Does Past Exposure to Stressful Life Events Moderate Treatment Effectiveness of an Evidence Based STI Prevention Intervention for Justice-Involved Couples?

Background & Purpose: People under community supervision have disproportionately high rates of HIV/STIs and stressful life events (SLEs). SLEs in this population are also known to exist in latent classes, which have been linked to sexual behaviors associated with HIV/STIs. Protect and Connect (PACT), a 5-session HIV/STI prevention intervention, is one of the few evidence-based interventions for this population. While PACT was designed to reduce sexual risk by addressing unique stressors of males on community supervision, and their female partners, it is not known how past exposure to past SLEs may moderate effects. This study includes a nuanced examination of how latent classes of SLEs moderate treatment effects of PACT.

Methods: Data came from participants who completed a randomized clinical trial of PACT (2013-2016). The sample included substance-using males on community supervision in New York City and their main female sexual partners (n = 394 individuals within 197 couples). Participants were randomized to either the intervention group (PACT) or a control condition which included 1-session of counseling, testing and referral. Latent classes of SLEs included: Class-1 (low exposure), Class-2 (high exposure to street-based violence/deprivation), Class-3 (high exposure to household-based violence/deprivation) and Class-4 (high exposure). Measures of sexual health risk in the past 90 days were collected via self-report at baseline and 12-months follow-up. Moderation effects of SLEs on PACT effects were tested with multilevel Poisson regression models (for count-based outcomes) or logistic regression models (for dichotomous outcomes) both with random effects for couples. Models estimated moderation effects of SLE class on the overall intervention effect at 12-months using interaction terms between SLE classes and PACT exposure. Covariate adjustments were made for outcome measures at baseline, and gender, age, race, education, marital status and unemployment. Analysis was done in Stata 16.

Results: Main effects of SLE class on sexual health risk indicated that, as compared to individuals in Class-1 (low exposure to SLEs), individuals in Class-2 (high exposure to street-based violence/deprivation) had significantly higher rates of condomless intercourse (aIRR = 2.25) and discussed condoms significantly less; individuals in Class-3 (high exposure to household-based violence/deprivation) discussed condoms and HIV significantly less (aIRR = 0.33 & 0.10). As compared to same class individuals in the control condition, PACT was significantly more effective than the comparison arm for individuals with certain classes of SLEs (p < 0.05). PACT participants in: Class-1 had lower numbers of sexual partners (aIRR = 0.72); Class-2 had lower rates of condomless sex (aIRR = 0.40); Class-3 had lower rates of condomless sex (aIRR = 0.43), lower numbers of sexual partners (aIRR = 0.65), greater condom use self-efficacy (aIRR = 1.42), and discussed HIV more frequently (aIRR = 6.33).

Conclusions: The reversal of the direction of the main effects for certain classes of SLEs across multiple measures indicates that PACT was more effective than comparison treatment. This provides evidence that PACT addresses both active stressors related to community supervision and past SLEs. Research provides evidence for how different classes of SLEs differentially impact sexual health, and response to interventions.