Counselling behavioural interventions for HIV, STI and viral hepatitis among key populations: a systematic review of effectiveness, values and preferences, and cost studies

Introduction: Key populations (sex workers, men who have sex with men, people who inject drugs, people in prisons and other closed settings, and trans and gender diverse individuals) are disproportionately affected by HIV, sexually transmitted infections (STIs) and viral hepatitis (VH). Counselling behavioural interventions are widely used, but their impact on HIV/STI/VH acquisition is unclear.

Methods: To inform World Health Organization guidelines, we conducted a systematic review and meta-analysis of effectiveness, values and preferences, and cost studies about counselling behavioural interventions with key populations. We searched CINAHL, PsycINFO, PubMed and EMBASE for studies published between January 2010 and December 2022; screened abstracts; and extracted data in duplicate. The effectiveness review included randomized controlled trials (RCTs) with HIV/STI/VH incidence outcomes; secondary review outcomes of unprotected sex, needle/syringe sharing and mortality were captured if studies also included primary review outcomes. We assessed the risk of bias using the Cochrane Collaboration tool, generated pooled risk ratios through random effects meta-analysis and summarized findings in GRADE evidence profiles. Values and preferences and cost data were summarized descriptively.

Results: We identified nine effectiveness, two values and preferences, and two cost articles. Meta-analysis of six RCTs showed no statistically significant effect of counselling behavioural interventions on HIV incidence (1280 participants; combined risk ratio [RR]: 0.70, 95% confidence interval [CI]: 0.41-1.20) or STI incidence (3783 participants; RR: 0.99; 95% CI: 0.74-1.31). One RCT with 139 participants showed possible effects on hepatitis C virus incidence. There was no effect on secondary review outcomes of unprotected (condomless) sex (seven RCTs; 1811 participants; RR: 0.82, 95% CI: 0.66-1.02) and needle/syringe sharing (two RCTs; 564 participants; RR 0.72; 95% CI: 0.32-1.63). There was moderate certainty in the lack of effect across outcomes. Two values and preferences studies found that participants liked specific counselling behavioural interventions. Two cost studies found reasonable intervention costs.

Discussion: Evidence was limited and mostly on HIV, but showed no effect of counselling behavioural interventions on HIV/VH/STI incidence among key populations.

Conclusions: While there may be other benefits, the choice to provide counselling behavioural interventions for key populations should be made with an understanding of the potential limitations on incidence outcomes.