Introduction We explored patterns of sexual risk behaviour in a population of heterosexually-identified Peruvian men using latent class analysis, a technique that creates classes based on response patterns to questions related to a latent variable. In this study, the latent variable of interest is sexual risk behaviour.

Methods We used data from the Peru site of the National Institute of Mental Health (NIMH) Collaborative HIV/STD Prevention Trial to investigate patterns of high-risk behaviour among heterosexually-identified men (n=2109). Seven sexual risk behaviours were analysed in latent class analysis to group men into risk classes. These self-reported behaviours included using drugs in the past month, using alcohol or drugs prior to sex, having 3 or more sex partners in the past 6 months, having concurrent sex partnerships, exchanging sex for money, and having a male sex partner in the past 6 months.

Results Four latent classes of risk were identified, of which two classes had lower probabilities of these risk behaviours and two classes had higher risk classes probabilities of these risk behaviours. Increasing probability of risk behaviours by classes was significantly correlated with increasing reports of unprotected sex (p value for trend test <0.001). HIV/STI prevalence was elevated in all classes and prevalence increased with increasing probability of reporting the risk behaviours, although the differences were not significant. The risk behaviours in two of the latent classes identified were primarily related to alcohol and drug use.

Conclusions Risk behaviours in this population of heterosexually-identified men vary by latent class; however, given HIV/STI prevalence in each class all require prevention interventions. Interventions should also focus on alcohol and drug use, as these are important risk factors for this population. Future behavioural interventions for the prevention of HIV/STI infection may benefit from this information by tailoring messages to fit the observed patterns risk.