Improving health systems requires good governance alongside technical interventions. Evidence suggests that strengthening governance results in more responsive health systems. Despite receiving increased attention, governance is still a relatively new area of health systems and policy research. There is no universally agreed way to define, measure or assess health system governance. Little information exists regarding governance at sub-national (health policy implementation) level. Assessing governance at different health system levels (national, county and health facility) in Kenya is the aim of this study. The specific research objectives explored are:
1) factors influencing health system governance in Kenya
2) key stakeholders’ perspectives on health system governance
3) whether the status of health system governance differs with the functionality of health facilities.
The study employed qualitative research methods via interviews with 39 key informants from three levels of the Kenya health system. The study used a conceptual framework adapted from previously published tools for assessing governance and draws on “institutional analysis” theory to help analyse and interpret the findings.
Key factors that influence governance in the health system include devolution, rapidly changing political context, constrained health financing and challenges in managing the health workforce. The most notable influence appeared to be the impact of devolution and frequent health workers’ strikes. Stakeholders shared their views on all six principles of governance and these revealed opportunities for abuse of the system, weak enforcement of policies and accountability measures, and a lack of participation in policy development. They also commented on the lack of improvement in equity in the health system.
Careful analysis using the new institutional economics theory showed that there were observed differences in governance at facility level: fully functional versus not fully functional. The most surprising difference was that staff responding to the lack of weakness of formal institutions by creating informal arrangements that might circumvent or support the goals of the formal system. Fully functional facilities had accountability mechanisms that they self-enforced; by contrast, facilities that were not fully functional lacked both self-enforcement and effective third-party enforcement mechanisms. Norms and practices for controlling corruption were clear in fully functional facilities but confused in some not fully functional facilities.
This study provides an in-depth exploration of what factors influence institutional arrangements for good governance and how these were enforced or not. Analysis guided by theory, with a strong emphasis on context, is an important contribution to the existing literature on governance. This study critically evaluated existing frameworks to assess health system governance from a cross-disciplinary perspective, which can inform future research on governance. The findings highlight research implications for Kenya at policy and operational levels, particularly on the need to monitor health system governance over time due to rapidly changing political and socioeconomic circumstances, especially concurrent devolution.
You can read published papers related to this research here:
- Pyone, T., Smith, H. & Van Den Broek, N. 2017. Implementation of the Free Maternity Services Policy and Its Implications for Health System Governance in Kenya. BMJ Global Health, 2(4).
- Pyone, T., Smith, H., & van den Broek, N. (2017). Frameworks to assess health systems governance: a systematic review. Health Policy and Planning, 32, 710-722.
Thidar also presented an abstract at the International Health Economics Congress in Boston, USA in July 2017, titled: “Are there any differences in governance between fully functional and not fully functional facilities?” The topic was presented as part of the panel discussion titled “Health Sector Corruption: Studies on measurement, policy and strategy.” chaired by William Savedoff. Read more here: