Boosting health provider performance with non-financial incentives: A cluster-randomized controlled trial in Tanzania

Background: Non-financial incentives are frequently used to improve performance among healthcare providers, capitalizing on mission-driven intrinsic pro-social motivation. However, the effectiveness of incentives varies across settings and may depend on whether they are provided privately or publicly. Using a cluster-randomized controlled trial (ClinicalTrials.gov: NCT05525533) among drug shopkeepers in Tanzania, we designed and evaluated the effectiveness of non-financial incentives in boosting provider performance. Methods: We developed a non-financial incentive that involved providing shopkeepers with monthly reports of aggregated customer feedback compiled from anonymous surveys from young women (15–24) to appeal to shopkeepers’ pro-social motivation for helping these customers. We randomized whether the feedback was provided privately (via a report) or publicly (displaying certificates of the customer feedback ratings). We estimated linear regression models on provider performance as measured by sales across different categories to young women from administrative point-of sales data over 12 months and estimated whether performance measures were correlated with shopkeepers’ pro-social motivation and concern for social image measured by surveys at Baseline. Results: Young women customers completed 9,108 anonymous surveys across 99 shops. Shops receiving non-financial incentives privately did not increase performance. However, shops receiving non-financial incentives publicly reported an increase in sales to young women customers (58%, 95% CI: 20%, 97%), most notably for sexual and reproductive health products (96%, 95% CI: 4%, 187%), specifically oral contraception (154%, 95% CI: 9%, 306%) and pregnancy tests (75%, 95% CI: 8%, 143%). Performance measures were correlated with concern for social image but not pro-social motivation at baseline. Conclusions: Publicly provided non-financial incentives increased performance among drug shopkeepers in Tanzania serving young women. Performance was strongest among those with higher concern for their social image at baseline, rather than those with stronger pro-social motivation. Future interventions using non-financial incentives to motivate healthcare providers should consider leveraging providers’ social image concerns to amplify the effectiveness of incentives.