|
Activity Number:
|
363
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Wednesday, August 6, 2008 : 8:30 AM to 10:20 AM
|
|
Sponsor:
|
Biometrics Section
|
| Abstract - #301947 |
|
Title:
|
Bayesian Spatial Analysis of Sexual Behaviors Among HIV High-Risk Individuals in Sexual Networks
|
|
Author(s):
|
Robin Jeffries*+ and Robert E. Weiss and Pamina Gorbach and Steve Shoptaw
|
|
Companies:
|
University of California, Los Angeles and University of California, Los Angeles and BERG and University of California, Los Angeles
|
|
Address:
|
3220 Midvale Ave., School of Public Health, Los Angeles, CA, 90034,
|
|
Keywords:
|
autoregressive spatial prior ; hierarchical model ; random effects ; repeated measures
|
|
Abstract:
|
We develop Bayesian statistical methodology to analyze multivariate self-reported sexual behaviors in a sample of injecting drug users and men who have sex with men. Subjects were sampled by respondent driven sampling: current subjects recruit the next wave of subjects from among their sexual partners and people they use drugs with. This induces correlation between subjects and their recruits but provides information on the sexual and drug-using networks of people at risk for HIV. Risk behavior information is collected on up to three sex and three drug partners for each individual surveyed, giving correlated repeated measures data on subjects in a network that we model with a multivariate spatial random effects model. By jointly analyzing unprotected and protected anal intercourse, we can determine how drugs affect risky behaviors and how safe people are with different partner types.
|
