A Latent Class Analysis of the Pattern of Substance Use among Men Who Have Sex with Men and Its Relationship to HIV Status and Risk Behaviors

Purpose: Numerous studies have found an association between substance use among MSM and participation in sexual behaviors that increase their risk for HIV infection. Studies of this relationship, however, have been methodologically limited as to how substance use was modeled (e.g., as a single drug or as multiple drugs but with the effect of each drug modeled independently). Consequently, there is not yet a good understanding of the pattern of substance use among MSM and its relationship to sexual risk behaviors. This study sought to address these issues through modeling substance use as a latent class variable and then examining the relationship between substance use latent class and sexual risk behavior among MSM. Because of the empirically demonstrated association of HIV status (i.e., seropositive or seronegative) and increased risk for both substance use and risky sexual behavior, we also examined the effects of HIV status in the latent class model.

Methods: Data analyzed for this study were collected in Chicago for the Centers for Disease Control (CDC) National Behavioral Surveillance (NHBS) system, which tracks HIV risk behaviors among high-risk populations such as MSM. Participants were randomly selected using time-space sampling and interviewed face-to-face. The analytic sample for this study included 1,258 Chicago MSM, 18 years or older. The NHBS questionnaire items analyzed included: past-year frequency and use of 8 substances, self-reported HIV status, and past-year HIV risk behaviors such as unprotected anal intercourse with serodiscordant and/or multiple partners. Latent class analyses (LCA) were conducted with Mplus software and evaluated using standardized fit indices (e.g., Bayesian Information Criterion [BIC]).

Results: The best fitting LCA model was composed of four latent classes: a large (68%) “normative” class characterized by no substance use or infrequent use of alcohol or marijuana; a recreational/experimental use class (26%) characterized by occasional use of poppers, cocaine, and methamphetamine; a “party-drug” use class (4%) that used club drugs, ecstasy, poppers and methamphetamines on a weekly or greater basis; and a small class of heavy poly substance users (2%) characterized by regular use of marijuana, cocaine, and tranquilizers and at least moderate use of virtually all other substances. Sexual risk behaviors were strongly associated with latent class with MSM in the two most intense substance use classes having significantly higher rates of risk behaviors on all indicators. MSM who were seropositive had a greater likelihood of being in the party-drug use class and evidenced higher levels of sexual risk behaviors as well.

Implications: A small but epidemiologically important group of MSM regularly use multiple substances, particularly club drugs and methamphetamine. They coincidently and frequently engage in sexual behaviors that place them at high risk for HIV infection. Prevention and treatment of substance use among MSM need to consider how the use of various substances clusters in this population rather than targeting any single substance. The pattern of intensified substance use and sexual risk behaviors among some HIV positive MSM may also be an important transmission vector and a target for prevention and treatment activities.