Document 2072
 DOCN  M94A2072
 TI    Constructing a sampling framework for homosexual gathering places as an
       alternative to convenience samples to determine risk behaviors in
       Tijuana, Mexico.
 DT    9412
 AU    Rangel-Gomez G; Izazola-Lioea JA; del Rio C; Bustamante J; El Colegio de
       la Frontera Norte, Mexico.
 SO    Int Conf AIDS. 1994 Aug 7-12;10(1):388 (abstract no. PD0161). Unique
       Identifier : AIDSLINE ICA10/94370503
 AB    OBJECTIVES. To describe a method to build a sampling framework of gay
       gathering places to conduct probability surveys instead of convenience
       samples, to determine populations at risk for HIV, and other risk
       estimators. METHODS. A bi-dimensional multi-step sampling framework for
       mobile populations was built. Each step had selection units of place,
       time (day and hour) that had a given probability of being selected.
       Weighing factors of the sampling units come from the probability to
       select: A) persons (gay and bisexual men), B) geographic units (bars)
       and C) time (day and hour). The main assumption is that gay and bisexual
       men are mobile units, moving in place and time according to the moment
       they visit gay gathering places (bars). This methodology has been
       previously used to determine a) an estimate of the number tourists in a
       resort area in Baja California, Mex., and 2) to estimate the number of
       migrants of Mexicans to the United States. 221 questionnaires were
       administered to determine risk behaviors and blood was drawn to
       determine HIV prevalence. RESULTS. The original model of analysis has
       been modified and the number of people attending gay gathering places in
       Tijuana will be finally estimated as an absolute number. The
       distribution of risk behaviors will be estimated from the relative
       frequencies of the sample estimates. There were additional problems to
       those found in the other two studies, for instance, there was a lot of
       movement from one bar to another in the same day and hour (some people
       visited more than 3 bars in one hour), initial non-responses were
       delayed and people looked for being interviewed later, and even if they
       were not in the selected time and place, etc. Sample estimates have not
       yet been translated to the population level. Some sample estimates are:
       18% reported to have used condoms always in the past year, 12% have
       never used it in that period. The overall prevalence rate of HIV
       infection is 10.5%, and there is a significant higher rate for those who
       did not used a condom in the past year (23%). CONCLUSIONS. This
       methodology allows the direct estimation of the size of the population
       under study: gay and bisexual men attending gathering places, with
       statistical representativeness.
 DE    Condoms  *Developing Countries  *Homosexuality  Human  HIV
       Infections/EPIDEMIOLOGY/*PREVENTION & CONTROL/TRANSMISSION  HIV
       Seroprevalence  Knowledge, Attitudes, Practice  Male
       Mexico/EPIDEMIOLOGY  Risk Factors  Sampling Studies  *Sex Behavior
       *Social Environment  MEETING ABSTRACT

       SOURCE: National Library of Medicine.  NOTICE: This material may be
       protected by Copyright Law (Title 17, U.S.Code).