Identifying Syndemic Factors and Their Effect on HIV Sexual Risk Behavior Among Gay, Bisexual and Other Men Who Have Sex with Men in Vancouver, Montreal and Toronto

Abstract

We set out to identify syndemic factors associated with HIV sexual risk behavior among HIV-negative gay, bisexual and other men who have sex with men (GBM), and examine associations between syndemic factors and HIV sexual risk behaviors using models of co-occurring epidemics. We used baseline data from 1571 sexually active GBM from the Engage Cohort study, recruited using respondent-driven sampling in Montreal, Toronto, and Vancouver. We explored associations between potential syndemic factors and HIV sexual risk behavior, and used oblique component analysis to reduce the number of variables. Path analysis and logistic regression with interaction terms were used to evaluate the association of syndemic factors with HIV sexual risk behavior. In the mediation model, childhood sexual abuse (CSA) was associated with HIV sexual risk behavior (total effect: c = 0.14, 95% CI 0.06-0.21). This effect was mediated by any intimate partner violence (IPV), transactional sex, polysubstance use, and sexual compulsivity. We observed a multiplicative interaction between CSA and transactional sex: GBM with a history of both had 14 times the odds of HIV sexual risk behavior (aOR of 14.0 95% CI 5.6-35.2) compared to those with neither. The appropriate application of a syndemic approach allowed a nuanced assessment of syndemic factor associations with HIV sexual risk behavior. IPV, transactional sex, sexual compulsivity, and polysubstance use mediated the effect of CSA on HIV sexual risk behavior. Additionally, we found a significant multiplicative interaction between CSA and transactional sex.

Keywords: HIV sexual risk behavior; Interactions; Mediation; Men who have sex with men; Syndemics.

Fig. 1

Framework of the proposed analysis

Fig. 2

Significant associations between syndemic factors and sociodemographic characteristics (as indicated by a significant Somers’ D with p-values < 0.05) (n = 1458). The thickness of the line is proportional to the absolute value of Somers’ D. ADD/ADHD Attention deficit disorders/attention deficit and hyperactivity disorder; CSA childhood sexual abuse, FSI financial strain index, IPV any intimate partner violence, IPVP intimate partner violence perpetration, IPVV intimate partner violence victimization, PSTD Post-traumatic stress disorder, RMBAS Revised male body attitudes scale

Fig. 3

Significant associations between syndemic factors and sociodemographic characteristics (as indicated by a significant Somers’ D with p-values < 0.05) (n = 1458). The thickness of the line is proportional to the absolute value of Somers’ D. CSA childhood sexual abuse, IPV any intimate partner violence, IPVP intimate partner violence perpetration, IPVV intimate partner violence victimization

Fig. 4

Significant associations between syndemic factors and sociodemographic characteristics (as indicated by a significant Somers’ D with p-values < 0.05) (n = 1458). The thickness of the line is proportional to the absolute value of Somers’ D. CSA childhood sexual abuse, IPV any intimate partner violence

Fig. 5

Serially causal epidemics model using mediation analysis [Standardised estimates (95% CI)]. * p-value < 0.05. Note: a, effect of exposure on mediator; b, effect of mediator on outcome; c, total effect of exposure on outcome; c’, direct effect of exposure on outcome. The model was adjusted by city, age, sexual orientation, ethnic group, education, marital status, income and RDS weights