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

XIV International AIDS Conference

Abstracts WePeC6183-WePeC6186 127

Results: Men between 30 and 39 years of age maintained the highest rates
during the period studied, with a rising incidence in the last decade (63.1 per
100,000 men in 1990 and 121.75 in 1996), maintaining high rates in 1997 and
1998 (112.50 and 111.86 respectively). Cases in individuals between 20 and 29
have decreased (from 80.28 in 1992 to 53.96 in 1998). No major differences were
observed in rates during the 90s. An increasing rate trend was observed for females, from 22.42 cases per 100,000 in 1994 to 46.99 in 1998, in the 30 to 39 age
group. Rates for women between 40 and 49 increased significantly (4.40 in 1990
and 23.63 in 1998). Among heterosexual women there was a decrease in the
proportion of cases in the 20-29 year old group after 1993 and an increase in the
30-39 group. For heterosexual men, the major case ratio remained in the 30-39
year olds during the whole period. The case ratio in injecting drug user men (UDI)
in 20-29 year olds prevailed until 1995 (48.06%), rapidly decreasing after 1992.
The 30-39 year-old group became the most frequent after 1996. The ratio of UDI
women between 20 and 29 years was also the most frequent until 1996, when
it was surpassed by older women (30-39). Among men who have sex with men
(MSM), the prevailing age during the whole period was 30 to 39 year olds and
there was a decreasing trend among cases diagnosed among 20-29 year-olds.
Conclusion: Based on the data obtained, the initial proposition has not been
confirmed. On the contrary, an increase in the rates of older age groups was
observed.
Presenting author: Angela Tayra, R Santa Cruz 81- Vigilancia Epidemiol6gica,
Vila Mariana, S.o Paulo- SP, CEP 04121 -000, Brazil, Tel.: [phone], Fax:
[phone], E-mail: [email protected]
WePeC6183 A decade of aids-related knowledge, attitudes, and
beliefs in the United States-BRFSS, 1990 - 2000
A. Lansky, RL. Fleming, D. Holtzman, R.L. Frey. CDC, 1600 Clifton Road,
Mai/stop e-47, Atlanta, GA 30333, United States
Background: General population surveys complement behavioral surveillance
targeted to high risk groups to inform for HIV prevention activities. With low prevalence of risk among the U.S. general population, these surveys also are important
for monitoring HIV-related knowledge, attitudes, and beliefs.
Methods: We analyzed data from the Behavioral Risk Factor Surveillance System
(BRFSS), a state-specific, population-based random telephone survey of adults
aged 18 - 64 years. We examined three categories of questions: AIDS Knowledge (1990 - 1992), AIDS Stigma (1990 - 1994), and AIDS Education for Children (1990 - 2000). From weighted data aggregated across states, we calculated
annual prevalence estimates.
Results: Knowledge increased over time: the proportion reporting that persons
infected with HIV can feel well and healthy increased from 72.4% to 83.0%; similarly, knowledge of the availability of drugs to treat AIDS increased from 51.6%
to 57.9%. Stigma and Education for children did not change: questions indicated
that many respondents would not be willing to work near a person with AIDS
(annual prevalence estimates ranged from 13.9% to 18.2%), and a high proportion would not eat in a restaurant where the cook had AIDS (range: 57.0%
to 64.5%). About half believed that AIDS education should begin in elementary
school (range: 47.4% to 50.2%).
Conclusions: Knowledge about AIDS increased in the early 1990s, but stigmatizing beliefs about people with AIDS persisted. Many changes in the epidemic
have occurred since the mid-1990s when these data were collected; thus, they
are now useful as a baseline for future assessments. Items on general population
surveys should be rotated periodically to assess public opinion on current developments in the epidemic, yet repeated at intervals to be able to monitor trends.
Presenting author: Amy Lansky, 1600 Clifton Road, Mailstop e-47, Atlanta, GA
30333, United States, Tel.: [phone], Fax: [phone], E-mail: alansky
@cdc.gov
WePeC6184 China's first HIV/AIDS behavioral surveillance
study of sex workers, gatekeepers, men and
condom outlets
V. Yuan1, K. Hardee2, T. Manchester3. 1Horizon Research Group, Beijing,
China; 2 Futures Group International, Washington DC, United States; 3The
Futures Group Europe
Background: China has conservatively estimated 800,000 people infected with
HIV. Contaminated blood, unsafe injections, drug use, sex work and migration
are considered principal transmission avenues, but little quantitative behavioral
information exists on at-risk groups, including sex workers/clients. Prevention and
care Interventions are based on scant behavioral data. Patterns of condom use,
crucial for helping to stop the spread of HIV/AIDS, are little understood.
Methods: As part of a DFID-funded China UK HIV/AIDS Prevention and Care
Project, Futures Group Europe worked with Horizons Research Group to conduct a modified behavioral surveillance survey (BSS) (China's first) in Yunnan
and Sichuan provinces in late 2001. Samples included 830 sex workers, 101 sex

work establishment owners or managers, 111 condom sales outlets, and 817 men
ages 18-55. Data collection took place despite attempts by authorities to suppress
commercial sex activities.
Results: Date collection is complete. Currently under analysis (to be completed
early February 2002), the data will provide China's AIDS Programme with information on patterns of sex (paying and non-paying), condom use, STDs and
treatment seeking behavior, knowledge and where people get information, opin
ions and attitudes regarding STD/HIV/AIDS, exposure to interventions, stigma
and discrimination, drug use, and gender and power relations among sex workers, gatekeepers and clients.
Conclusions: Quantitative data collection on sex work in China was thought to
be impossible; therefore the few existing interventions have been based on conjecture. This BSS develops a portrait of an established and growing sex industry
and a growing HIV/AIDS epidemic, expanding within an ancient culture. While the
sex industry in China has many similarities to other countries, this survey provides
many remarkable insights that will provide critical information for future targetted
interventions.
Presenting author: Karen Hardee, 1050 17th St NW, Suite 1000, Washington,
DC, 20036, United States, Tel.: [phone], Fax: [phone], Email: [email protected]
WePeC6185 HIV infection among agricultural employees in Rift
Valley Province, Kenya: prevalence, risk factors,
and suitability for vaccine efficacy trials
W.B. Sateren1, P.O. Renzullo2, C. Mason3, M. Wasunna4, M. Robb',
L. Langat5, L. Elson5.1 Walter Reed Army Institute of Research, Walter Reed
Army Institute of Research, 1 Taft Court, Suite 250, Rockville, MD 20850, United
States; 2Henry M. Jackson Foundation, Rockville, MD, United States; 3Armed
Forces Research Institute for Medical Sciences, Bangkok, Thailand; 4Kenyan
Medical Research Institute KEMRI, Nairobi, Kenya; 5Henry M. Jackson
Foundation, Kericho, Kenya
Background: The HIV pandemic in Africa requires a range of interventions. One
such intervention is a preventive vaccine. In order to test these in phase III trials, the epidemiology of the epidemic must be established in potential trial sites.
A pilot cross-sectional seroprevalence survey was conducted in an agricultural
company situated near a major national highway in the Rift Valley Province of
Kenya. The company supports a population of 70,000 employees and dependents belonging to four major tribes.
Methods: After providing informed consent, volunteers completed a risk behavior
questionnaire and provided a blood sample. Blood samples were tested with two
rapid tests. Positive and discrepant samples were confirmed by ELISA. Results
were confidential and not disclosed to the employer.
Results: A total of 845 individuals volunteered. HIV status was determined for
821. Of these, 81 (9.9%) were positive. HIV prevalence varied by: gender (females 17.4%, males 8.0%); age (17-19 year olds 3.5% increasing with age and
highest among 40 to 44 year olds, 16.7%); marital status (single 5.0%, married
11.8%, widowed 40%); and education (none 25.0%, primary 10.5%, secondary
8.5%, post-secondary 2.9%). Luo tribe members were found to have a significantly higher HIV prevalence (24.8%) than Kalenjin tribe members (4.5%) (OR =
7.04, 95% CI 3.62 - 13.7). These tribes reported different rates of male circumcision (Luo 18.6%, Kalenjin 99%). Overall, circumcised men were significantly less
likely to be HIV+ (OR = 0.15, 95% CI 0.08 - 0.28). Adjusting the tribal OR for
circumcision rates significantly reduced the differences between the tribes, OR =
2.90, 95% CI 0.81 - 10.4. Over 97% of cohort members expressed a willingness
to participate in a future HIV vaccine trial.
Conclusions: This study provides valuable information for the characterization
of the HIV epidemic in a population which may be suitable for phase III vaccine
efficacy trials in Africa.
Presenting author: Warren Sateren, Walter Reed Army Institute of Research, 1
Taft Court, Suite 250, Rockville, MD 20850, United States, Tel.: +[phone],
Fax: +[phone], E-mail: [email protected]
WePeC61 86 Indexing risk for HIV: Maximizing the effectiveness
of behavioral surveillance data
D.L. Sykes, S.R. Truax, H. Rasmussen. California Office of AIDS, Office of AIDS,
RO. Box 942732, Sacramento, CA 94234, United States
Background: The proliferation of behavioral surveillance in HIV prevention has
resulted in an abundance of data about HIV risk behaviors. The resulting challenge is to compile the data in a meaningful and informative way. This presentation delineates a method of devising a "risk index" based on behaviors of survey
participants and discusses benefits and challenges to such an approach.
Methods: A method for calculating a Behavioral Risk Index (BRI) was developed
using a modified Bernoulli model to combine data relevant to HIV exposure and
transmission. Using this model, risk of HIV exposure is multiplicatively combined
with risk of transmission for each behavior and partner type to create a realistic index of risk for each participant. Exposure probabilities are estimated based
on HIV prevalence for specified partner types, and transmission rates are based
on estimates of per-contact transmission rates for specific behaviors. Combining
these factors using a probability model results in an estimate of the likelihood of
infection, limited only by data collected and the current state of knowledge regarding prevalence and transmission probabilities.

Results: The BRI provides a method for comparing levels of risk among individual participants or groups, and it allows for a more complete examination of the
factors associated with different categories of risk behaviors. For example, in a
data set delineating behavioral risks for HIV in a sample of sex workers, the BRI
was used to demonstrate that although sexual behaviors with risky main partners
contributed significantly to HIV risk in this group; sex workers' estimation of risk
was not affected by their behaviors with these partners.