Background
Decentralisation is argued to promote community participation and accountability, and enhance technical efficiency and equity in the management of public resources. Within the health sector, decentralisation has been a recurring theme in health system reforms for several decades [
1,
2]. The implementation of decentralisation polices within the health sector has adopted a wide range of modes and forms, determined by the nature and structure of the sub-national level entity to which responsibility is transferred. However, irrespective of the form, the final effects of decentralisation reforms have been influenced by many internal and external factors including the reasons or drivers for decentralisation, and the country’s political context [
1,
3‐
6]. In practice decentralization involves shifting power and authority over the management of public resources from national to sub-national levels of government. This makes it a highly political reform, though its political nature and context are rarely analyzed in empirical studies [
1,
4,
7,
8].
Human Resources for Health (HRH) and Essential Medicines and Medical Supplies (EMMS) are two critical building blocks of health systems. Considering that the two attract a substantial amount of total health system funding, they often generate contention during the design and implementation of health sector decentralization policies [
2,
9]. However, even with the acknowledgement of the central role of HRH and EMMS, decentralization policy formation and debate mainly focuses on financial resource allocation, financial management and reporting, with HRH and EMMS management plans rarely featuring [
2,
10].
In June 2015, we carried out a systematic search of published empirical studies on the effects of decentralization on HRH and EMMS management in LMICs, published in English language between 1983 and December 2014 which identified 14 articles on HRH, and 7 on EMMS. The studies described a wide range of both positive and negative effects on HRH and EMMS management in LMICs. On HRH management, several studies reported decentralization being associated with better attraction and retention of lower cadre staff, but poor attraction of specialized health workers [
11‐
15]. In Tanzania for example, after undertaking decentralization for all HRH management functions to the district level, rural districts were unable to attract and retain highly skilled staff such as medical specialists, leading the country to re-centralize some of the HRH management functions [
12,
16]. Some studies suggested that certain HRH management functions, including recruitment and distribution of highly specialized health workers, in-service training, and management of staff salaries, are best managed centrally [
11‐
13,
17]; while other functions like staff appraisals, promotions, recruitment and deployment of lower cadre health workers are best handled in decentralized units [
12,
14]. Another commonly reported HRH management problem linked with decentralization has been frequent delays and disruptions in payments of staff salaries; and challenges in managing in-service training and other career progression initiatives [
12,
14,
18]. In addition, several studies identified challenges in the management of the responsibility transfer process from central level to decentralized units, in the early stages of decentralization. This has often been associated with confusion, fear and anxiety on the part of health workers. In many instances, these HRH management challenges have resulted in low staff morale, industrial action like strikes and mass resignations [
12,
13].
On EMMS management, the literature shows that many countries with decentralized health systems retained most EMMS management functions under central control. In most cases it was argued that the central level had better capacity to undertake quantification of EMMS, obtain economies of scale associated with bulk purchases, and monitor and reinforce quality of drugs and commodities supplied [
10,
11,
17]. However, where EMMS management was decentralized, there was been some documentation of better budgetary allocation for commodities, leading to better servicing of commodity [
19] orders at facility level, for example in Ghana and Guantemala [
10].
As part of the implementation of the 2010 constitution, the Government of Kenya in 2013 adopted a devolved government system with 47 semi-autonomous county governments, with significant decision making autonomy, and minimal central level control [
20‐
22]. The design and implementation process of the devolved government system was largely driven by a political push to address real and perceived long-term political challenges of marginalization and inequitable resource allocation in the country [
19,
23]. Within the health sector, the constitution outlined that all health service delivery functions, including the procurement of EMMS and management of HRH, would be assigned to county governments, while the national Ministry of Health (MoH) was assigned the roles of health policy and standards formulation, pre-service training for health workers, and management of national referral services. A detailed breakdown of the functions assigned to national and county governments is found in the schedule 4 of the constitution [
20].
The constitution outlined a five-year plan for establishing county government structures and progressive transfer of functions. This was to begin with the establishment and capacity building of county level structures between 2010 and 2013, and progressive transfer of functions over a 3 year period from 2013 [
20], facilitated by a Transition Authority [
24] (Tsofa B, et al.: How does decentralisation affect health sector planning and financial management? A case study of early effects of devolution in Kilifi County, Kenya, submited). However, once they were elected into office in early 2013, the county governors began to agitate for an immediate transfer of all county level functions. The president in June 2013 succumbed to the pressure from the governors and directed that all county functions be devolved immediately, though at that time most counties had not established structures to undertake these functions (Tsofa B, et al.: How does decentralisation affect health sector planning and financial management? A case study of early effects of devolution in Kilifi County, Kenya, submited) [
25].
Kenya thus provides an ongoing opportunity to examine devolution of these key health sector management functions of HRH and EMMS. In addition to contributing to the literature on decentralization effects on HRH and EMMS management, this paper uniquely analyses the broader political context within which the devolution reform was implemented, and analyses health sector devolution effects as they played out during the process of implementation.
Discussion
Omar [
4] argued that the political drivers and context that push a country to adopting decentralized governance arrangements have a major bearing on how decentralization gets implemented in that setting [
4]. From our findings, the
political context in Kenya caused the transfer of devolved health sector functions to be done faster than had been anticipated by most health sector players (Tsofa B, et al.: How does decentralisation affect health sector planning and financial management? A case study of early effects of devolution in Kilifi County, Kenya, submited) [
25]. This happened at a time when the county governments had not set up their
organizational structures and built capacity for these structures to undertake their functions, causing major challenges and disruptions in public sector service delivery. In this paper, we have specifically illustrated how this rapid transfer of functions caused major challenges in the management of HRH, and EMMS at county level.
On HRH management, the rapid transfer of functions before counties had established their HRH management structures meant that they could not undertake key HRH management roles, including payroll management and payment of salaries. However, an interim arrangement was agreed, where national MoH continued to pay staff salaries up to December 2013, and invoiced county governments for reimbursements. When the counties eventually took up this role, payment of staff salaries was often delayed, with numerous pay-roll inconsistencies and discrepancies and some staff totally missing from the payroll. There were also some reported cases of political interference in HRH management across the country. Other challenges included a lack of clarity over key HRH management roles by different players, including management of inter-county transfers, in-service training and career progression for health care workers. These challenges led to observed and reported fear, anxiety, and mass resignations of health care workers across the country; and eventually culminated into a protracted health workers’ strike that crippled the health sector country-wide for several weeks in late 2013.
Similarly, the EMMS management function was affected by the rapid transfer of responsibilities. In the early days after devolution, there were arguments and contestations between national MoH and CDoH country-wide on the role of KEMSA. In the interim phase, national MoH, with funding from donors, supplied 6 months’ worth of buffer stock of drugs for all public health care facilities countrywide as a stop-gap measure to allow for counties to set up their procurement and distribution systems. When the county governments eventually took up this role, there was widespread politicization of the drugs and commodities procurement and distribution within the health sector. Nevertheless, health facilities reported better fill rates whenever drugs were supplied.
Though our findings show that the implementation of the devolved government system in Kenya significantly increased
the decision space for HRH and EMMS management at county level, the ability of counties to claim, and utilize this space was undermined by an initial lack of proper structures and capacity to fully undertake all the HRH and EMMS management functions. However, with time as counties established their structures and built their capacity they did increase their ability to utilize the expanded decision space over these roles. Table
2 below illustrates our analysis of the the shifting county level health sector decision space over HRH and EMMS management functions over time. This happened in response over time during that period with the progressive building of
organizational structure and capacity at county level. From these findings, it is evident that decision space of decentralized units can be compromised by lack of capacity to undertake the decentralized functions.
Table 2
Shifting county level decision space over HRH and EMMS management functions corresponding to improvement in organizational structure and capacity of overtime
HRH Management |
Employment of staff | + | ++++ | ++++ |
Deployment/distribution of staff | + | ++++ | ++++ |
Payment of salaries | + | + | ++++ |
EMMS Management |
Commodity quantification | + | ++ | ++++ |
Commodity procurement | + | ++ | ++++ |
Commodity allocation/distribution to health facilities | + | ++ | ++++ |
In a study of health sector decentralization in Pakistan, Bossert and Mitchell (2011) reported that the de facto decision space over decentralized health sector management functions was always different from the
de jure decision space; and that the difference was often due to the capacity of the individuals and institutions tasked to undertake the decentralized functions. Our findings agree with those of Bossert and Mitchell. They highlight the importance of ensuring that appropriate peripheral level capacity to undertake decentralized functions, is in place in decentralized units if the benefits of health sector decentralization are to be realized (Tsofa B, et al.: How does decentralisation affect health sector planning and financial management? A case study of early effects of devolution in Kilifi County, Kenya, submited) [
33]. In addition, our findings also highlight the need to be ready to develop interim measures when this capacity is not yet available, and for central government and development partners to support this. The findings also highlight the need for clarity of roles of actors of different HRH and EMMS management functions across the different levels.
In our analysis, we note that the increase over time of county level decision space over HRH and EMMS management functions led to several positive effects. For HRH, the county gained the ability to determine the actual number of staff based on its budget and decide where to deploy them within the county. For EMMS, decentralized procurement led to a reported better fill-rate in health facilities, and the county was able to ensure all facilities were supplied with EMMS irrespective of registration status. This in turn allowed previously non-functioning facilities to operate, and thus previously underserved areas to have access to a facility. It is therefore likely that increased decision space at county level enhanced local level equity in the allocation of health resources, and health service provision at county level, and ensured services reached previously underserved populations.
However, the increase in county level decision space also led to perverse negative effects. With regards to HRH management, for example, the transfer to counties of decisions over the number and type of health workforce led to complications over transfers to other counties. In addition, the increase in county level decision space led to political interference over recruitment and deployment of staff, as local politicians began to demand that only health workers from within their county and tribe should be employed within the county. Political interference over HRH management within decentralized settings has also been reported in the Philippines by Grundy et al. (2003) [
11]. Though not reported in our study, decentralized HRH recruitment has also been associated with inability to attract highly skilled health workers in rural remote areas in rural districts in Uganda [
12]. Though our study was conducted in the early days of devolution implementation, this challenge may be less likely to occur in Kenya as the counties have more decision space for HRH management than their Ugandan counterparts, including the power to create special incentives to attract and retain staff.
For EMMS, as reported in Ghana and Guatemala [
10], it could be argued that decentralized procurement led to the loss of economies of scale associated with a centralized procurement system. The increased decision space over EMMS management also fueled county level prioritization of the procurement of highly visible commodities such as ambulances at the expense of much needed drugs for Primary Health Care services (PHC). These observations are consistent with those made in Uganda and South Africa, where decentralized units prioritized allocations to curative services at the expense of PHC, as the former were more visible to the community [
36‐
38]. This might be expected given the political nature of decentralization, leading local level political decision makers to prioritize issues that are more politically visible, and which will resonate with the electorate in order to maintain political support [
39].
In relation to
accountability structures and practices our study also found that these influenced decision space among health sector actors. For HRH, for example, we found that senior managers of the county government, including the County Department of Health, had to undergo public vetting at the County Assembly for their suitability for office, before being appointed. Similarly, we found that that the health facility managers had to involve their respective FMCs in the EMMS quantification process, and later inform them once the supplies are received at the facility. In both these cases, we find that the devolved government system was deliberately designed with increased public participation and accountability mechanisms with an intention to guard against potential decision space excesses by management decision makers. In addition to being an accountability mechanisms, enhanced public participation, especially in health resources allocation also does enhance community responsiveness of health service prioritization and thus promoting equity [
22,
40].
In general, a recurring theme in the early days of Kenyan devolution affecting the design of the devolved government systems, the transfer process of county level functions generally, and the early effects on HRH and EMMS management process, was the politicization of the process. This observation underscores the arguments that by the mere fact that it involves the shifting of power and control from the center to the periphery, decentralization is a highly political process in its own right, and any attempt to analyse health sector effects of decentralization policies should always include an analysis of the political context [
4,
7].
Study strengths and limitations
The primary focus on only one county out of the 47 in the country could be considered a limitation of this study. However, the decision to use one county was deliberate, as it allowed for a deeper exploration of the issues under focus, by involving extended engagement with a broad range of stakeholders. Kilifi County is also part of the health systems governance learning sites for the Resilient and Responsive Health Systems (RESYST) consortium [
21,
27]. This allows for longer-term tracking of the decentralization effects in this sites beyond the time of this study. The learning site setting also provided an opportunity for regular feedback to the county managers and national MoH thus increasing the potential of this study to inform the progressive implementation of devolution in the county and country.
Conclusions
Decentralization has been an important element of the health system governance reform agenda for many years owing to its perceived importance in creating opportunities for strengthening local level management efficiency over ever-scarce health sector resources. For these reason, many health systems decentralization initiatives have included the goal of increasing local level decision space over management or resources.
The implementation of a devolved government system in Kenya has significantly increased county level decision space over HRH and EMMS management. This increased decision space created great potential in allowing for targeted recruitment and deployment of health workers, and procurement and distribution of EMMS based on local level priority needs. However, in practice this potential was undermined by organizational structure and capacity limitations, particularly in the early stages of implementation. Political interference also played a key role over HRH and EMMS management at county level, and was exacerbated by the combination of increased decision-space at sub-national levels at a time when structures and systems were not in place, and capacity was inadequate. That political interference was evident is hardly surprising, given that devolution itself is highly political, involving the transfer of power over management of public resources from national to sub-national levels of government.
Recomendations
With these findings, we recommend the need for specific interventions to strengthen county level capacity over specific HRH and EMMS management functions so as to harness the potential positive effects of the increased decision space at county level brought about by the devolved government system in Kenya. These interventions should include creating clarity over HRH management roles between the CDoH and the CPSB, and improving the county payroll management system to stabilize payment of salaries for health care workers. On EMMS, there is need for strengthening the capacity of both health facility managers and CDoH managers in undertaking specifications and quantification of EMMS in order to streamline and speed up the ordering and procurement processes to avoid long periods of stock outs. At a national level, there is need for all stakeholders to come together to deliberate and build consensus on how certain HRH management roles including in-service training and inter-county transfers should be conducted across the country. For LMICs with similar settings to Kenya, we recommend that individual and institutional capacity considerations should always made when allocating functions between the center and the periphery during the design and implementation of health sector decentralization policies.
Acknowledgements
We would like to acknowledge the support of Edwine Barasa, Mary Nyikuri, Evelyn Waweru, Jacinta Nziga and Kelly Muraya (from the Kenyan learning site) and the larger RESYST Health Systems Governance theme members. We would also like to acknowledge our colleagues and collaborators in the County Department of Health, Kilifi and the national Ministry of Health in Kenya.
The preparation of the paper benefitted from discussion at an April 2016 writing workshop organized by the Consortium for Health Systems Innovation and Analysis (CHESAI) to generate deeper Southern-led perspectives on health systems and governance issues, CHESAI is funded by a grant from the International Development Research Centre, Canada.