Does Criminal Justice Involvement Influence Subsequent Sexual HIV Risk for Substance Abusing Men?

Background:

Many studies have linked criminal justice involvement (CJI) and various sexual HIV risk behaviors. However, whether CJI contributes to a confluence of HIV risk factors is relatively unknown; most studies have utilized cross-sectional designs, which lack the ability to detangle associations or make causal inference. Some researchers have argued that CJI exacerbates the instability of social networks, economic vulnerability and stigma, which increases the likelihood of HIV and other health risk behaviors. Others suggest that CJI is a proxy for disproportionately poor and minority populations who tend to have poorer health outcomes, and that it is social and economic disadvantage that contributes to both issues.

This presentation builds on a previous cross-sectional study which found associations between CJI and several sexual HIV risk behaviors among men on methadone. To improve upon that study, longitudinal data from the same sample is used to establish temporality, and HIV risk behaviors and pretreatment confounders are measured with greater specificity. The study tests the temporal relationship between CJI (arrest and incarceration) for men on methadone and their subsequent sexual HIV risk behaviors, controlling for empirically and theoretically driven confounders. A better understanding of the role of CJI in HIV risk will inform targeted HIV prevention interventions with CJI populations as well as public health and criminal justice policy.

Methods:

A random sample of 356 men in methadone maintenance treatment programs in New York City was assessed at baseline (wave 1) and at 6 months (wave 2), and 12 month (wave 3) follow-ups. Propensity score matching, logistic and poisson regression were employed in a longitudinal design to isolate the effect of CJI (arrest and incarceration at wave 2) on subsequent HIV risk behaviors (wave 3). The propensity score matching process employed numerous individual and structural level variables at wave1 to control for pre-treatment variation. Multiple imputation was used to reduce the bias of missing data and differential attrition.

Results:

The sample was primarily Latino (45%) and African American (38%) with a mean age of 44. Of the five sexual HIV risk behaviors assessed at wave 3 (unprotected vaginal and anal sex, multiple partners, risky partners, and sex trading), the only behavior significantly associated with arrest at wave 2 was number of unprotected vaginal sex acts in the past 30 days (average increase of 4.5 unprotected acts, p<.05) and no sexual risk behaviors were associated with incarceration at wave 2.

Conclusions:

Past research on this sample demonstrates numerous sexual HIV risk behaviors associated with CJI. However, when isolating the effect of arrest and incarceration while controlling for multiple confounding variables in a longitudinal design, many of these associations are no longer significant. Results support a social selection hypothesis, whereby social and economic hardships associated with criminal involvement may also contribute to vulnerability to HIV. The significant finding of increased unprotected vaginal sex after arrest highlights the importance of HIV prevention intervention aimed at reentering and community-based criminal justice populations. Men with both substance abuse problems and CJI should remain a priority for HIV prevention interventions.