Abstract
BACKGROUND: This study tests behavioral economics incentives to improve adherence to antiretroviral treatment (ART), with 1 approach being low cost. SETTING: Three hundred twenty-nine adults at Mildmay Hospital in Kampala, Uganda, on ART for at least 2 years and showing adherence problems received the intervention for about 15 months until the study was interrupted by a nation-wide COVID-19 lockdown. METHODS: We randomized participants into 1 of 3 (1:1:1) groups: usual care ("control" group; n = 109) or 1 of 2 intervention groups where eligibility for nonmonetary prizes was based on showing at least 90% electronically measured ART adherence ("adherence-linked" group, n = 111) or keeping clinic appointments as scheduled ("clinic-linked"; n = 109). After 12 months, participants could win a larger prize for consistently high adherence or viral suppression. Primary outcomes were mean adherence and viral suppression. Analysis was by intention-to-treat using linear regression. This trial is registered with ClinicalTrials.gov, NCT03494777 . RESULTS: Neither incentive arm increased adherence compared with the control; we estimate a 3.9 percentage point increase in "adherence-linked" arm [95% confidence interval (CI): -0.70 to 8.60 ( P = 0.10)] and 0.024 in the "clinic-linked" arm [95% CI: -0.02 to 0.07 ( P = 0.28)]. For the prespecified subgroup of those with initial low adherence, incentives increased adherence by 7.60 percentage points (95% CI: 0.01, 0.15; P = 0.04, "adherence-linked") and 5.60 percentage points (95% CI: -0.01, 0.12; P = 0.10, "clinic-linked"). We find no effects on clinic attendance or viral suppression. CONCLUSIONS: Incentives did not improve viral suppression or ART adherence overall but worked for the prespecified subgroup of those with initial low adherence. More effectively identifying those in need of adherence support will allow better targeting of this and other incentive interventions.