Background
Assuring equitable universal access to essential health services without exposure to undue financial hardship requires adequate resource mobilization, efficient use of resources, and attention to quality and responsiveness of services. The way providers are paid is a critical part of this process because it can create incentives and patterns of behaviour related to supply. The objective of this work was to describe provider behaviour related to supply of health services to insured clients in Ghana and the influence of provider payment methods on incentives and behaviour.

Methods
A mixed methods study involving grey and published literature reviews, as well as health management information system and primary data collection and analysis was used. Primary data collection involved in-depth interviews, observations of time spent obtaining service, prescription analysis, and exit interviews with clients. Qualitative data was analysed manually to draw out themes, commonalities, and contrasts. Quantitative data was analysed in Excel and Stata. Causal loop and cause tree diagrams were used to develop a qualitative explanatory model of provider supply incentives and behaviour related to payment method in context.

Results
There are multiple provider payment methods in the Ghanaian health system. National Health Insurance provider payment methods are the most recent additions. At the time of the study, the methods used nationwide were the Ghana Diagnostic Related Groupings payment for services and an itemized and standardized fee schedule for medicines. The influence of provider payment method on supply behaviour was sometimes intuitive and sometimes counter intuitive. It appeared to be related to context and the interaction of the methods with context and each other rather than linearly to any given method.

Conclusions
As countries work towards Universal Health Coverage, there is a need to holistically design, implement, and manage provider payment methods reforms from systems rather than linear perspectives, since the latter fail to recognize the effects of context and the between-methods and context interactions in producing net effects.