Universal test and treat and the HIV epidemic in rural South Africa: a phase 4, open-label, community cluster randomized trial

Replication paper
Original publication: The Lancet
Original researchers: Collins C Iwuji MRCP, Joanna Orne-Gilemann PhD, Joseph Larmarange PhD, Eric Balestre MPH, Prof Rodolphe Thiebault PhD, Prof Frank Tanser PhD, Nonhlanhla PhD, Thembisa Makowa BA, Jaco Dryer NDip, Kobus Herbst Msc, Prof Nuala McGrath ScD, Prog Till Barningh
Replication researchers: Sridevi Prasad, Morgan Holmes
Current status: Replication Plan Published

The Original Study

The ANRS 12249 Treatment as Prevention trial (TasP) conducted by Iwuji et al. (2018) aimed to determine whether a test and treat program would be effective in reducing HIV incidence at the population level. The study, conducted in rural South Africa, was the first of four trials of its kind to report results (Havlir et al., 2019; Hayes et al., 2014; Moore et al., 2013). It aimed to treat the population by providing ART to all in a randomized setting, where HIV-positive individuals received ART no matter what their CD4 levels were. The control group received ART once their CD4 levels dropped to 350 or less at first, and 500 or less after January 2015 (Department of Health, 2014). The ANRS 12249 TasP found a null effect on HIV incidence rates at the end of the six-year testing period in an area with an estimated 30% HIV prevalence. A deeper examination of the results showed poor linkage to care outcomes and high in- and out-migration, which likely contributed to the lack of clear program effects.

The Replication

In this study, we propose to replicate Iwuji et al. (2018) (ANRS 12249). The study was the first of four cluster randomized trials aimed at understanding whether treatment as prevention (TasP) programs were effective at the population level, rather than at the individual level. The original authors show that TasP in rural South Africa was ineffective due to poor linkage of care and high in- and out-migration in the study area. This replication study will use the raw data to reproduce the results in the original Iwuji et al. (2018) study. Additionally, it will apply different empirical methods to test the null result by examining the change in CD4 guidelines in 2015, the effects of the high migration in the study area, the effects of proximity to the nearest highway, and by using survival modeling techniques to look at time to HIV incidence.