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

XIV International AIDS Conference

Abstracts ThOrD1493-ThPpD2145 493

The ten-year follow-up showed first a sharp increase in condom use suggesting
that sexual health promotion messages may have had some impact on reported
risk behavior. This recent decrease in condom use may be explained, among
other hypotheses, by difficulties in maintaining both strict individual long term prevention behaviors and a high level of awareness in the population. This may be
the case especially among the young people who began their sexuality after new
treatments have arrived and drastically changed HIV social representation. Finally the decrease in large HIV prevention campaigns in the general population
may also be responsible for the decrease in condom use.
Presenting author: Nathalie Beltzer, ORS Ile-de-France, 21-23, rue Miollis,
75732 Paris cedex 15, France, Tel.: [phone], Fax: [phone], E-mail:
[email protected]
ThOrD1493 Social capital as a predictor of AIDS cases, STD
rates and adolescent sexual risk behavior
prevalence: A state-level analysis, U.S.A., 1999
D.R. Holtgrave, R.A. Crosby, G.M. Wingood, R.J. DiClemente, J.A. Gayle. Emory
University Center for AIDS Research, Emory University School of Public Health,
1518 Clifton Road NE, Room 540, Atlanta GA 30322, United States
Background: Social capital includes trust, reciprocity and cooperation among
community members working together to achieve common goals; it has been
correlated with public health variables including violence and mortality. However,
the relationship between social capital and infectious diseases has received little
attention. Here, we examine the relationship between social capital (as well as
poverty and income inequality) and AIDS case rates, STD rates and adolescents'
HIV-related sexual risk behavior prevalence at the state level (USA, 1999).
Methods: Bivariate and multivariate linear regression techniques were employed
in this state-level, secondary data analysis. Predictor variables were Putnam's
social capital index (www.bowlingalone.com), poverty and income inequality. Disease outcomes were rates of AIDS cases, gonorrhea, syphilis, and chlamydia
assessed by CDC's surveillance system. Behavioral outcomes (as assessed by
CDC's Youth Risk Behavior Survey) included 7 aspects of adolescents' sexual risk
behavior (e.g., early sexual debut, and multiple partners) analyzed separately for
males and females (for a total of 14 behavioral measures).
Results: At a significance level of p <0.01 (one-tailed test), social capital was a
significant bivariate predictor of every disease outcome (explaining between 25%
and 45% of the variance in each disease), and 10 of 14 behavioral variables (explaining between 19% and 52% of the variance in these 10 measures). Poverty
was a significant predictor of one disease outcome and two behavioral variables.
Income inequality was a significant predictor of two disease outcomes and one
behavioral variable. Overall, social capital was a stronger predictor variable than
the other two. Conclusions: Social capital is a key societal- level construct for
predicting sexually transmitted diseases and risk behaviors, and should be considered in the development of structural HIV and STD prevention interventions.
Presenting author: David Holtgrave, Emory University School of Public Health,
1518 Clifton Road NE, Room 540, Atlanta GA 30322, United States, Tel.: +1 404
727 5401, Fax: [phone], E-mail: [email protected]
ThOrD1494 Community Influences on Indiviudal level HIV
serostatus in Rakai District, Uganda
P. Patil, D. Bishai, R. Rakai Project Group. Johns Hopkins University, 116 West
University Parkway, Baltimore, MD 21210, United States
Background: The current HIV/AIDS research and prevention paradigm in dominated by the study of individual risk factors. Recent literature suggests the importance of re-focusing on contextual and social influences on individual level HIV
risk behaviors and status. This study examined the influences of community level
measures (social and economic) on individual level HIV status.
Methods: We used data from 56 communities with a sample of 15, 247 individuals gathered between 1994-1998 in the Rakai District of Uganda. Using both
community and individual level measures, we applied logistic regression to examine the effects of community means of AIDS knowledge and condom use as well
as community electrification, roads, bars, and discos on HIV serostatus.
Results: A chi-square test of joint significance revealed that community level social and economic variables significantly influenced individual HIV serostatus,
even when adjusting for individual level demographics and HIV risk behaviors.
Community means of AIDS knowledge and condom use were more strongly
linked to HIV serostatus than economic development indicators. Simulations suggest that interventions to increase community mean condom use would have a
larger effect on an individual's risk for HIV that changes in economic development
of a community.
Conclusions: Our study indicates that examining community contextual effects
on HIV risk behaviors and HIV status is a useful approach to HIV prevention research. And, in using community level variables we are better able to understand
the salient components of risk-environments that put individuals at greater risk

of HIV acquisition. Statistical simulations were run that demonstrate how policy
changes may influence HIV risk behaviors, and HIV status.
Presenting author: Priya Patil, 116 West University Parkway, Baltimore, MD
21210, United States, Tel.: [phone], E-mail: [email protected]

ThOrD1 495 Alcohol intoxication and HIV high-risk behaviors
among injection drug users (IDUs) in Puerto Rico
T.D. Matos, R.R. Robles, C.A. Marrero, J.C. Reyes, H.M. Colon. Universidad
Central delCaribe, Puerto Rico, Puerto Rico
Background: The association between alcohol use and HIV risk behaviors remains unclear. This study examined if alcohol use is associated with both injection
and sex risk behaviors among IDUs.
Methods: A total of 555 IDUs not currently in a treatment program were recruited
between 1998-2000 in a suburban area in Puerto Rico as participants of a randomized prevention trial.
Results: Alcohol intoxication during the last 30 days was reported by 12.4% of
participants. Compared to IDUs who did not report alcohol intoxication, those
who did were more likely to report 4 or more injections per day (39.7% vs. 55.2%,
p=.034), needles sharing (12.3% vs. 26.1%, p=.005), pooling money to buy drugs
(67.5% vs. 83.8%, p=.007), had higher mean scores in the Addiction Severity Index (0.12 vs. 0.16, p=.008), were more likely to report casual sex partners (18.3%
vs. 42.9%, p=.001), paid sex partners (7.7% vs. 19.0%, p=.038), and unprotected
sex (71.7% vs. 88.1%, p=.024). After adjusting for demographic characteristics,
injection settings, frequency of injection, and pooling money to buy drugs, IDUs
who reported alcohol intoxication were more likely to share syringes (OR = 2.0;
95% CI: 1.0-4.1).
Conclusions: IDUs who also consume alcohol to intoxication are at greater risk
of exposure HIV than IDUs who did not. Preventive interventions need to address
alcohol consumption and it relation with HIV risk behaviors among IDUs.
Presenting author: Tomas Matos, P. O. Box 60327, Bayamon, Puerto Rico,
00960-6032, Puerto Rico, Tel.: [phone], Fax: [phone], E-mail:
[email protected]
ThOrD1496 Effect of knowledge of partner HIV status on
condom use among voluntary counselling and
testing clients in Uganda
E. Tumuhairwe1, F. Baryarama2, R. Bunnell2, J. Mubangizil, J. Kalule1,
J. Mermin2, C. Hitimana-Lukanikal, R. Ransom2. 1AIDS Information Centre,
CDC-Uganda, PO Box 49, Entebbe, Uganda; 2CDC-Uganda, Entebbe, Uganda
Background: The AIDS Information Centre (AIC), has been providing voluntary
counselling and testing (VCT) to both individuals and couples since 1990. 75%
of couples with at least one HIV-infected member are discordant. Although AIC
has been promoting VCT for couples, the majority (77%) of clients still present
as individuals. We assessed levels of knowledge of partner HIV status and their
association with condom use among first time individual VCT clients.
Methods: We analysed data from 18,795 first time individual VCT clients on
their sexual partners during six months prior to HIV testing including partner type
(spouse, steady, casual), condom use and knowledge of partner HIV status. Consistent condom use was defined as "always used condoms with the partner in the
last 6 months".
Results: Out of the 18,795 first time testers who tested as individuals, 11,403
(61%) reported having sex within the last 6 months. Of 12,034 partners reported,
60% were spouses, 23% were steady partners and 17% were casual partners.
Over 90% of clients had no knowledge of their partners' HIV status. Consistent
condom use was 15% but varied by partner type; 4% with spouses, 31% with
steady partners and 35% with casual partners. Only 12% of clients who knew
their spouses were HIV positive used condoms consistently compared with 3%
for those not knowing their partners' status.
Conclusions: Consistent condom use rates were low even when clients knew
that their partners were HIV-infected. Partner type appears more important than
partner HIV status in determining condom use. VCT counsellors must stress the
importance of couple testing, mutual disclosure of test results, and condom use.
Presenting author: Rebecca Bunnell, CDC-Uganda, PO Box 49, Entebbe,
Uganda, Tel.: [phone], Fax: [phone], E-mail: [email protected]
ThPpD2145 AZT,3TC, NVP and Caesarean section in Prevention
of MTCT in Chennai,lndia
S. Shyamprasad1, P. Madivanan1, M. Dhanaraj2, D. Geetha1, S. Solomon1,
S. Cu-Uvin3, K. Meyer3. 1 YRG Care, CSI Rainy Hospital 45, G.A.Road,
Chennai 600 021, India; 2CSI Rainy Multi Specialty hospital, Chenna, India;
3Brown University ProvidenceRI, United States
Background: A combined approach of antiretrovirals and elective caesarean
section has been shown to perinatal transmission rates to 2% or less. A pilot
study was undertaken in Chennai, India to assess the efficacy and safety of this
type of intervention on pregnancy outcome.
Methods: 30 women were enrolled in this study from 9/99 to 1/01 The average
age was 22.1 yrs and 70% were primigravidae.86.7% had acquired the infection
heterosexually, 10% by blood transfusion and 3.3% unknown. The average CD4

count was 778 cells/cmm at the time of enrolment - ranging 182 to 2071. AZT
300mg bd and 3TC 150mg bd was given from 34 weeks of gestation and patients scheduled for elective Caesarean section at 38wks gestation. One dose
of 200mg NVP was given 6 hours before surgery. Women were counseled not to
breast feed. The infants were given one dose of NVP and AZT / 3TC until RT PCR
results were known to be negative. CD4 was repeated at 6 weeks postpartum and
infant's HIV PCR done at least twice over 6 months.