Predicating HIV risk behaviour secondary to substance use

Background: Strong associations exist between substance use and sexual risk-behaviours. However, less is known about why substance use is a risk factor for UAI (unprotected anal intercourse) for some gay/bisexual men and not others. It also remains unclear which drug types are more strongly associated with UAI. Beliefs about the effects of alcohol/drug use (‘expectancies’) and risk-taking personality traits have been hypothesised as key factors underlying both substance use and sexual risk-taking. This research was conducted to better understand the relative contributions of: (i) drug and alcohol use; (ii) substance-related expectancies (utilising the SEP-MSM); and (iii) risk-taking personality traits on UAI across various drug types (alcohol, cannabis, stimulants, amyl). No known studies have examined the relative contributions of these factors to predict UAI across the range of substances commonly used among gay/bisexual men, using expectancy measures that have been specifically developed for men who have sex with men (MSM). Method: A cross-sectional design was used involving a community sample of gay and bisexual men (n = 277) throughout Australia, recruited through GLBT (gay, lesbian, bisexual and transgender) media/networks. Participants completed online questionnaires regarding demographics, substance use, sexual activity (e.g. UAI), substance-related expectances and risk-taking personality traits. Discriminant function analyses (DFA) were conducted to determine the predictive utility of the model tested across substance types, and to determine which factors most uniquely contributed to predicting UAI in conjunction with substance use. Results: Models were predictive for alcohol and stimulants (not cannabis and amyl). Substance use and expectancies uniquely contributed to predicting UAI, while risk-taking personality traits did not. Significant models predicted 14–20% of the variance for UAI secondary to substance use; with a correct classification for UAI of 70%. Other modifiable risk factors (based on social cognitive theory) are likely to further contribute to the predictive model (e.g. self-efficacy) and should be examined in future research. Discussion: The current research fills an important niche to better substantiate hypotheses regarding the role of expectancies and substance use on risk-taking, and helps to better understand which drugs are associated with greater risk. The SEP-MSM showed significant utility in predicting UAI among MSM who may be at risk of sexual risk taking in conjunction with substance use patterns, particularly for alcohol and stimulant users. Future applications of the SEP-MSM in health promotion, clinical settings and research are likely to contribute to reducing harm associated with substance use among MSM (e.g. HIV transmission). Individual and community level interventions focussed on empowering MSM to engage in safer sexual practices whilst under the influence, challenging commonly held beliefs about substance use, and improving self-esteem and assertiveness skills are also likely to be beneficial