Abstract Book Vol. 2 [International Conference on AIDS (14th: 2002: Barcelona, Spain)]

182 Abstracts WePeD6394-WePeD6397

XIV International AIDS Conference

Among them, 48.9% reported >1 six month periods with unprotected anal sex
with a partner of unknown or seropositive status (RAS). Logistic regression analyses were used to identify determinants of risky anal sex.
Results: Among regular bathhouse users, RAS was linked (univariate analyses)
with the recurrence between TO-T4 of greater preoccupation/dissatisfaction with
their emotional/sexual lives*; regular alcohol/drug use during sex**; associating
less risk with receptive/insertive anal sex*; seeing anal sex more as the ultimate
pleasure***, an act of communion* and a proof of virility/sexual maturity*; anticipating more sexual excitation from unprotected anal sex*; a more sensationseeking personality (SS)**and less perceived behavioural control (PBC) over condom use***(*p<0.01;**p<0.001;***p<0.0001). In the multivariate analysis, the
best predictors of RAS were SS (Odds ratio adjusted(OR):1.64; 95% Confidence
Interval (95%CI):1.04-2.60) and PBC (OR:0.009;95%CI:0.001-0.064). PBC can
signal a lack of personal/interpersonal skills in using condoms in various contexts.
Conclusion: Bathhouse norms of interaction/seduction (ex: non-verbal cues,
brief exchanges, substance use,...) and a focus on and anticipation of sexual sensation/satisfaction, combined with a reduced sense of risk towards anal
sex among regular bathhouse users may form the context in which they have
RAS. Should prevention focus on adapting to bathhouse norms and/or on personal/interpersonal skills building?
Presenting author: Jean Lawrence Roy, Hopital Saint-Luc du CHUM, 264, boul.
Rene-Levesque Est, suite 203, Montreal, Quebec H2X 1P1, Canada, Tel.: +1 514
890 8306, Fax: [phone], E-mail: [email protected]
WePeD6394 The socio-economic determinants of HIV
prevalence in India
M.P. Kumar. Kerala State AIDS Control Society, TC 19/2390, Poojappura,
Trivandrum, Kerala, Pin Code 695012, India
Background: The antenatal HIV prevalence in the various states of India varies
widely from 0% to 2.6%. A study using multivariate linear regression analysis was
done to explain this variation on the basis of a number of socio-economic and cultural parameters which might determine the sexual/ risk taking behaviour of the
local population.
Methodology: The mean antenatal HIV prevalence values obtained in three
rounds of sentinel surveillance (1998-99) conducted nationwide was used as the
dependent variable. A variety of aggregate socio economic variables of each of
the states such as literacy, life expectancy, gender inequality, female empowerment, migration status, urbanization, circumcision status, income, condom use
etc were chosen as the independent variables. Their values were obtained from
published governmental sources. Some states and union territories with low population and states with zero HIV prevalence were excluded. Finally, 20 states
which accounted for 84% of India's population were included in the regression
analysis.
Results: Five aggregate variables showed significant correlation with the HIV
prevalence in the states with the regression coefficients of three being significant
at the 0.01 level. Higher values of life expectancy at birth, literacy rate of men and
the mean age of effective marriage of women were shown to be protective while
higher State Domestic Product/capita and a higher fraction of literate women who
are aware of AIDS were shown to indicate higher HIV prevalence. The higher SDP
leading to higher HIV infection is probably related to more industrial development
leading to more migration and also more commercial sex.
Conclusion: This study suggests that socioeconomic variables may be able to
explain much of the variation seen in the HIV prevalence among the states in
India. Policies formulated to change critical socio-economic factors may offer additional means to control the HIV epidemic.
Presenting author: Prasanna Kumar, TC 19/2390, Poojappura, Trivandrum, Kerala, Pin Code 695012, India, Tel.: [phone], Fax: [phone], E-mail:
[email protected]
WePeD6395 Does personal knowledge of someone with AIDS
influence the use of condoms in casual sexual
encounters? Analysis of data from the 1998 South
Africa Demographic and Health Survey (SADHS)
C.S. Camlin, C.E. Chimbwete. Africa Centre for Health and Population Studies,
africa centre, p.o. box 198, mtubatuba, 3935, South Africa
Background: Prior research suggests that personally knowing a person diagnosed with HIV/AIDS or who died of AIDS (PWA) may positively influence condom use by countering denial and increasing perceived HIV risk. Rising AIDS
mortality in South Africa warrants re-examination of this hypothesis.
Methods: Data were analyzed for 8,112 sexually active women whose last sexual
encounter was with a casual partner, do not want children, and do not report that
their partner disapproves of condom use. Logistic regression analysis was performed to test the hypothesis that knowing a PWA positively influences condom

use at last sex, controlling for confounders including knowing that condoms prevent AIDS, rural/urban residency, province, education, age, ethnicity and marital
status. Odds ratios were determined for independent effects of each variable.
Results: Descriptive analyses showed that women who personally knew a PWA
were more likely to have used a condom at last sex with a casual partner (22%)
than their counter parts who did not know a PWA (15%), but multivariate analyses
showed no significant difference in condom use between those who knew and did

not know a PWA [OR=1.09; -.12-0.30 (95% CI)]. Women who knew that condoms
prevent HIV/AIDS had twice the odds of those who did not of having used condoms at last casual sex [OR=1.95; 0.34- 0.99]. Odds of condom use were seven
times higher for women with postgraduate education than for those who had not
completed primary education [OR=7.26; 1.32-2.65]; 1.5 times higher for women
ages 20-29 vs. 40-49 [OR=1.53, 0.14-0.71]; 1.6 times higher for urban vs. rural
[OR=1.56; -0.66,-0.23]; and 1.7 times higher for women with regular, non-marital
partners vs. married women [OR 1.7; 0.35-0.78].
Conclusions: Findings suggest that communities need not wait until AIDS mortality is high to initiate prevention campaigns. Messages that reinforce the effectiveness of condoms for HIV/AIDS prevention can influence positive behavior
change.
Presenting author: carol camlin, africa centre, p.o. box 198, mtubatuba, 3935,
South Africa, Tel.: [phone], Fax: [phone], E-mail: ccamlin
@mrc.ac.za
WePeD6396 Recent experience of sexual dysfunction is
associated with high-risk sexual behaviours - A
cross-sectional study of HIV-positive gay men
attending a London HIV clinic (SHARP - Sex,
Health and Antiretrovirals Project)
J. Imrie1, J.M. Stephenson1, M.D. Davis', G.J. Hart 2, O.R. Davidson3,
I.G. Williams'. 'Dept of STD, Royal Free & University College Medical School,
Dept of STD, Royal Free & University College Medical School, Mortimer Market
Centre, off Capper Street, London WCIE 6AU, United Kingdom; 2MRC Social &
Public Health Sciences Unit, University of Glasgow, Glasgow, United Kingdom;
3Mortimer Market Centre, Camden & Islington Community Health Services NHS
Trust, London, United Kingdom
Background: Behavioural surveillance consistently shows that HIV-positive men
are more likely to engage in unprotected anal intercourse than HIV-negative or
untested men, but how physical factors may contribute to these behaviours is
still poorly understood. We explored HIV-positive gay men's experience of sexual
dysfunction in relation to self-reported high-risk sexual behaviours.
Methods: Cross-sectional study (N=413) of gay men attending an HIV outpatient clinic. Self-reported sexual dysfunction symptoms (e.g. erectile dysfunction,
retarded ejaculation, loss of erection when applying a condom) and sexual risk
behaviour data (e.g. anal intercourse (AI) with new partners, inconsistent and
non-condom use) were collected using computer-assisted self-interview.
Results: Of all respondents, 93% were sexually active in the last year and 90%
reported sexual dysfunction symptoms lasting 2 weeks or more during the same
period. These included: loss of libido 73%; erectile dysfunction 54%; loss of erection while applying a condom 42%; and retarded ejaculation 36%. Of sexually
active men, 55% reported Al with at least one new partner in the last month and
24% reported inconsistent or no-condom used. Compared to men who consistently used condoms for Al with new partners, men reporting inconsistent or noncondom use experienced more sexual dysfunction symptoms, particularly erectile
dysfunction (p=0.07) and loss of erection while applying a condom (p=0.001).
Conclusion: In this sample of HIV-positive gay men, a very high proportion reported sexual dysfunction symptoms. Experience of sexual dysfunction was associated with higher sexual risk behaviours and may contribute to increased sexual risk behaviours. However, the contribution of antiretroviral therapy to the high
prevalence of sexual dysfunction symptoms in this population is still poorly understood.
Presenting author: John Imrie, Dept of STD, Royal Free & University College
Medical School, Mortimer Market Centre, off Capper Street, London WC1E 6AU,
United Kingdom, Tel.: [phone], Fax: [phone], E-mail: jimrie
@gum.ucl.ac.uk
WePeD6397I What explains the inconsistency between risky
behavior and risk perception? And, how we can
we use the answer to this question for better
policy design?
T.A. Falba, M. Jofre-Bonet. School of Public Health, Yale University, New Haven,
CT United States
Although the predictors of HIV risk behavior have been studied extensively and
policy has been designed accordingly, little is known about perceived HIV risk.
Given the potentially essential role of subjective beliefs on behavior decisions,
more needs to be learnt about the relationship between behavior and perceived
HIV risk and how social demographic variables might mediate this association. In
this paper, we explore what explains apparent inconsistencies between HIV risk
behavior and perceived HIV risk.
We select 2805 individuals whose behavior places them at relatively high risk of
contracting HIV from waves 1996, 1998 and 2000 of the Behavioral Risk Factor
Surveillance System. Selection is based on objective measures of risky behav
ior such as having multiple partners and no condom use during last intercourse.
From this high HIV-risk subpopulation, only 131 (or 5%) think they are at high risk
of contracting HIV; 399 (or 14%) think they are at medium risk; 1450 (or 52%)
believe they are at low risk; and 825 (or 29%) think they are at NO risk at all. We
use econometric analysis and descriptive statistics to study the determinants of
HIV risk perceptions. Preliminary results show that the likelihood of self-perceived
high risk for HIV is lower for females and for those with higher education. The like