Short-term effectiveness of a community health worker intervention for HIV-infected pregnant women in Tanzania to improve treatment adherence and retention in care: A cluster-randomized trial

Introduction Community health workers (CHWs) are lay workers who have the potential to enhance services to prevent mother-to-child HIV transmission (PMTCT) and improve the health of women living with HIV infection. We conducted a cluster-randomized trial of an intervention to integrate CHWs with 'Option B+' PMTCT services in Shinyanga Region, Tanzania. Methods The intervention was implemented for 11 months and included four integrated components: 1) formal linkage of CHWs to health facilities; 2) CHW-led antiretroviral therapy (ART) adherence counseling; 3) loss to follow-up tracing by CHWs; and 4) distribution of Action Birth Cards (ABCs), a birth planning tool. We cluster-randomized 32 facilities offering PMTCT services, within strata of size, to the intervention (n = 15) or comparison (standard of care, n = 17) groups. Intervention effectiveness was determined with a difference-indifferences strategy based on clinical and pharmacy data from HIV-infected postpartum women at baseline (births in 2014) and endline (births April-Oct 2015). The primary outcome was retention in care between 60 and 120 days postpartum. Secondary outcomes included ART initiation, timing of ART initiation (as measured by week of gestation), and ART adherence 90 days postpartum, measured using the medication possession ratio (MPR >= 95%). Results Intervention and comparison facilities were similar at baseline. Data were collected from 1,152 and 678 mother-infant pairs at baseline and endline, respectively. There were no significant differences in retention in care, ART initiation, or timing of ART initiation between the intervention and control groups. Adherence (MPR >= 95%) at 90 days postpartum was 11.3 percentage points higher in the intervention group in ITT analyses (95% CI: -0.7, 23.3, p = 0.06), though this effect was attenuated after adjusting for baseline imbalance (9.5 percentage points, 95% CI: -2.9, 22.0, p = 0.13). Among only sites that had the greatest fidelity to the intervention, however, we found a stronger effect on adherence (13.6 percentage points, 95% CI: 2.5, 24.6). Conclusions Despite being feasible and acceptable, the CHW-based intervention did not have strong effects on most PMTCT indicators. CHW involvement in PMTCT programs may improve ART adherence among HIV-infected postpartum women, however, and success appears heavily dependent on program implementation.