The Impact of 6-Month ART Dispensing (6MMD) on Retention in Malawi's HIV Program: A Target Trial Emulation Study

Background: Six multi-month dispensing (6MMD) of antiretroviral therapy (ART) for HIV treatment clients has expanded rapidly in the past decade, but its effect on individual outcomes in routine (non-trial) care is still poorly documented and based on observational programmatic data. Malawi launched 6MMD in April 2019 and scaled-up implementation in 2020. We compared retention in care for clients who received 6MMD to those who did not using a target trial emulation (TTE) approach to minimize bias. Methods: We used routine clinical data from Malawi's Electronic Medical Record (EMR) system to identify ART clients eligible for 6MMD in 27 districts from 01/2020- 12/2021. Eligible participants were non-pregnant adults (≥18 years), on ART for ≥6 months, clinically stable as evidenced by a dispensing duration of 3 months (3MMD), and with no prior 6MMD exposure. We created four six-month trials, defined eligibility at the start of each trial period, and classified participants as either receiving 6MMD or non-6MMD (dispensing duration of 1-3 months) within the six-month interval. Follow-up started at 6MMD enrollment for the 6MMD arm or the first visit in the trial enrollment period for the non-6MMD arm. Retention at 12 and 24 months was defined as having a clinic visit within 12-24 (trial 1-4) and 24-36 (trial 1-2) months from trial enrollment. Using an intention-to-treat approach, we estimated adjusted risk differences (aRD) with 95% confidence intervals (CI) using a Poisson regression model with an identity link function and robust standard errors adjusting for age, sex, duration on ART, facility type, regional location, WHO clinical stage at ART initiation. Pooled RDs were estimated by accounting for within-subject variation in a Poisson regression model using data from all trials. Results: Of the 159,801 unique patients eligible for this study (65% female, median age 37 years), 74% (118,910) were ever enrolled in 6MMD. Retention rates at 12 months (trials 1-4) and 24 months (trials 1-2) were consistently higher in the 6MMD group than the non-6MMD group. The pooled risk for retention was 3% higher in the 6MMD vs non-6MMD groups (aRD 3.0%; 95% CI: 2.8%-3.3%) at 12 months and 2.0% higher (aRD: 2.0%; 95% CI: 1.7%-2.4%) at 24 months. Conclusions: We observed slightly higher retention in care rates in Malawi at 12 and 24 months among patients on 6MMD compared to those receiving shorter medication dispensing intervals. Future work to assess the impact of 6MMD on visit burden and resource use would offer a comprehensive view of the benefits to both ART clients and the health system.