What works - what does not work for pay-for-performance models for maternal health program- Experience from an operations research in Bangladesh

Introduction. Maternal health care suffers from sub-optimal performance, client extortion and low quality of care in Bangladesh. Supply side barriers with a synergistic effect on demand side constraints make services inaccessible for poor women living in rural, remote areas.
Theoretical focus. Inspired by Pay-for-performance (P4P) that ties performance with financial incentives, the government implemented two P4P models- Strategy-I incentivized both providers and clients, and Strategy-II only the providers.
Methods. With a quasi-experimental design, interviews with 2,124 exit clients and 473 providers were conducted; monthly service statistics, quarterly quality assurance visit measurements, and auditor’s validation reports were collected for evaluation.
Findings. Institutional deliveries and quality of care score were increased significantly. Overall client satisfaction was highest in strategy-I for providing free medicines and not requiring extra payment. Cost per service unit was US$8. Although overall client satisfaction increased, staff behavior was not improved. Targets could not be set for complications services. Facilities not having minimal staff tend to fail.
Conclusions. The study tested a team based incentive mechanism. Design adjustments are to be made to make providers accountable for their behaviors.