UC Berkeley

A key aim of health policy is to ensure that patients are able to access high quality care when appropriate while minimizing the over-use of unnecessary treatments. Achieving this balance requires aligned decision-making between patients and providers, which can be hindered if incentives are incompatible with this aim. I study how patient and pharmacist decision-making in the context of malaria case management is affected by financial incentives. I test whether incentives for rapid tests and high quality antimalarials (ACTs) only to malaria-positive cases targeted to patients (demand-side), providers (supply-side), or both (hybrid) improve malaria case management and align incentives towards socially-optimal antimalarial use. Using a cluster randomized trial in 140 pharmacies in malaria-endemic zones in Kenya, I randomize patient discounts and pharmacist performance incentives and compare their effectiveness and cost-effectiveness to the status quo standard of care. I find that both patient subsidies and pharmacy incentives for diagnostic testing significantly increase usage of testing and may encourage malaria positive individuals to purchase ACTs, and report results from unadjusted models here. Patient subsidies increase the likelihood that a symptomatic patient takes a rapid test by 25 percentage points (p=0.046), from a control group mean of 8 percent. Provider incentives and the hybrid approach increase the likelihood of rapid test uptake by 20 (p=0.008) and 25 (p<0.001) percentage points, respectively, which is indistinguishable from the demand-side approach. I also find that all arms improve ACT targeting by more than 200%: patient subsidies, pharmacy incentives, and the hybrid arm increase the likelihood that a patient purchases ACTs with a diagnostic test by 7 percentage points compared to a control group mean of 6 percent (p=0.244, 0.088, 0.032, for patient subsidies, pharmacy incentives, and hybrid arms, respectively). I find that both demand- and supply-side incentives reduce the likelihood that a patient purchases malaria treatment without a confirmatory diagnosis by between 17-19 percentage points, compared to 81 percent in the control group. Taken together, these results suggest that appropriately calibrated and targeted financial incentives are promising for improving malaria case management in a pharmacy setting. Additionally, the fact that supply- and demand-side incentives lead to comparable effects on diagnostic testing and treatment targeting suggests that both patients and pharmacists are effective channels for incentive targeting.