University of California, San Francisco | Search UCSF | AIDS Research Institute

Research Findings > Scientific Presentations - 1998 International AIDS Conference

CAPS presentations at the 1998 International AIDS Conference - Geneva, Switzerland

Abstracts are listed in alphabetical order, by author.

Oral presentations

[133/33269] Voluntary HIV counselling and testing (VCT) reduces risk behavior in developing countries: Results from the voluntary counselling and testing study
Thomas Coates1, Gloria Sangiwa2, D. Balmer3, S. Gregorich1, C. Kamenga4, T. Coates1. 1USCF Center for AIDS Prevention Studies, 74 New Montgomery, San Francisco, CA; 4FHI/AID SCAP Arlington, VA, USA 2Muhimbili University College Dar Es Salaam, Tanzania 3Kenya Assn. of Professional Counselors, Nairobi, Kenya

Objectives: To determine the impact of VCT on sexual risk behavior in 3 developing countries. Methods: Multicenter randomized controlled trial in Nairobi, Kenya, Dar es Salaam, Tanzania, and Port of Spain, Trinidad. Participants who enrolled individually (N = 3120) were randomized to receive VCT (pre-and post-test client-centered counseling) or a standardized health information (HI) intervention (culturally appropriate AIDS risk reduction video in local language w/6 month wait for VCT). Standardized behavioral interviews were administered at baseline and 6-month follow-up. Intention-to-treat analyses were conducted. Results: Data are presented collapsed across sites. 81% of those enrolled were retained at the 6-month follow-up. 96% of those assigned to VCT were tested and 79.5% received their test results. The results for three self-reported sexual risk behaviors are summarized below; unprotected intercourse with: any primary partner (UI-PP), any non-primary partner (UI-NP), and any commercial (UI-CP) partner. Reports of each risk behavior decreased over time for both treatment groups, p's < .01. However, risk reduction with non-primary partners was significantly greater among those assigned to VCT compared to HI, p < .01. The VCT group also reported greater risk reduction with commercial partners, but the difference was not statistically significant, p = .09. Women reported higher levels of risk with primary partners and lower levels of risk with non-primary and commercial partners, p's < .01. Participants from the two African sites reported lower levels of risk with primary partners and higher levels of risk with non-primary and commercial partners, p's < .01.

+-------------------------------------------------------------------------+
| | | UI-PP | UI-NP | UI-CP |
+--------+----+-------------------+-------------------+-------------------+
| | |Baseline |6-Months |Baseline |6-Months |Baseline |6-Months |
+--------+----+---------+---------+---------+---------+---------+---------+
|Overall |VCT |39% |30% |27% |16% |10% |5% |
+--------+----+---------+---------+---------+---------+---------+---------+
| |HI |38% |28% |27% |23% |11% |8% |
+--------+----+---------+---------+---------+---------+---------+---------+
|Women |VCT |46% |35% |23% |13% |8% |3% |
+--------+----+---------+---------+---------+---------+---------+---------+
| |HI |43% |34% |24% |19% |9% |6% |
+--------+----+---------+---------+---------+---------+---------+---------+
|Men |VCT |32% |25% |31% |19% |12% |7% |
+--------+----+---------+---------+---------+---------+---------+---------+

Both VCT and HI reduced sexual risk behaviors among study participants. Compared to HI, VCT produced greater reductions in the prevalence of unprotected intercourse with non-primary partners and was marginally more effective in reducing the prevalence of unprotected intercourse with commercial sex partners. This demonstrates the efficacy of VCT as a intervention that can help reduce HIV transmission.

[593/24323] Confidentiality and couple HIV counselling encourage client disclosure of serostatus and risk behavior: Results from the Voluntary HIV counselling and testing study
Olga Grinstead1, A. Van Der Staten1, G. San Giwa2, M. Hogan2. 1Center for AIDS Prevention Studies - UCSF, 74 New Montgomery, San Francisco, CA 94105, United States 2Muhimbili University College, Dar Es Salaam, Tanzania

Objectives: To describe the experience of HIV counseling and testing in East Africa from the perspective of clients receiving services and from the perspective of counselors providing services. This report focuses on disclosure of serostatus and risk behavior. Methods: A multi-center randomized controlled study was conducted in Nairobi (Kenya), Dar es Salaam (Tanzania) and Port of Spain (Trinidad). Study participants were randomized to receive HIV C&T or a standardized health information intervention (with 6-month wait period for C&T). Focus groups and semi-structured interviews were conducted with counselors and counseling supervisors in Nairobi and Dar es Salaam. Participants who received C&T at baseline were also interviewed in Tanzania. Interviews included: problems and benefits of study participation, disclosure of test results, social support and how counseling reduces sexual risk behavior. Findings are based on 11 counselor interviews, 2 counselor focus groups and 39 client interviews. Results: 1) Disclosure of risk behavior to the counselor depended on the establishment of a trusting relationship. Confidentiality was central to establishing trust, distinguished counseling from other forms of communication and allowed clients to keep control over sensitive information. 2) Counselors saw client disclosure of serostatus to their sexual partners as central to risk reduction. All felt that couple counseling facilitated disclosure more than individual counseling. Counselors unanimously felt that couple counseling was more difficult, because of the ethical issues raised when one partner refuses to disclose serostatus. 3) Clients' fear of and experiences of stigma impaired disclosure of serostatus to family/social network. Clients and counselors perceived that confidentiality facilitated trust, trust facilitated disclosure of serostatus and disclosure facilitated risk reduction. Couple counseling was favored over individual counseling because it was believed to facilitate disclosure and to reduce disruption of relationships. Client concerns about stigma were common. Findings underline the importance of training counselors to protect client confidentiality. Findings also support the practice of HIV testing and counseling with couples.

[592/24321] Counselling strategies: they work! Results from the voluntary HIV counselling and testing (VCT) study
Olga Grinstead1, Don Balmer2, E. Van Praag3, O. Grinstead1, S. Gregorich1, G. Sangiwa4, C. Furlonge5. 1UCSF Center for AIDS Prevention Studies, 74 New Montgomery, San Francisco, CA 94105, United States 2Kenya Assn. Of Professional Counselors, Nairobi, Kenya 3UNAIDS/World Health Organization, Geneva, Switzerland 4Muhimbili University College, Dar Es Salaam, Tanzania 5Voluntary Counseling & Testing Centre, Curepe, Trinidad WI

Objectives: To describe the counseling process used in the VCT study and to report on utlization of counseling services in the study. Methods: The multicentre study was conducted in a low income area in Nairobi and at the teaching hospital in centre Dar Es Salaam. Participants were randomized to receive VCT or health information and return in 6 months. A client-centered counseling approach was used which emphasised a confidential relationship in which clients assessed their behaviour, formulated a risk reduction plan, addressed worries and emotions and learned to cope with test results. Counsellers (nurses and health workers) were retrained during 5 days and were the ones to break the news of the test result. Clients were encouraged to return for more counselling sessions. 716 clients in Tanzania and 761 in Kenya received VCT at baseline. Results: 95% of those assigned to VCT were actually tested; 75% (Tanzania) and 85% (Kenya) of those tested returned for their test results. In Tanzania, those who enrolled as part of a couple (16%) and those who were HIV sero-positive at baseline (21%) were less likely to return for their test results (OR = 0.64 & OR = 0.60, respectively); none of the explanatory variables (i.e., age, gender, couple/individual, serostatus) reached significance in Kenya. The numbers of post-test sessions among those returning for test results ranged from 1 to 8, median 1. Using more post-test counseling sessions was predicted by being seropositive (Tanzania OR = 11.11: Kenya OR = 8.54) and by being older (Tanzania, OR = 1.04 per year of age). Group and individual supervision for counselors was appreciated, reduced stress and enabled handling emotional situations. Participants accepted counseling and testing-most randomized to VCT were tested, and most returned for their test results. When allowed unlimited access to counseling, most clients chose to attend only two sessions. In Tanzania, seropositives were less likely to return for test results, but when they did return used more counseling sessions which strengthened coping and allowed for an early entry into care and support. Findings support the feasibility of client-centered VCT in developing countries.

[35/42201] Placebo control trials of short-course antiRetroviral regimens to reduce mother-to-child HIV transmission are essential to establish standard of care in Africa
JamesG. Kahn1, E. Marcelle2. 1Inst. for Health for Stud. UCSF Box 0936 San Francisco, CA; 2Inst. for Health Policy Studies USCF, San Francisco, CA, USA

Objectives: Trials in Africa of feasible short-course antiRetroviral (ARV) drugs to prevent mother-to-child HIV transmission have been criticized for using placebo-control designs instead of equivalence comparisons to AZT/ACTG 076. We compared the value of these 2 designs for determining local standard of care. Design/Methods: We review how statistical, design, and implementation issues affect trial validity, generalizability, and rapidity in African settings. We estimate potential HIV infections prevented with faster trials. Results: Validity: An equivalence design requires a comparison with known efficacy. However, the efficacy of 076 cannot be confidently extrapolated to Africa due to differences in risk for HIV transmission that might affect efficacy, including immune status, HIV disease state, and breastfeeding practices. Natural history studies cannot serve as historical untreated controls due to the wide variability of observed transmission rates (25%-48%). If 076 were found to be better than short-course ARV, it would be unclear if the alternative were better than nothing. A placebo trial provides direct information on efficacy. Generalizability: An equivalence design has limited generalizability because the long 076 regimen requires women to enroll long before they usually seek prenatal care. These women may have atypical health practices, so that results may not apply well to the general population of pregnant women. Placebo trials enroll at the time when women usually seek prenatal care. Rapidity of completion: An 076 equivalence trial is more time-consuming than a placebo trial, needing to recruit women by 30 weeks gestation (when 15-20% of women seek prenatal care) rather than by 34 weeks (when about 75% seek care) and with a 50% larger sample size. Study completion could take months to more than a year longer. A fast trial may save lives by speeding implementation of a perinatal ARV program. Africa has 200,000 mother-to-child HIV transmissions per year. Once implemented, a program reaching 50% of care sites, with 75% of women appearing by 34 weeks gestation, and 50% ARV efficacy would prevent 37,500 mother-to-child HIV transmissions per year. Conclusion: We believe that using placebo controls is essential to reliably establish an appropriate local standard of care. A placebo control trial of ARV is far more likely than an equivalence trial to produce results that are valid and generalizable for the African countries where these trials are being conducted. It also requires less time and therefore supports earlier implementation of ARV programs, which may prevent nearly 40,000 mother-to-child HIV transmissions per year.

[482/44241] Expansion of US Medicaid system to cover HIV drugs will prevent thousands of deaths and AIDS diagnoses, and is affordable
James G. Kahn1, B. Haile1, S.W. Chang2. 1Center for AIDS Prevention Studies, UCSF, Inst. for Health Pol Stud., UCSF, Box 0936, San Francisco, CA; 2HJ Kaiser Family Foundation, Menlo Park, USA

Objectives: To assess the health and federal fiscal implications of expanding Medicaid, the US insurance system for the poor, to cover HIV care for those without access to HIV medications. Design: Computer spreadsheet modeling of HIV disease progression, treatment costs, and financing. Methods: We compared projected outcomes with and without an expansion, over 5 years. The expansion assessed is for all individuals lacking medical insurance for antiRetroviral medications and with income below $10,000 per year (125% of poverty for a single person). It covers medications and outpatient care. The analysis incorporates estimates of the number of people living with HIV infection by disease stage; current access to antiRetroviral therapy, insurance status, and income; likelihood of qualifying for and enrolling in the expansion; a natural history model of HIV disease progression, calibrated to cohort studies; the efficacy of combination antiRetroviral therapy in reducing HIV disease progression, based on clinical studies; the costs of HIV medications and care; and federal government costs for the expansion. Results: An estimated 37,100 individuals would enroll in the expansion, 83% pre-AIDS and 17% with AIDS. Over 5 years, there are 5,200 fewer deaths with the expansion than without. 6,700 fewer individuals with early HIV disease progress to an AIDS diagnosis or early death. The expansion increases total life years by 14,500 over the 5 years. Total federal costs for the expansion itself are $1.3 billion over 5 years; these are offset by decreased federal expenditures of $765 million: Medicaid, due in part to slowed progression to AIDS ($385 million), AIDS Drug Assistance Program ($240 million), and other programs ($140 million). Federal budget neutrality, currently required to approve Medicaid waivers to states, can be achieved by an 18% reduction in Medicaid prices for HIV drugs. Despite this price reduction, total spending for HIV drugs increases by $175 million. Conclusion: This analysis suggests that expansion of the US Medicaid system to provide wider access to combination antiRetroviral therapy would prevent thousands of deaths and AIDS diagnoses, leading to 14,500 more years of life for persons with HIV disease over 5 years. The program is affordable from a federal perspective, with budget neutrality achievable if the expansion is accompanied by an 18% reduction in HIV drug prices for Medicaid.

[471/44123] Additional unethical aspects of vertical transmission studies in developing countries
Peter G. Lurie, S.M. Wolfe. Public Citizen, 1660 20th St., NW Washington, DC, 20009, USA

Issues: After the publication of the ACTG 076 study, which demonstrated that zidovudine caused a 2/3 reduction in HIV transmission from pregnant women to infants, attention turned to identifying an affordable regimen for developing countries. In April 1997, we identified a total of 15 studies involving over 17,000 developing country women that sought to identify such regimens (e.g., regimens using zidovudine for 4-6 weeks antepartum) by providing placebos or interventions not yet proved effective to at least some women. We initiated a campaign, both in medical journals and in the popular press, to redesign the studies so that all women had access to at least some antiRetroviral drugs. So far, one study in Ethiopia has been redesigned to eliminate its placebo group. Project: To determine whether the studies violate additional ethical guidelines. Results: There are three additional unethical aspects to these studies. 1. The informed consent form in at least one study fails to state that a pre-planned subanalysis of data from ACTG 076 showed a 2/3 reduction in HIV transmission among women receiving an average of only 7 weeks of zidovudine. 2. There is no provision for the women to continue on antiRetrovirals after the study is completed, even though noncompliance with antiRetrovirals (in this case, forced noncompliance due to zidovudine being unaffordable in most developing countries) is a known cause of HIV resistance. 3. There is also an observational study of 125 HIV-positive pregnant women, conducted by the Thai and US Armies, Thai researchers and Johns Hopkins University, which has continued to provide the subjects with no zidovudine, even though zidovudine is so available that Thai researchers terminated their own placebo-controlled trial in January 1997. Conclusion: The studies are unethical in a variety of ways beyond the provision of placebos. This highlights the need for the studies to be redesigned and for a renewed commitment to the conduct of ethical studies in developing countries.

[252/34356] Cochrane Collaborative Review Group on HIV infection and AIDS: Systematic reviews of HIV and AIDS prevention and treatment
George Rutherford, G.E. Kennedy, L.A. Bergo. UCSF 74 New Montgomery St, Ste 600 San Francisco, California, USA

Issue: To describe the purpose, structure and products of the Cochrane Collaboration and the new Cochrane Collaborative Review Group (CRG) on HIV Infection and AIDS. Project: The Cochrane Collaboration is an international, interdisciplinary effort to review published and unpublished studies of the effects of health care which have been evaluated in randomized controlled trials and other types of controlled studies and to prepare, maintain and disseminate systematic reviews of these studies. Systematic reviews are commissioned and managed by 60 CRGs (e.g., diabetes, pregnancy and childbirth). The reviews, along with an annotated bibliography of other meta -analyses and a registry of randomized controlled trials, are published electronically each quarter in The Cochrane Library on CD-ROM and via the Internet. We have undertaken the establishment of a CRG on HIV Infection and AIDS. Results: We held two meetings in 1997 attended by scientists and advocates from 7 countries and several disciplines. We have decided to focus the CRG initially on conducting systematic reviews in four areas - treatment, diagnosis and prognosis; behavioral, social and policy prevention interventions; biomedical prevention; and health services delivery - and identified 8 editors from Australia, Canada, Peru, the United States and the United Kingdom to oversee the reviews in these areas. While primarily relying on randomized controlled trials, we will also include in our reviews other types of trials. We have identified 36 systematic reviews which we will initially commission (e.g., post-exposure prophylaxis for needle sticks, community-level STD control to decrease HIV transmission); several of these will be updates of previous meta analyses. We anticipate publication of the CRG's work in The Cochrane Library will begin in 1998. Lessons Learned Systematic reviews are integral to evidence-based medicine and public health. This CRG extends the Cochrane Collaboration to studies of the prevention and treatment of HIV infection and AIDS. We welcome additional collaborators as reviewers, peer referees and journal hand searchers.

[580/33277] HIV counselling and testing in Tanzania and Kenya is cost-effective: Results from the voluntary HIV counselling and testing study
Michael Sweat1, G. Sangiwa2, D. Balmer3. Voluntary HIV C & T Testing Team; 1Johns Hopkins University, Baltimore, MD, USA 2Muhimbili University Dar Es Salaam, Tanzania 3Kenya Assn. of Professional Counselors, Nairobi, Kenya

Objective: To estimate the cost per HIV infection averted from HIV voluntary counseling and testing (VCT) among persons seeking this service in Dar es Salaam, Tanzania and Nairobi, Kenya in association with a randomized controlled trial of the impact of VCT. Methods: Study participants were randomized to receive HIV VCT or a standardized health information intervention (w/6 mt. wait period for VCT). The number of HIV infections averted for a one year period among those randomized to receive VCT were estimated using a Monte Carlo based probability model with values for changes in behavior from baseline to 6-month follow up. Per-client costs (in US $ and independent of research costs) were calculated for a service reaching 3000 persons at each site per year with a similar demographic and risk profile as those in the study. Results: The cost per client for HIV VCT was $29 in Tanzania and $27 in Kenya. In both sites these total costs were composed of ~ 74% labor and infrastructure, ~ 2% start up, and ~ 24% commodity expenses. The single largest cost was for counselor and supervisor salaries and benefits (~36%). The cost per HIV infection averted averaged $303 in Tanzania and $241 in Kenya. In Tanzania many seronegative men who received VCT did not change risk behavior; the intervention was not cost-effective for this group. The cost per HIV infection averted from VCT by sub-group were:

+----------------------------------------------------------+
| |Avg |Couples |Individuals |Males |Females |
+--------------+-----+--------+------------+------+--------+
|Tanzania HIV+ |$124 |$406 |$120 |$127 |$105 |
+--------------+-----+--------+------------+------+--------+
|Tanzania HIV- |$482 |$445 |$1117 |$3731 |$718 |
+--------------+-----+--------+------------+------+--------+
|Kenya HIV+ |$75 |$58 |$100 |$73 |$74 |
+--------------+-----+--------+------------+------+--------+
|Kenya HIV- |$464 |$451 |$591 |$442 |$480 |
+--------------+-----+--------+------------+------+--------+

HIV VCT is a labor intensive intervention due to interpersonal counseling, laboratory procedures, supervision and quality control. VCT is more cost-effective when targeted to HIV+ persons, couples, and women. HIV VCT compares favorably to other interventions in cost per HIV infection averted in East-Africa ($251 for enhanced STD services in Mwanza, Tanzania) and the US ($60,000 for VCT in acute care settings with HIV prey GT 1%). HIV VCT services in East-Africa result in significant reductions in HIV transmission at lower costs.

[190/13347] Long-term non-progressors (LTNP): Prevalence and predictors through 18 years of follow-up
Eric Vittinghoff1, M.I. Feinberg3, T. Elbeik2, S. Staprans2, M. Carrington4, G. Colfax1, S. Buchbinder1. 1Department of Public Health, Suite 500, 25 Van Ness Avenue, San Francisco; 2University of California, San Francisco, CA; 3Emory University, Atlanta, GA; 4National Cancer Institute, Rockville, MD, USA

Background: To describe HIV disease progression, nonprogression, and predictors of loss of LTNP status among men who have sex with men (MSM) in the San Francisco City Clinic Cohort. Methods: LTNP were initially defined as HIV infected >10 years with CD4 counts persistently >500 cells/mm3 in the absence of antiRetroviral therapy. Kaplan Meier methods were used to estimate time to 1993 AIDS diagnosis, death, and time to loss of LTNP status after 10 years. Participants initiating antiRetroviral therapy prior to a decline in CD4 < 500 cells/mm3 were censored at initiation of therapy. Cox proportional hazards models were used to evaluate predictors of loss of LTNP status. Results: Among 622 MSM with well-characterized dates of seroconversion, median time to AIDS was 9.8 years. At 18 years after seroconversion, the estimated probability of developing AIDS was 86% (95% CI 80-89%). Median time to death was 11.8 years; 18 years after seroconversion, estimated mortality was 75% (95% CI 72-79%). Of 596 men infected >10 years, 67 (11%) were initially LTNP and an additional 15% were alive, AIDS-free, but not followed. In Kaplan-Meier analysis of those 67 men, only 14% (95% CI 3-26%) were estimated to remain LTNP after 18 years of infection. There was considerable heterogeneity of plasma viral levels among LTNP (range <500-129,000 copies/ml by bDNA assay), and plasma viral level was the strongest independent predictor of LTNP status (relative hazard 2.8 per log bDNA, 95% CI 1.42-5.54). We were able to isolate virus from PBMC culture without CD8+ depletion on at least one occasion in 14/15 LTNP tested. Of 19 LTNP with genetic testing of chemokine alleles, 32% were heterozygous for the CCR5D32 allele, 8% for the CCR2641 allele, and 37% for either. Conclusion: Long-term nonprogression was a relatively rare phenomenon in our cohort of long-term HIV infected MSM, accounting for only 2% of all HIV-infected MSM 18 years after seroconversion. Plasma viral level predicts loss of LTNP status, suggesting that LTNP who have maintained high CD4 counts in the presence of moderate plasma viral loads should initiate antiRetroviral therapy. Current described variants in chemokine receptors account for a minority of LTNPs.

[154/14135] Self-acceptance of gay identity decreases sexual risk behavior and increases psychological health in US young gay men
Craig R. Waldo, S.M. Kegeles, R.B. Hays. USCF Center for AIDS Prevention Studies, 74 New Montgomery St., San Francisco, CA 94105, USA

Objectives: To examine the relationships between self-acceptance of gay identity and sexual risk-taking behavior and mental health in gay men aged 18-27. A secondary objective was to explore the role of gay community involvement, social support for gay identity, and sexual orientation disclosure to these variables. Design: Cross-sectional analysis of a longitudinal prospective cohort from 3 mid-sized US cities. Methods: Data reported by 301 young gay men are analyzed via structural equations modeling in LISREL VIII. Constructs examined are displayed in Figure 1. HIV Sexual Risk-Taking is defined as unprotected receptive or insertive anal intercourse or receptive oral sex to ejaculation. Each latent construct was estimated with three composite indicators with maximum likelihood procedures. A measurement model analysis was performed to ensure that all constructs were measured adequately (see fit statistics in Figure). Results: All standardized path coefficients in the structural model reached statistical significance (p < .01) and all fit statistics indicated an excellent model fit (see Figure). Results indicated that: (1) Gay community involvement predicts social support for and self-acceptance of gay identity; (2) Social support for gay identity is related to self-acceptance of gay identity and psychological health; and (3) Self-acceptance of gay identity predicts increased psychological health, sexual orientation disclosure, and decreased sexual risk behavior. Self-acceptance of gay identity can be important for reducing sexual risk behavior in US young gay men. Men who accept their sexuality most are also likely to be more psychologically healthy. Gay community involvement and social support for gay identity increase self-acceptance of gay identity. These findings suggest that HIV preventive interventions for young gay men can be more effective if they facilitate personal and cultural environments that increase self-acceptance of gay identity.

Poster presentations

[14290] Why the prevention works differently among men and women? Effects of an aids prevention program for young adults in Brazil
Maria Cristina Antunes1 Maria Cristina Antunes2 R. Stall3 N. Hearst3 C.A. Peres4. 1R. Cel. Gordinho Filho, Sao Paulo; 2NEPAIDS-University of Sao Paulo, Sao Paulo; 3CAPS-UCSF, San Francisco; 4NEPAIDS-University of Sao Paulo, Sao Paulo, Brazil

Objectives: To understand the gender differences and obstacles to adopt consistent safer sex practices among youth who had participated in a AIDS Prevention Program in Sao Paulo inner-city. Methods: A longitudinal using a wait-list control design with pre and post-intervention measures. Students in two schools participated in four 3-hour workhops about Safer Sex, Reproduction, and AIDS. The workshops used an interactive approach to talk about issues such as sexuality, gender sexual scripts, AIDS symbolism, risk perception, knowledge, sexual scenes, and condom negotiation. Of 394 young adults who participated in the baseline survey, 304 completed a post intervention questionnaire, and 136 attended the workshops sessions after the first wave. We interviewed 10 students that participated in the workshops to explore aspects of perceived norms, sexual practices, risk perception, and barriers to condom use. Results: At baseline we found statistically significant differences among men and women. Men use to know more about sexual practices and HIV transmission through anal sex. Women feel more responsible to the sex consequences, to avoid pregnancy; sex is related to love and they don't have large experiences in different sexual practices. They talk more to avoid pregnancy and HIV test with partners than men, that conduct more the sexual act and say what will be done. We found statistically significant effects from the workshops for females, who improved communication with partners about sex and AIDS, increased belief on condom efficacy, improved negotiation of sex/pleasure with their partners and reported less unprotected sex with non-monogamous partners after the interventions. The unique statistically significant effects to males were to decrease the perceived self-efficacy to protect against HIV. Preliminary data from deep-interviews showed that the denial of the AIDS risk infection and the gendered sexual scripts are important barrier to adopt safer sex practices, and that it is essential to develop continuous intervention and involve all the community on the AIDS prevention program. Conclusion: To be in contact with sexuality and to break the prohibition and pleasure taboo makes these women able to think, decide, communicate, and negotiate in the sexual context. They feel less guilt and more confident with their own desires. Male gender norms in this community dictate that any man should be sexual experts and so less interested in learning about the sexual issues in which they are unfamiliar. They are also oppressed by very rigid gender norms that make difficult for them to think about risks for HIV transmission, condom use, or responsibility about sex.

[43107] Repeat HIV antibody testing among men who have high-risk sex with men in San Francisco
James Dilley1 B. Adler1 W. McFarland2 W. Woods 2 J. Sabatino 1 J. Rinaldi1 T. Lihatsh1. 1Univ Cal San Fran AIDS Health Project Box 0884 San Francisco CA 941439-0884; 2Center For AIDS Prevention Studies, USA

Background: Repeat HIV antibody testing has been noted among men who have sex with men (MSM) in several recent studies. However, the role of repeat testing in reducing high risk behaviors or maintaining safer behaviors is not well understood. We sought to further describe patterns of testing and risk related behaviors among repeat anonymous testers as part of a prevention intervention study. Methods: MSM (N = 93) were recruited for a counseling intervention study when scheduling an anonymous HIV antibody test. Participants were non-IDU, had at least one previous test, were HIV negative at their prior test, and had unprotected anal intercourse with another male who was positive or of unknown status within the last year. Subjects were asked about their prior testing history and risk related behaviors. Results: On self-report, subjects identified themselves as "regular" (59%) or "non-regular" (41%) repeat testers. Among regular testers, the most common testing schedule was every 6 months (58%) followed by every year (33%). Compared to non-regular testers, regular testers had tested more times in the past (median 7 vs. 4 times, p < 0.001), were older (median 34 vs. 32 years, p = 0.009), were more likely to be white (82% vs. 59%, p = 0.012), and were less worried about already being infected (p < 0.001). While both groups reported equal numbers of episodes of unprotected anal sex, regular testers reported a greater number of partners (median 5 vs. 3, p = 0.004) in the last 12 months. Regular repeat HIV testers are not simply worried, but continue to be at risk for infection. Recognizing that repeat testing is not necessarily inappropriate, we recommend: 1) determination of whether repeated counseling and testing is effective in maintaining safer behavior; and 2) evaluation of targeted counseling interventions for this population.

[23439] Knowledge of AIDS and risky sexual practices of adolescent female hawkers in bus and truck stations in Ibadan, Nigeria
Ademola Ajuwon1 K. Osungbade2 F. Fawole3 P. Lurie4 N. Hearst4. 1Dept. of Preventive & Social Medicine, University College Hospital, Ibadan;2University College Hospital, Ibadan;3College of Medicine, Univ. of Ibadan, Ibadan;4UCSF Center for AIDS Prevention Studies, San Francisco, CA, USA

Objectives: In Nigeria, risky sexual activities among adolescents increasingly place this population at risk of HIV. Female adolescents who hawk in bus and truck stations in cities may be more at risk for HIV than others because their poor economic status and relative youth make them vulnerable to sexual exploitation by the older men operating in these settings. The objectives of this study were to assess AIDS knowledge; identify sexual risk practices, and plan appropriate interventions. Methods: Data were collected in Nov. 1997, by trained female interviewers who used a questionnaire to conduct face-to-face interviews in 4 Ibadan bus and truck stops used for long distance travel. 228 female hawkers, aged 10-19 years, agreed to participate in the study (response rate 97%). Results: The mean age of the hawkers was 17.1 years and the items sold were mainly non-alcoholic beverages, snacks and food, in that order. 48% hawked for parents, 21% on their own and 17.7% were hired by older women traders. The STDs known were AIDS (74%) and gonorrhea (43%). The hawkers averaged 3.2 on a 7-point AIDS knowledge scale; 57% believed that HIV can be transmitted through sharing toilet seats with an infected person and 27% believed that there is a cure for AIDS in Nigeria. 42% did not know what to do to prevent STD's; others would abstain from sex (12.4%), use a condom (9.2%), have one sexual partner (6.0%), wash the vagina after intercourse (3.2%), use drugs (2.8%), pray (0.9%), and be careful in selecting partners (0.9%). Forty-two percent have had sex; 58% had not. An older male was mostly (57%) mentioned as the first sexual partner. 16% used a condom during the first episode of sex, 84% did not. Of the 59 who had sex in the past three months, 10% used a condom during their last sexual encounter; 90% did not. 3.9% reported that they had been raped in the past in the course of trading in the stations. 87% of those sexually active reported that their last episode of sex was with an older male partner. The hawkers averaged 9.8 points on a 21-point self-reported assertiveness skills scale. Conclusion: The low level of condom use by the hawkers and the frequency of having older partners places them at high risk for HIV. Barriers to condom use include poor knowledge about AIDS and limited assertiveness skills. An intervention is now underway to address these problems.

[389/32406] Protease inhibitors (PI) in the HIV+ homeless and marginally housed (H/M): Good adherence but rarely prescribed
David Bangsberg12, M. Robertson3, E. Charlebois3, J. Tulsky3, F.M. Hecht3, J. Bamberger4, A.R. Moss3. 1995 Potrero Ave. Ward 95 Room 513 San Francisco California 94110; 2Center for AIDS Prev Studioes SFGH/UCSF San Francisco CA; 3Univ. of California San Francisco San Francisco CA; 4Department of Public Health San Francisco CA, USA

Objectives: Combination therapy with PIs is sometimes withheld from poor or marginalized populations because of concerns about adherence to therapy. We report on the prevalence of PI use and adherence to PI therapy in the REACH cohort, a prospective cohort of HIV-positive H/M persons. Design: Prospective cohort study. Methods: We recruited a representative cohort of 154 HIV-positive persons from lunch lines, shelters and hotels charging <400/mo in San Francisco. We characterized antiRetroviral (ARV) use as (a) combination therapy with a PI and 2 reverse transcriptase inhibitors (PI + 2RTI), (b) RTI therapy alone and (c) no therapy. Adherence was measured by self-report of doses missed. We validated self-reported adherence by drug plasma levels. Results: 87% of eligible subjects agreed to be followed. Cohort retention was 82% at one year. At baseline, 7% were on PI/RTI therapy and 25% on RTIs alone. There was no increase in baseline use PIs over time. Among those in the cohort, PI use increased to 30% at one year of followup. The median drug exposure was 4.5 months. Prevalence of each PI was: nelfinavir-43%, indinavir-37%, saquinavir-17%, nelfinavir/saquinavir-2% and ritonavir-0% Of these, 20 subjects had >6 months of PI exposure (median = 10.2 mo). Street and shelter dwellers were less likely to receive PIs at baseline than hotel dwellers (3% vs. 8%; p = .05) and women less likely than men (0% vs. 9% p = .03). There was marginally less PI use at baseline in injection drug users (IDU) compared to non users (5% vs. 11%). (p = .06). 80% of subjects on PI therapy report missing less than 2 doses per week (n = 23). 88% of adherent subjects had detectable drug in their plasma (n = 18). Conclusions: 1. Baseline access to PIs in the H/M population was poor (7%) compared to levels of 50-70% reported in standard clinical settings. PI use was increased by being followed in the REACH cohort. 2. H/M persons prescribed PIs report relatively good adherence, validated by plasma drug levels. 3. Access to PIs should be expanded in the homeless and marginally housed.

[32390] Protease Inhibitors (PI) are associated with viral load suppression in HIV+ Homeless and Marginally housed (H/M) adults
David Bangsberg1 2 A.R. Zolopa3 E. Charlebois4 F.M. Hecht4 M. Holodniy5 T.C. Merigan3 A.R. Moss4. 1995 Potrero Ave. Ward 95 Room 513 San Francisco California 94110; 2Center for AIDS Prev Studies SFGH/UCSF San Francisco CA; 3Stanford University Stanford CA; 4University of California San Francisco San Francisco CA; 5VA Palo Alto Healthcare System Palo Alto CA, USA

Objective: Combination therapy with PIs is sometimes withheld from poor or marginalized populations because of concerns about adherence to therapy. We examined the extent of viral load reduction associated with adherent and nonadherent PI use in the H/M. Design: Prospective cohort study. Methods: We recruited a representative cohort of 154 HIV-positive persons from lunch lines, shelters and hotels charging <400/mo. We characterized antiRetroviral (ARV) use as (a) PI and 2 reverse transcriptase inhibitors (PI + 2RTI), (b) RTI therapy alone and (c) no ARV therapy. Viral load was measured with the Roche Amplicor assay (ND < 400 copies/ml). Adherence was measured on the day of viral load determination by number of self-reported missed doses in the prior week (=missed doses/wk = nonadherent). Drug plasma levels were consistent with self reported adherence. Results: The cohort (n = 154) is 61% nonwhite; 40% are current injection drug users. 58% live in low income hotels, 35% in streets and shelters. Mean viral loads in the PI/RTI, RTI and No ARV groups were 2.7, 3.9, and 4.3 copies/ml respectively (p = 0.03; fig 1. bar = SE of mean). 30%, 16% and 6% of subjects in the respective three groups had undetectable viral RNA. Viral suppression was confined to the adherent subjects (fig 2. Ad = adherent;NA = nonadherent) Fig 1. Viral Load by Treatment Group Fig 2. Viral Load by Treatment Group and Adherence Conclusions: 1. PI therapy in HIV+ homeless and marginally housed people is associated with viral load suppression. 2. Viral load suppression was confined to those reporting good adherence. 3. Viral load among the non-adherent did not differ from that among those on no ARV therapy.

[39/43170] Sexual health in a young city: A community-based intervention on youth sexual health in Lima
Carlos F. Caceres, C. Cabezudo, O. Jimenez, R. Valverde, G. Perez-Luna. Dept. of Public Health, Cayetano Heredia, Univ. AV Honorio Delgado 430, Lima 31, Peru

Objective: To design, implement and evaluate a community-based and multi-sectoral intervention on youth sexual health (SH) and HIV/STD prevention in two lower-middle class districts in Lima, Peru. Design: A community intervention trial using mostly qualitative methods. Methods: A program promoting community mobilization, development of local resources and youth empowerment to improve youth sexual health/rights was designed, implemented and evaluated in two districts of Lima. It involved 3 components (i.e. strengthening/articulation of youth SH services; mobilization of relevant social actors such as the municipalities, the local health and education sectors, youth organizations, CBOs and the press; and a multi-strategy campaign). Evaluation data were collected at onset (i.e. a home-based survey of youths; youth focus groups, and interviews with community stakeholders) and during/immediately after the intervention (i.e. interviews with community leaders and youth peer educators, discussion groups, post-workshop surveys and an evaluation workshop at the end). Results: Most planned activities (i.e. 2 music concerts, 2 community fairs, 4 train-the-trainers worshops, 18 sexual health workshops for youth, a poster design contest for World AIDS Day, and several community stakeholders meetings) were implemented despite the limited budget during the 8-month intervention. The young found the program strategy very innovative and praised its proximity to their values and its participatory nature. The program generated great expectations amongst social actors involved, particularly among health care providers (who were united in a pilot network), teachers, CBOs and some youth organizations, which contrasted with difficulties in assuring its continuity due to the limited priority this issue had for the municipalities and health and education authorities, as expressed in their lack of initiative to incorporate this strategy into their policies, and to allocate resources and staff time to this program. Conclusions: This community-based and multi-sectoral youth sexual health program succeeded in raising worthy expectations and interest among intermediate-level. community stakeholders and the youth themselves, but its continuity was threatened by the local authorities' perception of low applicability to their dynamics (due to a lack of immediate political outputs) and difficulty of multisectoral coordination. This strategy seems very promising for urban contexts in Latin America, but local authorities' understanding of, and commitment to it should be assured at initial stages.

[13217] Evidence for Evolution of HIV-1 within Behavioral Risk Groups in San Francisco 1984-1997
Edwin D. Charlebois1 D. Osmond2 R. Grant2 A. Moss2 W. Winkelstein3. 1UCSF-CAPS, Suite 600, 74 New Montgomery, San Francisco; 2University of California San Francisco, CA; 3University of California Berkeley, CA, USA

Objectives: To compare HIV-1 genetic diversity in San Francisco seroconverters to investigate the evidence for segregation of viral strains by behavioral risk group and to look for evidence for evolutionary change overtime. Methods: HIV-1 RNA was extracted from stored or fresh plasma taken within 6 months from the time of seroconversion. Virus was successfully isolated from 9 gay men from the San Francisco Men's Health Study, who seroconverted during 1984-1985, eight gay men from the San Francisco Young Men's Health Study - seroconverting during 1994-95, and 9 heterosexual injection drug users from a study of HIV and TB in the San Francisco homeless seroconverting recently. Sequencing was performed using an ABI automated sequencer. DNA distances were calculated using a Kimura 2-parameter model within the DNAdist program contained within the PHYLIP 3.0 package for UNIX. Results: All isolates were found to cluster with the Clade B consensus sequence. Neighbor joining trees and maximum-parsimony methods placed the heterosexual isolates in among the samples from gay men taken in the 80's and 90's. A significant difference in the mean genetic distance between the three groups was observed (p < .0001 Wilcoxon test) Conclusion: Unlike the situation in Amsterdam, isolates from injection drug users did not uniformly cluster apart from isolates from gay men. Genetic diversity was relatively low in isolates from gay men at the start of the San Francisco HIV epidemic. Higher genetic diversity was seen in samples from gay men more recently infected with HIV, consistent with evolutionary divergence. Heterosexual injection drug users were found to have an intermediate level of diversity, suggesting a later introduction of HIV into this population.

[33152] Real experience: The key to vaginal microbicide acceptability
Margaret Chesney1 M.E. Bentley2 K.M. Morrow3 K. Mayer3. 1UCSF Prevention Sciences Group 74 New Montgomery San Francisco, CA; 2Johns Hopkins University Baltimore MD; 3Brown University Providence RI, USA

Objectives: To determine whether acceptability of a vaginal microbicide is influenced more by initial attitudes about microbicides or experience with actual products. Methods: As part of a Phase I Study of BufferGel, a novel vaginal microbicide, 27 low risk abstinent or monogamous women in Rhode Island responded to self-report questionnaires assessing acceptability of vaginal microbicides. Mean age of the women was 31 years; 63% were White, 15% Hispanic, 11% African American, and 11% Other. Acceptability data were collected at enrollment and after daily use of the product for four weeks. At the end of the study, women were asked about their willingness to try BufferGel if approved for vaginal use. With no differences between abstinent and monogamous women on key variables, the data were combined for this analysis. Results: At enrollment, women indicated that they would be more willing to try microbicides that are gels rather than creams. While the majority (85%) were moderately or very willing to try gels, 15% were only slightly willing. The women had no preference for reusable or disposable applicators. After using the product, 100% of the women indicated liking BufferGel's color. Conversely, although 100% of the women indicated a willingness to use reusable applicators prior to the study, after experience with the reusable BufferGel applicator, 7 of these women reported not liking the applicator, particularly the fact that it had to be washed between uses. After daily use of BufferGel, 75% of the women said that they would use the microbicide if it were approved for vaginal use. This willingness was not related to any sociodemographic factors, or any attitudes or preferences assessed prior to use. This willingness was supported by a positive change in attitudes about clear gel after experience with the product. There was also a trend indicating that not being willing to use BufferGel is associated with reports of not liking reusable applicators, after having had experience with them. Conclusions/Lessons Learned Actual experience with vaginal microbicides rather than attitudes prior to use are associated with willingness to these products in the future. Strategies that encourage women to try these products are more likely to increase vaginal microbicide use than educational campaigns designed to change pre-existing attitudes.

[23112] HIV prevention programs must address environmental influences to reduce risk behavior among young Asian men who have sex with men
Kyung-Hee Choi E. Kumekawa. Center for AIDS Prevention Studies 74 New Montgomery St. #600 San Francisco CA 94105, USA

Objectives: To explore environmental issues affecting HIV risk among young Asian men who have sex with men. Methods: We conducted in-depth interviews with 51 individuals who were knowledgeable about young Asian men who have sex with men (e.g., providers from health departments and community-based organizations, bartenders, shopkeepers, gay Asian community leaders and members) using the Community Identification Process, an ethnographic technique designed to identify hard-to-reach populations and to develop HIV prevention programs for these populations. The interviews were conducted in San Diego, Anaheim-Garden Grove, Oakland-Berkeley, and Seattle during May-September 1997. Study participants were asked about the influence of the family, the general Asian community, and the mainstream gay community on sexual risk among young Asian men. Results: We found 6 environmental factors related to sexual risk among young Asian men who have sex with men: * cultural expectations of the parents (parental pressure to get married, have children, carry on the family name and traditions, and not bring shame on the entire family); * family silence about sex (little communication about homosexuality because of its personal, private, and sexual nature); * stigmatization of homosexuality in the Asian community; * the contrast between self-image and the ideal image of male beauty in the gay community (e.g., a "glamour[ous], chiseled, healthy-looking White image"; the "young, [with] blond hair with buff bodies"); * negative stereotyping of Asian men in the gay community (e.g., monogamous, subservient, being the receptive partner in anal sex, and at low risk because of the perceived low incidence of HIV in the population); and * related emotional difficulties experienced by Asian men including negative identity, low self-esteem, self-image, and self-worth, alienation, and depression. Conclusions: The data suggest that future HIV prevention strategies must consider the strong influence of environmental factors. These strategies should be directed at the family, the Asian community, and the gay community with programs such as family counseling, support groups for families with gay children, mass media campaigns to educate the community about sexual and ethnic diversity, and forums to discuss homophobia and negative stereotyping of Asian men.

[33268] Serodiscordant married couples undergoing couples counseling and testing reduce risk behavior with each other but not with extra-marital partners
Thomas Coates1 G. Coates2 G. Sangiwa2 D. Balmer3 C. Furlonge4 C. Kamenga5 S. Gregorich1. 1USCF Center for AIDS Prevention Studies, 74 New Montgomery San Francisco, CA; 5FHI/AIDSCAP Arlington, VA, USA 2Muhimbili University Dar Es Salaam, Tanzania 3Kenya Assn. of Professional Counselors Nairobi, Kenya 4Voluntary Counseling & Testing Centre, Curepe, Trinidad WI

Objectives: To determine the impact of voluntary counseling and testing (VCT) on seroconcordant and serodiscordant couples' sexual behavior with themselves and with secondary partners. Methods: Multicenter randomized controlled study in Nairobi (Kenya), Dar es Salaam (Tanzania), and Port of Spain (Trinidad). Couples (N = 586; married couples = 429) were randomized to receive VCT or Health Information (HI) and followed for 6 months. Standardized interviews were administered at baseline and follow-up in Kiswahili or English. Couples assigned to CT received results individually and then were encouraged to share results with their partner in the presence of the counselor. Results: 95.2% of those assigned to CT (N = 233) agreed to counseling and testing and 74.4% received test results. 72.3% of the couples were concordant seronegative (M-F-), and 6.5% were concordant seropositive (M+F+). 21.3% were serodiscordant (11.3% M-F+; 10% M+F-). 87% of couples were retained at 6 month follow-up. The table presents the percent reporting sex with spouse and unprotected sex with spouse during the previous six months at baseline and the six month follow-up for men; data for women are comparable.

+----------------------------------------------------------------+
| | Sex with Spouse |Unprotected Intercourse with Spouse |
+-----+---------------------+------------------+-----------------+
| |Baseline |Six months | Baseline | Six Months |
+-----+---------+-----------+------------------+-----------------+
|M+F+ |100% |52% |75% |38% |
+-----+---------+-----------+------------------+-----------------+
|M+F- |100% |55% |100% |45% |
+-----+---------+-----------+------------------+-----------------+
|M-F+ |84% |71% |79% |50% |
+-----+---------+-----------+------------------+-----------------+
|M-F- |97% |79% |93% |70% |
+-----+---------+-----------+------------------+--------
Relative to M-F-, M+F+ decreased significantly in sex with spouse (p < .05); trends for M+F- and M-F+ were non-significant (p < . 12). Relative to M-F-, the other three groups decreased significantly in unprotected intercourse with spouse. Unprotected sex with non-enrollment partners increased from 11% to 15% and there were no differences between groups. Couples counseling and testing reduced unprotected intercourse among spouses, especially among serodiscordant and seropositive concordant couples. Rates of unprotected intercourse with secondary partners was low at baseline and increased slightly at 6 months. VCT may be a useful strategy to prevent HIV transmission in serodiscordant couples.

[13116] Sexual risk behaviors, knowledge, and attitudes in a population-based probability sample of Dar es Salaam, Tanzania: Results from the voluntary HIV counseling and testing study (VHCTS)
Thomas Coates1 Japhet Killewo2 S. Gregorich1 G. Sangiwa2 T. Coates1. 1UCSF Center for AIDS Prevention Studies 74 New Montgomery San Francisco, CA, USA 2Muhimbili University College Dar Es Salaam, Tanzai

Objectives: To estimate and describe the level of sexual risk behavior, HIV transmission knowledge and related attitudes in the population of Dar Es Salaam, Tanzania. To empirically assess the need and demand for voluntary HIV counseling & testing (VCT) services in the population of Dar es Salaam. To complement the findings of the VHCTS--a multi-site randomized controlled trial of the efficacy of VCT as a preventive intervention. Methods: A multistage area sampling design was developed. Dar es Salaam is administratively divided into districts, wards, and ten-cells. Of the 32 urban wards, 10 were selected with probability proportional to estimated size. Within each selected ward, 20 ten-cells were selected at random. Within selected ten-cells, 5 households (HH) were selected at random and within each HH one eligible member was selected at random as the respondent. Results: 859 (86%) of all interviews were completed. We will present data regarding participants' demographics, sexual history, current sexual behaviors, condom use, HIV transmission knowledge, HIV transmission attitudes, and HIV testing history. Reports on sexual behaviors include prevalence of protected and unprotected sexual intercourse with both "primary" and other partners. We will also compare these data to those from participants in the VHCTS. These are unique data that exist only for a few countries. This population-based probability sample is important in: (1) mapping sexual risk behaviors in Dar es Salaam, (2) providing a population-based comparison to the risk-profiles of participants in the VHCTS, which has already demonstrated the efficacy of VCT as a preventive intervention, (3) informing policy decisions regarding the delivery of HIV-related services in Dar es Salaam--specifically, decisions regarding the need for services and their optimal placement.

[14107] Validation of self-reported sexual risk behavior with STD incident rates: Results from the voluntary HIV counseling and testing study
Thomas Coates1 Colin Furlonge2 D. Mwakagile3 C. Kamenga4 J. Schacter1 S. Gregorich1 T. Coates1. 1UCSF Center for AIDS Prevention Studies, 74 New Montgomery San Francisco, CA; 4FHI/AIDSCAP, Arlington, VA, USA 2Voluntary Counseling & Testing Centre, Curepe, Trinidad WI 3Muhimbili University College Dar Es Salaam, Tanzania

Objective: To assess the validity of self-reported sexual risk behavior by measuring STD incidence as a biological marker of unprotected sexual intercourse. Method: A multicenter randomized controlled study was conducted in Nairobi, Kenya, Dar es Salaam, Tanzania, and Port of Spain, Trinidad. Study participants were randomized to receive voluntary HIV counseling and testing (VCT, N = 2152) or a standardized health information (HI) intervention (and to return for VCT in 6 months, N = 2141). Diagnosis and treatment of STDs was offered at 6-months, not at baseline. Urine specimens were collected at baseline and 6-months. All urine samples were stored and tested by Ligase Chain Reaction for gonorrhea (GC) and chlamydia (CT). Standardized interviews assessing sexual risk behaviors were administered at baseline and at 6 months. Here we report on results from the Kenyan site (N = 1515). Results: Incidence rates for GC; 1.6% for men and 4.0% for women; for CT: 2.2% for men and 3.8% for women. Participants assigned to VCT were less likely to report having unprotected sexual intercourse with a non-primary partner (UI-NP: OR = .69, p < .025). Respondents who reported UI-NP at 6-months were 2.5 times more likely to have an incident STD at 6-months compared to those reporting no UI-NP (p < .01). After controlling for UI-NP at 6 months, those who reported UI-NP at baseline were nearly twice as likely than their less risky counterparts to have an incident STD (OR = 1.9, p < .025). Self-reported risk behavior at baseline and at 6 months were associated with incident STDs. These findings support self-reports as valid measures of sexual risk behavior.

[23126] Whose responsibility is it to stop the spread of HIV - Is it mine, is it his, it ours?
Michael Crosby. UCSF/Center for AIDS Prevention Studies, 74 New Montgomery, Ste. 502 San Francisco, CA 94105, USA

Objectives: Responsibility for the sexual transmission of HIV is assumed to be shared by both HIV positive and HIV negative men. However, since gay men are continuing to seroconvert, it is clear that some men are not taking that responsibility. The following is an exploration of the ways that both groups of men conceptualize responsibility for the spread of HIV. Methods: Between May and November 1997 we conducted 91 semi-structured, in-depth, interviews lasting 90-120 minutes with middle class African American (41%) Latino (25%) and Caucasian (34%) gay men in San Francisco all of whom engaged in anal intercourse in the previous six months. The centerpiece of the interview included narratives of anal sex during that time and included questions about responsibility for HIV transmission. Interviews were recorded, transcribed and coded for thematic content. Results: The following results are based on 97% of the sample who reported engaging in anal sex with someone other than their primary partner. Responsibility for HIV infection was percieved and experienced in the following three ways. 1) Individual Responsibility for condom use. HIV pos: "I always use a condom because I don't want to infect someone." HIV neg: "I always use a condom because I don't want to get sick and die." 2) Mutual Responsibility for serostatus disclosure and condom use. HIV pos: "We talked about serostatus and decided not to use a condom." HIV neg: "We talked about serostatus and decided it best to use a condom." 3) Abdicating Responsibility for condom use. HIV pos: "If he was negative he would have said something, I left it up to him." HIV neg: "If he was positive he would have said something, I left it up to him." Conclusions: Some men are taking responsibility for HIV while others are abdicating it leaving open the potential for new infections. The meanings for both groups are quite different. Prevention agencies must address these differences through multi-level strategies that could include community forums, media campaigns and workshops.

[23130] Improved antiRetroviral treatment does not affect sexual decision-making among the majority of men who have high risk sex with men
James Dilley1 W. Woods2 W. Mc Farland2 J. Sabatino1 J. Rinaldi1 B. Adler1 T. Lihatsh1. 1Univ CA San Fran AIDS HLTH PROJ; Box 0884 San Francisco CA; 2Center AIDS Prevention Studies, San Francisco CA, USA

Background: Recent advances in antiRetroviral (ARV) treatments of HIV infection have prompted concern that individuals at risk may reduce their commitment to safer sex practices. Methods: 93 men were recruited for a counseling intervention study when they phoned to schedule an anonymous HIV antibody test. Participants were non-IDU, had at least one previous negative HIV test, and had unprotected anal intercourse (UAI) with another male within the last 12 months. The entire sample completed a structured interview and a randomly selected subgroup (n = 47) was enrolled in a counseling intervention that focused on cognitions at their last incident of unprotected anal intercourse. Results:

+--------------------------------------------------------------------------+
|Table 1: Responses with regards to the availability of new ARV treatments.|
+---------------------------------------------------------------+----------+
| General Attitudes/Intensions (n = 93) | % | |
+-------------------------------------------------------+-------+----------+
|I am much less concerned about becoming HIV+. | 18 | |
+-------------------------------------------------------+------------------+
|I am much more willing to take a chance of getting | 8 | |
|infected when having sex. | | |
+-------------------------------------------------------+------------------+
|I have already taken a chance of getting infected. | 11 | |
+-------------------------------------------------------+------------------+
|I am less likely to get infected from a guy on the new | 8 | |
|treatments than from a guy who is not on them. | | |
+-------------------------------------------------------+------------------+
(continued table)
+--------------------------------------------------+-------+
| Cognitions at Last UAI (n =47) | % |
+--------------------------------------------------+-------+
|New treatments make HIV manageable. | 19 |
+--------------------------------------------------+-------+
|I knew his viral load was low. | 4 |
+--------------------------------------------------+-------+
|A guy on treatment is unlikely to infect me. | 7 |
+--------------------------------------------------+-------+
|I'll take new treatments to prevent my infection. | 17 |
+--------------------------------------------------+-------+

The men whose behavior had been affected by the new treatments did not report a greater number of partners, more past STD, less education, nor were they more likely to report a primary partner who was HIV+. Conclusions: The initial success of new ARV treatments has not reduced concern about infection or the perception of risk of infection among the majority of these men and most have not changed their sexual behavior as a result of recent treatment successes. However, recent treatment advances are currently effecting the sexual decision making for some men who have high-risk sex. Prevention efforts must address the changes in risk perception that result from improvement in the medical management of HIV disease.

[14209] Male STD patients in Mumbai, India are in urgent need of culturally specific and feasible AIDS prevention programs
Maria Ekstrand1 C. Lindan1 G. Bhave2 P. Gupte2 D. Javeri2 E. Hudes1 J. Mandel1. 1Ctr for AIDS Prevention Studies UCSF S-600 74 New Montgomery SF CA 94105, USA 2GS seth Medical School KEM Hospital Mumbai, India

Objectives: 1) To examine rates of unprotected sex and sociocultural issues among male STD clinic patients in Mumbai India; 2) To develop, implement and conduct a process evaluation of a culturally-specific AIDS/STD prevention intervention for this population. Methods: We interviewed 362 primarily low income male outpatients in a Mumbai municipal STD clinic. We used the interview data to develop and process evaluate a pilot AIDS prevention intervention Results: 32% of this sample was HIV positive, Prevalence of self-reported risk behaviors are as follows:

+--------------------------------------------------------+
|lifetime history of sex with sex workers: |95% |
+---------------------------------------------------+----+
|lifetime history of male-to-male sex |12% |
+---------------------------------------------------+----+
|frequent intoxication before sex worker visits: |54% |
+---------------------------------------------------+----+
|often/always use condom with sex worker: |8% |
+---------------------------------------------------+----+
|often/always use condom with wife/primary partner: |3% |
+---------------------------------------------------+----+

Virtually all men reported that it is very important for an Indian man to be able to have and take care of his family. Based on the interview data, we designed a culturally specific intervention that focused on risk reduction as a way of improving one's ability to get married, have children, and take care of one's family. This one-day, 6 hour program included a combination of lectures, multimedia, presentations, and interactive components. Misconceptions regarding transmission routes, effective prevention strategies, and fears associated with carrying condoms were addressed. An HIV infected patient described the impact of living with HIV on himself and his family. Participants discussed obstacles to condom use and practiced putting on condoms while blindfolded, since they reported that much sex happens quickly and in the dark. The participants rated each program component as 3 = "good", 2 = "neutral" or 1 = "not very good". Mean scores for the 6 components ranged from 2.6 to 2.9. Conclusions: Their high rates of HIV and unprotected sex put male STD outpatients in Mumbai in urgent need of culturally specific HIV prevention programs that can be easily incorporated into the current clinic structure. Presenting risk reduction as a way of achieving culturally valued goals appears to be acceptable to this population. Process data show that the pilot intervention was well received, It now needs to be tested in a rigorous clinical trial, using both behavioral and biological outcome indicators.

[23116] Increasing rates of unprotected anal intercourse among San Francisco gay men include high UAI rates with a partner of unknown or different serostatus
Maria Ekstrand R. Stall J. Paul D. Osmond T. Coates. UCSF Center for AIDS Prevention Studies S-600 74 New Montgomery SF 94105, USA

Objectives: 1) To examine the prevalence and individual patterns of sexual risk taking in a prospective longitudinal cohort of San Francisco gay men; 2) to measure the prevalence and correlates of unprotected anal intercourse (UAI) with partners of unknown of different serostatus in 1996. Methods: San Francisco gay men between 18-29 years were recruited using household sampling (n = 408) and participant referrals (n = 622) and followed annually. Behavioral and psychosocial measures were administered using a self-administered survey. In 1996, participants were asked whether any UAI had occurred with a partner of different or unknown serostatus. Results: Cross-sectional rates of UAI have increased significantly during the 4-year study period:

+----------------------------------+
| |1993|1994|1995|1996|
+--------------+----+----+----+----+
|Receptive UAS |31% |31% |37%*|41%*|
+--------------+----+----+----+----+
|Insertive UAS |25% |28% |35%*|39% |
+--------------+----+----+----+----+
|All UAS |37% |37% |45%*|50%*|
+--------------+----+----+----+----+
|*significantly (p < .05) greater |
|than in previous wave |
+----------------------------------+

An examination of individual behavior patterns shows that 51% of participants practiced unprotected anal sex occasionally (i.e. in 1, 2, or 3 years). In 1996, we examined in detail the conditions under which UAI occurred. 22% of the participants indicated that they had engaged in UAI with a partner of different or unknown serostatus, Compared to all other men, participants who reported UAI with a partner of an unknown or different serostatus also reported having more male sexual partners, more sex in bathhouses and sexclubs, more sex under the influence of alcohol, and were more likely to perceive UAI as "inevitable". Conclusions: The increasing rates of unprotected anal sex in this population can not be attributed solely to negotiated safety since about half of the men reporting UAI stated that this occurred with partners of unknown or different serostatus. This figure is especially alarming in light of the high rates of HIV prevalence among San Francisco gay men. Intervention programs targeting this population are urgently needed to avoid ongoing waves of new HIV infections in San Francisco.

[43116] Home collection HIV tests in the US: Those who intended to use are not using
Susan Fernyak1 K.A. Phillips1 B. Branson2 J. Catania1. 1UCSF/Center for AIDS Prevention Studies 74 New Montgomery, #510, SF CA 94105; 2Center for Disease Control Atlanta GA, USA

Objective: Prior to the Food and Drug Administration's approval of the first home collection HIV test (HCT) in 1996, several surveys of the US population indicated that an HCT would be widely used. To determine if barriers to its use exist, the numbers and characteristics of those who indicated their intention to use the HCT prior to its approval was compared to those who have actually used HCTs. Methods: National probability samples from the National Health Interview Survey ('92 NHIS), the Kaiser Family Foundation Survey on HIV Public Knowledge ('95 KFF) and the Family of AIDS Behavioral Surveys ('96 FABS) provided data on the characteristics of those who stated that they were "very likely" or "somewhat likely" to use HCTs. These data were compared to data on actual users, provided by the US Centers for Disease Control. Results: Prior to approval of the HCT, 29% of NHIS, 43% of KFF, and 31% of FABS respondents stated they were "very likely" or "somewhat likely" to use the HCT. However, during the first 12 months following approval, only 152,000 HCTs were used (0.6% of the population). Intended users were younger, non-white, less educated, had lower income, had tested previously (except for NHIS) and were at-risk for or concerned about HIV. Actual users were more likely to be male, white, heterosexual, 25-34 years of age, and not previously tested.

+----------------------------------------------------------------------------+
|Characteristics |"Likely" to use |Actual users (%) |
| Of users |(% of each identifier) | |
+--------------------+-----------------------+-------------------------------+
| |NHIS |FABS |KFF |US Centers for Disease Control |
+--------------------+-------+--------+------+-------------------------------+
|Male |30 |31 |42 |63 |
+--------------------+-------+--------+------+-------------------------------+
|Female |27 |31 |44 |37 |
+--------------------+-------+--------+------+-------------------------------+
|White |27 |26 |39 |84 |
+--------------------+-------+--------+------+-------------------------------+
|African-American |37 |43 |65 |5 |
+--------------------+-------+--------+------+-------------------------------+
|Latino |36 |53 |59 |5 |
+--------------------+-------+--------+------+-------------------------------+
|No prior test |75 |27 |40 |60 |
+--------------------+-------+--------+------+-------------------------------+

Conclusions: HCTs are being used predominantly by whites and first-time testers; prior to approval, African-Americans and Latinos indicated they were more likely to use the tests. In addition, actual use of HCTs has been much lower than expected. These data indicate there are barriers to use of HCTs, particularly among non-whites and those at high risk for HIV. Efforts should be made to identify and address these barriers.

[43275] Evidence suggests a social change-oriented approach to prevention with injecting drug users
David R. Gibson1 N.M. Flynn2 R. Anderson2 L. Clancy2 M. Acree3 T.J. Coates3. 1UCSF Center for AIDS Prevention Studies, 74 New Montgomery St. San Francisco, CA; 2University of California, Davis, Davis CA; 3University of California, San Francisco, San Francisco, CA, USA

Objectives: To examine associations between 1) peer norms and peer communication concerning safer injection/safer sex, and 2) drug users' injection-related and sexual risk. Peer norms and peer communication are hypothesized to be the mechanism by which a community intervention to prevent HIV produces community-level behavior change. The intervention is being evaluated in two West Coast US cities. Design: Cross-sectional survey of 604 injecting drug users. Methods: Drug users were administered a 12-item behavioral risk assessment and four scales, two measuring the extent to which peers are concerned with practicing safer injection and safer sex, and two the extent to which they communicate with peers regarding these issues. The behavioral assessment was used to classify drug users as being at low versus high risk of infection with HIV with regard to injection-related and sexual practices. Results: Norms and communication were measured on four- and three-point scales, respectively. Results: Norms and communication were highly predictive of injection risk, but less so of sexual risk. Counterintuitively, communication predicted higher sexual risk.

+-----------------------------------------------------------+
|Injection risk |Norms1 |Communication1 |
+-----------------------+---------------+-------------------+
|Simple odds ratio |0.26 |0.14 |
+-----------------------+---------------+-------------------+
|Adjusted odds ratio |0.38 |0.39 |
+-----------------------+---------------+-------------------+
|Sexual risk |Norms1 |Communication1 |
+-----------------------+---------------+-------------------+
|Simple odds ratio |0.78 |1.182 |
+-----------------------+---------------+-------------------+
|Adjusted odds ratio |0.67 |1.58 |
+-----------------------+---------------+-------------------+
|1 All 4 scales had alpha reliability >.8 2 Non-significant statistically (all other findings significant). |
+-----------------------------------------------------------------------------------------------------------------+

Conclusion: Given the strong association between injection norms and communication and behavior, social change-oriented interventions (e.g., community interventions) may be an especially effective approach to preventing HIV among injecting drug users. While peer norms for safer sex are associated with reduced sexual risk, peer communication may not support the practice of safer sex.

[14223] Culture counts: Understanding the context of unprotected sex for HIV positive men in a multi-ethnic urban sample in the US
Cynthia A. Gomez1 P. Halkitis2. 1UCSF Center for AIDS Prevention Studies 74 New Montgomery San Francisco Ca 94105; 2Cntr for AIDS/HIV Education & Training Jersey City NJ, USA

Background: HIV-seropositive (+) men who have sex with men (MSM) play a key role in preventing further transmission of HIV. Understanding ethnic differences is vital to intervention development for HIV+ men. Methods: A cross-sectional sample of 255 HIV+ MSM was recruited from San Francisco and New York City. Men participated in a semi-structured interview, and completed a survey exploring health status, sex and drug use practices, HIV status disclosure, and comfort with same-sex behaviors. For the present analyses differences were examined among African-Americans (AA) (28%), Latinos (L) (24%), and Whites (W) (31%). Results: Ethnic differences among HIV+ men were found for unprotected sex reported in the:

+--------------------------------------------------------------------------+
| | AA | L | W |p Value |
+-------------------------------------+-------+---------+---------+--------+
|Previous Year (N = 211): |(N=73) |(N=61) |(N=77) | |
+-------------------------------------+-------+---------+------------------+
| ï Anal insertive with HIV-negative |20% |35% |21% |.09 |
| (-) or Unknown status men (Ukn) | | | | |
+-------------------------------------+-------+---------+---------+--------+
| ï Anal receptive with HIV- or Ukn |26% |33% |35% | |
+-------------------------------------+-------+---------+------------------+
| ï Oral insertive with HIV- or Ukn |56% |67% |79% |.01 |
+-------------------------------------+-------+---------+---------+--------+
|Past 3 Months (% among men with |(N=17) |(N = 24) |(N = 30) | |
| HIV- partners, N = 71): | | | | |
+-------------------------------------+-------+---------+------------------+
| ï Anal insertive with HIV- men |29% |13% |0% |.009 |
+-------------------------------------+-------+---------+---------+--------+
|Past 3 Months (% among men with |(N=38) |(N = 39) |(N = 55) | |
| HIV Ukn partners, N = 132): | | | | |
+-------------------------------------+-------+---------+------------------+
|ï Anal insertive with HIV Ukn men |32% |31% |16% | |
+-------------------------------------+-------+---------+------------------+
White men reported a larger number of male sex partners (M = 19.3) in the past three months than did L (M = 10.7) or AA (M = 10.3, p < .05) men. AA were more likely to report sex with women (23% versus 7% L, 4% W, p = .001), were more likely to self-identify as bisexual (24% versus 8% L, 0% W, p < 001); and were less comfortable with their same sex behavior (M = 18.3) than were L (M = 19.8) and W (M = 20.4; p < .01) men. Interview data suggest that AA and L are less likely to consider oral sex as a substitute for penetrative sex. Conclusion: HIV+ L and AA men reported higher rates of UAI with HIV- or Ukn status partner than did W men. Ethnic differences warrant closer examination of cultural factors in determining unprotected sexual behaviors for HIV+ MSM with HIV- or Ukn status partners. Cultural messages regarding definitions of sex and same-sex behavior may contribute to sexual risk for some men. Interventions for HIV+ MSM of all ethnic backgrounds are critical and may significantly decrease the spread of HIV among HIV- MSM.

[24328] Being tested for HIV does not increase the incidence of negative life events in three developing countries: Results from the voluntary HIV counseling and testing study
Steve Gregorich1 Olga Grinstead1 M. Hogan3 S. Gregorich1 D. Balmer2 G. Sangiwa3 C. Furlonge4 T. Coates1. 1UCSF Center for AIDS Prevention studies 74 New Montgomery San Francisco, CA, USA 2Kenya Assn. of Professional Counselors, Nairobi, Kenya 3Muhimbili University College, Dares Salaam, Tanzania 4Voluntary Counseling & Testing Centre, Curepe, Trinidad WI

Objectives: To describe positive and negative life event outcomes among individual and couple participants in a randomized clinical trial testing the efficacy of voluntary HIV counseling and testing (VCT) in Dar es Salaam, Tanzania (N = 1427), Nairobi, Kenya (n = 1515) and Port of Spain, Trinidad (N = 1351). Methods: Study participants were randomized to receive VCT or standardized health information (HI) and to return for VCT in 6 months. Positive and negative life events were measured at the 6-month follow-up; occurrence of each life event was compared across treatment group, gender, enrollment status (couple versus individual) and baseline serostatus among those who received VCT at baseline. Results: The most commonly endorsed life events were positive: strengthening of a sexual relationship (39%, more common among couples) increased emotional support from peers (29%) and increased emotional support from family (23%). With the exception of break-up of a sexual relationship (22%), negative life events were uncommon (1%-4% overall). There were no significant differences between those who received VCT versus HI in reported levels of positive or negative life events. Among those assigned to VCT at baseline, seropositives were more likely than seronegatives to be estranged by peers (1% v. 2%, p < .01), discriminated against by employers (1% v. 2%, p = .06) and neglected by their family (2% v 4%, p < .05). HIV+ women who enrolled in a couple reported higher rates of physical abuse (18% versus 5% overall, NS) and break-up of marriage (15% versus 3% overall, NS). Women in female positive serodiscordant couples were the most at risk. There was no evidence that participating in VCT increased negative life events. With the exception of break-up of a sexual relationship, negative life event outcomes were rare in both treatment groups; ending a sexual relationship may be a risk reduction strategy rather than a negative outcome. Women who enrolled in couples are at increased risk for physical abuse and break-up of the marriage when they were in female positive discordant couples; additional services may be needed for women in relationships who test positive for HIV.

[14314] Sex is more than HIV risk: The many roles sex plays in the lives of young gay men
Robert B. Hays S.M. Kegeles T. Slama M. Chesney T.J. Coates. UCSF Center for AIDS Prevention Studies 74 New Montgomery San Francisco CA 94105, USA

Background: Interventions to change young gay men's sexual behavior must recognize the profound significance sex has for them and the personal context of their sexual acts. We identified the meanings sex has for young gay men and examined men's subjective interpretations of why they engaged in unsafe sex. Methods: 137 young gay men (age 18-27; 22% Black, 22% Latino, 36% White, 18% Asian), recruited from bars, streets, community events and public sex venues, participated in in-depth, semi-structured qualitative interviews which examined the personal meanings sex had for them and the roles sex played in their lives. They also described in detail their most recent safe and unsafe sexual interactions and discussed factors which influenced how the acts transpired. Interviews were transcribed and content-analyzed. Results: The multi-layered functions sex serves for young men were striking; 73% identified at least 3 different meanings sex had for them. The physical pleasures of sex were frequently cited (72%), but the interpersonal roles of sex were equally prominent (69%), i.e., serving as a way to express affection (46%), as a relationship-building tool - to get to know someone, make friends or test the potential for a relationship (33%), and as a way to please a partner (7%). Sex also served important psychological functions, i.e., alleviating loneliness (25%), boosting self-esteem/validating one's desirability (23%), releasing stress (12%), self-exploration (12%), gaining power over others (9%), sensation-seeking (6%), spirituality (4%), and fulfilling a social norm (5%). 7% acquired money or other material rewards from sex. 4% described sex as an addiction/compulsion. Frequently cited reasons for unsafe sex were: assuming the partner was "safe", i.e., HIV- or monogamous (39%), influence of drugs/alcohol (37%), physical enjoyment (24%), feeling invulnerable (22%), losing control in "heat of moment" (20%), non-assertiveness (16%), desire to please partner (12%), pressure from partner (10%), curiosity (10%), and no condoms available (10%). Most men knew their behavior was risky for HIV, but felt their personal needs had overwhelmed their health concerns. Conclusions: To be effective, HIV prevention programs must respect the complex, multi-determined nature of sex for young gay men and understand their personal meanings of sex. Helping men to identify their reasons for having sex in particular situations, explore alternative means of filling those needs and/or empower them to have sex safely may prove valuable. Focusing on the disease transmissibility of sex should not overshadow an appreciation of the important interpersonal and psychological roles sex serves for young gay men.

[43546] How would gay men decide whether or not to participate an HIV vaccine efficacy trial?
Robert B. Hays S.M. Kegeles B.W. Ralston T. Slama. UCSF Center for AIDS Prevention Studies 74 New Montgomery San Francisco CA 94105, USA

Objectives: Large numbers of gay men will need to volunteer for Phase 3 HIV vaccine trials in order for them to occur in the US. Given the history of distrust and ambivalence of gay men toward the government and biomedical institutions, and the physical and social risks of trial participation, the challenge of enlisting sufficient numbers of men is daunting. To help guide the design of trial recruitment strategies, we examined: 1) the processes gay men would go through in deciding whether to volunteer; 2) sources of information and advice they would seek; and 3) factors they perceived to be critical in influencing their decision. Method: Project LinCS is a qualitative study of attitudes toward vaccine trials/biomedical research. 39 HIV-gay/bisexual men (age 20-47; 23% Black, 21% Latino, 18% Asian, 36% White) in San Francisco participated in semi-structured, open-ended interviews about how they would make their decision about volunteering for an HIV vaccine trial, and whose opinions and what information would be most influential for them. Results: Decision-making styles varied but the modal pattern was to first gather "as much information as possible" about the vaccine and trial, discuss it with significant others and those with expertise (physicians, AIDS service organizations-ASOs) and very carefully weigh the perceived risks/benefits. They would closely attend to gay media, but also actively seek out information/advice from: friends (cited by 33%), trial staff (28%), ASOs, personal physicians and family (all 20%), other media, (10%), library (10%), health department (10%), health workers (8%), Internet (5%) and community forums (5%). Friends and family opinions would carry the most weight, followed by physicians and ASOs. Factors which would make men more likely to participate were: assurance of minimal physical harm; complete information about side effects; competent trial staff they "could relate to"; nondisruptive trial procedures; safeguards from social harms; endorsements from community leaders; and friends also participating. They would be most discouraged by: fears of side effects, doubts about vaccine's efficacy; culturally insensitive procedures; and patronizing or evasive staff. Conclusions: Volunteering for a vaccine trial is not a decision gay men would make lightly. Recruitment efforts must enlist the support and collaboration of ASOs and physicians and recognize the critical role of the men's social networks as forums for information-processing and support. The men want mutually respectful dialogue with trial staff including detailed presentations that "give all sides of the story" (particularly about side effects), with opportunity to air and discuss concerns and hear the opinions of trusted community leaders.

[14176] HIV+ MSM in discordant relationships report sexual risk-taking with primary and non-primary partners
Colleen Hoff B. Faigles C. Gomez R.B. Hays. UCSF Center for AIDS Prevention Studies, 74 New Montgomery St., San Francisco, CA 94105, USA

Objective: To describe differences in sexual risk-taking behavior with primary versus non-primary partners among HIV+ men who have sex with men (MSM). Method: A cross-sectional sample of HIV+ MSM (N = 255) was recruited from two US urban centers (New York and San Francisco). Participants were recruited from public sex environments, AIDS service organizations and gay venues. Men participated in a semi-structured interview and completed a self-administered survey. The interview included detailed narratives regarding specific sexual encounters. The survey explored a broad spectrum of variables including: sexual behavior (past 3 months), relationship issues, and health status. Men in primary relationships (N = 94) were stratified by their primary partners HIV status (+/+, N = 45 and +/-, N = 49) to describe sexual behavior with primary and non-primary partners. Results: HIV+ men in discordant relationships (+/-) were more likely to have unprotected anal intercourse (UAI) with an HIV- or HIV status unknown (HIV?) non-primary partner than HIV+ men in concordant relationships. Yet, men in HIV+ concordant relationships were more likely to have unprotected anal sex with their primary partners.

+------------------------------------------------------------+
| |Concordant |Discordant |
+------------------------------------+-----------+-----------+
|UAI with primary partner |56% |37% |
+------------------------------------+-----------+-----------+
|UAI with non-primary HIV-/? partner |15% |28% |
+------------------------------------+-----------+-----------+

Guilt free sexual gratification emerged as a theme in the interview data explaining why some men in discordant relationships engaged in UAI with non-primary partners. For example, a respondent said, "I go outside of my relationship to have sex, well, the way I like it." Another stopped using condoms with his primary partner saying, "...I felt like I was some sort of specimen or something, that I had to be covered in certain areas, to protect this person." Conclusion: Risk for HIV transmission from HIV+ men in discordant relationships to their primary and non-primary HIV-/? partners is substantial. Sex with HIV-/? partners is likely stressful for HIV+ men and may lead some men to rely on dangerous prevention strategies. Prevention programs that target couples must address risk with non-primary partners as well as risk within the relationship.

[60785] The production function is an economic tool that can inform the design of HIV prevention strategies
James G. Kahn1 Natalie D. Beltzer2. 1Center for AIDS Prevention Studies, UCSF, Inst. for Health Pol Stud., UCSF, Box 0936, San Francisco, CA 94143, USA 2INSERM U379, Marseille/CAPS, UCSF, Marsielle, France

Background: HIV prevention planning groups must design prevention strategies (ie, the mix of prevention programs) to maximize HIV infections averted when resources are limited. The production function (PF), an economic tool, describes how inputs (eg, labor) relate to outputs (eg, manufactured goods). We explored the use of PF to describe HIV prevention programs and to inform the design of HIV prevention strategies. Design/Methods: We reviewed effectiveness and economic evaluations of HIV prevention programs from the perspective of PF. For each of four HIV prevention interventions, we examined inputs (eg, labor and materials) and output dynamics (eg, how safer sex averts HIV infections). We developed PF curves to describe likely relationships between inputs (expressed as dollars) and outputs (HIV infections averted). Results: The key derminant of how quickly HIV infections are averted as spending rises is fixed cost (eg, recruiting volunteers). Marginal rates of return, estimated from epidemic models, determine curve shape. Maximum benefit is a function of client population size and HIV risk and prior exposure to prevention. Conclusion: Production function curves concisely portray the relationship between resources for HIV prevention and the intended output, HIV infections averted. To use production functions to inform the design of prevention strategies, these curves can be confirmed and refined with local cost, HIV risk, and effectiveness data. Then, a joint production function for multiple interventions could be calculated to show the relationship between overall spending and HIV infections averted.

[43547] How should large-scale HIV vaccine efficacy trials be conducted? Recommendations from US community-members likely to be targeted
Susan M. Kegeles1 R.P. Strauss2 D.S. Metzger3 T. Slama1 B.W. Ralston1 R.B. Hays1 K.M. MacQueen4. 1UCSF Center for AIDS Prevention Studies, 74 New Montgomery, San Francisco, CA 94105; 2University of North Carolina, Chapel Hill, NC; 3University of Pennsylvania, Phildelphia, PA; 4Centers for Disease Control and Prevent, Atlanta, GA, United States

Background: Individuals needed for Phase III HIV vaccine efficacy trials come from communities that historically have been marginalized and have strong distrust of the government and public health research. This study examines community members' attitudes about vaccine trials and biomedical research. Methods: In-depth semi-structured interviews were conducted with San Francisco gay men; Philadelphia injection drug users (IDUs); and African-Americans from Durham, N.C.; these US sites were selected because they represent communities likely to be targeted for recruitment into HIV vaccine efficacy trials. Respondents were asked how trials should be conducted so that the studies would be considered ethical, sensitive and results would be trustworthy. A thematic analysis of responses was conducted. Results: The sample (N = 91) was 78% male; 53% gay/bi; 29% IDUs; 48% African-American, 29% white, 9% Asian, 9% Latino. Although there are some differences among the communities in the rank order of the themes, the similarity across diverse groups of people is striking. Major themes that emerged were: * Scientists should be completely open about all aspects of the trial ("everything on the table" "expected rates of success" "how this candidate compares with others") (40% of respondents mentioned) * Participants provided diverse recommendations about who should participate in trials ("all people" "people at risk for HIV" "scientists" "homeless" "politicians" "not just us") (21%) * Disclosure to trial participants should go beyond minimum informed consent procedures (fully explain side effects, risks, potential discomfort - don't just list; explain for range of literacy levels) (19%) * Participants were very concerned about the potential for social, physical, or psychological harm (15%) * Participants used strongly stated negative invocations to researchers about conducting trials ("don't lie" "don't hide scandals" "don't use loopholes to hide information) (10%) * If harm occurs, trial participants stated that they should be compensated/cared for (9%) Conclusions: For successful implementation of HIV vaccine efficacy trials to occur, effective strategies for enhancing community collaboration must be developed. Strategies must include: work closely with communities, proactively disclose all aspects of trials, minimize all forms of harm, ensure compensation if harm occurs, be connected with and committed to the communities by having diverse research teams. Scientists, trial participants and communities must be partners in the research.

[43283] The whole is greater than the sum of its parts: An analysis of the intervention components of an HIV prevention program for young gay men
Susan M. Keegeles R.B. Hays C.R. Waldo L. Pollack. UCSF Center for AIDS Prevention Studies 74 New Montgomery, San Francisco CA, USA

Background: The Mpowerment Project has been shown to be effective among US young gay men. This study examines the effectiveness of various program components on post-intervention sexual risk-taking. Methods: Independent of the intervention, we recruited a longitudinal cohort of gay/bi men aged 18-27 in Eugene, OR and Santa Barbara, CA prior to intervention, and resurveyed them post-intervention (N = 172 from the 2 communities). Data were combined to increase statistical power since we found no significant differences between the communities on risk behavior, demographics, and psychosexual variables. Scores were computed to indicate how much respondents participated in each component. Participants self-selected which activities they participated in. Involvement in one activity did not preclude involvement in other activities. T-tests were used to compare component scores of men who did vs. did not engage in unprotected anal intercourse (UAI) with nonprimary partners in the previous 2 months. Results: There were no significant differences between men who did vs. did not engage in UAI at baseline in their subsequent participation in various program components. However, men who did vs. did not engage in UAI post-intervention varied significantly in participation in program components (shown below).

+----------------------------------------------------------------------------+
|**p< 01 |% participated |mean participation among men who did/ |
| |in 1+ |did not have UAI post-intervention |
+-----------------+------------------+---------------------------------------+
|*p< 05 t<.10 |component activity| |
+----------------------+-----------------------------------------------------+
| | |UAI |no UAI | t |effect size (ES) |
+----------------------+-------------+-----+-------+-------+-----------------+
|Core Group or |31% |.38 |.81 |2.2* |.40 |
| volunteer | | | | | |
+----------------------+-------------+-----+-------+-------+-----------------+
|small groups |33% |.12 |.36 |3.2** |.59 |
+----------------------+-------------+-----+-------+-------+-----------------+
|large social events |59% |.88 |1.64 |2.0* |.51 |
| (safer sex promoted) | | | | | |
+----------------------+-------------+-----+-------+-------+-----------------+
|outreach (formal & |77% |1.92 |2.50 |1.7t |.36 |
| informal) | | | | | |
+----------------------+-------------+-----+-------+-------+-----------------+
|total Mpowerment |82% |4.46 |6.95 |2.5** |.49 |
| Project activities | | | | | |
+----------------------+-------------+-----+-------+-------+-----------------+

Conclusions: The small groups had a large ES, but reached substantially fewer men than social events and outreach. Though not as powerful, the social events and outreach were critical to the effectiveness of the program as sources of recruitment to the small groups and as a means of reaching men disinclined to attend small groups. The effectiveness of program components are not independent; the synergy created by the whole program makes the net effect of the intervention activities greater than the sum of its parts.

[13438] An effective model AIDS prevention program for port workers in Santos, Brasil
R. Lacerda1 N. Gravato1 A. Mello2 E. Hudes3 R. Stall3 N. Hearst3. 1Nucleo de Educacao E Prevencao-DST/AIDS, Praca Rui Barbos A, 23-40 Andar, Santos, SP; 2ASPPE Programa Municipal AIDS, Santos, Brazil 3UCSF Center for AIDS Prevention Studies, San Francisco, CA, USA

Objectives: The port of Santos is one of the original epicenters of the AIDS epidemic in Brazil and is leading the trend towards an increasing caseload attributed to heterosexual transmission in that country. We undertook a worksite-based AIDS prevention program among port workers in Santos and measured its impact on heterosexual risk behavior for HIV infection. Methods: A representative sample of 226 male port workers was interviewed in 1994, 1995, and 1996 in a three-wave prospective cohort study. An intervention costing US $90,000 directed towards the entire community of 20,000 port workers was conducted between Waves 2 and 3 of data collection. The multifaceted intervention involved outreach workers, training of selected port workers as "multipliers," and distribution of educational materials and condoms. No organized AIDS control program had targeted this population previously. Results: Sexual risk behavior showed no sign of decline between Waves 1 and 2 (before the intervention) but decreased substantially between Waves 2 and 3 (after the intervention). This resulted from both a decrease in the number of non-primary partners and an increase in condom use. * not measured Wave 1

+-------------------------------------------------------------------------+
| |Wave1 |Wave2 |Wave3 | p value |
+----------------------------------------+------+------+------+-----------+
|mean # nonmonogamous partners |.58 |.58 |.41 |.01 |
+----------------------------------------+------+------+------+-----------+
|% with nonmonogamous partners |22.1% |24.9% |20.7% |.07 |
+----------------------------------------+------+------+------+-----------+
|% unprotected sex with nonmonogamous |16.1% |20.3% |11.1% |< .0001 |
| partners | | | | |
+----------------------------------------+------+------+------+-----------+
|% sex with "casual" partner |* |20.7% |12.0% |.002 |
+----------------------------------------+------+------+------+-----------+
|% unprotected sex with casual partner |* |12.0% |3.2% |< .0001 |
+----------------------------------------+------+------+------+-----------+
Conclusions: This study indicates that a worksite-based AIDS control program can produce marked behavior change at a modest cost. Furthermore, it demonstrates that it is feasible to measure the impact of such a program using a longitudinal design. We hope these results will encourage other prevention programs in similar populations elsewhere.

[13452] Truck drivers in Brazil: Prevalence of HIV and other STD'S, risk behavior, and potential for spread of infection
Regina Lacerda1 N. Gravato1 W. McFarland2 G. Rutherford2 K. Iskrant2 R. Stall2 N. Hearst2. 1Nucleo De Educacao E Prevencao-DST/AIDS Praca Rui Barbos A, 23-40 Andar, Santos, SP; 2UCSF Center for AIDS Prevention Studies San Francis

CURRENT RESEARCH

RESEARCH FINDINGS

PREVENTION TOOLKIT

CAPS presentations at the 1998 International AIDS Conference - Geneva, Switzerland

Oral presentations

Poster presentations