Bridging the Gap: Conference Record [Abstract book, International Conference on AIDS (12th: 1998: Geneva, Switzerland)]

12th World AIDS Conference

Abstracts 43424-43428 929

43424 Further decline in HIV prevalence rates in Uganda
Joshua Musinguz1, G. Asiimewe-Okiror2, A. Opio2. 1c/o STD/AIDS Control
Programme, Minestry of Health, Box 8, Entebbe Uganda; 2Epidemiologist,
STD/AIDS Control Programme, Uganda Entebbe, Uganda
Objectives: 1. To monitor the trend of the HIV epidemic in Uganda through
sentinel surveillance. 2. To monitor the magnitude of the HIV epidemic in Uganda.
Design: Serial analysis of HIV sentinel surveillance data from 1989 to date.
Methods: - Analysis of HIV trends in pregnant women from ANC sentinel
surveillance data collected from 1989-1997. - Analysis of ANC surveillance data
by age group comparing 1990-93 data with 1994-97.
Data Collection Methods: (1) From 1989 to date the STD/AIDS Control Programme has been collecting blood samples from antenatal mothers in 19 sites
and STD patients from one site all distributed geographically to cover all parts of
the country. (2) 250-300 samples of blood are collected from every site biannually using unlinked anonymous procedures and tested in a central laboratory at
Uganda Virus Research Institute (UVRI) for HIV. In 3 of the sites where second
generation sentinel surveillance has started 500-600 samples are collected.
Results: HIV prevalence rates have continued to decline in urban areas of
Uganda. In Kampala sites, rates have fallen from 15% in 1996 to 14.9% in 1997.
In Mbale, Eastern Uganda, rates continued to fall from 8.4% to 6.9% in the same
period.
Conclusions: - Rates in the rural areas seem to be on the decline as well.
Rates fell from 7.7% to 5.3%, 10.4% to 8.5%, 12.7% to 9.0% in Soroti, Kilembe
and Hoima respectively between 1996 and 1997. - Desegregated by age; the
young age group 15-19 years experienced the most consistent declines in all the
sites.
We have previously reported a decline in HIV prevalence in urban areas. This
decline has continued and we're beginning to observe similar trends in rural areas
[43425 The relationship between HIV incidence, prevalence,
mortality, migration and compliance in a population
cohort in Uganda
James Whitworth, A. Ruberantwari, A. Kamali. Medical Research Council
Programme On AIDS PO. Box 49, Entebre, Uganda
Objective: To examine the contribution of HIV prevalence and incidence, and
other factors, to the pool of seropositive individuals in a cohort in SW Uganda
followed over a 7 year period.
Methods: A population cohort of about 5000 adults (aged over 12 years) in
15 neighbouring villages has been followed by annual census and serosurveys
since 1989. We categorized all seropositive individuals at each survey into prevalent, incident, immigrant, not previously bled resident, and maturing child cases.
Losses of seropositive cases from a previous survey were categorized as dead,
emigrants, not bled at that survey round, or matured out of the age group.
Results: There were 1947 seropositive records out of a total of 24,020 records
(8.1%). For adults of all ages there were no significant trends in incidence, prevalence, migration or mortality, but there was substantial improvement in follow-up
rates. For young adults (aged 13-24 years) there was a significant reduction
in prevalence, but not incidence. On average prevalent cases made up 54% of
seropositive cases, incident cases 15%, immigrants 21%, residents not previously
bled 8%, and maturing children 2%. Losses each year averaged 5% of cases
died, 13% emigrated, 16% were not bled at the subsequent survey round, and
18% matured into another age-group.
Conclusions: We have demonstrated how HIV seroprevalence can fall without a concomitant fall in incidence. These data, collected over 7 years, are a
substantial contribution to understanding of HIV population dynamics in a mature epidemic. Retrospective and prospective projections of the Ugandan HIV
epidemic can be modelled using this information.
43426 Evaluation of a European project on monitoring HIV in
sentinel populations of sexually transmitted diseases
(STD) patients
Veronique Batter, J. Van Der Heyden, A. Sasse, A. Stroobant. The European
Study Group; 14 J. Wytsman street 1050 Brussels, Belgium, Scientific Institute
of Public health Brussels, Belgium
Objectives: To evaluate the feasibility and the outcomes of a collaborative European study as a prototype for a pan European surveillance system for HIV among
high risk populations.
Methods: This European study involved 19 surveillance networks of 16 countries. Sixteen networks used voluntary confidential HIV testing and 3 used unlinked
anonymous testing. Evaluation included the analysis of data on 157,476 STD
cases registered between 1990 and 1996, and a questionnaire survey of collaborators in all networks. Analyses included assessments of the stability of the
networks, completeness of reporting, stability of study population and potential
biases due to test refusals. Networks were compared with respect to clinic accessibility, range in STD diagnoses, counselling offered and data management

procedures.
Results: Some variations were observed in reporting and population characteristics over time in some networks, but the impact of the changes on the validity
of measured HIV trends was assessed as minimal. More important, was the high
proportion of persons with no HIV test results (24%) in the 16 networks with
voluntary HIV testing, either as a result of test refusal (12%) or failure to offer
a test (12%). Analysis of the data and results of the survey also showed that in

some networks data quality have been adversely affected by heavy workloads
and lack of extra resources. The survey indicated that in most countries this
project resulted in the initiation or improvement of STD and HIV surveillance
in important target groups. Many other benefits were reported and in fact most
participants indicated their wish to continue data collection beyond the end of the
funded collaborative project.
Conclusions: This surveillance system has proven to be easily implemented
in many networks in Europe, including Central Europe. Whilst overall the study
provided robust trend data, in several networks the reliability of observed trends
was questioned because of the high proportion of patients with no HIV test result.
This project has provided important national and European information in populations at greater risk for HIV infection and offers a prototype for future surveillance
of the HIV epidemic in high risk groups.
143427 Use of condoms at last sexual intercourse: An example
of advantages and pitfalls of standardizing questions in
surveys for HIV and STD risk behavior
Cornelis A. Rietmeijer, R.R. Fichtner. Behavioral Surv. Workgroup;
CDC-NCHSTP/OD Mail Stop E08, 1600 Clifton Road, Atlanta, GA 303333, USA
Objectives: There is general consensus that the value of surveys for the development and evaluation of HIV prevention programs would be enhanced by
standardization of questions on HIV risk behaviors. We evaluated the utility of a
standardized question on the use of condoms across different surveys and survey
methods.
Design: A comparative analysis of recent surveys in the US.
Methods: As a pilot evaluation, we chose condom use at last sexual intercourse as this is the only condom measure for which a national objective exists
in the US (by the year 2000, 50% of sexually active, unmarried people should
have used a condom at last sexual intercourse). This measure was identified as a
standard question on 4 national surveys. We contrasted this data with data from
a 5-city, community-level, survey among high-risk individuals, including injection
drug users (IDU), men who have sex with men (MSM), female sex traders (FST),
and youth in high-risk situations (HRY).
Results: In national surveys, condom use at last sex varied with age across
surveys (63% for youth in 9th grade, 50% for youth in 12th grade, 38% among
18-24 year-old college students, and from 37% at age 25-29 to 15% at age
45-49 in the general population). Condom use was higher when last sex occurred outside a primary relationship (62% vs. 19%), however only 1 national
survey made this distinction. Distinction by type of sex (vaginal, anal) was also
made by 1 national survey, but yielded limited data on anal intercourse. Among
community-recruited high-risk individuals, however, condom use at last sex varied
by risk group, type of sex, and type of partner: from 11% (IDU) to 43% (HRY) for
vaginal sex with a main partner; from 26% (IDU) to 70% (FST) for vaginal sex
with non-main partners; from 16% (FST) to 50% (MSM) for anal sex with main
partners; and from 26% (IDU) to 73%(MSM) for anal sex with non-main partners.
Conclusion: Standardization of questions, in theory, allows for direct comparisons across surveys. However, lack of specificity may limit the usefulness of
such comparisons. For condom use at last sex, specifying type of partner (main
vs. non-main) and type of sex (oral, vaginal, anal) may enhance the utility of this
question in making cross-survey inferences.
S43428 Reliability of self-reported unprotected sex among
partners in heterosexual HIV-discordant couples in
northern California is influenced by serostatus
Nancy S. Padian1, A. van der Straten2, J. Quan2, S. Glass2, C. Gomez2.
SUCSF, 505 Parnassus Ave. M-1485, San Francisco, CA; 2University of
California, San Francisco, CA, USA
Objective: To ascertain the reliability of self-reports of unprotected sex reported
among both partners in heterosexual HIV-discordant couples.
Design: Cross-sectional analyses of baseline data for a prospective intervention study
Methods: Eligibility criteria for entry into an intervention for HIV-discordant
couples included practice of unprotected penetrative (penile-vaginal or anal) sex
over the past six months. Each partner received a short interview. The last time
sex occurred was categorized as <6 months ago, >6 months ago, or never.
Results: Data from 98 couples (or 196 individuals), approximately half of the
anticipated sample, have been analyzed. The majority of the couples were African
American, between thirty to forty years of age, and approximately 20% had at
least one partner who was a current injection drug user. Overall, agreement as to
the last time penetrative contact with and without a condom occurred was good
(kappa = 0.7 for both). However, over a third of the couples disagreed about
always using condoms with each other:

Self-Reported Behavior With Partner
Always used condoms in the past month
Always had sex ever with a condom
Never used a condom in last 6 months
Never had sex with a condom

% of Couples Who Disagree
17%
35%
10%
25%

Disagreement did not differ depending on partner age, ethnicity, IDU status,
or number of other sexual partners. Couples with an infected female partner
were more likely to disagree (p = 0.02) regardless of the behavior (these women
reported more risk), and overall, infected individuals reported more recent unpro