Abstract
Payment for performance (P4P) is not a uniform intervention, with programme effect dependent on several interconnected factors. In this study, a system dynamics model was developed to explore the pathways to improved outcomes and how changes in the design, implementation and context of a P4P programme affected maternal and child health (MCH) service delivery outcomes in Tanzania. A previously developed causal loop diagram of the programme effects was used to inform model development, with further data sources (including an impact evaluation of programme, health surveys, stakeholder feedback and relevant literature) used to build the model. A number of pathways were identified to improved services under P4P, with increased availability of drugs underpinning the content of care outcome (intermittent preventative treatment during ANC), which together with increased supervision, enhanced health worker motivation. This in turn increased perceived quality of care at the facility which improved the coverage of services outcome (facility-based deliveries), and with increased outreach, increased awareness of services also boosted demand. Minor delays in payment reduced provider purchasing power for medicines, with severe delays driving erosion of provider trust and motivation for programme participation. Allocating a larger share of funds for facility operations can enhance performance effects, particularly for those services that rely on efficient drug administration. Contextual factors including limited baseline provision of essential medications, lower community awareness of facility services and dispersed/distant populations reduced programme effect. This paper demonstrates the feasibility and the potential of such models to inform the design of effective health system interventions.