Background
Financial incentives are known to help initiate routine healthcare behaviors. However, a more nuanced understanding of behavioral responses to incentives over time, particularly after they are removed, could guide more effective interventions. In our study, we explore the dynamic effectiveness of small, short-term monthly financial incentives on HIV care-seeking behaviors during and after incentive removal.
Methods
We conducted a 32-site cluster randomized controlled trial (NCT04201353) among adult ART initiates (<30 days) in four regions of Lake Zone, Tanzania. HIV care and treatment clinics (CTCs) were randomly assigned 1:1 to the intervention or the standard of care. Participants at intervention clinics had the opportunity to receive up to six monthly (>=25 days apart) cash transfers of 22,500 TZS each (~10 USD), conditional on appointment attendance. Using time-to-event analysis, we assess the effect of incentives on three HIV-care seeking behaviors (missed appointment by >4 days, <90% ART adherence, and ART interruption of >28 days) over the entire study period, during and after incentive delivery, and for each month of the follow-up period.
Results
Among 1990 participants (mean age: 37, 60% female, mean days on ART: 3.5) included in the final sample, we found large and significant reductions in rate of missed appointments (adjusted hazard ratio (aHR), 95% CI: 0.51, 0.39-0.68), <90% ART adherence (aHR, 95% CI: 0.58, 0.41-0.83), and ART interruption (>28 days without ART) (aHR, 95% CI: 0.54, 0.35-0.83) during the six months of incentives, but no effects after their removal. Hazard ratios by month reveal similar findings, with dramatic and consistent reductions in risk concentrated in months two through six that disappear in month seven, immediately after incentives are discontinued.
Conclusions
Our study provides a detailed assessment of dynamic incentive effects that are often hidden in cross-sectional evaluations of incentive interventions. In this setting, financial incentives significantly improved HIV care-seeking behaviors over standard services while they were delivered, but effects returned to levels in the standard of care group immediately after incentives were removed. These results highlight the value of short-term incentives for initiating new care-seeking behaviors, and emphasize a need for innovative design choices, such as escalating rewards or incorporation of time- or context-based cues, that may help motivate sustained behavior change after their removal.
Table or Figure
