Decentralization of health systems is a common pillar of health sector reform initiatives and is often viewed as a means to improve efficiency and quality of services as well as promoting accountability and local governance of the health system. Despite many years of health decentralization initiatives in many countries worldwide, our knowledge of the relationships between particular forms of decentralization and performance of health systems remains limited. For instance, does decentralization help, hinder or have no impact on equity in resource allocation? Or are there relationships between decentralization of logistics systems and health sector performance? Part of the difficulty in addressing such questions lies in the inherently contextual relationship between decentralization and health systems performance, as well as the difficulty in untwine decentralization from other health sector reforms. Yet as importantly, we continue to lack an adequate framework to analyze decentralization both within and across countries.
IHSP faculty developed and have applied widely the innovative “Decision Space” approach to comparatively analyze health sector decentralization (see (Bossert 1998)) Based on principal-agent theory, the Decision Space framework focuses on the range of choices accorded to decentralized authorities over a number of different functions. The range of choice includes both formally-permitted actions (such as laws and regulations which govern the nature of the decentralization) and unwritten lines of authority. The range of choices may or may not differ by health functions, the latter of which include financing, service organization, human resources, and governance. This approach has been implemented through survey questionnaires at municipal and district levels in Nicaragua, Morocco, Pakistan and India to determine the actual range of choice that local officials have been able to exercise. These detailed questionnaires are tailored to specific situations in each country and include questions on decision space, capacities and accountability.
One of the first studies conducted by IHSP faculty found that the Decision Space framework was able to comparatively describe decentralization across four countries (Zambia, Ghana, Uganda and Philippines; see (Bossert and Beauvais 2002)). The studies found considerable variation in decision space across the different financing, service delivery, human resources and governance functions of the health systems. For instance, none of the countries were found to grant wide choice over revenues, expenditures, and fees, most granted a moderate range of choice over revenues and all grant a moderate range of choice over expenditures, though Zambia had little choice over alternative revenue sources. There also is commonly variation in decision space by function within a particular country. Zambian health officials, for example, were found to have a moderate range of choice over certain functions, such as expenditures, user fees, and contracting, but relatively limited over choice over others, such as salaries and allowances (see (Bossert, Chitah et al. 2003).
More recently, two major new studies have begun in Pakistan and India using an expanded decision space approach to study the decision space, capacities and accountability of health and local officials at district levels. Preliminary findings from Pakistan suggest there is a surprisingly wide variation in decision space from district to district, with districts having greater capacities generally exercising wider decision space. In some crucial areas, districts had sufficient capacity to take on even greater decision space than currently sanctioned.
In additions to providing a means to describe decentralization, the decision space methodology has been used to analyze relationships between decentralization and health sector performance. While some believe that decentralization of authority may contribute to inequity in resource allocation (e.g., wealthier localities will allocate more to health than poorer localities can), our studies have not found this to necessarily be the case (see (Bossert, Larranaga et al. 2003)). Although wealthier localities in Chile and Colombia were able to spend more per capita on health, the 1998 study found that the gap between rich and poor was declining over the period of decentralization, and there was increasing equity of per capita expenditures at the municipal level. Indeed, a possible explanation for these findings was that that national-level mechanisms such as earmarking and population-based formula funding may have encouraged local authorities to assign similar amounts from their own source revenues is a local choice that is encouraging.
In another line of research, an innovative assessment of decentralization of logistics system for vaccines, essential drugs and contraceptives in Guatemala and Ghana indicated the nuanced relationship between decentralization and health sector performance. In these countries, it was found that that allowing local officials more decision space for some functions (e.g., forecasting, procurement and planning) was associated with better performance measures for those functions, while other functions were more effective if they remained centralized (e.g., logistics information systems and warehousing practices) (see (Bossert, Bowser et al. Forthcoming)).
Finally, decision space analyses have also attempted to examine relationships between decentralization and health outcomes. In Zambia, utilization of health services, immunization coverage and family planning activities appeared to bear little relationship to the degree of decision space at the district level, found little variation during the period 1995-98 except for a decline in immunization coverage, which may have also been affected by changes in donor funding.
Our research findings suggest that more work needs to be done to assess the relationship between decentralization and health sector performance. While, in terms of resource allocation, we have demonstrated at least for two health systems, forms of decentralization that include mechanisms to improve equity, such as earmarking and population-based formula funding, may improve equity of resource allocation, it is necessary to assess this finding in other country contexts and to assess the impact of different forms of decentralization on efficiency, quality as well as equity of services. Our findings for logistics systems suggest that more sophisticated assessments are needed to provide policy advice for decentralizing specific programs and systems. More research in information systems, supervision systems, and other management control systems is needed.
As our recent research in Pakistan and India suggests, it is necessary also to assess the capacities and the accountability processes at local and central levels to complete the picture of decision space and decentralization. Studies of decentralization should develop policy-relevant evidence for both the types and degrees of choice allowed to local authorities and the contexts — both existing capacities and accountability processes — that foster more effective performance and outcomes of health systems.
Bossert, T. (1998). "Analyzing the decentralization of health systems in developing countries: decision space, innovation and performance." Soc Sci Med 47(10): 1513-27.
Bossert, T. J. and J. C. Beauvais (2002). "Decentralization of health systems in Ghana, Zambia, Uganda and the Philippines: a comparative analysis of decision space." Health Policy Plan 17(1): 14-31.
Bossert, T., M. B. Chitah, et al. (2003). "Decentralization in Zambia: resource allocation and district performance." Health Policy Plan 18(4): 357-69.
Bossert, T. J., O. Larranaga, et al. (2003). "Decentralization and equity of resource allocation: evidence from Colombia and Chile." Bull World Health Organ 81(2): 95-100.
Bossert, T., D. Bowser, et al. (Forthcoming). "Is decentralization good for logistics systems? Evidence on essential medicine logistics in Ghana and Guatemala." Health Policy and Planning.
Click here to return to the IHSP Research home page.