Characterizing network-based HIV testing interventions to guide HIV testing and contact tracing at STI clinics in Lilongwe, Malawi

Understanding heterogeneity across patients in effectiveness of network-based HIV testing interventions may optimize testing and contact tracing strategies, expediting linkage to therapy or prevention for contacts of persons with HIV (PWH). We analyzed data from a randomized controlled trial of a combination intervention comprising acute HIV testing, contract partner notification (cPN), and social contact referral conducted among PWH at 2 STI clinics in Lilongwe, Malawi, between 2015 and 2019. We used binomial regression to estimate the effect of the combination intervention vs. passive PN (pPN) on having any (1) contact, (2) newly HIV-diagnosed contact, and (3) HIV-negative contact present to the clinic, overall and by referring participant characteristics. We repeated analyses comparing cPN alone with pPN. The combination intervention effect on having any presenting contact was greater among referring women than men [prevalence difference (PD): 0.17 vs. 0.10] and among previously vs. newly HIV-diagnosed referring persons (PD: 0.20 vs. 0.11). Differences by sex and HIV diagnosis status were similar in cPN vs. pPN analyses. There were no notable differences in the intervention effect on newly HIV-diagnosed referrals by referring participant characteristics. Intervention impact on having HIV-negative presenting contacts was greater among younger vs. older referring persons and among those with >1 vs. ≤1 recent sex partner. Effect differences by age were similar for cPN vs. pPN. Our intervention package may be particularly efficacious in eliciting referrals from women and previously diagnosed persons. When the combination intervention is infeasible, cPN alone may be beneficial for these populations.