Examining the level and variation in the efficiency of county health systems in Kenya, and how it can be improved

Universal Health Coverage (UHC), which means that everyone has access to the care that they need without getting into financial difficulties, is a global health priority. Kenya, like other low and middle income countries (LMICs) has made a commitment to achieve UHC. However, like other LMICs, Kenya's UHC aspiration is impeded by the twin problem of low public funding of healthcare, and wastage of available resources. Achieving UHC will require additional funding, that will only be justified if current resources are not wasted. Efficiency refers to the optimal use of resources by maximizing healthcare outcomes and outputs, given available resources. That is, efficiency is about reducing resource wastage. Improving the optimal/efficient use of available resources is one of the feasible ways of increasing the availability of resources in the health sector, especially in LMICs such as Kenya. Improving the efficiency of the Kenyan health system is therefore an important research and policy question. However, research in this area is scarce, and has only focused on the facility level (hospitals and primary healthcare facilities). The available research shows that there is variation in the level of efficiency in the Kenyan health sector. Kenya transitioned into a devolved system of government in 2013, creating 47 semi-autonomous county governments. These county governments are now responsible for delivering healthcare services to Kenyans. This research aims to measure the level of efficiency of county health systems in Kenya, and examine the reasons for the differences in efficiency between counties. It also aims to explore how the efficiency of county health systems can be improved, and how this can in turn result in additional resources for the health sector. Further, the research aims to test the application, and refine available methods for measuring efficiency me, so that they can be readily applied to LMIC settings, and at sub-national levels of the system (such as counties/districts) rather than health facility level (such as hospitals).

To carry out the research, we will employ mixed methods. This includes applying existing quantitative methods (data envelopment analysis and stochastic frontier analysis) to measure efficiency, and regression methods to examine the reasons for variability in the efficiency of county health systems. We will then employ qualitative case study methods to examine these reasons for variation in more depth in selected (well and poor performing) counties. We will develop models to explore the potential for unlocking additional health sector resources by improving the efficiency of county health systems in Kenya. The findings of this research will be relevant not only to Kenyan but also similar LMIC health system policy makers in informing strategies for efficiency improvement and ultimately resource mobilization for UHC. The study will also contribute to literature and knowledge building on the methods for efficiency measurement in LMICs, and for sub-national health system units (counties/districts), beyond healthcare facilities (such as hospitals).

This research aims to measure the level of efficiency of the 47 county health systems in Kenya, and to examine its determinants. It also aims to examine the potential for efficiency gains and how this can in turn result in increased fiscal space for the Kenyan health sector. Further, the research aims to test the application, and explore methodological refinements of current efficiency measurement methods to LMIC settings, and at sub-national levels of the system (such as counties/districts) rather than health facility level (such as hospitals). We will use two frontier methods to measure the technical efficiency (and its determinants) of county health systems. First, we will use the double bootstrap Data Envelopment Analysis technique. Data envelopment analysis (DEA) is a non-parametric mathematical linear programming technique that has been widely used in low and middle income countries (LMICs) due to its flexible nature. Stochastic frontier analysis (SFA) is a parametric econometric approach that uses regression analysis to estimate a production or cost function, that accounts for error, but can be computationally complex. Bootstrap DEA will generate bias corrected efficiency scores, which will then be subjected to truncated regression analysis to examine the determinants of efficiency. To further explore the determinants of efficiency, we will select 2 counties that are found to be efficient and 2 inefficient counties that are lower than the 75th percentile in efficiency rank, and carry out in-depth qualitative inquiry through in-depth interviews and document reviews. This component of the study answer questions about how identified factors interact with and influence health system efficiency, and how these might be leveraged on to improve health system efficiency. Lastly, we will examine the potential for efficiency gains using DEA, and excel based models

This study will benefit 4 main groups of beneficiaries: 1) Kenyan national ministry of health policy makers, 2) Kenya county level health policy makers, 3) outside Kenya, in global health institutions and health system policy makers from other LMIC settings, and 4) the local and international academic community. In addition, the findings stand to benefit patients and populations through the effects of more efficient use of resources on overall resource availability and ultimately, service delivery.

National Ministry of Health
Under the devolved system of government in Kenya, the national ministry of health is responsible for formulating policy and legislation, and for monitoring and evaluating health system performance. Two main policy maker groups will benefit from the findings of this study.
First is the health financing unit. This unit has been tasked to develop and spearhead the implementation of a health financing strategy whose aim is the country's attainment of UHC by 2030. A strategy that has been identified in this strategy is efficiency gains aimed at increase fiscal space for health. Implementing this strategy therefore requires evidence on the level of health system efficiency in Kenya, as well as its determinants. Knowledge on the factors that predict health system efficiency will provide levers for policy interventions to improve the efficiency of the Kenyan health system. Second is the monitoring and evaluation unit of the national ministry of health. This unit has expressed a need to develop a framework for performance measurement for county health systems so as to track the country's progress in achieving its health goals. Further, this unit is keen on using performance measurement and feedback as a tool to promote county health system improvements. Efficiency has been identified as one of the key indicators for measuring and ranking county health system performance. This study will enrich the work of this unit by contributing to the conceptualization of health system efficiency in the Kenyan health system, and providing evidence on where to intervene to improve performance.

Kenya county level health policy makers
Under the devolved system of government, county governments have been assigned the service provision function (secondary and primary, preventive and promotive healthcare services). County governments own and operate healthcare facilities, recruit and manage healthcare workers, and provide essential supplies to healthcare facilities, as well as provide budgetary support to healthcare facilities. Evidence on the relative efficiency of county health systems will enable the benchmarking of performance against peers and promote health system improvements. Evidence on determinants will provide policy levers for counties to target with interventions to improve their efficiency.

Global health institutions and health system policy makers from other LMIC settings
Global health institutions and health system policy makers in similar LMIC settings have a need for evidence on the determinants of efficiency in sub-national/regional health systems. However, the efficiency of sub-national/regional health system and in LMICs is understudied. Further, evidence on the determinants of efficiency, and the mechanisms by which these determinants influence health system efficiency is scant. This study will provide evidence on these issues and hence find utility in this group of users.

Local and international academic community.
This study will impact on local and international academic community by contributing to application and suitability of existing efficiency measurement methods in sub-national/regional health systems and in LMIC settings, the application of mixed methods research to not only examine determinants of efficiency but to further elucidate potential mechanisms by which these factors influence the efficiency.