar findings are a result of the varying
proponions of non-whites in these samples. The percentage of non-whites was 8%
in the study from Adib et al. (1991), 4% in Valdiserri et al. (1988), 21% in Hays et al.
(1990), while non-whites made up one third of the sample of Linn et al. (1989). Low
proponions of non-whites result in relatively wide confidence intervals around
association estimates which may account for the nonsignificant results.
Two demographic factors, age and geographical location, seem related to AIDS
risk behavior. They will be discussed below.

GEOGRAPHICAL LOCATION­
Several authors have suggested that gay men in low AIDS prevalence areas
engage in more risk-taking behavior (Jones et al.. 1987; Kelly. St. Lawrence, Brasfield,
Stevenson et al., 1990). While these studies looked at low-prevalence areas only,
there are two studies which compared data from low-prevalence areas with data
from high-prevalence areas: one in the United Kingdom (Weatherburn et al., 1991),
the other in the United States (St. Lawrence et al., 1989). The results of the British
study showed that men in London were significantly more likely to use condoms
compared with men who live outside London (49% compared to 37% ). The results
from the American study showed significantly more occurrences of unprotected
anal intercourse, and significantly fewer occurrences of low-risk sexual practices in
low-prevalence cities.
The authors of the abovementioned studies relate these findings to the fact that
many health education initiatives are targeted at homosexual men who live in AIDS
epicenters, while men who live outside these areas have less access to organizations
for information and support. Furthermore, both St. Lawrence et al. ( 1989) and Jones

AGE
Several studies have found younger age to be related to risk-taking sexual
practices. Ekstrand and Coates (1990) found in a longitudinal study that younger

’Since there are only two studies available that compare low-prevalence areas with high-prevalence areas,
this determinant is not shown in Table 2.

II
I

I

HOSPERS AND KOK

84

83

1

I

DETERMINANTS OF SAFE AND RISK-TAKING BEHAVIOR

85



et al. (1987) presume that the relatively low number of persons with AIDS in these
areas results in an underestimation of risk of infection which in tum leads to
persistence of high-risk behavior.

86

HOSPERS AND KOK

none. In other studies the relationship between drug use ant
brisk sexual
behavior was absent (Gold et al., 1991; Seage et al., 1992). A nu...uer of studies
found that alcohol and/or drug use was associated with more risk-taking sexual
behavior (Kelly et al., 1991) and less condom use (Valdiserri et al., 1988), while two
other studies did not find an association (Adib et al., 1991; Connell et al., 1990).
With respect to the relationship between alcohol use and sexual behavior, the
results are not straightforward either. Some studies have found that alcohol use prior
to or during sex was related to risk-taking sexual behavior (c.g., Stall et al., 1986),
other studies could not discover this relationship in their data (e.g., Hays et al.,
1990). In a detailed analysis of sexual diaries of 430 gay men, Wcatherburn (1992)
found that in 250 sexual encounters that involved alcohol use during sex, 58
included anal intercourse. Of these encounters, 46 were with the regular partner
(35 exclusive), and in 10 of the 12 remaining encounters condoms were used. The
authors reject the hypothesis that alcohol use and risk-taking behavior are related.
One study found, contrary to expectations, that higher alcohol consumption at the
first measurement was related to the adoption of safer sex practices at follow-up
(McCusker, Stoddard, Zapka, Zorn & Mayer, 1989). Another study asked subjects to
recall their most recent safe encounter and their most recent risk-taking encounter.
Analyses showed that being under the influence of alcohol at the start of sex, and the
amounts of alcohol used during sex, did not distinguish between type of encounter
(Gold et al., 1991).
Investigations into the relationship between the use of inhaled nitrites and
protective sexual behavior have found either a negative relationship (Ostrow et al.,
1990; Seage et al., 1992; Stall et al., 1986) or no relationship (Gold et al., 1991; Kelly
et al., 1991; McCusker, Stoddard et al., 1989).
It is clear that we do not yet fully understand the relationship between sub­
stance use and risk-taking sexual behavior. Some authors who found a relationship
suggest that this.relationship might be causal (e.g.. Stall et al.. 1986). The same point
of view can be found in several publications on HfV-prcvcntion policy (c.g., Adib &
Ostrow, 1991; Peterman, Cates, & Wasserheit, 1992; Peterson, Ostrow, & McKirnan,
1991; Shernoff & Bloom, 1991). However, other authors have expressed doubts
about a direct causal relationship between substance use and engaging in risk-taking
sex (Gold et al., 1991; Siegel et al., 1989). They suggest that substance use and
risk-taking sex might be manifestations of some unknown variable, such as desire for
excitement or a predisposition to risk-taking behavior. Finally, a number of authors
explicitly doubt whether there is any association at all (Wcatherburn et al., 1993).
Summarizing, research on substance use shows equivocal results with regard to
its influence on risk-taking sexual behavior. Although a disinhibition explanation
seems appealing (it is in fact the reason that many men give for having engaged in
risk-taking sex), there is still uncertainty whether there is a direct causal relationship
between substance use and subsequent risk-taking sexual behavior.

SITUATIONAL AND BEHAVIORAL VARIABLES

RELATIONSHIP STATUS

With respect to safe and risk-taking sexual behavior three forms of relationships
are usually distinguished: 1) one regular partner, exclusive; 2) one regular partner,
non-exclusive; and 3) casual partners only.5 Studies that include measurements of
relationship status generally find that men who are in a primary relationship show
significantly lower levels of risk-taking behavior with casual partners than with their
regular partner (Bochow, 1990; Fitzpatrick et al., 1989; Fitzpatrick et al., 1990;
McKusick et al., 1990; Valdiserri et al.. 1988). More specifically, Doornbos (1992)
reports that 47% of the men who had one regular exclusive partner, 28% of the men
who had both a regular partner and casual partners, and 9% of the men who only
had casual partners, had engaged in unprotected anal intercourse in the previous six
months. De Wit, de Vroome et al. (1991) have shown that, while 37% of their
respondents quit having anal intercourse with casual partners, 91% continued this
practice with their regular partner. Thus, sexual behavior partly depends on the type
of partner involved.
One other point to keep in mind is that in some instances, the labeling of sexual
behavior as unsafe is unjustified. If men entered a monogamous relationship in the
pre-AIDS era or if both men in a monogamous relationship have been tested HIV
negative, it is clear that unprotected intercourse within their relationship carries no
risk of HIV infection.
Summarizing, it appears to be the case that a considerable proportion of gay
men have not only limited their number of sexual partners but have also reserved
unprotected anal intercourse exclusively for their primary relationship. Although
this might seem to be a sensible way to cope with the threat of HIV infection, it also
should be noted that as long as the HIV status of the partner is unknown, this coping
strategy does not guarantee that one will not get infected.
DRUG AND ALCOHOL USE

A large number of studies have examined the relationship between substance
use (alcohol, drugs, alcohol and/or drugs, and nitrite inhalants) and risk-taking
sexual behavior. As can be seen in Table 2, almost all studies (there is one excep­
tion) either report a negative relationship with protective behavior or fail to find a
relationship.
With respect to drug use. both Siegel and associates (1989). and Stall and
colleagues (1986) reported a relationship between marijuana use during sex and
subsequent risk behavior. Men who used marijuana were more likely to fall in the
high-risk category. Ostrow et al. (1990) found that men who used three or more
drugs were more likely to continue risk-taking sexual behavior than men who used

PSYCHOSOCIAL VARIABLES

In most instances psychosocial variables are explicitly derived from prevailing
(health) behavior theories like the Theory of Reasoned Action (Fishbein & Ajzen,
1975), the Health Belief Model (Janz & Becker, 1984), Social Learning Theory
(Bandura, 1986), and the Theory of Planned Behavior (Ajzen, 1988). In the follow­
ing section we will summarize research that included one or more of the following

is important to note that there are. of course, more wap of classifying the variety of gay male
relationships. The level of refinement depends, in part, on the kind of research questions under study. For
example, with respect to non-regular partners it might be useful to distinguish between anonymous and
non-anonymous sex partners.

1

I

1

i•«

?

DETERMINANTS OF SAFE AND RISK-TAKING BEHAVIOR

88

87

HOSPERS AND KOK

(1987) found that the risk estimates of respondents who practice
t-taking sex
were only slightly higher in comparison with safe and low risk groups in their
sample. Furthermore, they found that 51% of their sample rated their risk as be­
low average. The group mean of the estimated risk of getting AIDS relative to
other gay men was -.35 (range -1 to +1), showing the optimistic bias described
above.
Although the finding that risks are underestimated is worrisome in itself, the
implications for prevention are not ven’ clear. Mixed results have been found when
looking at the relationship between risk perception on the one hand and behavior or
behavior change on the other hand. Several studies found that risk perceptions
differed between"lovSTandTiigh risk-taking groups. Hays ct al. (1990) found that
high-risk takers, compared to low-risk takers, perceived unprotected anal in­
tercourse to be less risky, while they estimated that their current behavior placed
them more at risk for HIV infection. Kelly, St. Lawrence, Brasfield. Lemke ct al.
(1990) report similar results, that is, men who practiced unprotected anal in­
tercourse with multiple partners regarded their sexual behavior as more risky than
men who practiced safe sex. They also found that the risk-estimate decreased as the
proportion of protected anal intercourse (i.e., condoms were used) increased.
McCusker, Stoddard, et al. (1989) found that greater perceived susceptibility at
pretest was the strongest predictor of the adoption of consistent safe behavior at
posttest.
A number of studies found that risk perceptions did not distinguish between
safe andjisk-taking behavior (McKusick eF5!^ 1990; Siegel et al., 1989) or changes in
risk-taking behavior (Fitzpatrick et al., 1990; McKusick ct al.. 1990). In these studies
men who showed risk-taking behavior underestimated their risk to such extent
that their actual estimates did not differ significantly from men who practiced
safe sex.
A review of risk perception research does not clarify whether a relationship
with behavior or behavior change exists. There arc several possible explanations for
the variation in results. First, the operational definitions of risk perceptions vary
across studies. While in some studies respondents were asked to rate the riskiness of
their current sexual behavior (e.g.. Kelly et al.. 1991). other studies asked respon­
dents to rate their absolute chance of developing AIDS (Fitzpatrick ct al., 1990).
or their chance of developing AIDS compared with other gay men (Siegel et al.,
1989). Although the number of studies is limited, it seems that respondents have
a more realistic appraisal of their risks when they arc asked to judge their current
or recent sexual behavior than when they are asked to judge possible long-term con­
sequences of their behavior (i.e., developing AIDS). Also. motiyatjon_tQ..protect
oneself should be framed in terms of susceptibility to the risk behavior (Janz &
Becker, 1984; R. W. Rogers, 1983; Ronis, 1992). Thus, when studying AIDS-protec­
tive behavior, a question could be: “What is your risk of contracting HIV if you
engage in unprotected anal intercourse". If questions are framed in this way, one
would hypothesize that those respondents who engage in unprotected anal inter­
course have lower risk estimates than those respondents who do not engage in
anal intercourse or who consistently use condoms (which, in turn, explains cur­
rent behavior).
Second, the majority of studies have used cross-sectional designs. As Weinstein
(1989) has argued, even if a positive relation between risk perception and behavior
is found, the direction of causation is unclear. It could well be that behavior change
influences risk perception rather than vice versa.

psychosocial variables: knowledge, risk perception, attitudes, social norms, and
self-efficacy.

KNOWLEDGE

AIDS knowledge has been assessed with respect to a variety of domains. Most
frequently, studies have looked at knowledge of modes of HIV transmission related
to sexual practices, public health guidelines, and risk-reduction measures. Some
studies included measures of knowledge regarding the interpretation of HIV test
results, misconceptions about HIV and AIDS, and prevalence of HIV and AIDS among
gay men.
AIDS education efforts have undoubtedly had a positive influence on gay men’s
knowledge about the disease. Without exception, studies show very high levels of
knowledge (Dallas, 1990; Emmons et al. 1986; Joseph et al., 1987b; Kelly, St.
Lawrence, Brasfield. Lemke et al., 1990; McKusick et al., (1990); Siegel et al., 1989;
St. Lawrence et al., 1989; Valdiserri et al.. 1988). However, the available data on the
relationship between knowledge level and risk behavior do not show one clear
tendency. Although some studies have shown that knowledge is related to initial
behavior change, especially when AIDS was first discovered, it seems that at present
it has little predictive value for behavior change or maintenance. Due to ceiling
effects, knowledge levels can hardly change, which is probably why an increasing
number of researchers have stopped including knowledge as a variable in their
studies.
In sum. gay men arc well informed about the modes of HIV transmission, public
health guidelines, and risk reduction measures. However, the most important con­
clusion is that, given the complexity of the behavior change that is required, it is not
surprising to find that knowledge alone, though a prerequisite for change, is not
sufficient to enable behavior modification and maintenance. As has been demon­
strated with other health related behaviors as well, within any target group the first
distinction one has to make is between adopters and non-adopters (E. M. Rogers,
1983). Among adopters a further distinction can be made with respect to behavior
change. A number of people will indeed change their behavior on the basis of health
information that gets to them. However, the majority of this group requires more
than information to change their behavior. Supportive social norms and the acquisi­
tion of necessary skills arc examples of what is needed for them to change (E. M.
Rogers, 1983).
RISK PERCEPTION

Some behavioral theories, especially the Health Belief Model, explicitly
hypothesize that a sense of personal susceptibility to a disease (i.e., a perception of
risk), among other factors, influences decisions to undertake actions. Several studies
have investigated this presumed influence (see Table 2).
There are two types of risk estimates that are generally asked for in this type of
research: 1) absolute risk estimates (e.g., “What are your chances of getting infected
with HIV?"); 2) relative risk estimates in comparison to others (e.g., "What are your
chances of getting infected with HIV, compared to an average gay man your age?").
In both instances it is found that people systcma.lically underestimate their risk. This
finding is usually referred to as 'optimistic bias’ and is found with regard to many
(health-related) behaviors (Weinstein, 1987). For example, Bauman and Siegel

W

f

I
1

wmcKMlAAMi Ub SAFE AND RISK-TAKING BEHAVIOR

90

89

ATTITUDES

HOSPERS AND KOK

influence of social networks is the swiftness with which safe sex r* imendations
became known and adhered to by large numbers of gay men.
s also been
demonstrated that norms that prevail in peer groups influence the degree of AIDS
preventive behavior. A group norm prescribing safe sex has been found to relate to a
decrease in number of sex panners (Emmons et al., 1986). avoidance of anonymous
partners (Joseph et al., 1987a), safer sexual behavior (Kelly. St. Lawrence. Brasfield,
Lemke et al., 1990; Kelly, St. Lawrence, Brasfield. Stevenson et al., 1990). an inten­
tion to behave more safely in the future (Fisher, 1988). avoidance of anal in­
tercourse (de Wit, Sandfort et al., 1991), an intention to use condoms (Ross &
McLaws, 1992), and consistent condom use (de Wit, Sandfort et al., 1991). Adib and
colleagues (1991) report that the absence of peer support was a-significant pre­
dictor of relapse to unprotected anal intercourse. Prieur ( 1990) found that the most
noticeable difference between men who do not engage in high-risk sex and those
who do, is that many of the men tha£belong to the latter group have a loose social
network. Similar results are reported by Connell et al. (1990). They report that
involvement with gay community organizations is low for men who engage in
unprotected anal intercourse with casual partners. Fishbein et al. (1992) in­
vestigated the influence of attitudes and social norms on the intention to engage in
11 sexual behaviors, varying in AIDS risk. The study was conducted in three cities in
the U.S. (Seattle, Denver, and Albany). They found that the influence of social norms
was greatest in Seattle and smallest in Albany. They attribute this difference to the
fact that the gay community in Seattle is larger and better organized, which results in
well-linked interpersonal networks and more interpersonal interactions.
Hence, stimulating the development of social networks as well as integration in
a social network that enhances safe-sex norms may be a very powerful safeguard
against HIV infection.
Influencing social norms is easier said than done. Continuous efforts on a
community level, advocating adherence to safe sex recommendations are necessary
but not sufficient. Concurrent to this, small-scale activities have to be carried out to
support men seeking help in changing their high-risk behavior. This can be done
through small ^oupjnteryentions or face-to-face counseling. An example of this is
the “Stop AIDS” community discussion groups that have been conducted in many
AIDS epicenters (Miller et al., 1990). Another promising approach that Kelly et al.
(1992) have applied is the use of peer leaders who are trained to communicate
about safe sex with their peers. The results show impressive reductions in un­
protected anal intercourse.

Research on attitudes related to risk behavior shows the importance of anal
intercourse for many gay men. Comparisons between men who engage in risk-taking
sexual behavior and those who do not, show that the former regard anal intercourse
as more important and enjoyable (Connell et al., 1990; Hays et al., 1990; Kelly et al.,
1991), consider anal sex more as “ultimate intimacy" (Dallas, 1990), experience oral
sex as less satisfying (Dallas, 1990; Hays et al., 1990), and are less inclined to avoid
anal intercourse (de Wit, Sandfort et al., 1991). Men who reported that unprotected
anal intercourse is their favorite technique, were more likely to engage in un­
protected anal intercourse (McKusick et al., 1990). Gold and colleagues (1991)
showed that the desire to have intercourse without a condom, to have exciting sex,
and to have excitement in any way possible, were significantly greater in subjects’
most recent unsafe encounter compared with subjects’ most recent safe encounter.
Finally, Connell and Kippax (1990) report that 54% of their sample rated in­
tercourse without a condom as the most physically satisfying form of sex. Thirty-six
percent of their sample rated it as the most emotionally satisfying practice.
In sum, men who have positive attitudes toward anal sex are more likely to
practice unprotected anal sex. The qualitative studies of Prieur (1990; 1991) show
that anal sex is more than just an alternative from the whole gamut of sexual
activities. Whenever it is incorporated in the sexual repertoire it becomes valuable
to and partotwbat might. be odledagay identity.. Connell et al. (1990) argue that for
many men anal intercourse is a central part of being gay.
Similar results have been found with regard to attitudes toward condoms. That
is, men who have negative attitudes toward condom use are more likely to engage in
unprotected anal intercourse (Valdiserri et al., 1988). Using factor analysis, Ross
(1988) interprets the most prominent factor (accounting for over a quarter of the
total variance) as a negative attitude toward condoms. Items with high loadings on
this factor reflect the view that condoms are unerotic, unreliable and uncomfortable.
Clearly, neither a positive attitude towards anal intercourse, nor a negative attitude
toward condom use promotes safer sex.
There are two possible avenues one may take if one wants to promote safer sex
through attitudinal change. First, one might try to change these positive attitudes
toward anal sex. The aim of such interventions would be to have gay men refrain
from anal sex. Second, one might try to change attitudes toward condom use in a
positive direction and preserve positive attitudes toward anal sex. The obvious
advantage of the former is that refraining from anal sex would eliminate the risk for
HIV infection (assuming that no other risk behavior is prevalent). However,.givxn
th^sigHificance of anal intercourse, many have questioned the feasibility of this,
approach and have relied on the latter strategy, that is, promoting condom use. The
advantage of this approach is that the positive attitude toward anal intercourse can
be left unaffected. This is in line with E. M. Rogers’ (1983) argument that inno­
vations have to be compatible with existing values, past experiences, and needs of
potential adopters to facilitate adoption. The disadvantage of this approach is that
men might judge their behavior as completely safe, while it is known that condom
failure rates are substantial.

SELF-EFFICACY
Self-efficacy is a person's evaluation of the extent to which he is capable of
exerting a certain control over his behavior (Bandura, 1986). The value of selfefficacy estimates is more than a reflection on one’s past behavior; it also has a
predictive value for future behavior. With respect to AIDS related behavior, selfefficacy is a person’s evaluation if, and to what extent, he has the skills to exercise
control over sexual situations (Bandura, 1989). The relationship between selfefficacy and AIDS risk behavior has repeatedly been demonstrated. De Wit, Sandfort,
et al. (1991) found that self-efficacy in using condoms was the strongest predictor of
consistent condom use. Other studies have shown that a non-assertive attitude
predicted relapse to unprotected anal intercourse (Adib et al., 1991), and that men
who engaged in unprotected anal intercourse rated their sexual communication

SOCIAL INFLUENCE AND NORMS

Norms and values that are supported by peers can have a substantial influence
on one’s behavior (see Fishbein & Ajzen, 1975). One of the best examples of the

i
fi
1

I
I

DETERMINANTS OF SAFE AND RISK-TAKING BEHAVIOR

91

HOSPERS AND KOK

92

relationship between age and risk-taking behavior. Wheth<
mg gay men
constitute a high-risk group that needs special attention is - question that
needs to be addressed.
2. More research is needed among ethnic minorities and gay men with lower
SES status.
3. Future research should try to answer questions on the possible (causal) role
that substance use may play in AIDS-related behavior. If substance use is
related, cither causal or non-causal, it becomes important to understand the
underlying processes of the relationship.
4. Future research should answer unresolved questions with respect to the
influence of AlDS-risk perceptions on AIDS-related behavior.
5. The nature of the disease demands behavioral changes that have to be
maintained for long periods of time. It may well be that determinants that arc
related to initial behavior change differ from determinants that arc related to
long-term maintenance or relapse. For this reason, more research should be
conducted that provides insight in factors that enable gay men to maintain
behavior change or cause gay men to return to risk-taking sexual behavior.
This implies long-term monitoring of sexual behavior and possible determi­
nants.

skills lower than men who only practiced safe sex (Hays et al., 1990). A comparison
of gay men who consistently practiced safe sex with gay men who consistently
engage in risk-taking behavior, showed that the latter group expressed more difficul­
ty in changing their sexual behavior (Siegel et al., 1989). Dallas (1990) asked
respondents to rate their self-efficacy with respect to behaving safely for the subse­
quent ten years. Of the respondents who did not engage in anal intercourse, 27%
stated they thought it would be difficult, while 47% of the respondents who actually
engaged in anal intercourse suted likewise. One important point that has been made
is that AIDS risk behavior is special in the sense that it always involves an interaction
with one (or more) other individual (Davies & Weatherbum, 1990). In connection
to self-efficacy Bandura has noted:

Translating health knowledge into effective self-protection action against
AIDS infection requires social skills and a sense of personal power to
exercise control over sexual situations.. . . Problems arise in following
safer sex practices because self-protection often conflicts with in­
terpersonal pressures and sentiments. . . . The weaker the perceived
self-efficacy, the more such social and affective factors can increase the
likelihood of risky sexual behavior. (1989, p. 129)

Thus, interventions designed to enhance self-efficacy should be well aware of the
interpersonal nature of AIDS risk behavior. One potent strategy is the development
of social skills to cope with these interpersonal situations. The desired effects of
learning social skills are the ability to cope effectively with high-risk situations and to
build up self-assurance. Affcr assessment of existing coping skills for situations
where self-efficacy is low (i.e., high-risk situations) adequate coping skills are
identified and acquired skills are practiced, preferably by means of role plays. Finally,
ample feedback should be given on achievements.
Again, small group interventions or face-to-face counseling are suitable settings
to apply these techniques (Kelly, St. Lawrence, Betts, Brasfield, & Hood, 1990).

RECOMMENDATIONS WITH RESPECT TO METHODOLOGICAL ISSUES

1. In our opinion it is vital for the generalizability and comparability of results
of studies that greater consensus is reached on the operational definitions
of behavioral measures. In addition, as mentioned before, efforts should
be made to increase the validity and reliability of self-reported sexual be­
havior.
2. The use of theories that explain behavior or behavior change is recom­
mended. The results of determinant studies are extremely valuable for the
development of effective interventions. As Fisher and Fisher (1992) have
demonstrated in their review of AlDS-risk-reduction interventions, in­
terventions that are conceptually based appear to have the greatest impact.
Therefore, intervention development based on theories will most likely
benefit from determinant studies that are also conceptually based (sec also
Kelly & Murphy, 1992). In addition, we recommend that researchers use
and test complete theories instead of isolated concepts, as most behavior is
determined by a combination of different determinants. Only complete
theories contain a sufficient number of determinants to explain complex
behavior such as AIDS-related sexual behavior.
3. Researchers are encouraged to provide ample information on the measures
they used in their research.

CONCLUSIONS

To a certain extent the threat of HIV infection has prompted changes toward safer
sexual behavior among gay men: many have reduced their number of partners and
have increased their use of condoms. Despite these changes, however, a significant
number of gay men still put themselves at risk. This is similar to other health related
behaviors. That is, some people adopt protective behaviors faster than others, and
some people experience more difficulty in maintaining changes they have made in
their behavior. Continuous efforts to improve, and innovate interventions are war­
ranted. In order tcTbe able to develop effective interventions, a good understanding
of the determinants of safe and risk-taking sexual behavior is needed. We will
conclude with general recommendations for future research on determinants of
AIDS-related sexual behavior and related methodological issues.

In this contribution we have reviewed the literature on determinants of safe and
risk-taking behavior among gay men. Our recommendations show that efforts have
to be made to expand our understanding of AIDS-related sexual behavior and
enhance the quality of the research in this area. At the same time we hope that
concurrent to this endeavor, efforts will also be made to develop, implement, and
evaluate prevention interventions for gay men that are based on the theoretical
concepts that seem to pertain to risk behavior change and maintenance.

RECOMMENDATIONS FOR RESEARCH ON DETERMINANTS

1. To benefit the development of prevention activities, studies need to be
conducted that increase our knowledge with respect to the observed

1

II

; •••
-

93

DETERMINANTS OF SAFE AND RISK-TAKING BEHAVIOR

94

altb Education
Belief Model: A decade Is
AIDS. Health Education Quarterly, IJ, 331Quarterly. 11. 1-47.
345.
Jones. G C. Waskin. H.. Gcrety. B.. Skipper. B. J..
Fishbein. M.. & Ajzen, I. (1975). Belief, attitude,
Hull. H. F.. & Mertz. G. J. (1987). Persistence of
intention, and behavior. An introduction to
high-risk sexual activin- among homosexual
theory and research. Reading. MA: Addisonmen in an area of low incidence of the Acquired
Wesley.
Immunodeficiency Syndrome. Sexually Trans­
Fishbein. M.. Chan, D. K S.. O'Reilly. K. Schnell. D..
mitted Diseases. 14. 79-82.
Wood. R.. Becker. G. & Cohn. D. (1992). Altitu­
Joseph. J. G.. Adib. M.. Koopman. J. S.. & Ostrow.
dinal and normative factors as determinants of
D. G. (1990). Behavioral change in longitudinal
gay men's intentions to perform AIDS-related sex­
studies: Adoption of condom use by homosexual behaviors: A multisitc analysis. Journal of
ual/biscxual men. American Journal of Public
Applied Social Psychology. 22. 999-1011.
Health. KO. 1513-1514.
Fisher. J. D. (1988). Possible effects of reference
Joseph. J. G., Montgomery. S. B.. Emmons. C-A.,
group-based social influence on AIDS-risk be­
Kessler. R. C.. Ostrow. D. G.. Wortman. C B..
havior and AIDS prevention. American Psy­
O'Brien, K. Eller. M.. ft Eshleman. S. (1987a).
chologist. 43. 914-920.
Magnitude and determinants of behavioral risk
Fisher. J. D.. & Fisher. W. A. (1992). Changing
reduction: longitudinal analysis of a cohort at
AIDS-risk behavior. Psychological Bulletin.
risk for AIDS. Psychology and Health. I. 73-95.
III. 455-474.
Joseph.
J. G.. Montgomery. S. B.. Emmons. C-A..
Fitzpatrick. R.. Boulton. M.. Hart. G.. Dawson. J.. &
Kessler. R. C.. Ostrow. 0. G.. Wortman. C- B..
Mclx-an.J. (1989). High risk sexual behaviour
O'Brien. K. Eller. M.. & l-shleman. S. (1987b).
and condom use in a sample of homosexual and
Perceived risk of AIDS: Assessing the behavioral
bisexual men. Health Trends 21. 76-79.
consequences in a cohort of gay men. Journal
Fitzpatrick. R.. Mclxan, J.. Boulton, M.. Hart. G.. &
of Applied Social Psychology. 17, 231-250.
Dawson. J. (1990). Variation in sexual be­
Kauth. M. R.. St. lawrence. J. S.. & Kelly. J. A.
haviour in gay men. In P. Agglcton. P. Davies. &
(1991). Reliability of retrospective assessments
G. Hart (Eds ), AIDS: Individual, cultural and
of sexual HIV risk behavior: A comparison of
policy dimensions (pp. 121—132). London: Fal­
biweekly, three-month. and twelve-month selfmer Press.
reports. AIDS Education and Prevention. .5.
Glasner. P. D.. & Kaslow. R. A. (1990). The
207-214.
epidemiology of human immunodeficiency
Kelly.
J. A.. Kalichman. S. G. Kauth. M. R.. Kilgore.
virus tnJccivon.Journal of Consulting and Cli­
H. G.. H<xk1. H. V.. Campos. P. E.. Rao. S. M..
nical Psychology. 58. 13-21.
Brasfield. T. I.. & St. lawrence. J. S. (1991).
Gold. R. S.. Skinner, M. J.. Grant, P. J.. & Plummer.
Situational factors associated with AIDS risk be­
D. G (1991). Situational factors and thought
havior lapses and coping strategies used by gay­
processes associated with unprotected in­
men who successfully avoid lapses. American
tercourse in gay men. Psychology and Health.
Journal of Public Health. Hl. 1335-1338.
5. 259-278.
Kelly. J. A.. & Murphy. I). A. (1992). I*sychological
Griensvcn. G. J. P. van. & Vroomc, E M. M. de
intenentions with AIDS and HIV: Prevention
(1987). Failure-rate van kondooms gedurende
and treatment. Journal of Consulting and Cli­
anogenitaal scksucel kontakt bij homoseksuele
nical Psychology. 60. 576-585.
mannen (Failure rates of condoms during an­
Kelly. J. A.. St. lawrence.J. S.. Betts. R.. Brasfield. T
ogenital sexual contact among homosexual
U & Hood. H. V. ( 1990). A skills training group
men|. Tijdschrift voor Sociale Cexondheidsintervention model to assist persons in reduc­
zorg. 65. 703-705.
ing risk behaviors for HIV infection. AIDS
Griensvcn, G. J. P. van. Vroomc, E M. M. de.
Education and Prevention. 2. 24—35.
Ticlman, R. A. P.. Goudsmit, J.. Noordaa. J. van
Kelly.
J. A.. St. lawrcncc. J. S.. & Brasfield. T. I.
der, Wolf. F. de, & Coutinho. R. A. (198").
( 1991). Predictors of vulnerability to AIDS risk
Scksuclc gedragsverandcring in ccn groep op
behavior relapse. Journal of Consulting and
HlV-antistoffen geteste homoseksuele mannen
Clinical Psychology. 59. 163-166.
(Sexual behavior change in a group of gay men
Kelly.
J. A.. St. lawrcncc. J. S.. Brasfield. T. Itested for HIV|. Tijdschrift voor Sociale
Lemke. A.. Amidei. T.. Roffman. R. E.. Hood. H.
Cezondheldszorg 65, 345—350.
V.. Smith. J. E. Kilgore. H„ & McNeill. C.
Handsfield, H. H. (1985). AIDS and sexual be­
(1990). Psychological factors that predict AIDS
havior in gay men. A merican Journal of Public
high-risk versus AIDS precautionary behavior.
Health. 75. 1449.
Journal of Coiuulting and Clinical Psycholo­
Hays. R. B.. Kegeles. S. M.. & Coates. T. J. (1990).
gy. 58. 117-120.
High HIV risk-taking among young gay men.
Kelly. J. A.. St. lawrcncc. J. $.. Brasfield. T. L.
AIDS, 4. 901-907.
Stevenson. Y.. Diaz, Y. E.. & Hauth, A. G (1990).
Janz, N. K. & Becker. M. H. (1984). The Health

REFERENCES
Adib, S. M., Joseph, J. G„ Ostrow. D. G.. & James. S.
A. (1991). Predictors of relapse in sexual prac­
tices among homosexual men. AJDS Education
and Prevention, 3, 293-304.
Adib, S. M., & Ostrow, D. G. (1991). Trends in
HIV/A1DS behavioural
research among
homosexual and bisexual men in the United
States: 1981-1991- AIDS Care. 3. 281-287.
Ajzen, I. (1988). Attitudes, personality, and be­
havior. Milton Keynes. UK: Open University
Press.
Bandura. A. (1986). Social foundations of
thought and actions Englewood Cliffs. NJ: Pre­
ntice-Hall.
Bandura, A. (1989). Perceived self-efficacy in the
exercise of control over AIDS infection. In V. M.
Mays. G. W. Albee. & S. F. Schneider (Eds.).
Primary Prevention of AJDS (pp. 128-141).
Newbury Park, GA: Sage Publications.
Bauman, LJ., & Siegel. K (1987). Misperceptions
among gay men of the risk for AIDS associated
with their sexual behavior. Journal of Applied
Social Psychology, i 7. 329-350.
Becker, M. H.. & Joseph. J. G. (1988). AIDS and
behavioral change to reduce risk: A review.
American Journal of Public Health. 78. 394—
410.
Blake, S. M.. Sharp, E. S.. Temoshok. I.. & Rundeli.
J. R. (1992). Methodological considerations in
developing measures of HIX' risk-relevant be­
haviors and attitudes: An empirical illustration.
Psychology and Health. 6. 265-280.
Bochow, M. (1990). AIDS and gay men: Individual
strategies and collective coping. European
Sociological Review. 6, 181-188.
Bureau of the Census (1992). 1990 Census of
Population and Housing: Summary of Social.
Economic, and Housing Characteristics Wash­
ington, DC: Government Printing Office.
Catania, J. A., Gibson, D. R., Chitwood. D. D., &
Coates, T. J. (1990). Methodological problems
in AIDS behavioral research: Influences on
measurement error and participation bias in
studies of sexual behavior. Psychological
Bulletin, 108, 339-362.
Catania, J. A., Gibson. D. R., Martin, B„ Coates, T. J..
Be Greenblatt, R. M. (1990). Response bias in
assessing sexual behaviors relevant to HIV
transmission. Evaluation and Program Plann­
ing 13. 19-29.
Centers for Disease Control (1985). Self-reported
behavioral changes among homosexual and
bisexual men. Journal of the American Medic­
al Association, 254. 2537-2538.
Chmiel, J. S., Detels, R., Kaslow, R_, VanRaden, M..
Kingsley, L A., Brookmeyer, R.. & the Multicen­
ter AIDS Cohort Study Group (1987). Factors
associated with prevalent human immuno­

deficiency virus (HIV) infection in the multi­
center AIDS cohort study. AmericanJournal of
Epidemiology. 126, 568-577.
Coates. R. A., Calzavara, L M., Soskolne, G L,
Read. S. E.. Fanning. M. M., Shepherd, F. A..
Klein. M. H.. & Johnson, J. K. (1988). Validity of
sexual histories in a prospective study of male
sexual contacts of men with AIDS or an AIDSrelated condition. American Journal of
Epidemiology, 128, 71 $>-728.
Connell. R. W., Crawford, J., Dowsctt, G. W„ Kip­
pax. S.. Sinnott, V.. Rodden, P., Berg. R., Baxter,
D.. & Watson, I. (1990). Danger and context:
Unsafe anal sexual practice among homosexual
and bisexual men in the AIDS crisis. A ustralian
and New Zealand Journal of Sociology, 26,
187-208.
Connell. R. W„ & Kippax, S. (1990). Sexuality in
the AIDS crisis: Patterns of sexual practice and
pleasure in a sample of Australian gay and bisex­
ual men. TheJournal ofSex Research, 27, 167198.
Dallas. M. ( l$>90). Onveilige sex bij homomannen (Unsafe sex among gay men). Utrecht. The
Netherlands: University of Utrecht, Department
of Gay and lesbian Studies.
Daries. P. M.. Hickson, F. G I.. Weatherbum. P., &
Hunt. A. J. (1993). Sex. gay men and AIDS
London: Falmer Press.
Daries. P.. & Weatherbum. P. (1990). Towards a
general model of sexual negotiation. In P.
Aggieton. G. Hart, & P. Davies (Eds.), AIDS:
Responses, interventions and care (pp. 111125). lendon: Falmer Press.
Detels. R.. Fahey. J. F.. Schwartz, K. Greene. R. S.,
Visschcr. B. R.. & Gottlieb, M. S. (1983). Rela­
tion between sexual practices and T-cell sub­
lets in homosexually active men. Lancet, i,
609-611.
Doll. 1. SM Petersen, L R., White, G R., Johnson. E.
S., Ward, J. W., & The Blood Donor Study Group
(1992). Homosexually and nonhomosexually
identified men who have sex with men: A be­
havioral comparison. Journal of Sex Research,
29. 1-14.
Doornbos, R. (1992). Safe and unsafe sex be­
tween men. Utrecht, The Netherlands: Univer­
sity of Utrecht, Department of Gay and Lesbian
Studies.
Ekstrand. M. L, & Coates, T. J. (1990). Mainte­
nance of safer sexual behaviors and predictors
of risky sex: The San Francisco men's health
study. American Journal of Public Health, 80,
973^-977.
Emmons, C. A., Joseph, J. G., Kessler, R. G, Wort­
man, G B., Montgomery, S. B., & Ostrow, D. G.
(1986). Psychosocial predictors of reported be­
havior change in homosexual men at risk for

II
i

HOSPERS AND KOK

I

■V;

1

DETERMINANTS OF SAFE AND RISK-TAKING BEHAVIOR

AIDS risk behavior patterns among gay men in
small southern cities. American Journal of
Public Health, 80. 416-418.
Kelly, J. A, St. Lawrence, J. S., Stevenson. L Y.,
Hauth, A G. Kalichman, S. G, Diaz, Y. E.. Bras­
field, T. L, KoobJ. J., & Morgan, M. G. (1992).
Community H1V/AIDS risk reduction: The
effects of endorsements by popular people in
three Cities. American Journal of Public
Health, 82, 1483-1489Kingsley, L A, Detels, R., Kaslow, R., Polk. B. F.,
Rinaldo, G R-, Chmiel, J., Detre, K. Kelsey, S. F..
Odaka, N., Ostrow, D., VanRaden. M., & Visscher, B. (1987). Risk factors for seroconversion
to human immunodeficiency virus among male
homosexuals. Lancet, i, 343-349.
Kinsey, A G, Pomeroy, W. B., & Martin. G E.
(1948). Sexual behavior in the human male
Philadelphia: Saunders.
Linn, L S., Riegel. J. S.. Mathews. W. G. Leake. B..
Lien, R., & Brooks, S. (1989). Recent sexual
behaviors among homosexual men seeking pri­
mary care. Archives of internal Medicine 149.
2685-2690.
Mann, J. M., Tarantola, D. J. M., & Nctter, T. W.
(Eds.) (1992). AIDS in the world. Cambridge.
MA: Harvard University Press.
Martin, J. L (1986). AIDS risk reduction recom­
mendations and sexual behavior patterns
among gay men: A multifactorial categorical
approach to assessing change. Health Educa­
tion Quarterly, IJ. 347—358.
Martin, J. L (1987). The impact of AIDS on gay
male sexual behavior patterns in New York
City. American Journal of Public Health, 77.
578-581.
McCusker, J., Stoddard, A M„ Zapka. J. G.. Zom.
M.. & Mayer. K H. (1989). Predictors of AIDSpreventive behavior among homosexually ac­
tive men: A longitudinal study. AIDS 3. 443448.
McCusker, J., Zapka, J. G., Stoddard. A. M.. & May­
er. K. H. (1989). Responses to the AIDS
epidemic among homosexually active men:
Factors associated with preventive behavior.
Patient Education and Counseling, 13, 15-30.
McKusick, L, Coates, T. J., Morin, S. F.. Pollack. L.
& Hoff, G (1990). Longitudinal predictors of
reductions in unprotected anal intercourse
among gay men in San Francisco: The AIDS
Behavioral Research Project. A mericanjoumal
of Public Health. 80. 978-983.
McKusick, L, Conant, M., & Coates, T. J. (1985).
The AIDS epidemic: A model for developing
intervention strategics for reducing high-risk
behavior in gay men. Sexually Transmitted
Diseases, 12, 229-234.
Miller, T. E, Booraem, G, Flowers, J. V., & Iversen.
A E (1990). Changes in knowledge, attitudes
and behavior as a result of a community-based

96

95

of self-reported sexual liehavior risk factors lor
HIV infection in homosexual men. Public
Health Reports, 102, 692-697.
Schechter, M. T., Craib. K J. P., Willoughby. B.,
Douglas, B., McLeod. W. A, Maynard, M„ Con­
stance, P., & O'Shaughnessy, M. (1988). Pat­
terns of sexual behavior and condom use in a
cohort of homosexual men. American Journal
of Public Health. 78. 1535-1538.
Seage, G. R., Mayer, K H., Horsburgh, C. A. Cai. B..
& Lamb. G. A (1992). Corroboration of sexual
histories among male homosexual couples.
American Journal of Epidemiology. 133. 7984.
Seage. G. R., Mayer. K H., Horsburgh. C. R.. Holm­
berg. S. D.. Moon. M. W.. & Lamb, G. A. (1992).
The relation between nitrate Inhalants, un­
protected receptive anal intercourse, and the risk
of Human Immunodeficiency Virus infection.
American Journal of Epidemiology, 135. I—II.
Shcrnoff, M., & Bloom. D. J. (1991). Designing
effective AIDS prevention workshops for gay
and bisexual men. AIDS Education and Preven­
tion, 3. 31—46.
Siegel. K. Bauman, L J., Christ, G. H., & Krown. S.
(1988). Patterns of change in sexual behavior
among gay men in New York City. Archives of
Sexual Behavior, 17. 481-497.
Siegel. K, Mesagno. F. P.. Chen. J.-Y., & Christ. G.
(1989). Factors distinguishing homosexual
males practicing risky and safer sex. Social Sci­
ence and Medicine, 28. 561—569.
Solomon. M. Z., & Dejong. W. (1986). Recent
sexually transmitted disease prevention efforts
and their implications for AIDS health educa­
tion. Health Education Quarterly, 13, 301-316.
Sull. R. D., Coates. T. J.. & Hoff. C. (1988). Behavioral risk reduction for HIV infection among
gay and bisexual men. American Psychologist.
43. 878-885.
Stall. R. D„ Ekstrand, M.. Pollack. L. McKusick. 1..
& Coates. T.J. (1990). Relapse from safer sex:
The next challenge for AIDS prevention efforts.
Journal of Acquired Immune Deficiency Syn­
dromes, 3. 1181-1187.
Sull. R. D., McKusick, L, Wiley. J.. Coates. T. J.. &
Ostrow, D. G. (1986). Alcohol and drug use
during sexual activity and compliance with safe
sex guidelines for AIDS: The AIDS behavioral
research project. Health Education Quarterly.
13. 359-371.
St. Lawrence, J. S., Hood, H. V., Brasfield. T.. &
Kelly, J. A (1989). Differences in gay men's
AIDS risk knowledge and behavior patterns in
high and low AIDS prevalence cities. Public
Health Reports, 104, 391-395.
Valdiserri, R. O., Lyter. D.. Leviton, L G. Callahan.
G M.. Kingsley. L A, 4 Rinaldo. G R. (1988).
Variables influencing condom use in a cohort of
gay and bisexual men. American Journal of
Public Health, 78, 801-805.

AIDS prevention program. AIDS Education and
Prevention, 2, 12-23.
Ostrow, D. G., VanRaden, M. J., Fox, R., Kingsley,
L A, Dudley,)., Be Kaslow, R. A (1990). Recrea­
tional drug use and sexual behavior change in a
cohon of homosexual men. AIDS 4, 759-765.
Peterman, T. A., Cates, W., & Wasserheit, J. N.
(1992). Prevention of the sexual transmission
of HIV. In V. T. DeVita, S. HeUman, & S. A.
Rosenberg (Eds.), AIDS Etiology, diagnosis,
treatment, and prevention (3rd ed., pp. 443451). Philadelphia: Lippincott.
Peterson. P. L, Ostrow, D. G., & McKiman, D. J.
(1991). Behavioral interventions for the pri­
mary prevention of HIV infection among
homosexual and bisexual men. Journal of Pri­
mary Prevention, 12, 19-34.
Prieur. A (1990). Norwegian gay men: Reasons
for continued practice of unsafe sex. AIDS
Education and Prevention, 2, 109-115.
Prieur. A ( 1991). Mann—mdnnlicheLiebe in den
Zeiten von AIDS [Male—male love in the AIDS
era]. Berlin: Deutsche AIDS Hilfe.
Reiss. I. 1., & Lcik, R. K (1989). Evaluating strat­
egies to avoid AIDS: Number of partners vs. use
of condoms.Journal of Sex Research, 26, 411433.
Robert. B.. & Rosser. S. (1990). Evaluation of the
efficacy of AIDS education interventions for
homosexually active men. Health Education
Research, 5, 299-308.
Rogers. E. M. (1983). Diffusion of innovations
New York: Free Press.
Rogers. R. W. ( 1983). Cognitive and physiological
processes in fear appeals and attitude change: A
revised theory of protection motivation. In ). T.
Cacioppo, & R. E Petty (Eds.), Social Psy­
chophysiology A sourcebook (pp. 153-176).
Neu- York: Guilford.
Ronis. D. L (1992). Conditional health threats:
Health beliefs, decisions, and behaviors among
adults. Health Psychology, 11, 127—134.
Ross, M. W. (1987). Problems associated with
condom use in homosexual men. American
Journal of Public Health, 77, 877.
Ross. M. W. (1988). Attitudes toward condoms as
AIDS prophylaxis in homosexual men: Di­
mensions and measurement. Psychology and
Health. 2. 291-299.
Ross, M. W.. & McLaws, M.-L (1992). Subjective
norms about condoms are better predictors of
use and intention to use than attitudes. Health
Education Research, 7, 335-339.
Rotheram-Borus, M. J., & Koopman, C. (1991).
Sexual risk behavior, AIDS knowledge, and be­
liefs about AIDS among predominantly minority
gay and bisexual male adolescents. AIDS Educa­
tion and Prevention, 3, 305—312.
Saltzman, S. P., Stoddard, A M., McCusker, J.,
Moon. M. W., & Mayer, K. H. (1987). Reliability

1

II
i

HOSPERS AND KOK
ic diary mctluKl
Wcaihcrburn. I’. (1992). I
out the relation­
to enable causal inferen
ship between alcohol use and unsafe behaviour.
Paper presented at the Second European Con­
ference on Homosexuality and HIV, Amster­
dam. The Netherlands.
Weatherbum. P., Davies. P. M., Hickson. F. G L,
Hunt. A. J.. McManus. T. J., & Coxon. A. P. M.
(1993). No connection between alcohol use
and unsafe sex among gay and bisexual men.
AIDS 7. 115-119.
Weatherburn. P.. Hunt. A. J.. Davies. P. M.. Coxon.
A. P. M., & McManus. T. J. (1991). Condom use
in a large cohort of homosexually active men in
England and Wales. AIDS Care 3- 31-41.
Weinstein, N. D. (1987). Unrealistic optimism
about illness susceptibility: Conclusions from a
community-wide sample. Journal of Be­
havioral Medicine. 10. 481-500.
Weinstein. N. D. (1989). Perceptions of personal
susceptibility to harm. In V. M. Map. G. W.
Albee. & S. F. Schneider (Eds.). Primary pre­
vention ofAJDS (pp. 142-167). Newbury Park.
CA: Sage Publications.
Winkclstcin. W., Samuel. M., Padian. N„ Wiley, J.
A. lang. W._ Anderson. R. F... & l.evy,J. (1987).
Reduction in human immunodeficiency virus
transmission among homosexual/bisexual men:
1982-1986. American Journal of Public
Health. 76. 685-689.
Wit. J. B. F. de. Sandfort. T. G. M.. Vroomc, E M. M.
de. Griensven, G. J. P. van. Kok, G. J.. Coutinho.
R. A. & Tielman. R. A. P. (1992). Serieel crosssectionclc en longitudinale veranderingen in
seksucel gedrag in cen cohort homascksuele
mannen in Amsterdam in de periodc 19841988 [Serial cross-sectional and longitudinal
changes in sexual behavior in a cohort of gay
men in Amsterdam in the period 1984—19881.
Tijdschrift voor Sociale Cezondhcidszorg. 70,
457-462.
Wit. J. B. F. de. Sandfort. T. G. M., Vrtximc, E M. M.
de. Griensven. G. J. P. van. & Tielman, R. A. P.
(1991). Dctcrminanten van vcilig cn onveilig
seksucel gedrag bij mannen met homoscksuelc
contacten; ecn cross-sectionele verkenning
[Determinants of safe and unsafe sexual be­
havior among men with homosexual contacts: a
cross-sectional exploration |. In R. W. Meenens,
A P. Buunk. & R. van dcr Vlist (Red.). Sociale
psychologic & voorlicbting en maatschappelijke probIemen (pp. 29—40). Den Haag: VUGA
Wit, J. B. F. de, Vroomc, E M. M. de. Sandfort, T. G.
M.. Griensven. G. J. P. van, Metselaar, G, &
Tielman, R. A P. (1991). De relatie tussen coping-stijl, health locus of control en seksucel
gedrag in ecn cohort homoscksuelc mannen
[The relationship between coping style, health
locus of control and sexual behavior in a cohon
of homosexual men]. Gedrag en Gezondheid, 9.
88-92.

a

KJH " I % •

High HIV risk-taking among young gay men
Robert B. Hays, Susan M. Kegeles and Thomas J. Coates
Previous research has shown younger age to be correlated with greater HIV sexual

risk-taking among gay men. The purpose of this study was to identify variables
associated with HIV risk-taking among younger gay men. Ninety-nine gay men
aged 18-25 in three medium-sized West Coast communities completed self-report

questionnaires regarding HIV-related behaviors and attitudes. Of the respondents,
43% reported having engaged in unprotected anal intercourse during the previous 6

months. Men who engaged in unprotected anal intercourse reported greater enjoyment
of unprotected anal intercourse, perceived less risk of unprotected anal intercourse,

labeled themselves as more at risk for AIDS, reported poorer communication skills

with sexual partners, and were more likely to have a boyfriend/lover than men who

had not engaged in high-risk sex. In addition, respondents perceived the likelihood of

acquiring HIV from unprotected anal intercourse with young gay men to be significantly
lower than with older gay men. These findings highlight the need for HIV risk-reduction

interventions designed specifically for young gay men and identify critical areas to be
targeted in such interventions.

AIDS 1990, 4:901-907

Keywords: HIV risk-taking, gay men, sexual behavior, AIDS-related attitudes.

Introduction

While older gay men in urban areas have dramatically
decreased sexual behaviors that can transmit HIV infec­
tion, evidence from a variety of studies suggests that high
percentages of younger gay men continue to engage in
unsafe sex [1-3]. For example, data from the San Fran­
cisco Men’s Health Study, a population-based sample of
gay and bisexual men in San Francisco who have been
followed since 1984, show age to be the most consist predictor of high-risk behavior [1], Younger men
were significantly more likely to engage in unprotected
anal intercourse and to do so with more partners than
were older men. Similarly, in a survey of 526 gay bar pa­
trons in Seattle, Tampa and Mobile, younger men were
more likely to engage in unprotected anal intercourse
than were older men [2].

While a number of studies have documented the associ­
ation between age and HIV risk-taking, we are not aware
of any study that has specifically examined factors asso­
ciated with risk-taking among gay' men in the 18-25-year
age range. The experience of this age group is very differ­
ent from that of previous generations of gay men since
their sexual careers were initiated in an era when infor­
mation about AIDS was widely disseminated. AIDS pre-

vention programs that are effective for older gay men may
be less appropriate for younger gay men for whom issues
specific to their generation will need to be addressed.

A variety of factors may contribute to high HIV risk-taking
among young gay men. First, men in this age group may
still be in a ‘coming out' stage with regard to their sexual
identity [4,5]. They may not fully identify diemselves as
‘gay’, and therefore may not perceive diemselves to be
in a ‘risk group’ for AIDS. Due to their relative inexpe­
rience in personal and sexual relationships, young men
may be less socially skilled, and so may be less likely to
communicate openly about sexual matters and less com­
petent in negotiating low-risk sexual interactions. Young
people in general have heightened feelings of invulnera­
bility [6,7] which may cause younger men to engage in
more HIV risk behaviors than older men. Finally, since
the bulk of AIDS cases among gay men are in the 30-^0year age group [8], younger men may perceive AIDS to
be a problem of older gay men. For example, in focus
groups we conducted with young gay men in the San
Francisco Bay area, young men expressed a stereotypical
view of gay men who were likely to have AIDS as ‘older
men with moustaches who go to leather bars’. Young
men therefore may feel that it is safe to have unprotected
intercourse with other young men. The prolonged incu-

From the Center for AIDS Prevention Studies, Division of General Internal Medicine, University of California, San Francisco, USA.

Sponsorship: Support for this research was provided by NIMH/NIDA center grant no. MH42459.
Requests for reprints to: Center for AIDS Prevention Studies/Box 0886, University of California, San Francisco, CA 94143-0886, USA.

Date of receipt: 18 December 1989; revised: 24 April 1990.

(c) Current Science Ltd ISSN 0269-9370

901

I

$02

AIDS 1990, Vol 4 No 9

bation period of HIV infection feeds this misperception
since men who become infected in their early twenties
are not likely to show symptoms until they are in their
thirties.
The purpose of this study was to identify factors asso­
ciated with HIV risk-taking among gay men aged 18—25This information is essential in designing AIDS preven­
tion programs specifically targeted to rhe special needs
and situations of young gay men.

ous 6 months. Activities listed included anal intercourse
with and without condoms, vaginal intercourse with and
without condoms, mutual masturbation, oral-anal con­
tact and oral-genital sex. Numbers of male and female
sexual partners during the previous 6 months were also
requested. In addition, respondents were asked to report
the settings in which they' had met their male sexual part­
ners during the previous 6 months and the frequency
with which they had used alcohol or drugs before or dur­
ing sex in the past 6 months.
Individual characteristics

Method
Procedure

I.

The study was conducted in three medium-sized West
Coast cities with low to moderate AIDS prevalence rates
(Santa Cruz, California; Santa Barbara, California; and
Eugene, Oregon). Self-report questionnaires regarding
HIV-related behaviors and attitudes were distributed to
men between the ages of 18 and 25 as they left settings
identified by local gay leaders as frequented by young gay
men (for example, gay bars, adult bookstores, gay stu­
dent union campus events, public parks and beaches).
Gay male field workers approached men individually as
they left die target settings and invited them to participate
in the study, which was described as a university-based
research survey to examine ‘how young men were cop­
ing widt die AIDS epidemic’. Potential respondents were
handed the surveys to complete at their convenience and
mail back to the researchers in the stamped, addressed
envelopes provided. Two hundred and forty-three sur­
veys were distributed in this manner; 86 (35%) were re­
turned. In an attempt to increase the response rate, we
repeated die same procedure in one city (Santa Cruz)
using a shonened quesdonnaire and offered to mail re­
spondents $10 upon receipt of their completed question­
naire. Eighty-six surveys were distributed in this manner;
27 (31%) were returned. Since diere were no significant
differences in the data collected through the two proce­
dures, we pooled the data sets. The data from 14 men
who either (1) were not within the 18-25-year age range,
or (2) identified themselves as heterosexual and reported
never having engaged in sexual activity with a man were
not used in die analyses reported here.
Survey instrument

Respondents completed a 90-item self-administered
questionnaire wdiich assessed sexual behaviors, individ­
ual characteristics and attitudes hypothesized by the AIDS
Risk Reduction Model [9] to be relevant to HIV risk-tak­
ing behavior.
Sexual behaviors

Respondents were presented with a checklist of sexual
behaviors and asked to indicate which behaviors they had
engaged in with men and with women during the previ-

Respondents were asked to report their age, ethnic iden­
tification, amount of education, whether they currently
were involved in a boyfriend/lover relationship (and, if
so, whether either partner had sex with others), and
whether they had ever been tested for HIV antibodies
(and if so, the results). In addition, respondents were
asked to indicate whether they would describe them­
selves as gay, bisexual or heterosexual; the age at which
they first had sex with a man; and their degree of open­
ness with other people about their sexual feelings toward
men (rated on a five-point scale ranging from 'completely
out of the closet' to 'definitely in the closet’).
HIV-related attitudes

A series of brief scales (presented in the Appendix)
was used to assess the following HIV-related attitudes:
self-efficacy (beliefs :that one can successfully perform
safer-sex behaviors), perceived social norms (percep­
tions that one's friends support safer-sex behaviors), sex­
ual communication skills (perceptions that one can effec­
tively communicate about safer-sex behaviors with part­
ners), self-labeling (perception that one's current sex­
ual behavior places one at risk for HIV infection), safersex efficacy (belief that safer sex behaviors can effec­
tively prevent HIV transmission), interpersonal barriers
to safer-sex (perceptions of undesirable consequences
of attempting -co engage in safer-sex behaviors). Items
for each scale were rated on six-point Likert scales rang­
ing from ‘strongly agree’ to ‘strongly disagree’. Respon
dents were also asked to rate their degree of enjoyment
of high-risk activities (anal intercourse without condoms)
and low'-risk activities (mutual masturbation) on six-point
Likert scales ranging from ‘enjoy very much' to ‘dislike
very much'. Similarly, respondents rated their perceptions
of riskiness of. the same high- and low-risk activities on a
five-point scale wdiich ranged from 'not at all risky' to 'ex­
tremely risky’. Cronbach alphas for all scales ranged from
0.60 to 0.85. In addition, respondents were asked to rate
the chances of becoming infected with HIV by having un­
protected intercourse with young gay men (25 years old
or less) versus gay men in their thirties on six-point Likert
scales ranging from ‘no chance at all’ to ‘definitely would
happen’.
Reasons for sexual risk-taking

Respondents were asked to rate the likelihood that young
men in their community ‘might have unsafe sex (for ex­
ample, anal intercourse without a condom)’ in a variety
of circumstances using a five-point liken scale ranging
from ‘not at all likely’ to ‘extremely likely’.

2
HIV risk-taking among young men Hays et al.
Finally, a series of free-response questions asked for
the respondents’ perceptions of how their behavior had
been affected by the AIDS epidemic, reasons why young
gay men might have unsafe sex, and ideas on how to pre­
vent the further spread of AIDS among young gay men
in their community.

Results

previous 6 months; two had engaged in unprotected vagi­
nal intercourse. The median number of male partners in
the past 6 months was three (range 0-50); the median
number of female partners was zero (range 0-2). Sixtyone per cent reported using alcohol/drugs during sex in
the past 6 months. Thirty-seven per cent said they did
so half or more -than half the time they had sex. The
most common places the men reported meeting male
sexual partners were gay bars (65%), parties (44%), pub­
lic parks/beaches (30%), adult bookstores/bath-houses
(19%) and school (17%).

Sample description

Low versus high risk-takers

Usable questionnaires were returned from 99 men. Their
mean age was 22.57 (s.d. 2.02, range 18-25). The sample
was 79% white, 13% Hispanic, 4% Asian, and 4% other.
Thirty-five per cent were college graduates; an additional
50% had some college education. Eighty-eight per cent
described themselves as gay; 12% as bisexual. The aver­
age age of their first gay experience was 16.35 (s.d. 3.12).
Sixty-three per cent described themselves as mosdy or
completely out of the closet’ about their gayicy, while
37% described themselves as ‘in the closet’ at least half
the time. One quarter reponed having had sex with both
men and women during the past 3 years. Sixty-four per
cent reported that they had been tested for HIV antibod­
ies and six (6%) reponed that they were seropositive.
Two per cent had had gonorrhea and 2% syphilis during
the previous year. Forty per cent of the men reponed
currently having a boyfriend/panner, but 43% of those
relationships were not monogamous.

Table 2 represents a comparison of men who had en­
gaged in high-risk sex (unprotected insertive or receptive anal intercourse) during the preceding 6 months
versus men who had not. Multiple analysis of variance
(MANOVA) was used to test for differences between
high- and low-risk takers. The 10 HIV-related attitudes
and frequency of combining drugs/alcohol with sex
were included as dependent variables. The significant
MANOVA [F(l 1, 87) = 5.82, P< 0.001] was further ex­
amined with univariate analyses of variance (ANOVA).
Compared with men who did not engage in unpro­
tected anal intercourse, men who engaged in unpro­
tected anal intercourse: reported more enjoyment of
anal intercourse without condoms [F(l, 97) = 29.76,
P < 0.001]; perceived less risk*of unprotected anal inter­
course (especially of insertive anal intercourse) (F(l,97)
= 6.66, P < 0.02 ]; labeled themselves as more at risk
for AIDS (F(l, 97) = 13-22, P< 0.001] and reported
poorer communication skills with sexual partners (F(l,
97) = 4.04, P <0.05]. In addition, chi-square analy­
ses revealed that high-risk men were more likely to
have a boyfriend/partner (chi-square = 5.83, d.f. = I,
P<0.01).

Sexual activities

Table 1 presents the percentages of men who engaged
in each of the various sexual activities during the preced­
ing 6 months. Forty-three per cent of the respondents reported engaging In unproteaed anal intercourse during^
the preceding 6 months (32% receptive, 34% insertive).
Forty per cent had had oral-anal contact; 28% had en­
gaged in oral sex with ejaculation into the mouth. Six
per cent reported having had sex with a woman in the

Ten of the men who had engaged in unprotected anal
intercourse reported being HIV-negative and in mutually
monogamous relationships. Since one might argue that
unprotected anal intercourse for these men was not a
high-risk behavior, we repeated the above analyses ex-

Table 1, Percentages of young gay men (n = 99) engaging m various sexual activities during preceding 6 months.

Sexual activities with men

Sexual activities with women

Anal intercourse

Receptive without condoms

Vaginal intercourse

Insertive without condoms

32%
34%

Receptive with condoms

46%

Insertive with condoms

40%

Oral-genital sex

Receptive with ejaculation

Insertive without condoms
Insenive with condoms

2%

6%

Anal intercourse

Insertive without condoms
insertive with condoms

0%
0%

Insertive with ejaculation

19%
24%

Receptive with ejaculation

0%

Receptive without ejaculation

82%

Receptive without ejaculation

4%

Insertive without ejaculation
Oral-anal

83%

Cunnilingus

2%

Active

Passive

Oral-genital sex

Oral-anal

22%

Active

1%

31%

Passive

0%

Mutual masturbation

Mutual masturbation

Active

82%

Active

Passive

86%

Passive

3%
4%

903

AIDS 1990, Vol 4 No 9
Table 2. Comparisons between young gay men who engaged in unprotected anal intercourse (high risk-takers) versus men who did not Gow risk-takers),
including 95% confidence intervals.

Individual characteristics
Age
Years since first homosexual sex

Ethnicity (% non-white)
Education (years of school)
% having boyfriend/lover
Number of male sex partners (past 6 months)
% taken HIV-antibody test
% tested HIV-positive
% bisexual (self-labeled)
Openness with regard to homosexuality
HIV-related attitudes
Enjoyment of unprotected anal intercourse
Enjoyment of mutual masturbation
Perceived riskiness of unprotected anal intercourse
Perceived riskiness of mutual masturbation
Self-labeling
Sexual communication skills
Self-efficacy
Perceived social norms
Safer-sex efficacy
Interpersonal barriers to safer sex
Drug/alcohol use during sex

eluding those men. When these men were excluded from
the analyses described above, the results did not change.
Reasons for risk-taking
Respondents perceived die likelihood of acquiring HIV
from unprotected anal intercourse with young gay men
to be significandy lower than from older gay men (t =
7.59, df = 95, P < 0.001). To gain further insight into
factors contributing to sexual risk-taking among young
gay men, we asked respondents to rate the likelihood that
young men in their community might have ‘unsafe sex’
with another man in a variety' of circumstances. Table 3
presents the mean ratings of each circumstance for high
and low risk-takers. While high risk-takers rated the like­
lihood of each item higher dian did die low risk-takers
(reflecting their own tendency to engage in high risk ac­
tivities), die relative ranking of items by high and low risk­
takers was the same. On the average, respondents consid-

Low risk-takers
n = 56 (95 % Cl)

High risk-takers
n = 43 (95 % CD

22.7 (22.1 23.2)

22.4 (21.8. 23.1)

5.8 (4.8. 6.8)

6.4 (5.2. 7.6)

14% (1%. 27%)

30% (16%. 45%)

14.7 (14.3, 15.1)

14.4 (13.9. 14.9)

24% (10%. 39%)

62% (45%. 79%)

3.9 (2.6. 5.1)

6.8 (3.9. 9.8)

58% (45%. 71%)

72% (59%. 85%)

10% (0%. 27%)
11% (0%, 24%)

14% (0%. 28%)

3.5 (3.1 3.8)

3.8 (3.4. 4.2)

3.00 (2.5. 3.5)
5.6 (5.4. 5.8)
4.6 (4.4. 4.8)
1.3 (1.1 1.5)
2.7 (13. 3.0)

4.8 (4.4. 5.2)
5.3 (5.0. 5.6)

5.0 (4.7. 52)

4.5 (4.2, 4.9)
2.3 (2.0. 2.6)

4.5 (4.2. 4.9)
5.5 (5.3. 5.7)
4.8 (4.4. 5.2)
4.6 (4.2. 5.0)
2.6 (2.3. 3.0)

1.3 (1.0. 1.7)

1.5 (1.1. 2.0)

5.6 (5.4. 5.7)
5.0 (4.7, $.3)

10% (0%. 27%)

4.2 (3.9. 4.5)
1.4 (1.3. 1.6)

3.7 0.2. 4.2)

ered it 'very likely’ a young man would engage in unsafe
sex if he were ‘in love with his partner’ or using alcohol
or drugs prior to having sex. Unsafe sex was considered
‘somewhat likely’ if: both partners were HIX'-positive; he
was ‘too sexually turned on to stop’; his partner was ‘re­
ally good looking’; his partner said he was not infected
with HIV; the partner ‘talked him into it’; there were no
condoms available, or ‘he didn’t want to disappoint his
partner’.

Discussion
Tills study found a high rate of HIV sexual risk-taking
among gay men aged 18-25 and identified factors which
characterize young gay men who engage in high-risk sex.
Young gay men who engaged in unprotected anal inter-

Table 3. Perceived likelihood of engaging in unsafe sex in various circumstances for high- and low-nsk-taking men. including 95% confidence internals’

Being 'in love' with partner
Using drugs/alcohol
Being too 'sexually turned on' to stop
Both partners are HIV-positive
Partner is 'really good-looking'
Partner says he is HTV-negative
Partner talks him into it
No condoms available
Don't want to disappoint partner

*1, not at all likely; 3. somewhat likely; 5, extremely likely.

Low risk-takers
n = 56 (95 % Cl

High risk-takers
n = 43 (95 % Cl

3.6 (3.2. 3.9)
3.4 (3.0. 3.8)
3.1 (2.7. 3.6)
3.2 (2.8. 3.7)
2.9 (2.5. 3.3)
2.9 (2.6. 32)
2.9 (2.5. 3.3)
2.6 (22. 3.0)
2.6 (22. 3.0)

4.2 (3.8. 4.5)

4.1 (3.8. 4.3)
3.9 (3.6. 42)
3.7 O.1 4.1)
3.6 (33. 4.0)
3.4 (3.1, 3.8)

3.0 (2.7. 33)
3.4 0.1. 32)
3.0 (17. 33)

HIV risk-taking among young men Hays e( al.
course were found to differ from men who did not on
five dimensions: they reported greater enjoyment of un­
protected anal intercourse; perceived less risk of unpro­
tected anal intercourse (especially of insertive anal inter­
course); labeled themselves as more at risk for AIDS; re­
ported poorer communication skills with sexual partners,
and were more likely to have a boyfriend/lover. In ad­
dition, respondents rated the chances of acquiring HIV
from unprotected anal intercourse with young gay men
to be significantly less than with older gay men.
These findings have important implications for the design
of effective HIV risk-reduction interventions for young gay
men. First, HIV prevention programs must reduce the
perception that sex with younger men is ‘safer’ than with
older men. As expressed in the free-response comments
of one HI\z-positive 21-year-old, ‘Especially in people un­
der 18 there seems to be an attitude that "It can’t happen
to me". When I was that age I saw a lot of men in their
thirties who had AIDS but hardly ever heard about some­
one my age getting it’. Heightening awareness of young
gay men chat HIV infection does exist among their peers
is critical. Further, in contrast to older gay men who tend
to be highly knowledgeable about AIDS and safer sex (2 ],
younger men appear to need more didactic presentations
of AIDS information. As one man stated, ‘I meet a lot of
19-20-year-olds who don’t know what is or isn’t safe or
how- to use a condom’. Another wrote, ‘When I started to
experiment with sex with men 3 years ago, I never played
safe. But as I learned about AIDS, I began to play safe and
have since changed my behavior’. Young gay men, whose
initial ‘coming out’ period often includes a flurry of sexual
experimentation [4], cannot afford tliat early period of
AIDS ignorance. Several specific areas of misinformation
regarding HIV risk were identified: young men need to be
educated that insertive anal intercourse without condoms
is risky and diac safer sex guidelines should be followed
even if both partners are HIV-positive or in an emotion­
ally committed relationship. In addition, the mistaken be­
lief that anal intercourse without condoms was safe if they
withdrew before ejaculation was expressed by several in
their frce-response comments.

While important starting points, increasing awareness and
knowledge are clearly insufficient in preventing HIV infec­
tion among young gay men given our finding that men
who engaged in unprotected anal intercourse did recog­
nize that their behavior places them at risk for AIDS. To
some degree, this may reflect feelings of personal invul­
nerability characteristic of youth [6,7]. Young gay men
may recognize the riskiness of their behavior but feel
the negative consequences ‘won't happen to me’. However, knowingly engaging in high-risk sex may also reflect
motivational and/or skill deficits among young gay men.
As postulated by the AIDS Risk Reduction Model [9], an
analysis of the costs and benefits of engaging in safer sex
influences one’s commitment to low-risk activities. As we
found, a potent cost for young gay men is the reduced
enjoyment associated with using condoms. In the words
of one respondent, ‘Even if condoms are readily avail­
able, there is still a deep feeling that sex without con­
doms is more fun'. Another wrote, ‘I know a lot of guys

my age just coming out and are having too much fun
to worry about AIDS’. Effective HIV risk-reduction pro­
grams for young gay men must therefore help them to
develop ideas and skills to heighten the enjoyment value
of low-risk activities (for example, exercises designed to
eroticize safer sex activities).
Effoas to increase motivation for safer sex among young
gay men must also consider the broader social context
within which young gay men's sexual behavior occurs.
Sex within dating and primary relationships must be dis­
cussed, emphasizing that being ‘in love’ or committed
to one’s partner does not make unprotected intercourse
‘safe’. Using sex as a way to attract a partner or gain af­
fection must also be addressed. As one HIV-positive re­
spondent wrote, ‘gay youths are also incredibly' in need
of love and attention and oftentimes would do anything
(including unsafe sex) if they thought they were getting
that love and anention. I have a lot of biaemess over the
fact that as a gay teenager all I wanted was to be loved
and all I got was dick up the butt and the HFV infection.
I got over it but it happens all too often to gay youth’.
Feelings of isolation and alienation common among gay
youth [10] may also reduce their motivation to engage in
safer sex. As one respondent wrote, ‘It seems like nobody
cares if I die anyway'.

The finding that men who engage in unsafe sex rated
their communication skills significantly lower than men
who did not suggests that skill deficits are an additional
contributor to high risk-taking among young gay men.
Due to their relative inexperience in interpersonal rela­
tionships, young men may need training in communica­
tion skills necessary for negotiating safer sex. Assertive­
ness training in ways to resist being ‘talked into’ doing
something unsafe may also be critical. Further, our find­
ings suggest that issues of impulse control must also be
addressed for young men. Being 'too sexually turned on
to stop’ or being overwhelmed by a good-looking partner
were prominently cited as possible causes of unsafe sex.
In free-response comments, one man admitted occasion­
ally being unable to control himself during sex, adding
‘1 always feel guilty and emotionally disturbed once it is
over though’. Strategies which young men can use to di­
rect their sexual energies to low-risk activities would be
useful.
In view of die poor communication abilities of many
young men, skills training should also emphasize meth­
ods of structuring situations in which sex is likely in ways
that maximize the likelihood of safer sex. For example,
die high percentage of young men who reported combin­
ing drugs and alcohol with sex is problematic in view of
research showing drug and alcohol use as a contributor
to HIV risk-taking [11]. Young men must be made aware
of die dangers of engaging in sexual activity while un­
der die influence of artificial substances. Likewise, taking
responsibility for having condoms readily available must
be emphasized ['More than once 1 would have practiced
unsafe sex if not for the free condoms’ (given out at the
bar)].

There are several limitations of this study. First, our sam­
ple was not a random sample of the young gay men in

905

7

AIDS 1990, Vol 4 No 9

these communities. Although we distributed the surveys
in a variety of settings where young men who have sex
with ocher men congregate, the men who completed the
survey were predominantly gay-identified and described
themselves as fairly out of the closet’. While this pop­
ulation is vitally important for AIDS prevention efforts,
the findings obtained may not generalize to gayly active
young men who do not identify themselves as gay or bi­
sexual or who do not frequent the public settings we
targeted. Although an extremely difficult population to
reach, investigations of this subgroup would be ve^ im­
portant In addition, the sample was predominantly white
and college-educated; the results may not generalize to
gay men who are from ethnic minorities or less educated.
Further, the experience of young gay men in larger urban
areas where AIDS is more prevalent may differ in impor­
tant ways from that of men in smaller communities such
as those studied here. Finally, the 35% response rate ob­
tained in this study' is less than ideal. However, given that
the primary purpose of the study u-as to examine corre­
lates of high-risk sex among young gay men, since the
sample did include sufficient numbers of both high and
low risk-taking men, we feel the findings are useful in
identifying differences between young men who do and
do not engage in high-risk sex. Nonetheless, strategies
for increasing die response rates in future studies of this
nature would be invaluable.

In conclusion, this study shows that intervention with
young gay men must be a high priority in efforts to halt
the AIDS epidemic and identifies critical areas to be tar­
geted in such interventions. On die bright side, if young
gay men can be socialized to engage in safer sex from
die outset of their sexual careers, the likelihood of their
maintaining low-risk behavior is high since dieir risky
habits are likely to be less entrenched than among older
men. This study has identified the need and necessary
content of AIDS prevention interventions for young gay
men; research into the most effective formats for such
intervention is an important next step.

2.

3.

4.

5.
6.

H.

9.
10.
11.

Kelly ja. St Lawrence JS. Brasfield TI, tr al: Psychosocial
factors which predict AIDS high-risk versus AIDS precau­
tionary behavior. J Consult din Psychol 1990. 58:117-120.
Valdserri RO. Lyter LD. Ieviton CM. Cauauan CL, Kingsley
CL, Rinaiexd CR; Variables influencing condom use in a co­
hort of gay and bisexual men. Am J Publ Healdj 1988
78:801-805.
Beu. AP, Weinberg MS: Gayiiies. Neu* York: Simon and Schus­
ter. 1981.
RamaFEDI G: Gay youth; a challenge to contemporary soci­
ety. JAMA 1987. 258:222-225.
Elkind D: Tlx Child's Reality: Three Deielapmcntal Tlxmes
HillsadJe. NJ: Eribaum. 1978.
CVETKOVICH G. Grote B. Bjokseth a. Sarkissian J: On the
psychology’ of adolescent's use of contraceptives. J Sex Res
1975. 11:256-270.
Centers for Disease Control- HIV/aIDS sun-eillance. US De
panment of Health and Human Services. July 1939.
Catania JA, KEGEiES SM. Coates TJ: The AIDS Risk Reduction
Model (ARRM): a model for predicting high risk sexual
behavior. Healdj Educ
1990. 17:27-39.
Feidman DA: Gay youth and AIDS. J Cayity 1989. 17:185-193
Stall R. McKusick I, Whey J. Coates TJ. Ostrom DC: Alcohol
and drug use during sexual activities and compliance with
safe sex guidelines for AIDS: the AIDS Behavioral Research
Project. Health Educ
1986. 13:359- 371.

Appendix
Scales used co assess HIV-related attitudes (v/ith Cron
bach alphas).

Self-efficacy (alpha
(a)
(b)
(c)

(d)

0.67)

I can get a man Fm having sex with to use condoms i: ( want
him to.
If someone fm having sex with doesn’t want to follow sate sex
guidelines', there is little I can do about it.
It sonteone Fm having sex with does not want to use a condom,
there is little I can do about it.
If someone fm having sex with starts to do something unsafe,
there is little I can do about it.

Perceived social norms (alpha = 0.67)
(a)

Acknowledgements

(bl

(c)
We would like to thank Jo Kenny and David Beckstcin of the Santa
Cruz AIDS Project. Genic Cowan and Mark Grotke of Santa Barbara's
Gay/lesbian Resource Center and Sally Sheklow. Jim Clay and Tadd
Tobias of the Willamette AIDS Council for their help in planning
the data collection; Jim Koctska and Richard Wolitski for their assis
tance in survey distribution; and Bobby Hilliard for hdp with data
management and analysis.

Most of my friends think you should always use condoms when
having anal intercourse.
Most of my friends think that condoms are too much of a
hassle to use.
Most of my friends think you should avoid anal intercourse
without condoms.

Sexual communication skills (alpha = 0.61)
(a)
(b)
(c)

It is easy for me to tell a sex partner I won't have ana! inter­
course without a condom.
I find it difficult telling a sex partner not to do something i
think is unsafe.
It’s easy for me to tell a sex partner what I do or don’t like
to do during sex.

References

Self-labeling (alpha = 0.85)
i.

Ersikand ML, Coates 1J: Gay men in San Francisco arc
maintaining low-risk behaviors but young men continue to
be at risk. Am J Publ Health (in press).

(a)

(b)

There is little chance that I could catch or spread AIDS from
what I do sexually.
My sexual behavior is risky for catching AIDS.

HIV risk-taking among young men Hays el al.

(d

I don't do things that could cause me to catch AIDS.

Unprotected anal intercourse (alpha — 071)
Anal intercourse where:
you put your penis in your partner's anus without a condom;

(a)
(b)

your partner puts his penis in your anus without a condom.

Safer-sex efficacy (alpha = 0.64)
(a)

People who follow 'safe sex guidelines' can usually avoid getting

the AIDS virus.
People who always use condoms will probably not get the AIDS

Mutual masturbation (alpha = 0.78)

(b)

virus.

La)

you masturbate your, partner;

(b)

your partner masturbates you.

Interpersonal barriers (alpha

0.75)

(a)

If I suggested using condoms to a man I was having sex with,

(b)

he might be offended.
If I wanted to use a condom, my sex partner might think that

Masturbation where:

Perceived riskiness
Instructions: how risky (for you to get AIDS) do you think the various
sexual activities listed below are, if you were to do them with a man

whose antibody status was unknown to you?

I am infected with the AIDS virus.

(d

If I requested a man I was having sex with to use a condom,
he might think that I suspect him of being infected with the

AIDS virus.

Unprotected anal intercourse (alpha = 0.60)
Anal intercourse where:
you put your penis in your partner's anus without a condom;
(a)

(b)

your partner puts his penis in your anus without a condom.

Enjoyment
Instructions: for each sexual activity, circle the number that best shows

Mutual masturbation (alpha = 0.81)

how much you enjoy or think you might enjoy doing that activity with a

Masturbation where:

man. Please circle the number for each sexual activity whether you have

(a)

you masturbate your partner;

done it or not.

(b)

your partner masturbates you.

907

lo H - I 2 '
Review B

r

Sex-work harm reduction
Mie had I. Rekart

Sex work is an extremely dangerous profession. The use of harm-reduction principles can help to safeguard sex
workers l.ves in the same way that drug users have benefited from drug-use harm reduction. Sex workers are

Lancet 2005; 366: 2123 -34
Published online

work, and exploitation of migrants). Successful and
67732-X
st.ueg.es are available: education,
’ . , empowerment, prevention, care, occupational health and safety ''
dtcrmnnahs’t.on of sex workers, and human-rights-based approaches. Successful interventions include peci British Columbia Centre for
' . _.L. :!”,n,ng ,n
skills, safely tips for slreet-based sex workers, male and female condoms Disease Control, University of
the prevention-care- rsynergy,
------ occupat.ona health and safely guidelines for brothels, self-help organisations, and British Columbia, Vancouver
V5Z 4R4. BC, Canada
community-based child
-iild protection networks. Straightforward and achievable steps are available .(o •improve the day- (fiotM 1 Rekartml>)
to-day lives of sex 1
£p±rnlin"P l° 'VOrk- C0nCe'”',:,lisill8
del-ting sex-work harm reduction as a Coirespondence to:
new paradigm can hasten this
Prof Michael I. Rek.irl
mirhad.rekart(E-.bccrlc.ca

Sex work and injection drug use are among the most,
perilous activities worldwide. Harm reduction has
stimulated global debate about drug use, and the
application of harm-reduction principles to inter­
ventions such as needle exchange has reduced HIV
spread and improved the lives of drug users.’ Since drug
users might participate in sex work to pay for drugs,
drug-user harm reduction includes condom promotion,
and sex workers could use dings to cope with
psychological, emotional, and physical stress/ ’ Safe-sex
campaigns and social marketing of condoms are based
on harm-reduction principles. Tire process of harm
reduction is not new to the study of sex work. Harm­
reduction and risk-reduction strategies have been
adopted by health authorities, sex worker organisations
and sex workers themselves. This Review aims to
(1) examine studies of sex work, by concentrating on
peer-reviewed publications, and classify harms and
harm-reduction strategies into overall themes; and
(2) locus on simple, available strategies to improve sex
workers lives. Male and trans-sexual sex workers face
harms and can benefit from harm-reduction strategieshowever, this Review will not focus on these topics 01 the
pecific issues of clients outside of ihe general theme of
sex-work harm reduction.
Sex-work Imm reduction has been proposed by the
In.emauonal Harm Reduction Development (HIRD)
programme as a framework for discussion, action and
research." Sex-work harm reduction has also been
conceptualised rn newsletters, booklets, reports and
conference abstracts.7’"

Sex-work harms
Differences in social context need to be considered for
sex-work harms to be meaningful. In some societies, sex
work is legal or decriminalised; sex workers have access
to health and social services; and they are not heavily
stigmatised or economically destitute." Alternatively sex
dismut01^
3 SUrViVal taCtiC dUring SCVere so*tal
d.sruphon when no se,vices are available and life
necessities are scarce." Most societies exist between
Vk-ww.thelancet.com Vol 366

these extremes and sex-work harms thus vary from place
to place/ Poverty, war, globalisation, and neocolonialism
are important causes of the international sex-work trade'
but these issues arc beyond the realm of harm reduction.
Drug use

Injection drug use is common in sex workers in many
locations/
',V17 fs,-x workers who
■ inject drugs might use
;
condoms less consistently and, for more money, (hoy
might agree to unprotected sex or ana! sex.’-4-’6”'1'’
Individuals who share needles, <syringes, and ding
mjeclion paraphernalia are at risk of HIV, hepatilis B
and C, and syphilis.’071 Female sex workers could be in
relationships with male injectors who mix the drug and
inject the women, increasing their HIV risk/7 Physical
and sexual abuse by customers has been associated with
drug use in sex workers.Injection drug use tan cause
Search strategy and selection criteria

ialiOnS'1 Sei‘r-edMEDIJNE (from 19661 and EMBASE;

1Z1U’reVim

1980) us,ng the MeSH terms "prostitution" and "risk reduction". UN, UNAIDS and WHO
pubheatrons were searched online wit h "sex work", "sex wo,ker", "sex trade" '



prostitution", "prostitute", “survival sex", "transactional sex", "harm reduction" "risk
«urt,on , traffic king", “decriminalization"; arid, "human rights". The same terms were

Mra’iW
d °
S
i

work publications (from 2000)
I ClS' S0Cial 5*nce AbSt,MtS' allcl'110
Science.
social scientific and psychological review on sex

alS0used-

rroth selected articles and widely used

textbooks
.
on 5CXl’all>' transmitted disease were also reviewed. With the same kev words
earcbed non-peer-reviewed work using an online search engine (G X abs afls

Z : ernatl°T

and

COnreren"5

l,N' UNA'^ *"d WHO Z and

' OmBhOn ft0"' "“n-S°~^i organisations. Peer-revieweZ!

December 17/24/31. 2005
2123

j

Review

/

skin infections, thrombosis, sepsis, endocarditis,
overdoses, and other serious illnesses?"
Sex workers could use non-injection drugs such as
cocaine, crack, and crystal methamfetamine, '.16.17.I9.21-Z7
leading to poor judgment, unsafe sex, immune
suppression, cardiovascular and neurological disease
overdose, and addiction.— In Guyanese sex workers^
cocaine was significantly associated with inconsistent
condom use?7 Alcohol, probably the most important
drug in tlie sex-work industry,” "-7" * has been associated
with violence, abuse, unsafe sex, HIV infection, and liver
damage?-’"20-2’
Disease

Sex workers have an increased risk of sexually
transmitted infections (STIs), including HIV?'6’7’'*'6
Condom use varies among sex workers,2',nK'z',,zand the
decision to use condoms is often controlled by the
customer or brothel owner?’6”1^7 Descriptive and
analytical studies show that sex workers commonly use
condoms less often with regular partners, spouses, and
non-paying customers.‘! K, I7-,9J4 U,W 40
STI complications are common in sex workers,
including pelvic inflammatory disease
and ectopic
pregnancy?"”*
STIs
are cofactors
in
HIV
transmission ””-4' and frequent intercourse can cause
genital trauma, greatly increasing HIV risk'"■4M’ Sex
workers sometimes douche, or use drying or astringent
substances that remove the lubricating vaginal fluid to
increase a sense of tightness or induce "dry sex" These
praences have been associated with an increased risk of
STIs or HIV infection.'— Sex workers' could also
acquire hepatitis A or herpes through anal-oral contact."

Violence
Violence—against

sex
important issue in many communities. Violence
includes physical, verbal, and sexual abuse; gang rape
traumatic intercourse; emotional trauma;' robte w
confinement; and murder. Street-based sex worked
have an increased risk of violence.- Violence results in
sZss and low K emOti°naI
P^^^l
stress and low self-esteem. Significantly raised overall

X Vr

micide mortality have ^een sho™ hi

active and former sex workers/’ Violence by an intimate
ofmViJfect
n6™ aSS°Ciated With 3n increased risk
01 miv iniection.
Discrimination

Sex workers are easy targets for discrimination, the
TheV' aXPr^SS^d corollary of stigmatisation/JWi-”
I^ese ndtviduals are devalued in many societies and
often blamed for the breakdown of the traditional family
epidemics of STIs and HIV/AIDS, escalating crime and
t abu ''^7 Of^uth'W” StigmatisatiL can lead

health/•'•',,•", fSex workers with HIV/AIDS could be
• • K.
doubly stigmatised.
Debt
Young people sometimes enter sex work to support their
families but soon acquire personal debts for
transportation, accommodation, clothes, cosmetics,
condoms, food, medical care, drugs, and fines?•
Risk-taking in sex workers has been statistically
correlated with financial need?” Brothels can hold sex
workers in debt bondage, allowing them to keep a small
P’oPortlon of their earnings?"-24 As debts accumulate, the
likelihood of individuals leaving sex work falls?
Criminalisation

Prostitution, or some aspect of it such as soliciting, is
illegal in many countries, but the law is an ineffective
means of eliminating its negative aspects, often
resulting in the criminalisation of sex workers?-’0-U..W.S1.W
Even if prostitution is not illegal, sex workers can be
treated as criminals/-” Criminalisation leads to
violence; police harassment; increased HIV and STI
risk; reduced access to services; psychological disease;
drug use; poor self-esteem; loss of family and friends;
work-related mortality; and restrictions on travel
employment, housing, and parenting. lo.io.uijj.M.f.i-r,?
—..................
Estimated yearly occcurence
Adverse health effects in prostituted children*
Infectious disease
STDs
2 000 000
HIV infection
300 000
HPV infection
4 500000
HBV infection

Pregnancy
Maternal deaths
Spontaneous abortions
Induced abortions
Abortion-related complications
Abortion-related deaths
Mental illness
PTSD
Attempted suicide
Substance abuse
All substances
Violence
Physical assault
Rape
Murder
Malnutrition

500 000

4752
900 000
1224000
367 200
710
6 700 000
I64OOOO

9000 000

2500000
2500000
6900
Unable to estimate
Adverse health effects in infai
mts born to prostituted childrent
Infant deaths
190 080
Complication of STDs
237000
HIV infection
249480
Deaths from HIV infection
54 886
HBV infection

8316

boys prostituted per year, t Based on an estimated 2 Vf, mn ’f

> ”

h°"

and low V1° rnCe' criminalisation. denial of services
and low self-esteem, which affects sex workers1
2124
www.tlielancet.com Vol 366

December 17/24/31, 2005

Review

j

Exploitation

Child prostitution, human trafficking for sex work, and
Initiatives
Harms reduced
Education
the abuse of migrant sex workers are important
Peer education, outieadi programmes, accessible
Drug use, disease, violence,
and appropriate materials, sex worker involvement
examples of exploitation.1'- UNICEF has estimated that
debt, exploitation
Empowerment
Self-esteem, individual control, safe sex, solidarity,
1 million children enter the sex trade every year/*
Drug use, disease, violence,
personal safety, negotiating skills, refusal to clients,
debt, discrimination, exploitation
Children can be sold or led into prostitution by their
service access, acceptance by society
amihes."'-67 Customers frequently prefer young girls
Prevention
Male and female condoms, lubricant, vaccines,
Drug use, disease
behavioural change, voluntary HIV counselling
especially virgins, believing that there is less risk of
and testing, participation in research
diseases such as HIV/AIDS or that sex with virgins will
. Cate
Accessible, acceptable, high-quality, integrated
enhance their sexual potency, cure disease, or extend
Dni(| use, disease, violence,
care; prevention-care synergy; prophylaxis; STIs,
exploitation
heir h espan.1'— Children brought into prostitution
HIV/AIDS, and psychological care; social support
Occupational
ha.c little power to negotiate condom use,W7 and the
Control exposures and hazards, treatment for
Drug use, disease, violence, debt,
health and safety
injuries and diseases, employer duties, worker rights
c gH m'pf .|V,aglna and ccrvix are n,,:Je susceptible Io
exploitation
Decrirninalisation of
Sex worker organisations, sex work projects,
s. Child prostitutes are at high risk of HIV and
Ct iminalisation, disci imination.
sox woikers
non-governmental organisations
violence
SI Is, violence, sexual abuse, rape, substance use, mental
Righ ts-based approach Education, telephone hotlines, training
Exploitation (ie, child prostitution,
i ness, tuberculosis, hepatitis, malnutrition, suicide
targeted and user-friendly services,
human trafficking, exploitation of
and death (table
Pregn;)nt ado|esccm
'
government action, media, PREVENT/ refugee
mobile populations)
sex
package, community development
workers are at increased risk of p-----------pregnancy complicans, maternal morbidity and mortality',
and the
complications from safe and unsafe abortions 1
rhe UN defines human trafficking as "recruitment
L'Table 2: Interventions for sex-work harm reduction
~
-- -------------- --------- —-----------transportation, transfer, harboring or receipt of persons^
by coercion for the
‘aiT’ PurP°se of exploitation including
prostitution”.7'’ /Although
’ ’
trafficking and sex work raise on the>r own strategies, value their distinctive
d (Terences, not conflict with their culture and traditions
different issues, trafficking for
SeX WWk iS aSS0^ and increase lbeir option for seffide^;^'
with HIV infection, STIs, discrimination
autonomv6356
r'r
f°r selMcl«n>inalion.
I’3
auionomy,
and
control
Ti
conlro '
'.
immigration status, reduced access ^-aUrKHega!
..c-dicai ano .egat
J ue
social,
behavioural, and professional
assistance violence, and drug use.'1”-” Once trafficked
huiuiwgeneiiy
heterogeneity ot
of sex
gir^ might be reluctant to return home." Human worker subgroups often needs different individual
--------- J and
busmts"raS,eSt 8’0Wine intematte^l trafficking structural interventions.9VM V/;;
WHO’s Sex Work Toolkit
delineates the key principles
. 5 and issues for UJV
A migrant is an individual who is engaged in a prevention, care, and
empowerment, and the best
remunerated activity in a state where he or fh!^ not
practices against
the
inherent
challenges in
interventions fo:'E sex-work harm reductions (panel 2).
nahonal. M,grants can be at risk of discrim nafion
violence, HIV and STIs, criminalisation, poor medical’
Education
caic, and drug use
r
i

Migrant.....

bJ^XL

a

I

Panel 1: Personal coping strategies of sex workers

Ooing medical care?0-*"-*0’'5 An Australian ch 1 i
= higher risk of STIs and lower ZonXtT

• Keep working and personal lives separate"

■"‘ernational sex workers than for local sex workers"

• Prioritise positive roles, such as motherhood"

Sldi’lPBKiMM (e3, douching, condom

Strategies for sex-work harm reduction (table 2)
o

kissing

a/,^-lr-Pro9rammiclg,internaldia1ogue,andme1,cuIousn1anagenM>toftime

Maintain a positive and professional altitude towards work’

logic
1 i ves.

and

p*--

1

use
use

o

in

their

clients55

day-to-day

easier to negotiate with

Cy 5SeX

Practise good genital hygiene18-10
o

........ .’X;Skk
ww thelancet.com Vol 366

...
* Use two or three condoms at thi

... ............... .....

December 17/24/31, 2005

2125

Review

• Respect sex workers' views, knowledge, and life experiences
• Include sex workers, and, if appropriate, other community members in all stages of the
development and implementation of interventions

prevented.Outreach programmes delivered by
educators, social workers, nurses, and respected
community members have also had success.-1.1 1
Many groups associated with sex work can1 benefit from
education. 16'a’w-’x,',0‘,',,M Successful materials are simple,
clear,
consistent,
non-judgmental,
attractive,
and culturally sensilive.'w'-,,,‘ Positive reinforcement
can deal with prevailing practices, values, and
beliefs.'"’
Challenges include mobility, brothel
manager control, criminalisation, language, culture, and
traditions.

• Recognise that sex workers are usually highly motivated to improve their health and
wellbeing, and that sex workers are part of the solution

Empowerment

Panel 2: Principles and issues for effective HIV interventions in diverse sex work
settings

Key principles

• Adopt non-judgmental attitude
• Ensure that interventions do no harm

• Respect sex workers' rights to privacy, confidentiality, and anonymity
• Respect sex workers' human rights and accord them basic dignity

• Build capacities and leadership among sex workers to facilitate effective participation
and community ownership
• Recognise the role of clients and third parties in HIV transmission—ie, targeting the
whole sex work setting, including clients and third parties, rather than only sex workers

Sex work harms can be mitigated by empowerment—ie,
provision of the means and opportunity for selfassertion?'6,1'’’-107 Personal empowerment is the
awareness and strengthening of personal skills and
Recognise and adapt to the diversity of sex work settings and of participating individuals
options to control and improve sex workers' lives.
.^ey issues
Community empowerment strengthens the community’s
ability to participate in positive changes. Social
• Assessment: follow ethical guidelines and good research practice
empowerment enables sex workers to fight for their rights
• Planning: build local support, identify potential partners
and acceptance in society.,r’’’s
• Prevention outcomes: safer sex and increased condom use, increased sex worker
The aim of empowerment is to reduce vulnerability. Sex
participation and control over working and social conditions, reduced STI burden
workers could be vulnerable because of poor self-esteem,
• STI treatment: at a minimum, provide management of symptomatic STIs and either
lack of education and skills, negative societal attitudes,
screening for asymptomatic STIs or presumptive treatment for STIs if accurate
poverty, family responsibilities, poor health, mobility,
screening is not feasible
and cultural and legal restrictions.,i'’w"57'l':'",R-"’ This
• HIV testing and counselling: training to provide a sensitive, non-judgmental service;
vulnerability can result in difficulties for sex workers
strict confidentiality; pre-test and post-test counselling and informed consent; referral
to psychological support and clinical care if possible
accessing and using condoms, negotiating safe sex,
refusing clients, seeking redress, organising, parenting,
• HIV care: counselling and peer support; if possible, establish self-help groups and
using contraception, having abortions, and accessing
improve access to treatment, care, support, home care, and antiretrovirakreatment
public services?-'6^-1"-116 The sex-worker community
• Harm reduction for sex workers: discourage injection, needle-sharing, and overall use
of drugs
could be vulnerable because of invisibility and internal
competition. "*M
• Management: mentoring and support to adopt organisational transparency and open
communication,, cor
Successful initiatives have resulted in enhanced selfcommunity participation, clear policies, flexible and adaptable
structure
esteem; improved negotiating skills; ability to refuse
clients; access and use of condoms; training to recognise,
• Training: schedule training so that sex workers can attend, develop policies.on
incentives and payment for attendance, write reports for future use
\
avoid, and escape violence; STI and HIV preventive
n 11” rr ri nn and
nnd assessment: use
C— _ .11feedback information from stakeholder groups to
- Monitoring
services; safe houses; drop-in centres; and STI treatment
nroiects: use
use
.Sr.
through
,’',6'K
change, develop, and expand projects;
assessment results to lobby for funding,
through pharmacies.''
pharmacies?'
,’',6IO7'M'"'l07r’117
'"r’117
Civil
society
replication, expansion, or social or policy changes
oiganisations have promoted practical safety tips5 lO
empower street-based sex workers (panel 3).
Panel adapted from information in reference 82. with permission.
Tlieie aie structural examples of how policy and law can
empower sex workers. In Santo Domingo, Dominican
Republic, sex establishment support for condom use and
dispelling myths, and offering healthy lifestyle and work
HIV or STI prevention was a significant predictor of
options.*' Education can effectively reduce drug use,
consistent condom use (odds ratio 2-16; 95% CI
disease, violence, debt, and exploitation?'6'1'*6^-"’0
1 • 18-3-97).17 Thailand's 100% condom campaign
Peei education has resulted in substantial increases in
increased condom use in commercial sex from 14% to
STI and HIV knowledge, condom use, and safer sex
94% by making condoms freely available, sanctioning
practices, and reduced incidence of HIV and STIs,z.ig.'x^hh
Peer educators need training, support, protection, and against non-compliant brothels, and advising men
through the media to use condoms with prostitutes."1’ A
standards of conduct. Experienced sex workers can
icpoil
of significant decline in condom use by brothel­
counsel other, often younger, sex workers about how to
based female sex workers in Thailand underscores the
live safely. Peer education of sex workers in Chad was
shown to be die most cost-effective option for the need for interwentions to be sustained.'z,'
Community development has been successful in the
prevention of HIV/AIDS at under US$100 per infection
promotion of safe sex, identification of injustice,

2126
www.thelancet.corn Vol 366 December 17/24/31, 2005

Review

provision of child care, support for HIV-infected
workers, enhancement of self-esteem, co-operation with
police and controllers, provision of legal and financial
training, initiation of alternative income-generation
schemes, and support for migrants and human
nghts.’—’•’h.hmzmzz ln johannesburg>

Panel 3: Safety tips for sex workers

Appearance

Avoid scarves, necklaces, and bags that can be used to hold or
choke you

South Africai

hotel-based sex workers have united to reduce risk and to
educate newcomers?4 When dealing with authorities,
the community development model could be more
elective and safer than actions by individual sex
workers.16

Wear shoes that you can run in

Negotiations

Wear clothing that can be left on during sex in case you have to
run away
Stick to a price list and time limit
Pick your own parking spot or hotel
Have a supply of condoms and lubricant

Get money up front
Use the same stroll ■

Prevention

Male condoms reduce HIV and STI transmission in sex
workers"— and prevent STI complications such as
pelvic inflammatory disease?23 A reliable and accessible
supply of good-quality condoms is essential lf",w-U4
Condom promotion, distribution, and social marketing
5ult in increased condom use and reduced STI and
HIV infection rates, especially in female sex workers w
Local culture, language, and traditions should also be
considered?2'
Female condoms have successfully prevented
pregnancy and reduced STI transmission in analytical
s udres,1""" and there is in-vitro evidence and biological
plausibility for HIV prevention.1"’ Female condoms
empower women by enabling them to negotiate safe sex
by promoting healtliy behaviour, and by increasing self­
effectiveness and sexual confidence.1"’ A simulation
model m South Africa concluded that a well-designed
female condom programme for sex workers would be
high y cost effective - Female condoms do not need an

■ The car

Approach from the driver's side
Arrange service and location while outside car
Circle the car looking for other passengers

Take down the licence plate (or pretend to)
Do not fasten the seatbelt
Wave goodbye to someone and shout the time of your return
(or pretend to)

Oral sex

Learn to put on condom with your mouth

leakage lat,On, keCP PrGSSUre °n condom with y°ur Up* to prevent
Gargle with mouthwash or liquor afterwards, but do not brush
Vaginal sex

your teeth
Use birth control

Keep genital area well lubricated with water-soluble lubricant
Do not douche or use vaginal-drying substances

Position yourself on top, facing customer

Keep hand on base of penis to keep it hard and to avoid spillage
After ejaculation, remove penis from vagina immediately

Anal,sex

Try to negotiate out of it

Charge too much for the customer to afford
time PeT; Te.reusrable' and “n he inserted ahead of
time and left m after sex. Since they are made of
Use extra lubricant
polyurethane female condoms can be used with waterUse female condoms
Self-defence
Do not carry weapons
Kcented' 1
‘“Scants. Female condoms are
accepted by sex workers1"111 but major difficulties
Use your voice and speed (eg, scream, hit car horn)
include cost and poor availability.
Attack body areas that are easily injured (eg, throat, eyes, testicles)
DhomhaV|dSd°Win Significantly rcduccd breakage rates
Run away against traffic, towards lights and people
hout added slippage when more than one male
Work with friends
:
-idem was used.’1 When both male and female
Tell workmates about bad customers
condoms were available to brothel-based sex workers in
from infem,,.,-.,,,
jja,
Ihai and, unprotected sex fell by 17% (p=0• 16) and STI
■nedence by 24% (p-Q. 18).1'" Lubrication
is especially
Lubrication is
™I,ortant for female condoms.1'"1" Dental dams and
durl?15 hTare CUk lengthwise are Plausible barriers systems and inaPProP™te delivery
dur ng cunmhngus, but controlled trials are scarce The
pliability of an effective and safe microbicide wdl be an . ^‘a-analysis has shown that behaviour change
important advance in sex-worker safety. I341B
vaccffi'e 111 v' “'‘‘r
fr°m the
Use of “ HIV
accine. Vaccine-feasibility studies in Thai; ’ "
sex workers have shown ongoing high rates of HIV
HepaTfeB “
intereSt’ ’nd g°0d “mpliance.1"11’ S.COnd°mS
C0U'd PrediSP°- to H>V
Hepatitis B vaccination programmes for sex workers can
Voluntary HIV counselling and tesline Inq Im
be effective especially in the outreach setting and wh
the mterva between the second and thill do^e is associated with increased condom use, reduced number
shortened.™ However, coverage rates could be low clnl^This ff T85611 ,H'V and “ SeX WOrkers ai’d
bents. Hus effect results from behaviour change
www.thelancet.com Vol 366

December 17/24/31,2005
2127

A

Review

subsequent to education, suj
............................
pport,
and the knowledge of
one’s HIV status. Care programmes and participation in
research can have a similar enccl.nN' Integration of STI
and HIV services into family planning has been
espoused/—but there is little published evidence of
effectiveness.'^ Additional success factors include links
to community agencies, financial incentives, and
support for childcare, transportation, and meals."'1
Care

Se-X Workers need accessible, acceptable, and goods1ree'tyr"|er'?l! Care' Preve,,tion and “re are most
uccessful if delivered together, which is referred Io as
the prevention-care synergy.'-” '"' Integrated services are
important because sex workers could be exposed to
many health risks, and follow-up is difficult'*’5'"
Referral to specialised services such as those for safe
abortion and drug treatment is essential.’5"* Metaanalysis shows that STI treatment is highly effective in
die reduction of disease transmission.” ” "’
Accessibility, acceptability, and quality care for sex
worke,3 are challenging issues jn bo(h deve]ope[I anJ
criminT8 ,C°UntrieS because of mobility. discrimination
of heahh
’ P°Verty’ VulnerabiIiV, illegal status, lack
health insurance, and unfamiliarity with the local



clinics, roadside clinics at police checkpoints, drop-in

cenlies, and general clinics in sex-work areas.'6'""-"4"’*"'’
Acccplabilily often depends on staff altitudes/
which
can
be
improved
through
sensitivity
lainmg?**’-"' Childcare and the opportunity Io rest,
bathe, and talk with other sex workers enhance
acceptability.’5-'-' Waiting times and clinic distance are
also important. Sex workers will choose clinics that are
welcoming with appropriate testing and treatment."-"5-"'’
In Managua, Nicaragua, vouchers redeemable at private,
public, or non-governmental organisation clinics were
positively received by sex workers and clinics'"
Communication can be addressed by cultural mediators
and information in different languages. 16,57, my. i so
Care and support for sex workers with HIV/A1DS i«
HWAIDsJhe UNt'DS (,°int UN PWamme'on
H V/AIDS) basic package for HIV and AIDS includesvoluntary HIV counselling and testing, psychological
support, palhative care, treatment (for pneumonia, oral
hush
vagmal
candidiasis,
and
pulmonary

facilitTlmr5’' Pr°PbyhlXis With «-‘rimoxaz.ole, and
X t."g/Om’nUnlly aCtivifc tha‘ reduce the HIV
effect.
Antiretroviral prophylaxis during pregnancy
Chest radiographs, Mantoux PPD skin tests for

anguage and culture.' '*'" Sex workers should participate
houresC'SIfn maklng ab°Ut SerViCe location and 0Petling
hours of operation."''-’" Innovative access c
■ 8
include mobile delivery, hotel-room and home-based

possible,

°r be given viable options for leaving sex work.

Pone/4.- Australian health and safety guidelines for brothels

and the sex work industry
Employer duties

Protection and prevention

J

• ^**«er^d^n^ycS|^^^^^llh0^™'rd^is.3ndg^wd,tlleirsi|fedi;pis3|

: KsesEjj-s:

...... ■“■'*

:

...... .

......

„"'x““5'"»

„pl,WB

r™el,d.1pted(ramHorm,ltionir,refrtmcesis9_i6]wji(|^m^^

2128
www-thelancet.com Vol 366

December 17/24/31, 2005

Review

Sex workers and clients sometimes use antibiotics
before or after sexual contact to prevent STIs and
HIV/w.o.
Pre-exposure antibiotic
prophylaxis
warrants investigation/1 especially for individuals
heavily exposed for short periods such as seafarers on
shore leave and part-time sex workers. However,
prophylactic antibiotic use by sex workers has been
linked to unsafe sex and presumptive periodic treatment
of SI Is in female sex workers has shown only transient
success.*’”4 Sexually assaulted sex workers should be
offered postexposure prophylaxis.

Police are toften
"
blamed for criminalising prostitutes,
but education, training, and lobbying
. j can improve
relations so that sex workers view the police as
supportive and protective.’6'" "17 The courts
-...... - should assess
sex worker
5
testimonies objectively
f
~~j and sex workers
’’vuvkuio need
11CC
the opportunity to seek redress for rights violations K.S
Courts can iinterpret
'
• law

the
to improve the lives of sex
workers. In 2000,
t’
—, the
High Court of Bangladesh
declared that sex work was not illegal and that sex
workers had the right to earn a living.95 The Court
censured state agencies for closing brothels.
Incarceration and a criminal record
Occupational health and safety
— J can interfere with
housing,
social assistance, travel, employment,
Occupational health and safety refers to workplace
education, food aid, and parenting.6'-'6'1 Illegal
issues tliat can affect employees. These principles are
rarely applied to sex work, despite many occupational immigration status drives sex workers underground,
which results in poor access to health sei’vices,
exposures, hazards, injuries, and diseases, including:
violence,■ STI or HIV acquisition, and
harassment, violence, musculoskeletal injuries, bladder discrimination,
, . .
exp
oitation.
7

6
-6
’-— -< Decriminalisation of migrant sex
loblems, stress, depression, alcohol and drug use
workers would help them access services, seek redress
respiratory infections, latex allergy, the removal of
children, and death.™” Occupational health and lor rights violations, and protect themselves and their
safety standards are justifiable only if participation in sex customers from disease.
The health-care system can treat sex workers like
work is voluntap- and does not allow the participation of
ciimmals,
which affects access to services and health
children. ■ Health and safety guidelines for brothels and
education and leads to raised rates of HIV, STIs, hepatitis
disability, and death.— Mandato.y HIV testing is an
example.
■’ Educational and training efforts can be
he/!)"0 P,rOStl‘Ution is le8a1' Progress of occupational
health and safety could be hampered by owner or successful.1' '"”- The media can shape public attitudes to
upper either criminalisation or compassion ’■'w'“
manager dismterest and the so-called one-hazard
Society disapproval of sex workers could promote low self­
approach, focusing exclusively on STIs and HIV/AIDS
Employers argue that sex workers are independent esteem, risk-taking, drug dependency, and hopelessness
own h lib™ CaSUa' en,ployees responsible for their Lite^cy, education, empowerment, and unity can reverse
own health insurance, social security, pension and this downward spiral.'"7
benefits. However, workplace safety can be improved if
sound-policies and standards are in place “nd If sex Human-rights-based approaches
HWAmst”/-Pted 3 human-ri8hts-baSed approach to
workers are allowed to organise and lobby
Ettjttonmenta1 and structural support for condom use and
' utenS‘On °f thiS appr°ach 10 sex
, all°W " suPPortive environment
redicLr/r
35 “ imP°‘tant enahli
predictor of consistent condom pse in female sex enabling sex workers to participate in, contribute to and
orkers. Forced brothel closures and treatment of sex enjoy economic, social, cultural, and political
orkers as political scapegoats make the workplace development.'"”"" Child prostitution, human traffickin
for sex work, and exploitation of migrant and mobile sex
more danoprnne
dangerous.it*
161
workers
--- o aie serious abuses of human rights.
Peer education, outreach programmes
Decriminalisation of sex workers
and
appropriate
educational
materials
have
effectively
The u/Zn/i0/ refT t0 thC rem°Val ofcriminal L™*o 2 sex
it3 r WH° SUPP°rt decrimhialisation improved, the lives of women trafficked for sex work
01 adult sex work if no victimisation is involved '""' child prostitutes, and migrant sex workers '6-w"m.i5o.i7o.i7i’
however, no consensus exists among sex workers non­ Australia0'/ ”' WhiC1>
a .RliPina “x worker in
governmental organisations, and advocates F
Drug-use successfully, I
Xai/ng nSti»
harm reduction focuses on decriminalisation of d^

haVe been deVel°Ped in AuS,raIia

iXeU).’

shluldr

iUi/ drUg indUStry' SeX Worke^

bn

should not be treated as criminals. Sex-worker
rel/rch n0/' n0n’80^rnmenta' organisations, and
research pm ects have been effective in decriminalisine
mfioTal leT /,pro,ecti"8 their fog"1 rights, lobbying fol
protect them.''-"'7
and working at the grass roots to

WWW thelancet.com Vol 366

Mrmmas da
(Little girls of the night),- a collection
of investigalwe reports by Gilberto DiLnslein, exposes
child trafficking for sexual exploitation in the Amazon
region and northwest Brazil.'” Dimenslein exposes
sexual abuses of girls as young as 9 years and as small as

December 17/24/31, 200$

2129

...... .

Review

harbours, truck stops, and bus and train stations can
reach migrant sex workers.’"’■’7S A global moratorium
• Enact and enforce sex tourism laws
should be undertaken on mandatory HIV testing, which
• Establish national databases of child sexual offenders
increases lhe risk of discrimination, violence, exploita­
• Share information across jurisdictions and foster international collaboration
tion, and disease, and promotes a false sense of security
• Provide legal migration opportunities
among clients, controllers, and governments.u'7"
• Increase and enforce penalties for exploitation
Non-governmental and sex-work organisations and
• Provide legal visa options for victims of trafficking
their projects are at the forefront of the fight against
• Enact and enforce child pornography laws, including on the internet
exploitation.1-7'7’17-'^ CARAM Asia (Coordination of
• Monitor employment agencies
Action
Research on AIDS and Mobility) produces
• Facilitate photoshop reporting of pornographic pictures, especially of children
educational information, advocates local and national
• Provide w.tness protection for victims willing to testify against their exploiters
issues, and develops interventions throughout lhe
• Outlaw methods used to circumvent the illegality of trafficking (eg, fake marriages
migiation process.171 TAMPER (Transnational A1DS/STI
temporary wives, serial sponsorship, and the bride trade)
Prevention among Migrant Prostitutes in Europe
• prosdlu9™ernment a3endes t0 reP°rt on the status of human trafficking and child
Project) supports women, transvestites, and transexuals
from eastern Europe, Latin America, Africa, and
’ e?ptoitationiOna'
Pr°9reSS
® ProstitPtion, human trafficking, and
southeast Asia working as sex workers
in Eurof
Europe.17' The
workers in
Maiti Project in Nepal provides safe spaces for returned
Support a UN-sponsored international campaign to prevent child prostitution
trafficked women and educates the so-called sending
commumties to prevent other girls from being
trafficked.’7- In rural Cambodia, 52 villages have
Telephone hotlines provide confidential access to established a community-based child protection network
information for potential or actual victims of that educates lhe community about trafficking and
intervenes for children at risk.'77 The health needs of
exploitation and for family members and friends
in,° Prosli,uH™ * summarised as
Education and training are important for agencies,
individuals, and officials that interact with victims PPEVENi-psychoJogical counselling, reproductive
health services, education, vaccinations, early detection
including youth-serving agencies, health-care workers
nutrition, and treatment.67
P f’ .-P,°±a,1S’ tai drivers’ hotel
and tour
Sex work is a common survival tactic for refugees and
guides ■ . • ‘ Sex work customers can be educated
hrough the media, information at airports and travel disp aced people to earn money for food.'4 Women and
Clinics, and John School (educational classes for sex­ children refugees are highly vulnerable to sexual
work customers, focusing on STIs, HIV, and sex violence, rape, and trafficking. Refugee sex workers need
workers rights), where former victims educate condoms, protection, access to household bleach and
needle exchange, and basic H1V/A1DS and STI
oitenders to reduce recidivism.,lM
User-friendly drop-in clinics, open-door counselling information in the language of the refugee and host
centres, camps, and shelters have been successful. 121.170.171 community.'4 Radio is an important medium for
emces at high mobility sites such as transit stations and communication. Governments are in the best position to
border crossings
ossings and in high-risk zones such as markets, implement specific strategies (panel 5).

Panel 5^ Harm-reduction strategies best suited to government action

Sex work

______

/

\

.,y

/
>

A\

/

environment

/ '

/
Predisposed I

"I
.Hr"..
Vulnerability

\ IS ’ _
\

Diminished

Harm

Harm

w

\

Supportive
| erv<ronrnent

\

/

quality of life

f

i;--------- 1



|

pr"*"

t-pZZZjfl

Reduction |

'A;:/'

framework of sex-wo.k harm reduction

\

'^proved

If

qual,ty i',e H

Conclusions
The figure shows a cconceptual’ framework
'■
for sex-work
harm reduction. Poor determinants of health17’1
are often
prechsposing,factors for individuals entering sex work
Sex workers personal vulnerability might then act
synergistically with a risky environment, exposing them
larms t at lead to a reduced quality of ’ life 179
notability, a risky environment, sex work harms, and
dirnimshed quality of life often amplify each other in an
gomg cycle. An objective of harm reduction might
be to enable sex workers to move into a more positive
cycle of empowerment, supportive environment, harm
prevention and mitigation, and improved quality of life
?ostitSionC°U
tO eVentUally leave
Ibis summary of peer-reviewed, scientific work
substantiates the many serious harms of sex work and
pmsents sunpie. safe, and inexpensive strategies to avoid

2130

www.thelancet.com Vol 366

December 17/24/33,2005

Bi
Review i

risk, mitigate harm, and save lives. Sex-work harm
reduction should be viewed as a new paradigm to improve
the lives of sex workers through debate, discussion, and
action, in the same way that drug users' lives have been
improved by drug-use harm reduction.
the sex-work industry should not be condoned,
especially if n participates in victimisation. However the
8'°bal focus on the sex work industry could result in
ndividual sex workers becoming the unintended targets
ofelnmnation and control efforts. Civil society, especially
sex work orgamsat.ons, is deeply involved in improving
day-to-day lives of sex workers, and the scientifk

17
18

19

20

So,"
W' f’T M' Dur,can R- Sexrall!' ‘ra-smilted
iniccuons drug use, and risky sex among female sex workers in
Guyana. Sex Transm Infect 2000; 76: 318
worxcis in
Pauw I. Brener L. "You are just whores-you can’t be raped"Caoe To^'r h'n Ttf
WOrnen slreet sex work^s in
Cape Town. Cult Health Sexuality 2OO3; 5: 465-81.
R'.V.anwcscnb^ck I, van Zcssen G, Straver C. Visser J
Alcolml and drug use in heterosexual and homosexual
1995 7 ITV?
rClali°n l° l’IO,cdion behaviour. AIDS Care

21

D|Ol|dTainuL’ BXnneH L eds- Cecil textbook of medicine.
Philadelphia: WB Saunders, 2000.
Patrick DM, Tyndall MW, Cornelisse PG, et al. Incidence of

22

outbreak of HIV infection. Can Med Awe J 2001; 165; 889-95
Spillal PM, Craib K), Wood II, et al. Risk factors for elevated HIV

ouEk Cf h7v

community can take an active role by using evidence­
based research to pilot innovative initiatives, assess
existing strategies, and develop a database of proven
interventions, lire participation of sex workers fn this
effort will ensure its success.

23

"onflict of interest statement
declare that I have no conflict of interest.

24

Acknowledgments

25

No external funding was given for the writing of this Review
Gina oS i
aSriStauCe
SUPP°rl "Jacqueline Barnett
1-nim^Rekart Ellen
J°Sephine Rekarl* Edward Rekart,

26

southern Vietnam whoZed^X ti^^thV^^^

V"ong iniMion dr,,B USCTS d“ri"g =■>

§3KE=B=
WHO. Sex work in Asia. Gen,-va: World 1 tel,h Or8a„iza,i„n,

sSSSSSS-

27
IO sell sex: a

References

28

29

Track,' KF. Leigh BC. Alcohol consumption and unsafe sex- a

3
30
Saratov Oblast. Rusria.

31

in

control Of HIV.1 infection. A/DS 1991 5 ("upnl n S769 7F

es a damn if I

32

o.y Hanford, Co„i1Kticul. ^ AiX^X;
6

Open Society Institute. International Harm Reduction

33

7

SS'i^X^^dir^^olajo.Bwa^j,

WaXKMK- ■
sprea'd of HIV in four cities "
i c
tnercial sex and the
15 (suppl 4): S61 J&9.
‘n SUb‘Saharan Africa- ^DS 2001;

34

Amazon city. Sex Transm Dis 1999; 26- IW-or^^5

9

:x Work,
10

and chlarnydiahnfecUo^hlconniwrci7eValenCl0fgOn°COW

Making sex work safe. London:

Russell Press 1997

Government of Quebec. Prostitution de rue- avis Ou^k
Government of Quebec, 2004.
Quebec,

11

My 9-14. 2000: 241

" PerUvian

35
infections

=f asymptomatic infection. SCX Tran,m
36

n,ernallor|al Conference on AIDS (vol 2),

37

13 workers
'X"d^xXd
SH„e±ii. Mcron m- d°™™ »■
in New South Wale<s * 1 ar'd lransBcndf’r slreel sex
a

femalesex workers ,n the Oom,,,,™
38

Z“S N ZJ Puh ^allh 2001; 25-

84-89

Vermund Si t. eds.7™ ta '

O'Ch-rneme BJ.

N.w York, Kluwer Acad,,rnic^Zm,
39
16

una’^s drus r■sw

™’’onB a coh°r‘

SvX.Xz1’5-Gcn™,oim

www.thelancet.coni Vol 366

40

December 17/24/31, 2005
2131

lb.-..--.

A

Review

41

SJS=-s='--a

66

UNICEF East Asia and Pacific. Children on the edge.
July 15W2005j,UCCr-Org/vic,,’ani/,csol‘rc”-8%.i)lnil (accessed

67

Willis BM. Levy BS Child prostitution: global health burden
icseatch needs, and interventions. Lunccl 2002- 359 1417-22

68

dBSL„CMu
!’?" ?■" ?'.■rils' Se’™"y transmitted
PubSig. 1^"“ ",e PaCifiC An"W’k-

42
43

Sadhn M, Vogler M, Lee E, Lee YS, Dubin N. Heterosexual
^--on r^HV in a cohort of couples in New York Ci!y. AIDS

44

Saab a/Vadi'tion0?1

45

HIV
infect on in Malawian
M r’8'" women. AIDS
USC and1995; 9: 293-97.
with
Hiv infection
Ki^NR arSSeInIP’ Eschenbach DA- Paavonen J, Stevens Cl-

'D- ’--ba G. Canner JK,

69

Nguyen AT, Nguyen TH. Pham KC, et al. Intravenous druB use
<nnong street-based sex workers: a high-risk behaviour for HIV
transmission. Sex Transm Dis 2004; 31: 15-19.
70 cersonTZl '0 !?”"'■ SUPP?SS'and Punish trafficking in

acu. PeMc

and cMdren’ suPP'«nenHng the

46

71
47

48

^Van
reperloire Shuman efforts to avoid sexually
77(rA? rS
na,Sible diseases:PaSt
Pastand
a’ldPresent
present.Sex
Se: Transm
88-93
2000;
/v. oo—>3.
Church S. Henderson M, Barnard M I’
Hart G. Violence by clients

72

S8;^£?

Xt/M7200P™3S22“n33Cre';i '™rk
49

73

Potterat J), Brewer DD. Muth SO, et al Mortality in ■> in, .
cohor. of prosliiule women. A^ bpidr.3^;

74

50

75

76

xSSSS?SSssi""

78

Press, 1997.

lH nSkS'J Epicleniio1 Community I leall'h 2003; 57:

^ni>ed Nations Development Fund for Women (UNIFEM)

77

53

54

“kt»SWSSSS
“'
.-' ,\"1 K' 1C,‘ud;il<0V ». Women brothel workers and
809-l*5.,Ona

51

52

2SSX’,ransnalfonal <,reanized cri™-

John U,?i'Jd N°aH„0,',‘Jpe 8'°b111 ' "V/A,DS ePidemic 2002. Geneva:
jvini united Nations Programme on HIV/A1DS, 2004.
Kilmarx PH, Lirnpakarnjanarat K, MaslroTD, Saisorn S
Kaewkungwal J Kojattana S. H1V-1 seroconversion in a

o™.,), Adrien A. Migralinn and HIV. AIDS 1W; 11 (snppl A):

and the Pacific. Westport, CT,' USA: Greenwood

Wojcicki JM Malala J. Condom use, power and HlV/Ainc • l
80

..... -



55

Cult Health Sexuality 2002: 4- 459-73
81

82

57

llw

83

58

•>«"»» righis, and
59

84

307-22.

"uacis. / alient Educ Couns 1994; 24:

60

Sex Transm Infect 2002; 78: 309.

P

h?

61
J LaW Med Ethics 2002; la 548-^

Pohcy. and
dn ig users.

62

work, and th!> HlV/STl epidemic in^l’ln’R1'118 “'r C"rnrnerciaI sex

63

ZOoT^SS:’ Sl

Sex Transm Dis 2003; 30; 46-48.

64

Russian Federation.

85

86

o ’J c°n‘r“*l’t

HeMI, are 1999; 4: 16S-S0.

^3 s

87

88
W. Hargono R.

89

65

Hastings Law J 1994; 45: 1565-92.

3 Pr3,U“70d,scasc
": ™- rid's

rr’ “■s,mai i,eaiih

worker., in Sydney. Cu,ilmiri„ MrVi TOfa72."17‘:rr',1a',Ona, Sra

decnminalisalion of prostitution in Asia. Lancet

SX^He
3nd tSex
l,C Behav
1991;20:205-18.
8
of Prost|lut'on. Arch

........ .

90

i>nmunoderlcieX^vi1lus\t,ifMtanhrnda'\r:C'al' I IU"’a"
iransmilted diseases amone north' d‘^lf;lralrnent for sexually
1996; 23: 264-69
8
C™
'”Cn- S“ ^anstn Dis

^XVhS^^liiS-P^Vlaxis a,not,RCo.„,nereis,
PP-PtkrnsoreHieacS^X'Sr^'S0^^11

^32
•.vww.thelancet.com Vol 366

.1

December 17/24/31, 2005

3
*

Review

91

J Acquit Immune Defic Syndr Hum Rctrovirol J997; 14: 169-73.
92 Rojanapithayakorn W, Gocdken J. Lubrication use in condom
promotion among commercial sex workers and their clients in
Kalctiabun, Thailand. J Med Assoc Thai 1995; 78: 350-54
93 ?!anch?rc! |F’0’Neil/ Rair*esh BM, Bhattacharjee P, Orchard T
Moses S. Understanding the social and cultural contexts of female
*a,r”alaka' India: implications for prevention of
1 ,,V mfection. J Infect Dis 2005; 191 (suppl 1): $139-46.
94 Yadav G, Rcflcin S, Ngugi E, et al. Associations of sexual risk
akmg among Kenyan female sex workers after enrollment in an
329-34P
’°n tna J Aa}l,ir ,mmune Def'c Sy>'dr 2005; 38:
95

UNAIDS Summary booklet of best practices. Geneva: Joint United
Nations Programme on HIV/A1DS, 1999,
CQ
96 Ngugi EN, Wilson D, Sebstad J, Plummer FA. Moses S. Focused
peer-mediated educational programs among female sex workers to
virtisCtransmiss *ra,'SIdiS?Se and hu’nan immunodeficiency
174^"^): 524^7
Zimbabw^
D" ^96;
97

Fvah.nt’ W’rarWan DN■ Suastina SS, Reed BD. Muliawan P.
in B li I0”/3 PeCrrcducation P^amme for female sex workers
m Bab. Indonesia. Ini J STD AIDS 2000; IT 731-33

98

?zr™V3™a5"y55rmissible infKfons o,,'ct hiv99 Merson MH, Dayton JM, O'Reilly K. Effectiveness of HIV
in dCVdOpi"'’

u

2000;

115 Williamson C, Folaron G. Violence, risk, and wiva! strategies of
street prostitution. West J Nurs Res 20' ’
116 Ovcrs C. Seropositive sex workers ami :
-vtnlion a
1992- pr]’Calis,ic p0licy dcvdoP’r“-'nl. AIDS Ileallh Promol P.xch

117 l-ord N, Koetsawang S. A pragmatic intervention to promote
condom use by female sex workers in Thailand, indl World
Health Organ 1999; 77: 888-94.
118 ?ai Prl0iecl;T"ck.s °r thc lrac|e: health and safely lips for street
2005)°rke,S' 200°' hllp://www-hannredllchon.org (accessed July 13,

119 Hanenberg; RS Rojanapithayakorn W. Kunasol P, Sokal DC.
dedhw nf rha,l^d S HIV<ontrol programme as indicated by the
decline of sexually transmitted diseases. Lancet 1994; 344; 243-45
120 Buckingham RW. Moraros J, Bird Y, Meister E, Webb NC Factors
associated with condom use among brolhcl-bascd female sex
workers in Thailand. AIDS Care 2005; 17: 640-45.
121 Asthana S. Oostvogcls R. Community participation in HIV
P^e'ition: problems and prospects for community-based strategies
among female sex workers in Madras. Soc Sci Mcd WM; 43- 133^48
122 Bandyopadhyay N, Banerjee B. Sex workers in Calcutta organize
Jemselves to become agents for change. Sexual Health Ilxch 1999;
123 Ness RB, Randall H. Richter HE, et al. Condom use and the risk of
SS fP(1 V'C •innammal0’7 d’sease, chronic pelvic pain or
Am P I ■
3n ep‘SOde °rpelvic ’Hhammalory disease
Am j Public Health 2004; 94: 1327-29.
Z
124 UNAIDS. Tlie public health approach to STD control UNAIDS

SwBsSi

Wnl U"iled N3,io"s ’’'Wtne on

125 Gerofi J, Deniaud F, Friel P. Interaction of condom design and
W-Z?"’91108 and COmlOni acccPlabili,y- Contraception 1995; 52:

102

G; WySS.K’ N’Diekhor V. Prioritization of prevention
actnities to combat the spread of HIV/A1DS in resource

126 rfontanel AL. Saba J, Chandelying V. cl al. Protection against

sexually transmitted diseases
the choice of using the male
randomized controlled trial. AIDS 1998; 12- 1851-59
127 Denjtiud F. Current status of the female condom. Sante 1997; 7:
128t.French PP, Latka M. Gollub EL. Roi
’gers C. Hoover DR, Stein ZA.
se-effectiveness of the female versus i
male condom in preventing
sexually transmitted disease in women
. Sex Transm Dis 2003; 30: ’
433-39.

Xu
-

106

FOT “

in , s£th

s,,ca ““ h,v
J7

ws S"'

«r„Lrdal

Na,i0"s p'°8r~

ns
meva:

Gei

79°f

2-3 oXz’X*
135 KbaXomu^g^hX^or^; E;U™rakul S' Cdc">’™ DD.

i

sex
>12 Alexander P. Key issues jn SCx work-related HlVMmerrrr.
prc™..on i^entas. AWS
137 Mak R. Traen A. Claeyssens M, Van Renterghem I

...,
114

''««» expericKe ^eMri

'v',rk'',s:
•: the

sex

iex Transm Infect 2003; 79: 157-59.

X'T " ........
www.thelancet.com Vol 366

December 17/24/31, 2005
2133

J ^Revievv

139 Fonck K Kaul H. Keli F. et al. Sexually transmitted inreclions and
vaginal douching m a population of female sex workers in Nairobi
Kenya. Sex Trunsni Infect 2001; 77: 271-75.
’’’
140 ydkinS°n ?'
M. Ramjee G. Rutherford GW. Nonoxynol-

141 Hilher SL Moench T, Shattock R, Black R, Reichelderfer P
Vreonese F. In vitro and in vivo: the story ofnonoxynol 9 '
J Acquir Immune Defic Syndr, 39: l-g.
142 Ghys PD, Diallo MO, Ettiegne-Traore V. et al. Effect of
iXT""°nAuv ;2S:SeXUally tran^itted disease on the
1421-31 f H V lnfcCt,On ,n female sex workers. AIDS 2001; 15:
143 tes^iTserU1^17; ln,egra!in^ HIV volu»‘ary counselling and
testmg services into reproductive healtli sellings: stepwise
if°r progra'nmc Panners, managers and service

oSX3’ Plan"ed

Regional

157 ind^r1’ Dil>’S oWpbCr L Risky busi,,ess: h(;al,h
s^ely in the sex
y <)Vei 3 J yC<" PerK,d- Sex 1 rtu,::’n Infect 1999: 75: 340-43
158 UNAIDS H1V/A1DS and human rights. New York and GenevaUnited Nations Publications. 2001
’M X 200M,U^!' W:"“- '?allh a‘,d
guidelines In,
13, 2005)
’ hllp //www-workcov«’r-nsw.gov.au (accessed July
160 Australian Capital Territory (ACT) Workcover. ACT sc
sex industry
W B. 2005)^ 1989‘ h,lp://WWW■WO1 kcovcr.act.gov.au (accessed

161 Scarlet Alliance with the Australian Federation of AIDS
<= Jn nuatllonAs- A 8ulde t0 best practice: occupational health and
safe yin the Australian sex industry, 1999. http://www.
scarlclalhance.org.au/pub/ (accessed July 13. 2005)
162 ^Ir?ara1tn1a1S111'in,dan CP- Sirhorachai A. Mandel JS. "Superstar"
model brothel": developing and evaluating a condon?

163
164 UNDTM3li°? DevcloP’nenl f;u»d for Women (UNIFEM)

165 Hepburn M. ProsHnuinn: »1)ukl . ................... ...... „M, WJ;

■“gssss==s^

167

,M-,rar“nlola JM- i-™-’ vulnerability to human riehls
OxrXX^
in lhe -Id U. New York:
Nations ProgrX^e mi'I II V/AIDs" ZOOJ^' G™0™ J0"" United

148 ° Co',nor CA, Patsdaughter CA, Grindel CG, Taveira PF

AIDS I alienl Care STDS 1998' 12- 931-37

"=“=«=t=^=r

"’SXi.S'£SK“-v.
151 Gilks CF, Codfrey-Faussett P, Batchelor BI et al rU, ♦

cCytol" gy. cX«4y md PCR fo .“pv’“
Singapore. imy.STO AW5 2001

169 Aral SO Mann JM, Ramgbose O. Commercial sex work and STD-

’f'1’'of“rVtal



.... ......... .

"’SEJssnrs”—1

-Workersin

153 M
’,
a"smissi0™"iaft<'’"-apy
a
in Taiwan.
J mfiefoi. 20% IM 879-85aHt,v',ramTC!ro
LF-el’^-PP^ia.eness

154

among rural Zimbabwean

P surnplive periodic treatment

s-- "• - *

Asia^nd the Pacific. AIDSP}^:^1^1”,V/AJW“

78 He to?’5 7alrablC “"""““'y-

3

2001: 358; 2095-29

. --

organization. Geneva:

' ’ ^XaUlyX'^t^' 2U«.7™''8 S- Social and

exotic dancer.,. Cm J

>

-rk of

2134
www.thelancet.com Vol 366

December 17/24/31, 2005

MH~
Sholk-chll.qxd 31/10/03 12:55 Page 199

11
Etic and Emic Categories in Male Sexual Health:
A Case Study From Orissa
MARTINE COLLUMBIEN, NABESH BOHIDAR, RAM DAS,

BRAJ DAS, AND PERTTI PELTO

INTRODUCTION

This chapter presents a case study from intervention research on male sexual
health in eastern India. The research was commissioned by the Department for
International Development (DFID) who started funding contraceptive social mar­
keting in the state of Orissa in 1995. Making use of commercial distribution chan­
nels, the intervention sells branded condoms and oral contraceptives at subsidized
prices. The focus of the advertising campaign was on the promotion of the concept
of child spacing and the increased use of reversible methods of contraception.
Research among the target group in urban areas showed that men reported higher
levels of condom use than women (AIMS-Bhubaneswar 1996). Male use for extra­
marital sex and under-reporting by women due to cultural sensitivities was suggested
as reasons for this discrepancy. With the rising concerns about the growing AIDS
epidemic in India, there was need for more research on sexual behaviour, condom
use, and the perception and recognition of sexually transmitted diseases. Starting
from the public health paradigm of control of STD/HIV, a study was designed to
learn about the sexual health problems as perceived by men in the community and to
measure the need for condoms by identifying and quantifying sexual risk behaviour
among single and married men.
The data presented focuses on sexual health concerns, and on the way the local,
emic perspective of male sexual health is at odds with the biomedical model of sexual
health, which underlies current public health efforts to halt the spread of HIV/AIDS.
Based on focused ethnographic research which preceded a carefully constructed
quantitative survey, we show that men’s concerns about sexual health comprise a
complex array of symptoms, of which a major portion are not directly related to
sexual transmission, but rather a reflection of their worries about semen loss. On the
other hand, the data indicate general, widespread awareness of sexually transmitted
We gratefully acknowledge DFID India and the British Council for funding the research presented in this
paper and IUSSP for inviting and supporting our participation in the seminar.

Sholk-chll.qxd 31/10/03 12:55 Page 200

200

Martine Collumbien et al.

infection, including at least superficial knowledge of AIDS. The study suggests that
programmes of sexual health information, as well as clinical services for males,
should be equipped to provide counselling and other services that go beyond simply
concentrating on diagnosis and treatment of sexually transmitted infections (STIs).1
We start by describing the setting of the study, followed by a discussion of the
focused ethnography methods and the survey data used in the research. We then set
out the biomedical model of sexually transmitted diseases, as currently used in public
health interventions of STD control, with details on the Indian context. The data
collected reflect the emic perspective, using men’s own vocabulary and criteria for
categorizing illness. The symptoms and perceived causes of the main conditions are
described and the local categories of sexual health concerns are analysed. The promi­
nence of psychosexual concerns of semen loss in the qualitative data are confirmed
with use of the population based survey data and differentials in reported experience
of semen loss are presented. Semen anxiety is then discussed in the context of the
ethnophysiology of sex as understood in South Asia, drawing on the existing anthro­
pological and psychiatric literature. The chapter concludes challenging the categor­
ical paradigms adopted in international health, and calls for more holistic healthcare.

THE SETTING
Orissa is a state in East India, with a coastline on the Bay of Bengal. It borders West
Bengal and Bihar in the North, Madhya Pradesh in the West, and Andhra Pradesh
in the South. Its coastal plains have for centuries served as a link between north and
south India and are a more developed region than the mountainous areas. Of all
states in India, Orissa has the second highest concentration of tribal people: the
sixty-two different Scheduled Tribes make up 22 per cent of its population. Of
the total population of 32 million enumerated in the 1991 census, 86 per cent was
classified as rural.
Orissa is one of poorest states in India. Using the Planning Commission poverty
line, (Datt 1998) estimated that over 40 per cent of Orissa’s urban and rural popula­
tion was living in absolute poverty,2 compared with a national average of 35 per cent.
The adult literacy rate is one of the lowest at 49 per cent, though rates are higher in
the coastal developed districts (60 per cent). Women are less likely to be literate than
men and strong social norms on women’s mobility prevail. About 95 per cent of
the Orissa population is Hindu, with 2 per cent Muslim, and 2 per cent Christian.
The general condition of poverty in Orissa contributes to poor reproductive and
child health. The total fertility rate (TFR) in urban areas is 2.5, and around 3 for rural
women. Infant and child mortality rates in Orissa are among the highest in India.
Morbidity in small children is equally high, with diarrhoea and malaria as main
contributors (PRC Bhubaneswar and UPS 1993).
The study area for the research on sexual health and behaviour was limited to the
four coastal districts3 with low concentrations of tribal people: Puri, Cuttack,
Balasore, and Ganjam districts. The distinctly different cultures of the scheduled

Sholk-chll.qxd 31/10/03 12:55 Page 201

Male Sexual Health Categories in Orissa

201

tribes in Orissa suggest the need for a separate, in-depth study in the tribal areas.
Such study was not feasible in the project reported here.

METHODS AND DATA
The strategy adopted for the qualitative data collection broadly followed the guidelines
for doing Focused Ethnographic Studies (FES) (Pelto 1994; Pelto and Pelto 1997). The
focus of the data gathering was on information needed to answer programmatic
questions in sexual health and condom promotion interventions. The fieldworkers
were trained in in-depth interviewing, social mapping, and the various structured
interviewing techniques.
In total, 17 sahis (localities: colonies, neighbourhoods or hamlets) were studied
in-depth by a study team of four male and two to three female researchers. The
average number of days spent in one location varied from 7 to 10 days, depending on
the availability of the informants, and initial time taken for rapport building. Though
locations were studied in all four districts, nine were in Puri district, in and around
the state capital Bhubaneswar and Puri town, famous for its beaches and the Lord
Jaganath Temple, which attracts thousands of pilgrims and tourists throughout the
year. Certain sahis in Puri are known to have a high prevalence of casual sex and men
having sex with men. Sahis were thus selected according to expected variations in
sexual behaviour and access to condoms. Four localities studied were rural. Each sahi
was treated as a separate case and a detailed study of the various role players in each
location was carried out. Data gathering and analysis took about four months to
complete from May to September 1997.
Participatory mapping exercises were carried out with informal groups of men
and women in the community. It was used for introducing the research topic in a sahi
and for motivating local people to participate in ongoing activities. Mapping
involved drawing a social and resource map of the community with identification
and marking places of health providers, places of recreation, places where liquor is
available, where condoms are available and places where men go to find or meet
partners for sex (Pelto et al. 1998). Other informal group discussions were held to
obtain further situational data and to identify appropriate key informants for more
in-depth interviews.
In-depth individual interviews were done with both key informants, and case
study informants. Key informants were selected for their extensive knowledge about
local cultural beliefs and practices and the conversation focused on local percep­
tions and behaviours. They included ordinary community members, outreach
workers, medical practitioners, traditional healers, and retailers selling condoms.
Case study informants included individuals who had experienced sexual ill-health
and others who engaged in risky sexual behaviour. The conversation focused on
their personal lives, to elicit illness episodes, case histories, and sexual histories. Each
case was contacted several times in order to build up rapport and to permit prob­
ing of sensitive issues.

Sholk-chll.qxd 31/10/03 12:55 Page 202

202

Martine Collumbien et al.

Information on peoples’ own explanatory models of sexual illness and local
vocabulary was generated through various structured qualitative techniques like free
listing, pile sorting, rating and ranking, and time-lines (Weller and Romney 1988).
This chapter draws heavily on these data for understanding the cultural perceptions
of men’s sexual health problems.
Free lists were done to help isolate and define the domain of sexual illnesses. By
asking key informants and case study informants during in-depth interviews ‘what
are all the sexual health problems men experience in this community?’ a list of sexual
health concerns is defined by the informants in their own language, using culturally
relevant categories. Given that there is no exact Oriya term for‘sexual health concern’,
it was defined as ‘problems relating to, or affecting, the genital area’. The usual key
terms in Oriya language included asubidha jounanga (problem of sexual organs),
jouno rog (secret disease) and jouno sambandhya rogo (disease relating to sex).
Pile-sorting was the next step. A selection of conditions produced by the free lists
were written on slips of paper and informants were asked to sort the conditions into
separate piles of similar illnesses. The criteria for similarity were left to informants
to decide and the reasons mentioned for grouping revealed details of explanatory
models. After the pile sort exercises the same cards were used for severity ranking of
sexual health concerns. The men were asked to group the concerns/illnesses into
three groups ‘severe’, ‘intermediate’, and ‘mild, not severe’.
Fieldworkers took notes during the group and in-depth interviews. These were
expanded and written out immediately afterwards. Transcripts of all interviews were
Borgatti
coded and analysed in Ethnograph. The software program anthropac (Borgatti 1996) AU:
(1996) is
was used to analyse the free list and pile sort data. The Multi-dimensional Scaling not listed.
procedure was adopted to provide a visual representation of how conceptually close p|s- chkor far the different illnesses were. Cluster analysis was a second method used to inter­
pret the pile sort data.
The structured survey to estimate the extent of sexual risk behaviour and the need
for condoms in the general male population followed the qualitative fieldwork. The
findings from the qualitative study were used in the design and refinement of the
survey instrument, mainly in terms of using the correct local vocabulary and defin­
ing coding categories (e.g. categories of partners, locations where people have sex).
A free listing question on sexual health concerns in the community was also included
in the questionnaire. The free list was included towards the beginning of the inter­
view, mainly as a rapport-building question before moving on to the more sensitive
personal questions on sexual behaviour.
This survey covered a large population-based random sample (n = 2087) of single
and married men in urban and rural areas of the four coastal districts in Orissa.
A multi stage random sample was obtained by randomly selecting two community
development blocks and two urban areas in each district; the second stage listed the
villages in the blocks and the wards in the urban areas and from these four sampling
clusters (villages or wards) were selected with a probability proportional to size.
In ach village/ward all houses were mapped and numbered. After enumeration,
thirty-two households in each sample cluster were randomly selected. All members

Sholk-chll.qxd 31/10/03 12:55 Page 203

Male Sexual Health Categories in Orissa

203

age 15 or over in the household were listed. All eligible members for individual
interview—men age 18-35—were ranked by age. According to the number of eligible
men in the household, the youngest, second youngest, etc.... was selected in strict
rotation in consecutive households to ensure randomization.4 The selected respon­
dent was then invited for an interview in privacy outside his home. The pretest indi­
cated that the interview needed to be preceded by a rapport-building chat of about
half an hour. Refusal rate for interview was as low as 1 per cent.
The statistical analysis of survey data was done using SPSS software. The data file
was weighted according to urban/rural residence and the size of the district, to make
it truly representative of the four coastal districts.

THE ETIC PERSPECTIVE OF SEXUAL HEALTH IN
THE HIV/AIDS ERA
Sexual health has been defined as the integration of the somatic, emotional, intellec­
tual, and social aspects of sexual being (World Health Organization 1975). However,
this holistic definition of sexual well-being is in practice not adhered to by public
sector health programmers and health planners. Before the emergence of AIDS as
a serious public health problem, the sexual health of men in low income countries
received very little attention indeed. While sexual health gained attention since the
AIDS pandemic, the focus has been primarily on disease transmission. There has
been increased interest in the public health importance of the ‘classic’ STIs since they
are shown to facilitate the transmission of the human immunodeficiency virus
(HIV) (Laga et al. 1991; Grosskurth 1995). Control and prevention of these STIs are
also important goals in their own right. According to the World Development Report
of 1993, STIs are the second most important cause of loss of healthy life years in
women of childbearing age worldwide (World Bank 1993). In India, there are an
estimated 40 million new STI infections a year; prevention programmes and facili­
ties for diagnosis and treatment are seriously neglected (Ramasubban 1999).
While HIV/STIs are a major burden to health in India and elsewhere, the control of
these infections has been problematic, even from a purely clinical point of view (let
alone the social barriers to access of care). There are no simple accurate diagnostic
tests, and comprehensive laboratory services for microbiological diagnosis are very
expensive. Common reported symptoms can be caused by different pathogens. To
address the problems of clinical diagnosis and low specificity of symptoms, syndromic
management has been widely propagated, and WHO developed standardized
protocols for the treatment of men and women in basic health care services in less
developed countries (World Health Organization 1997; Mayaud et al. 1998). These
flowcharts of clinical algorithms offer diagnosis on the basis of syndromes (group
of symptoms and easily recognized signs) reported by the patient, and treatment is
recommended for several possible causative organisms.
International planners and donor agencies now tend to regard STD control and
syndromic management as a valuable strategy for HIV control, partly because of the

Sholk-chll.qxd 31/10/03 12:55 Page 204

204

Martine Collumbien et al.

challenges faced trying to enforce condom use (Lambert 1998). In India, the use of the
syndromic approach has been opposed by qualified medical practitioners who consider
it as not comprising ‘real’ medicine, and because of competitive concerns about
training practitioners who are not biomedically qualified (Lambert 1998). According to
Indian guidelines, clinicians diagnose cases on an aetiological basis only and they do
not recommend co-treatment when the aetiology is uncertain. A study on the adequacy
of STD case management in public and private health facilities in India (Mertens et al.
1998) shows that only 10 per cent of the STD patients were satisfactorily managed.
The study concludes that by promoting the syndromic approach to STD management
and thereby simplifying existing guidelines, doctors in India could provide better care.
With the introduction of STDs into the framework of reproductive healthcare, the
provision of syndromic STD management for women attending family planning
or antenatal clinics has received high priority on the international health agenda.
However, in general populations with relatively low STD prevalence the indiscrimi­
nate application of syndromic management may lead to overtreatment on a large
scale (Hawkes et al. 1999) and more recently the integration of family planning and
STD treatment has been critically reviewed (Lush et al. 1999; O’Reilly et al. 1999).
The diagnosis of STIs in women is very complex, not only because of the relatively
low specificity of symptoms of possible infections, especially vaginal discharge, but
also because of the high rate of asymptomatic infections. Fifty per cent of women
with STIs will not have any symptoms, whereas their infected male partners will
usually experience pain and other signs of infection (Hook and Handsfield 1999;
Stamm 1999). Consequently, it has been concluded that the resources for STD con­
trol should be concentrated on individuals at higher risk and men should be targeted
(Cleland and Lush 1998; Hawkes 1998). Symptoms are more specific in men com­
pared with women, and the cost associated with overtreatment is therefore likely to
be lower in men than in treating equivalent symptoms in women. Policies to provide
clinical services for men may as a consequence reach asymptomatic but infected
women through partner notification strategies. Moreover, it is reasonable to assume
that in South Asian societies, men are the ones who are more likely to initially con­
tract STDs and transmit them to their wives (Mundigo 1995).
From the discussion above it is clear that the rising concern and need for HIV/STI
control programmes provided a biomedical rationale to target men with sexual
health services (Collumbien and Hawkes 2000). However, the public health paradigm
focuses quite narrowly on the disease category of sexually transmitted infections.
Within the biomedical treatment framework subcategories are thus identified as
syndromes, or symptoms of disease: for example, ‘urethral discharge and pain passing
urination’ as the syndromic diagnosis of gonorrhoea and chlamydia, and ‘genital
ulcers’ as an indication of syphilis or chancroid. The incidence of these syndromes
a^e also used to monitor the success of AIDS programmes in reducing risk behaviour
by increased condom use and adoption of other safe sex practices. Two standard
questions on self-reported symptoms of STIs are routine in national HIV/AIDS
monitoring surveys. One question asks about the incidence of sores and ulcers on the
penis, the other about pain during urination with discharge from the penis.

Sholk-chll.qxd 31/10/03 12:55 Page 205

205

Male Sexual Health Categories in Orissa

This etic framework of STD case management, a problem-frame defined by
biomedicine and based on a symptom-based syndromic approach informed our
research design. Despite this predetermined focus on sexually transmitted diseases,
as relevant to the intervention of condom social marketing and STD treatment
seeking, we started the fieldwork by investigating the local categories of sexual health
concerns to ensure the correct use and understanding of the local terms for the
symptoms we wanted to capture.

THE EMIC CATEGORIES OF SEXUAL HEALTH IN ORISSA

To define the boundaries of the cultural domain of sexual health concerns in local
terms, free lists were obtained from thirty five male informants deemed knowledge­
able about sexual health related issues. The most commonly mentioned problems are
presented in Table 11.1. The concerns are ranked by the number of informants listing
them, and the local terms have been given an approximate English ‘equivalent’.
The number of respondents mentioning a condition—or the frequency—is one
indication of the importance of a health concern in that community. When a condi­
tion is mentioned first or second it is also more ‘on the person’s mind’ than when it
comes lower down the list. The frequency and rank order combine in a measure
Table 11.1. Most Commonly Mentioned Sexual Health Problems in Free Listing
Frequency Average Salience
(n = 35)
rank

Sexual health problem

Local term

English translation

Dhatu Padiba
Jadu
Swapnadosh
AIDS
Handling
Gonoriha
Linga-gha
Parishra-poda
Hernia
Fileria
Hydrocele
Katchu
Syphilis
Bata
Linga-ghimiri
HIV
Ulcer
Malakantaka
Chau

Semen discharge
Itching
Nocturnal emission
AIDS/HIV
Masturbation
Gonorrhoea /generic term for STI
Ulcer/sores on the penis
Burning during urination
Hernia
Swollen penis, scrotum, leg, and foot
Swollen scrotum
Itching—scabies
Syphilis or generic term for STI
Rheumatism
Eruptions on penis ~ herpes
HIV
Ulcer
Fistula
White patches—skin infection

28
22
19
17
14
13
9
9
7
6
6

6
4
4
3
2
2
2
2

2.750
4.636
3.211
2.529
5.000
2.231
3.222
3.333
4.714
5.500
5.333
3.667
3.500
3.250
1.000
5.000
4.000
2.000
3.500

0.424
0.195
0.242
0.353
0.106
0.285
0.106
0.095
0.093
0.035
0.066
0.090
0.061
0.040
0.086
0.024
0.014
0.046
0.029

Sholk-chll.qxd 31/10/03 12:55 Page 206

Martine Collumbien et al.

206

of salience. Dhatu padiba (semen discharge) was the most frequently mentioned
problem, with 80 per cent of informants listing it. On average, it was the second
or third concern mentioned by the individual informant and it is thus clearly the
most salient concern. Although jadu (itching) was the second most listed concern,
on average it was further down the lists than swapnadosh (excessive nocturnal emis­
sions) and AIDS. In fact, AIDS comes out as the second most salient concern.
From this simple listing exercise it is clear that the emic perspective of sexual health
in Orissa includes sexually transmitted infections. However, it is also evident that
there are a range of other conditions that seem important concerns. Most items on the
list refer to symptoms rather than specific diseases. To avoid mistranslation from local
into biomedical terms we need to explore the categories in terms of the local under­
standing of disease aetiology. How and why men perceive different conditions to be
related to one another can illuminate local explanatory models of disease.
Pile sort data give a better understanding of categories of ‘similar’ diseases. The
more often two illnesses or symptoms are grouped together by different informants,
the closer they are conceptually. In multidimensional scaling (MDS) similarities are
translated into distances and concerns considered very similar appear close to each
other while illnesses that are not related will be furthest apart. Figure 11.1 gives an
example of a cognitive map by two-dimensional scaling of sexual health concerns.
The encircling of the four groups of illnesses/conditions is based on cluster analysis
(Johnson’s hierarchical clustering). The multidimensional scaling picture (stress
<0.15) together with the cluster analysis indicate that the Orissa men make fairly
clear distinctions among the types of sexual problems, particularly separating the
infectious conditions (group C) from the non-contact, semen-loss problems in
group D. They also recognize that hernia/hydrocele (group A) are a different kind of
problem, with different aetiology.

A^.—
x.

Hernia
Hydrocele

E
-----------

Parishra band
Parishra poda

Swapna dosh dhatu padiba

Linga sugar
Linga bindhiba Lingzghimiri

c

Syphilis
linga gha

Figure 11.1. Multidimensional Scaling of Sexual Health Concerns in Orissa

Sholk-chll.qxd 31/10/03 12:55 Page 207

Male Sexual Health Categories in Orissa

207

These maps are a means of exploring patterns among illnesses and this is supple­
mented by the information given by the sorters on reasons for grouping items
together. In other localities other cards were sorted, which resulted in different
conceptual maps (not presented). One common feature to all maps was the clear
distinction men had made between sexually transmitted diseases and others. In their
mind they have no doubts that linga cancer (wounds on the penis) and jouna
gha/linga gha (jouna gha can be translated as sores on the penis) can be clubbed
together with syphilis or gonorihia.
Five broad categories emerged from these data: sexually transmitted conditions,
conditions related to semen loss, skin infections, anal conditions, and a final category
grouping other diseases which affect other parts of the body besides the genital area.
We discuss these categories in more detail, especially the first two, giving indications
of reasons why conditions were grouped together, mainly reflecting indigenous
explanatory models of disease aetiology.
Starting with the sexually transmitted diseases it is clear that local Oriya vocabu­
lary is influenced by allopathic terminology, with the use of medical terms like
Gonoriha and Syphilis. However, these terms do not necessarily translate directly
into the specific medical diseases. They more broadly indicate conditions which are
sexually transmitted, and can thus be understood as generic terms for STDs. When
probed with questions about sexual diseases (jouno rog), most informants directly
mentioned gonorrhoea’. They described the symptoms like linga-gha (ulcers in
penis) or ling-ghimiri (small eruptions on penis) which correspond more to the
medically defined infections syphilis and herpes, respectively. Men did attribute
symptoms within this group to sex with sex workers.
AIDS was frequently mentioned in the free lists and came up as the second most
salient concern. Sexual transmission was clearly understood, though many inform­
ants stressed the fact that it was due to sex with ‘many’ partners Tf a person is having
sex relation with many women then AIDS may be transmitted (jadi kehi adhika
mahila sange samparka rakhe, tebe aids heba)’. It is generally described as dangerous
and very serious. There are indications, however, that the term ‘AIDS’ is becoming
another generic term for sexually transmitted disease. This is illustrated by an
informant expressing anxiety about having AIDS:

I had Lingare Gha (sores on penis) which used to be painful. It also had pus. Now there is no
pus but the Gha is still there. It is not getting healed. I do not know what to do.... when I got
this disease, I went to Dr. Nanda, he said and wrote on my prescription that it was AIDS. He
then prescribed both medicines and injections and said it can be cured.
The contradiction of ‘curable AIDS’ suggest that the informant suffered another
STD which was labelled as AIDS. In locations where AIDS had been mentioned, most
informants in the pile-sorting exercise grouped it as a separate category. When it was
associated with other items, it was grouped with sexually transmitted illnesses, and
this was confirmed by in-depth interviews.
A second category of concerns are related to semen loss. The most salient condition
was dhatu padiba (or meha padiba), which is best described as involuntary semen loss.

|Sholk-chll.qxd 31/10/03 12:55 Page 208

208

Martine Collumbien et al.

Symptoms include secretion of semen during urination or defecation, and secretion
of semen during erection. It is associated with thinning of semen or the quality of
semen. Men mention dhatu padiba as a secretion of milky, or chalky watery or semi­
liquid substance from the penis. Several informants relate their personal experience
and the following extract points to the fact that it ‘overcomes’ men as they seem to lose
semen ‘without their knowledge’.

informant: During defecation dhatu comes out. Not all the time but some time even a
large amount of dhatu comes out during urination. Even I am having this problem.
Most of the time dhatu comes out without my knowledge.
interviewer: How long you are suffering from this problem?
informant: It will be around last seven year I have been suffering from this dhatu
padiba
In the same category, men mentioned swapnadosh or nocturnal emissions which
are thought to be abnormal if they occur more than two or three times a month. It
is a common concern among youths. Men associate swapnadosh with both dhatu
padiba and with handling (masturbation). The difference with dhatu padiba is that
swapnadosh occurs during sleep and after erotic dreams. An informant asked to
differentiate between the two said ‘in both the cases dhatu comes out spontaneously
without our knowledge, but night fall sometimes is related to dream problem
(Swapna pai)’. The link with masturbation follows from this quote ‘since boys watch
blue-films and always think about erotic acts, they indulge in masturbation when
alone. Excessive dwelling on sexual thoughts results in erotic dreams and seminal
emissions’. The result is weakness, loss of weight, and memory loss.
Most people attributed dhatu padiba to various physiological factors. The general
causes stated are: excessive heat in the stomach (peta garamf improper diet and
strain due to hard physical labour. Other informants attributed semen loss to exces­
sive masturbation. One of the key informants states ‘Excessive masturbation cause
widening of the urethral opening making dhatu padiba easier (Besi muthimariley
parisra dwara chouda hoi jaye o dhatu padiba sahaja huey)’. Another explanation for
dhatu padiba is the absence of sex because of which the accumulated semen gets
discharged. One informant states: ‘Long days of abstinence causes accumulation of
semen which produces heat in the body and semen discharge occurs during urina­
tion. (Bahuta din kichi nakaley bija jami jaye o deha heat hoijaye. Parisra kala bele
dhatu padey)’. The immediate result of this situation is irritation and pain during
urination, physical emaciation, weakness, body pain, head reeling, or even ‘death’.
One respondent states ‘Dhatu padiba results in complete loss of physical power
(Dhatu padiley deharu sabu bala palai jaye)’ and thus it interferes with a healthy sex
life. Many of the informants indicated anxiety over the loss of dhatu (semen) and
perceived it as affecting their married life.
Skin infections formed a third emic category of disease. Jadu was most frequently
mentioned as a very common skin infection, usually affecting the inner thigh and
groin. It also affects the testes and leads to severe itching. Other skin infections are
chau and kachhu. Men made a distinction between bayasa chau, which is common

Sholk-chll.qxd 31/10/03 12:55 Page 209

Male Sexual Health Categories in Orissa

209

among elder people, and not perceived as an illness and dhala chau, which are white
patches on the skin. Kacchu is scabies and this can affect any part of the body. Skin
infections generally are explained by unhygienic conditions Tf a person doesn’t clean
his body after the day’s hard work then there are chances of having these problems’.
A fourth group are anal conditions, such as malakantaka which are wounds
(fistula) and arsa which leads to severe pain during defecation. In contrast to
malakantaka, bleeding occurs in arsa and this agrees with the medically recognized
piles or haemorrhoids. The final group of concerns affecting the sexual organs are
filaria (elephantiasis) with swelling of the feet and legs which extends to swelling of
the penis and scrotum. In hydrocele, there is enlargement of the testicles due to water
accumulation.
The severity of illnesses was elucidated by ranking exercises. Informants were
asked to rank order the sexual health conditions in terms of severity as follows:
3 = ‘severe’, 2 = ‘intermediate’ and 1 = ‘mild, not serious’. All informants rated AIDS
as severe, so it has a mean severity of 3.00. Other conditions suggesting sexually
transmitted infections, such asjounagha, linga cancer, syphilis, gonorrhoea with mean
values of 2.00. The various psycho-sexual concerns received more varied severity
rankings. Among them dhatu padiba was considered most severe with a mean rang­
ing from 1.50 to 2.42. For swapnadosh it ranged from 1.50 to 1.92. Skin infections,
like jadu and kacchu were generally ranked as not severe (a mean ranging from
1.00 to 1.38. With average ratings between of 1.50 and 2.00, the anal conditions were
considered more severe than skin infections.
These relatively simple structured qualitative methods taught us a lot about the
emic categories of sexual health. It also drew our attention to conditions we had not
anticipated and indeed had not incorporated in our conceptual model of sexual
health concerns. The dominance of involuntary semen loss problems was so striking
that it needed further exploration in the survey data.
Relative importance of sexual health concerns
The informants contacted in the qualitative phase of the research are a convenience
sample’ of men who are probably more knowledgeable and approachable concern­
ing sexual matters. While we believe the main culturally constructed ideas to be
broadly applicable to the population, we also collected free lists from the respon­
dents in the quantitative survey. The results of this large-scale collection of free lists
was expected to be somewhat different from the qualitative phase, mainly because
less time was given to building of rapport and probing for further responses. Thus,
we anticipated shorter lists, but we still expected that the same general picture
would emerge concerning types of sexual health problems. Table 11.2 compares
the lists of items from the qualitative phase with results from four different groups
of respondents in the survey: single and married men in urban and rural areas.
The table shows the differences in the salience ranking (salience is the frequency of
mention of each item, weighted by the average rank order in which it appeared in
peoples’ lists).5

Sholk-chll.qxd 31/10/03 12:55 Page 210

Martine Collumbien et al.

210

Table 11.2. Rank Order of Salience of Sexual Health Concerns Among Informants
and Survey Respondents

Informants

Survey respondents

Rural

Urban

Dhatu Padiba
AIDS
Gonoriha
Swapnadosh
Jadu
Handling
Linga-gha
Parishra-poda
Hernia
Kachu
Linga-ghimiri
Hydrocele
Syphilis
Bata
Fileria
Malakantaka
Machala
Arsa

(n=35)

Single
(n= 196)

Married
(n = 203)

Single
(n= 189)

Married
(n=202)

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
>15
>15
>15

5
3
4
11
2
>15
12
>15
9
6
>15
1
14
>15
>15
7
8
10

8
5
3
15
2
>15
12
>15
9
10
>15
1
11
>15
>15
4
7
6

6
7
11
12
1
>15
10
14
8
5
>15
2
>15
>15
>15
4
3
9

6

11
8
>15
1
>15
10
13
7
5
>15
2
>15
12
>15
3
4
9

We note immediately that dhatu padiba, which was the most salient item in the
qualitative sample, falls to lower salience (from fifth to eighth on the lists of survey
respondents). AIDS, which was second in salience in the qualitative sample drops
slightly for urban single men, but drops more sharply among rural men (seventh and
eleventh). Thus, the survey results give a clear indication of the greater impact of
AIDS information programmes in the urban sector. In a similar way, gonoriha
salience drops quite sharply in the rural population. Overall, the free lists in the sur­
vey give more prominence to several types of itching (jadu, machala, and kacchu).
Jadu (itching) turns out to be the most salient item in rural populations, and ranks
second in the urban population. In addition the salience of hydrocele in both urban
and rural populations is notable. Since they affect the genital area, the inclusion of
hydrocele, hernia, and filaria among men’s reported sexual health problems is wide­
spread in India. Perhaps the most striking result of the triangulation is in the salience
of swapnadosh and handling (masturbation). These are somewhat more sensitive
or embarrassing, we believe, and most of the respondents in the surveys did not
mention them in their lists. In this part of the study we feel that the qualitative data
give a more realistic measure of salience of those items. The concerns and anxieties
about nocturnal emissions confirm the preoccupation about sex among youths and

Sholk-chll.qxd 31/10/03 12:55 Page 211

211

Male Sexual Health Categories in Orissa

unmarried men. This is understandable given the patterns of late marriage and the
fact that more than 70 per cent of men in the general population sample experience
their first intercourse at marriage. It was estimated from the data on age at first
intercourse that 80 per cent of men are still virgin at age twenty, and 44 per cent at
age twenty-five (Collumbien et al. 2000).

Reported experience of semen-related concerns
Some further questions on sexual health concern in the questionnaire included the
personal experience of dhatu padiba, swapnadosh, and jadu. These questions were
added because of the anxiety expressed among the qualitative informants about semen
loss and to test whether survey respondents would report on their own experience of
it. Jadu was added since it is a prevalent rash, which is ‘innocent’ and considered not
severe. When asking about the nocturnal emissions, the question referred to ‘excessive’
swapnadosh, that is, more than two to three night emission a month. Table 11.3
presents the differentials in reported life time experience of these three conditions.
The first observation is the relative lack of differential observed. On average
27.4 per cent of men reported personal experience of dhatu padiba. Though it was

Table 11.3. Differentials in Reported Personal Experience o/Dhatu Padiba,
Swapnadosh, and Jadu

Total
Marital status
Single
Married
Residence
Urban
Rural
District
Puri
Ganjam
Balasore
Cuttack
Education
No/primary education
Secondary education
Higher education
Household income
Low
Medium
Higher

Na

Dhatu padiba

Swapnadosh

Jadu

2087

27.4

52.3

40.1

1054
1033

26.7
28.2

55.6
49.0

39.6
40.5

296
1791

25.4
27.8

51.7
52.4

33.2
41.2

513
386
400
788

34.2
34.8
25.7
20.3

69.6
52.7
46.1
44.0

36.6
34.2
44.7
42.8

679
723
685

29.7
29.3
23.2

50.8
51.7
54.5

41.7
40.8
37.7

1110
658
319

29.0
26.6
23.6

52.2
52.1
53.3

40.9
39.2
39.0

a Weighted number in each category.

|Sholk-chll.qxd 31/10/03 12:55 Page 212

212

Martine Collumbien et al.

experienced less among the more educated, richer and urban, the differentials were
not stark. The district-wise variations were the largest: nearly 35 per cent of men in
Puri and Ganjam reported they had ever had dhatu padiba, compared to 20 per cent
of men in Cuttack. Since men in Cuttack have a better education and are more likely
to live in urban areas, multivariate analysis was done to look at the independent effect
of these factors (not presented). For dhatu padiba, district was the only factor which
had an independent effect. For swapnadosh, both marital status and district had an
independent effect, with single men reporting more problems of excessive nocturnal
emissions than married men. In Puri nearly 70 per cent of men had a lifetime expe­
rience of night emissions compared to an overall average of about 52 per cent.
Drawing on the existing psychiatric and anthropological literature, we will now
discuss semen anxiety in the context of the ethnophysiology of sex as understood in
South Asia, looking at Ayurvedic and folk explanatory models.

Ethnophysiology as explanatory model for semen loss
The ethnographic and psychiatric literature documents other studies which relate
the importance and associated anxieties of involuntary semen loss in South Asia
(Bottero 1991; Caplan 1987; Dewaraja and Sasaki 1991; Edwards 1983; Kakar 1996;
Malhotra and Wig 1975; Nichter 1981; Weiss et al. 1986). Semen leakage is often
referred to as ‘Dhat Syndrome’ and invariably associated with fears over weakness.
Virya, the Hindi word for semen, also means ‘vigour’. Among slum dwellers in
Mumbai, virya was also equated to money, and dhat girna (the local term for semen
loss) was referred to as ‘loss of money’ drawing a parallel between a poverty stricken
man without money and a sexually weak person without semen (Verma et al. 1998).
The sexual anxieties about loss of strength are encoded in the ethnophysiology of sex
in South Asia, which bears similarity to the Chinese concepts of health and sex
(Kleinman 1980; Edwards 1983). Semen is considered to be the ultimate vital force,
the source of physical as well as spiritual strength. Mahatma Gandhi, India’s great
political leader and social reformer, was preoccupied with the transformation of
sexual potency into psychic and spiritual power, a core issue of Hindu metaphysics.
For this reason Gandhi took the vow to observe celibacy at the age of thirty (Caplan
1987). The loss of virya through sexual acts or imagery is thought to be harmful both
physically and spiritually (Nag 1996). Contemporary beliefs about the power associ­
ated with seminal fluid are based on Vedic scriptures. They all link food to health and
sex through transformational processes. It is believed that the consumption of 60 lb
of food is needed to replace the loss of semen in each ejaculation. Another variant of
this age-old belief is that each coitus is equivalent to an energy expenditure of 24 h of
concentrated mental activity or 72 h of hard physical labour (Kakar 1996).
In Ayurvedic medicine, health depends on the balance of three body humours, and
sexual health may be seen to depend both on a proper diet and an appropriate use of
semen (Edwards 1983). Bottero, who studied Ayurvedic doctors in an Oriya town,
reports that the main cause of dhatu padiba is overheating due to an unbalanced diet
of too many heating food, such as meats, fish, garlic, pepper, and eggs. These foods

Sholk-chll.qxd 31/10/03 12:55 Page 213

Male Sexual Health Categories in Orissa

AU: Pls.
chk. ellipses
at the end

213

increase the internal fire which burns the semen and melts it as ghee (clarified butter). As a
result of becoming liquid, the semen is discharged spontaneously, without the patient being
aware of it. And semen is all the more vulnerable to this overheating as it is not localized in the
testicles but stored throughout the body....Thus they believe that ‘sperm is in the body as
butter is in milk’, i.e., as a dissolved constituent which is expelled through a churning-like
process at the moment of the ejaculatory convulsion;... (Bottero 1991: 307)

The general cultural ideas of relationships between pervasive hot-cold qualities/
events and conceptualizing of semen as distributed widely in the body, give us a clearer
perspective on the indigenous models of semen loss. Pool (1987) has pointed to
the central importance of the hot-cold belief system in organizing understandings of
physiological processes, particularly those related to disease. While the folk medical
discourse is only partly related to the Ayurvedic theory of humors they are rooted in
the same Hindu cosmology (Lambert 1996). In our data from Orissa—as presented
above-multiple aetiologies were suggested for semen loss, with excessive heat, improper
diet and masturbation as frequent explanations.
Semen as a vital essence and refined form of life energy exists in both sexes and
indeed there are strong parallels between Dhat Syndrome in men and vaginal discharge
(leukorrhea) in women (Nichter 1981). The local terminolgy for several discharge
complaints in women included the term dhatu or semen (Patel and Oomman 1999).
Dhatu is an essential body humor associated with vitality and a source of positive
health. According to Nichter, dhatu has a role in the control of emotions
In the body, dhatu controls heat and thus all processes of transformation. In the mind, dhatu
facilitates buddhi (intellect) which controls and gives direction to manas (desire) which is
provoked by a quantum of heat. In order to be able to think clearly, focus one’s attention, or
have control over one’s emotions, enough dhatu must be present to counterbalance heating
influences’ (Nichter 1981: 390).

Complaints of excessive heat often communicate an unbalanced emotional state
and serve as a sign of general distress, as explained in the psychiatric literature.
‘Emotions are exacerbated by heating influences such as heating foods, alcohol, and
sexuality. Conversely suppressed emotional states are expressed somatically by refer­
ence to overheat’ (Nichter 1981: 390).
The association of semen loss with excessive heat indicates that the emic cate­
gory of white discharge may be more related to psychosocial problems rather than
infection. Indeed, white discharge in both men and women are now recognized as
idioms of distress and associated with feelings of weakness. Complications of sup­
posed consumption by semen loss are mental exhaustion with constant negative
thoughts (or depression) and hypochondria, mainly due to the extreme anxieties a
diagnosed patient suffers about his condition (Bottero 1991). The complaints of
weakness are the same as the diagnostic symptoms of anxiety and depression (Patel
and Oomman 1999). Somatization seems an important idiom through which
both men and women communicate distress, but an important distinction is that
vaginal discharge does not interfere with sexual capacity, whereas semen loss does
(Edwards 1983).

Sholk-chll.qxd 31/10/03

214

12:55 Page 214

Martine Collumbien et al.

For the lay person, feelings of weakness and distress are attributed to or explained
by the loss of vital fluid. In a comparative study among psychiatric patients in
Sri Lanka and Japan, Dewaraja and Sasaki (1991) show how attitudes and cultural
beliefs become incorporated into patients’ explanations of their subjective feelings of
distress and anxiety, both to the therapist and to themselves. In Sri Lanka, psycho­
logical problems were self-attributed to semen loss, whereas in Japan stress and over­
work prevailed as explanations. Similarly, women presenting at psychiatric clinics in
Bangalore, frequently attribute their physical problems to the passing of vaginal
fluids leading to depletion of energy (Chaturvedi et al. 1993).
Although our data found little evidence of social differential in the ever experience
of semen loss (Table 11.3), Edwards (1983) reported that lower socio-economic
groups were more likely to consider semen loss as harmful to health. He notes that
recommended foods for maintaining semen and health are the more expensive ones,
while the prohibited ones are the staple foods of the lower castes (Edwards 1983: 61).
Malhotra and Wig (1975) describe semen loss as a specifically Indian culture­
bound syndrome, while Edwards (1983) sees it as a South Asian syndrome affecting
Muslims and Buddhists as well as Hindus across the subcontinent. Bottero (1991)
points to a much wider distribution of semen anxiety beyond South Asia, at least in
earlier history. He goes back to Hippocrates on ‘consumption of the back’, but the
most striking parallel he quotes from the work of Tissot in 1760. This French physi­
cian became famous for his work on diseases caused by masturbation. He describes
something very similar to dhatu padiba:

loss of‘a very liquid seminal liqueur’ during urination, defecation an nocturnal emissions, mas­
turbation and also through spontaneous discharges, which constitute ‘gonorrhoea simplex’,
‘a flow of semen without erection’, described in ‘true gonorrhoea’ as opposed to ‘bastard or
catarrhal gonorrhoea’ (which corresponds to our modern blennorragia, a purulent urethritis).
(Tissot in Bottero 1991: 312)

Edwards notes that sexual anxieties, regarding masturbation and semen loss are
still present in contemporary Western societies, but that ‘revisions in the “scientific”
medical interpretation of sexual physiology have reduced and altered its manifesta­
tion and severity’ (Edwards 1983: 60). One could argue that if AIDS had been around
in the eighteenth century with the corresponding interest for STD control, Tissot
might have made the point we want to make about syndromic management.

ETIC AND EMIC CATEGORIES: CONFOUNDING
INFECTION WITH DISTRESS?

So what are the implications of this emic category of semen anxiety for syndromic
management of STIs? Men in Orissa did not associate dhatu padiba with sexual
transmission, but another concern sometimes mentioned in conjunction with dhatu
padiba was parishra-poda denoting a burning sensation during urination. Both
conditions are believed to be caused by peta garam (heat in the stomach) as a result

Sholk-chll.qxd 31/10/03 12:55 Page 215

Male Sexual Health Categories in Orissa

215

of excessive heat. One respondent states ‘Parisra poda occurs due to peta garam.
Prolonged parisra poda leads to dhatu padiba’. As discussed above the cultural
hot/cold belief system underlie the physiology of leaking semen. Of course a white
discharge together with painful urination is the syndromic diagnosis for gonorrhoea
and chlamydia, raising the potential for confounding semen loss with pus discharge.
It is important to consider how dhatu padiba gets diagnosed and whether men
suffering from it will consult allopathic providers. As Bottero (1991) points out dhatu
padiba is diagnosed either directly or indirectly. In the first case the patient reports a
white discharge from his penis while urinating or defecating, or the discovery of
a stain on his clothes. In the other case the doctor diagnosed it on the basis of a set
of complaints about weakness, persistent fatigue, and skinny appearance, a combi­
nation of mental and physical weakening. Semen loss is thus implied during indirect
diagnosis. The indirect diagnosis of semen loss through complaints about weakness
was far more common than the direct diagnosis of reported white discharge (Bottero
1991). This leaves us pondering whether the 27 per cent of men in our survey
who reported personal experience of dhatu padiba actually observed a discharge or
attributed an episode of weakness and fatigue to semen loss.
When asked about self-reported symptoms of STIs in the survey, men who reported
painful urination together with discharge, were probed to distinguish between semen
discharge, dhatu padiba and pus discharge, pujo padiba. In total 132 men reported ever
experience of pain during urination (parishra poda) concurrent with discharge:
110 specified dhatu padiba with painful urination, 15 pujo padiba, and 7 reported both
pujo and dhatu. So when probed more on the nature of discharge, 87.7 per cent
reported semen discharge and only 16.7 per cent specified the discharge as pus. This
raises the question of whether there is a huge over-reporting of penile discharge, and
also the extent to which men confuse pus and semen in the discharge. Men with semen
loss were more likely to report having sought medical treatment (30 per cent versus
26 per cent for pus discharge). Among those going to a provider, men complaining
about semen loss were as likely as those with a pus discharge to have consulted an
allopathic provider (68 per cent: doctor or medicine store), while 8 per cent went to see
folk healers and 24 per cent Ayurvedic doctors. Modern Ayurveda is less involved in the
treatment of mental disorders than other healing traditions (Weiss et al. 1988) and
despite popular ethnomedical understanding based on humoral aetiologies, Western
biomedicine was often shown to be the first preference in the pluralistic system in rural
Rajasthan (Lambert 1996). Thus when clinical services for men become more widely
available in the public sector, men will present with these psychosexual concerns.
As we have seen before, more often than not semen loss is implied in the diagno­
sis of dhatu padiba. In the assumption of a move towards adopting syndromic
management in primary healthcare settings in India, the implications could be
twofold. There is the potential for both under- and overtreatment. Reported or
implied discharge could be treated as gonorrhoea and chlamydia, leading to
overtreatment. Good training of health workers with insistence on substantiating
evidence of discharge, that is, observed during consultation through milking of the
urethra (Mayaud et al. 1998) may overcome this. Another suggestion could be to

Sholk-chll.qxd 31/10/03 12:55 Page 216

216

Martine Collumbien et al.

include a ‘risk assessment step’ based on sexual behavioural risk factors similar to
those proposed by WHO to reduce the overtreatment for vaginal discharge (Mayaud
et al. 1998) though this has not proven successful among women in South Asia
(Hawkes et al. 1999). Men may also confuse discharge due to infections with semen
loss. If those men with white discharge diagnose themselves as having dhatu padiba
and they do not seek appropriate care, this would result in undertreatment. With
the benefit of the research experience on self-reported symptoms of gynaecological
morbidity over the past decade,6 further research on male sexual morbidities needs
to consider both pathological and somatic causes.

CONCLUSIONS

This study on male sexual health in Orissa started off with a clear assignment to learn
more about the local perceptions of sexually transmitted infections in order to
inform condom promotion and sexual health services. We discovered that the focus
of Oriya men was clearly more towards non-infectious conditions, that seem to
reflect psychosexual concerns and psychosocial distress. The salience of involuntary
semen loss in the minds of Oriya men is unmistakenable in these data. The frequency
and promptness with which dhatu padiba emerged in the process of free association
when men were asked about sexual health concerns is reinforced by the survey
results. More than a quarter of men in a representative sample reported personal
experience of the condition.
Although we started from the categorical frameworks of biomedicine and public
health interventions in HIV/STI control, the deceivingly simple structured qualit­
ative methods adopted in this study have proven invaluable in getting at the emic
categories of sexual health. The free listing and pilesorting exercises were very useful
for beginning to understand the local explanatory models. We also relied heavily on
the existing ethnopsychiatric literature to interpret the importance of semen loss,
linking it to the ethnophysiology of sex and understanding it as a somatic idiom
of distress. That we did not know this literature before the study reflects the narrow
paradigms adopted in international sexual health research.
The biomedical reductionism employed in current sexual health interventions,
especially when focusing on syndromic management is inadequate to understand
and treat sexual health problems of Indian men. From a clinical STD and biomed­
ical perspective dhatu padiba may not be seen as a ‘real’ disorder since there is no
organic pathology. From the psychosomatic perspective in contemporary psychia­
try it is real, as indeed from the more holistic viewpoint of Ayurvedic and folk
perspective of humoral balances in health. Public health interventions aimed at
HIV/STD control will benefit from considering the Indian holistic view of sex,
health, and well-being. It goes beyond treatable symptoms and draws attention to
the root causes of ill-health.
We thus challenge the categorical paradigms of sexual health in current pro­
grammes and clinical services. What Patel and Oomman (1999) point out for women

Sholk-chll.qxd 31/10/03 12:55 Page 217

Male Sexual Health Categories in Orissa

217

‘health needs increasingly involve problems beyond reproduction, and it is our
contention that mental health already is and will continue to grow to become a core
health issue for women. The intersection of reproductive health and mental health
provides an avenue for exploring these issues..is equally be valid for men. Health
needs of men go beyond prevention and treatment of infections, and addressing
semen anxiety may provide the opportunity to discuss issues of sexuality and sexual
behaviour. Because of the inherent conceptual link of health with sexuality commun­
ication about sex and sexual behaviour can most appropriately be done through the
language of health (Lambert 2001).
More flexible models of sexual healthcare incorporating physical symptoms as
well as their psychosocial context are called for. We would urge that programmes
that pay adequate attention to those other aspects of men’s sexual health concerns
will have a stronger likelihood of success is dealing with new ideas of safer sex, use of
condoms and other messages relevant to the campaign against STIs and the spread
of HIV infection. Presentation of information about these ‘facts of sex life’ can be
effectively integrated with HIV/AIDS health promotion. From within the STD/HIV
paradigm we will be tempted to advocate communication strategies that would bring
men’s concepts of sexual health problems into a more realistic relationship with
biomedical facts about male physiological/sexual processes. The understanding of
normal psychosexual development may take some fears away about masturbation,
nocturnal emissions, and involuntary semen loss.
It is important not to dismiss the concerns about semen anxiety since mental health
forms an integral part of (sexual) health. We are at the initial stages of recognizing the
magnitude of mental ill-health, and starting to think about potential interventions
(Patel and Oomman 1999). Since common mental disorders in women have been
shown to be associated with poverty—not only in India (Patel et al. 1999) but also in
Britain (see the pioneering study by Brown and Harris 1978)—cultural expressions of
psychosocial distress and the implications of mental ill-health needs further exploration.

Notes
1. For the purpose of this chapter the terms STIs and STDs (sexually transmitted diseases) are
considered synonymously and used interchangeably.
2. The poverty line was defined as the per capita monthly expenditure of Rs 49 for rural areas
and Rs 57 for urban areas at 1973/74 prices. There are Rs 42 to 1 US $.
3. This refers to four of the thirteen un-divided districts. In 1992, the thirteen districts were
divided into thirty new districts.
4. This does mean that we under-sampled men living in larger households, and effectively underrepresented single men in the sample: they make up 56% of the de facto household sample,
whereas only 50% in the sample of individual respondents. However, since most results are
presented separately for single and married men, this does not greatly influence the results.
5. Not all freelists from the survey were' entered, limiting the analysis to 50 per category
(urban/rural and married/single) in each district. Freelist are considered stable from 30 to
50 respondents.

|Sholk-chll.qxd 31/10/03 12:55 Page 218

218

Martine Collumbien et al.

6. Following a pioneer study in rural Maharashtra showing the hidden burden of reproduc­
tive tract infections (RTIs), including STDs, in rural women (Bang et al. 1989), several
community-based studies of women’s reproductive ill-health showed high level of self­
reported vaginal discharge, which were interpreted as an indication of RTIs. Levels of
gynaecological problems were believed to be underestimated by self-reported symptoms
due to the culture of silent suffering among Indian women (Koenig et al. 1998). However,
there is very poor agreement between reported morbidity and clinically diagnosed STI/RTIs
(Hawkes et al. 1999), and now part of the aetiology of ‘medically unexplained’ vaginal
discharge is suggested to be a somatic idiom for depression and psychosocial distress (Patel
and Oomman 1999; Trollope-Kumar 1999).

References
AIMS-Bhubaneswar (1996). ‘Barriers and opportunities to contraceptive social marketing’,
Unpublished report submitted to DFID Delhi.
Bang, R. A., Bang, A. T., Baitule, M., Choudhary, Y., Sarmukaddam, S., and Tale, O. (1989).
‘High prevalence of gynaecological diseases in rural Indian women’, Lancet, 1: 85-7.
Bottero, A. (1991). ‘Consumption by semen loss in India and elsewhere’, Culture Medicine and

Psychiatry, 15: 303-20.
Brown, G. W. and Harris, T. (1978). Social Origins ofDepression: A Study ofPsychiatric Disorder
in Women. London: Tavistock.
Caplan, P. (1987). ‘Celibacy as a solution? Mahatma Gandhi and brahmacharya’, in P. Caplan
(ed.), The Cultural Construction of Sexuality. London and New York: Routledge, pp. 271-95.
Chaturvedi, S. K., Chandra, P. S., Issac, M. K., and Sudarshan, C. Y. (1993). ‘Somatization misat­
tributed to non-pathological vaginal discharge’, Journal of Psychosomatic Research, 37: 575-9.
Cleland, J. and Lush, L. (1998). ‘Prevention of unwanted pregnancies and of sexually trans­
mitted infection: the linkages’, Paper presented to the International Conference on
Reproductive Health, March 1998, Mumbai.
Collumbien, M. and Hawkes, S. (2000). ‘Missing men’s message: does the reproductive health
approach respond to men’s sexual health needs?’, Culture Health and Sexuality, 2: 135—50.
___ Das, B., Bohidar, N., and Pelto, P. (2000). ‘Male sexual behaviour in Orissa’, Workshop on
Reproductive Health in India: New Evidence and Issues.
Datt, G. (1998). Poverty in India and Indian States: An Update. Washington: International Food

Policy Research Institute.
Dewaraja, R. and Sasaki, Y (1991). ‘Semen-loss syndrome: a comparison between Sri Lanka
and Japan’, American Journal of Psychotherapy, 45: 14-20.
Edwards, J. W. (1983). ‘Semen anxiety in South Asian cultures: cultural and transcultural
significance’, Medical Anthropology, 7: 51-67.
Grosskurth, H. (1995). ‘Impact of improved treatment of sexually transmitted diseases on HIV
infection in rural Tanzania: randomized control trial’, Lancet 346: 530-6.
Hawkes, S. (1998). ‘Why include men? Establishing sexual health clinics for men in rural
Bangladesh’, Health Policy and Planning, 13: 121-30.
Hawkes, S., Morison, L, Foster, S., Gausia, K., Chakraborty, J., and Peeling, R. W (1999).
‘Reproductive-tract infections in women in low-income, low-prevalence situations: assess­
ments of syndromic management in Matlab, Bangladesh’, Lancet, 354: 1776-81.
Hook, E. and Handsfield, H. (1999). ‘Gonococcal infections in the adult’, in K. K. Holmes,
P. E Sparling, P. Mardh, S. M. Lemon, W. E. Stamm, P. Piot, and J. N. Wasserheit, (eds.),
Sexually Transmitted Diseases. New York: McGraw-Hill Publications, pp. 451-66.

Sholk-chll.qxd 31/10/03 12:55 Page 219

Male Sexual Health Categories in Orissa

219

Kakar, S. (1996). The Indian Psyche. New Delhi: Viking Penguin.
Kleinman, A. (1980). Patients and Healers in the Context of Culture. An Exploration of the
Borderland Between Anthropology, Medicine and Psychiatry. Berkeley: University of California
Press.
Koenig, M., Jejeebhoy, S., Singh, S., and Sridhar, S. (1998). ‘Investigating women’s gynaecolog­
ical morbidity in India: not just another KAP survey’, Reproductive Health Matters, 6: 84-97.
Laga, M., Nzila, N., and Goeman, J. (1991). ‘The interrelationship of sexually transmitted
diseases and HIV infection: implications for the control of both epidemics in Africa’, AIDS,
5(Suppl. 1): S55-S63.
Lambert, H. (1996). ‘Popular therapeutics and medical preferences in rural north India’,
Lancet, 348: 1706-9.
------ (1998). ‘Methods and meanings in anthropological, epidemiological and clinical encoun­
ters: the case of sexually transmitted disease and human immunodeficiency virus control and
prevention in India [see comments]’. Tropical Medicine & International Health, 3: 1002-10.
------ (2001). ‘Not talking about sex in India: indirection and the communication of bodily
intention’, in J. Hendry and C. W. Watson (eds.), An Anthropology of Indirect
Communication. London: Routledge, pp. 51-67.
Lush, L., Cleland, J., Walt, G., and Mayhew, S. (1999). ‘Integrating reproductive health: myth
and ideology’, Bulletin of the World Health Organization, 77: 771-7.
Malhotra, H. K. and Wig, N. N. (1975). ‘Dhat syndrome: a culture-bound sex neurosis of the
orient’, Archives of Sexual Behaviour, 4: 519-28.
Mayaud, P., Hawkes, S., and Mabey, D. (1998). ‘Advances in control of sexually transmitted
diseases in developing countries’, Lancet, 351 (Suppl. 3): 29-32.
Mertens, T. E., Smith, G. D., Kantharaj, K., Mugrditchian, D., and Radhakrishnan, K. M. (1998).
‘Observations of sexually transmitted disease consultations in India’, Public Health, 112:123-8.
Mundigo, A. I. (1995). Mens Roles, Sexuality, and Reproductive Health. Chicago: John D. and
Catherine T. MacArthur Foundation.
Nag, M. (1996). Sexual Behaviour and AIDS in India. New Delhi: Vikas Publishing House.
Nichter, M. (1981). ‘Idioms of distress: alternatives in the expression of psychosocial distress:
a case study from South India’, Culture, Medicine and Psychiatry, 5: 5-24.
O’Reilly, K. R., Dehne, K. L., and Snow, R. (1999). ‘Should management of sexually transmitted
infections be integrated into family planning services: evidence and challenges’, Reproductive
Health Matters, 7: 49-59.
Patel, V. and Oomman, N. (1999). ‘Mental health matters too: gynaecological symptoms and
depression in South Asia’, Reproductive Health Matters, 7: 30-8.
___ Araya, R., de Lima, M., Ludermir, A., and Todd, C. (1999). ‘Women, poverty and common
mental disorders in four restructuring societies’, Social Science and Medicine, 49: 1461-71.
Pelto, P. J. (1994). ‘Focused ethnographic studies on sexual behaviour and AIDS/STDs’, The
Indian Journal of Social Work, 55: 589-601.
------ and Pelto, G. H. (1997). ‘Studying knowledge, culture and behaviour in applied medical
anthropology’, Medical Anthropology Quarterly, 11: 147-63.
Pelto, R, Lakshmi, R., Lakhani, A., Sharma, S., Das, B., and Das, R. (1998). Where do men go
for sexual contacts? a social mapping approach, Unpublished.
PRC Bhubaneswar and UPS (1993). National family Health Survey (MCH and Family
Planning) Orissa 1993. Bombay: Population Research Center, Bhubaneswar and
International Institute for Population Sciences.
Ramasubban, R. (1999). ‘HIV/AIDS in India: gulf between rhetoric and reality’, in S. Pachauri
(ed.), Implementing a Reproductive Health Agenda in India: The Beginning. New Delhi:
Population Council, pp. 357-86.

Sholk-chll.qxd 31/10/03 12:55 Page 220

220

Martine Collumbien et al.

Stamm, W. E. (1999). ‘Chlamydia trachomatis infections in the adult’, in K. K. Holmes,
P. E Sparling, P. Mardh, S. M. Lemon, W. E. Stamm, P. Piot, and J. N. Wasserheit (eds.),
Sexually Transmitted Diseases. New York: McGraw-Hill Publications, pp. 407-22.
Trollope-Kumar, K. (1999). ‘Symptoms of reproductive-tract infection—not all that they seem
to be’, Lancet, 354: 1745-6.
Verma, R. K., Rangaiyan, G., Narkhede, S., Aggarwal, M., Singh, R., and Pelto, P. (1998).
‘Cultural perceptions and categorization of male sexual health problems by practitioners
and men in a Mumbai slum population’, Paper presented to the International Conference
on male as supportive partners, June 1998, Kathmandu.
Weiss, M. G., Sharma, S. D., Gaur, R. K., Sharma, J. S., Desai, A., and Doongaji, D. R. (1986).
‘Traditional concepts of mental disorder among Indian psychiatric patients: preliminary
report of work in progress’, Social Science and Medicine, 23: 379-86.
------ Desai, A., Jadhav, S., Gupta, L, Channabasavanna, S. M., Doongaji, D. R., and Behere, P. B.
(1988). ‘Humoral concepts of mental illness in India’, Social Science and Medicine, 27: 471-7.
Weller, S. C. and Romney, A. K. (1988). Systematic Data Collection. Qualitative Research
Methods. London: Sage Publications.
World Bank (1993). World Development Report: Investing in Health. Oxford: Oxford University
Press.
World Health Organization (WHO) (1975). Education and Treatment in Human Sexuality:
The Training of Health Professionals. Geneva: WHO.
------ (1997). Management of Patients with Sexually Transmitted Diseases. Global Programme
on AIDS Geneva: WHO.

Position: 1797 (2 views)