Untitled Document

Improving Access to Maternal Health Care through Devolution in Western Kenya

Susan Kilonzo, Eunice Kamaara and Kitche Magak*

Abstract
Devolution was greeted with great anticipation in Kenya as a means of bringing services closer to the people. However, since the implementation of the recent devolution reforms, criticism has mounted, with evidence of corruption, poor management, late payment of county staff and considerable disaffection among service providers, especially health professionals. In this study, we examine health-care users' and providers' perceptions of the effect of devolved health services on referral maternal health-care access in Kisumu and Uasin Gishu counties in Western Kenya. Our findings suggest that while health workers are dissatisfied, there is considerable satisfaction among users of referral maternal health services. Users largely associate their satisfaction with devolution. However, closer analysis suggests that improved access is not only linked to devolved health services but also to other developments both at the national level (health campaigns, increased mobile telephony) and county level (improved transportation, relocation of available funds).

Keywords: devolution, Kenya, maternal health care, access, citizen perception.

1 Introduction
Devolution, especially by central governments, transfers power to lower levels of administration, thereby reducing top-down approaches in decision-making (Bossert and Beauvais 2002; Naeem et al. 2012). Article 174 of the 2010 Kenya Constitution1 stipulates the role of devolution as enhancement of self-governance, economic development and equitable sharing of local and national resources. Yet, according to Robinson (2007), the history of decentralisation reforms in Africa is full of bad examples, mostly due to the absence of meaningful local political process, over-centralisation of resources, weak local revenue base, lack of local planning capacity, and limited changes in legislation and regulations. Since the 1990s, there has been a shift in the aims of decentralisation – to enhance democratisation and reduce the role of central governments – as well as in its form, from deconcentration to devolution (Conyers 2007). Still, the problem remains, with most central governments reluctant to devolve enough power to local governments for them to have a significant impact on service delivery.

Devolution affects every aspect of governance, which in turn affects the totality of development, including health. Recent devolution reforms in health care across Africa have displayed some interesting outcomes, although a key undercurrent remains: central governments' reluctance to let go of power. For instance, although health-care services have been devolved in Uganda, the national government still controls the budgets and provides conditional grants for the promotion of primary health care (Bossert and Beauvais 2002: 24). The implication of such controls is that the communities and administrators at local levels have limited say over the operations of health-care services that affect them. Looking at cases of what he calls 'deep democratic decentralization', Mehrotra (2006: 270) points out to the more successful cases of Benin, Guinea and Mali, where decentralisation of primary health-care services led to increased access to affordable health services, which in turn increased immunisation rates and reduced infant mortality. Part of this success was the fact that in all three cases, the state remained a relatively weak central authority, that could not support public health services in the early 1980s. More recently, Inkoom and Gyapong (2016) highlight that despite advances in the implementation of more ambitious health-care decentralisation plans in Ghana, Malawi and Tanzania, in all three countries policymaking is still based at the centre and local governments report a high dependence on central government for funds (usually earmarked), allowing for central government interference.

As a signatory to most international and continental agreements on improving access to maternal health, Kenya has launched several health campaigns. However, the country failed to achieve UN Millennium Development Goal 5 (to reduce maternal deaths by 75 per cent between 1990 and 2015) and Kenyan mothers still experience many maternal health-care challenges including complications during pregnancy and childbirth, as well as HIV/AIDS-related issues (Maoulidi 2011). This problem is even more acute in some areas, with the centralised decisionmaking system in Kenya being blamed for regional health disparities (MoH 2006). Still, there is a noted reduction of maternal deaths and the country remains committed to the improvement of maternal health through the implementation of various strategic initiatives including the Free Maternity Services programme and Beyond Zero Campaign.2

Before the promulgation of the 2010 Kenyan Constitution, the Ministry of Health (MoH) had already committed to decentralising its services in order to provide increased local-level authority in decision-making, resource allocation and management of health care. The aim was to involve communities in the management and implementation of essential clinical and public health services (MoH 2002, 2006). The 2010 Constitution, which transferred the health service delivery function to the newly created 47 counties in 2013 (Baker et.al 2014), provided an opportunity through a constitutional framework for an ongoing ministerial devolved health initiative. The health-care responsibilities that were devolved include community health, primary health care and county referral services. The national government retained the responsibility of national referral services. Unfortunately, there have been various challenges including an unprecedented number of strikes by health-care workers: there have been more than two dozen strikes since the devolution of health services in 2013 (Masika 2016). Cheeseman, Lynch and Willis (2016) caution that time is vital in assessing the full impact of devolution in Kenya. Studies on devolved health services in Kenya have mainly focused on the job satisfaction of health workers, devolution as a strategic approach to health care, and the implementation processes (Baker et al. 2014; Oyugi 2015; Mwamuye and Nyamu 2014). Overall, most studies provide a general overview of the effects of devolved governments on health care. Our literature review did not find any study that focused on access to referral maternal health care.

This study fills this gap by exploring the perceptions of maternal healthcare users, health providers and administrators on how devolution of health services affects access to referral prenatal and postnatal maternal health care in Kisumu and Uasin Gishu county hospitals. A United Nations Population Fund (UNFPA) report (2014) indicates that 15 of the total 47 counties in Kenya account for 98.7 per cent of total maternal deaths in the country. Within these, Kisumu County ranks seventh in this list with an infant mortality rate of 249/100,000 and a maternal mortality ratio of 597. While Uasin Gishu county is not within the top 15, it was chosen alongside Kisumu because both are urban counties in the Western Kenya region and they are both cosmopolitan. Besides, Uasin Gishu and Kisumu are the only two counties with a teaching and referral hospital to which the county hospitals across Nyanza, Western and North Rift regions of Kenya refer maternal users. The main maternal health challenges in Kisumu County include poor access to reproductive health services, poor quality of services, a high number of unskilled birth deliveries (Maoulidi 2011) and HIV/AIDS. The county referral health facility is Kisumu County Hospital, but the county also hosts the Jaramogi Oginga Odinga Referral Hospital, a national referral facility. The Uasin Gishu County Hospital is the county referral facility and the Moi Teaching and Referral Hospital hosts the Riley Mother and Baby Hospital, a national maternal referral facility. The main maternal health challenges in Uasin Gishu include poor health systems, poor management skills of health workers and poor community structures.

In this article, we examine health-care users' and providers' perceptions of the effect of devolved health services on access to referral maternal health care in these two counties. We start by describing what we mean by 'access', the relationship between health-care users and healthcare systems, breaking this concept into four dimensions – availability, accessibility, affordability and acceptability – which guided us in data collection, analysis and interpretation. In the subsequent section, we outline our methodology, followed by a discussion of our findings across the four dimensions of access. Opposite to findings from previous studies on users' perceptions of the quality of local government service provision (see Conyers 2007), and health providers having generally a negative view of devolution, we found that users' perceptions are overwhelmingly positive. We conclude by stating that while other health-related developments at national level (namely national health campaigns) have improved health-care access and played a role in users' perception of the effects of devolution, county-level decisions have also contributed significantly to the perceived positive effects of devolution in referral maternal health care.

2 Accessing health care: a framework
Worldwide, access is an essential element to the performance of health-care systems (Levesque, Harris and Russell 2013). In this study, we use a conceptual framework that builds on longstanding descriptions of access to health services, drawing on the work of Penchansky and Thomas (1981) and further developed by Peters et al. (2008) and McIntyre, Thiede and Birch (2009). Like Penchansky and Thomas (1981), we think of access as a multidimensional concept that summarises a set of more specific dimensions describing the fit between the user and the health-care system. The interaction between these dimensions determines access, even though each dimension is distinct and focuses on a set of clearly distinguishable issues (McIntyre et al. 2009). Ultimately, this framework provided us not only a structure to help our data collection, but also a conceptual lens to interpret the data collected on how devolved health services are perceived by health providers and users to impact access to referral maternal health care.

In this framework, we incorporate four dimensions of access: availability, accessibility, affordability, and acceptability. By availability, we mean having the right type of health services available to those who need them, that is, supplying the appropriate health-care providers and/or services in the right place and at the right time to meet the prevailing needs of the population. In accessibility, we refer to the relationship between the location of supply and the location of users, taking into consideration user transportation resources and travel time, distance and cost. Affordability deals with the relationship between the price of services and the willingness and ability of users to pay for those services. Finally, acceptability deals with how responsive health service providers are to social and cultural expectations of individual users and communities.

3 Methodology3
For this study, we adopted a mixed methods approach combining initial observation of hospitals with qualitative interviews with a range of stakeholders and a focus on perceptions. For the former, we engaged in participant observation to study and record the everyday life in spaces and places of interface between health-care users and providers, and for the latter we mostly made use of in-depth semi-structured interviews. As such, we relied mostly on ethnographic methods. Ethnography has evolved from its traditional application in anthropology to be used in fields such as policy, design and engineering, helping us provide rich explanations of the hows and whys of peoples' lives (Button 2000: 322). Ethnographic health studies focus on the health challenges that face people and provide frameworks of how and why people access or do not access health care. We adopted this approach to understand the whys and hows of existing perceptions of the effect of devolution on access to referral maternal health care. Using key ethnographic methods such as participant observation and qualitative interviews allowed us to provide narratives, verbatim reporting, and descriptions of what we heard and observed regarding the effect of devolved health services on access to maternal health care in the two counties. Ultimately, what emerged were the peoples' interpretations of their own health world.

We collected data sequentially in three phases, with each phase lasting about three weeks. The first phase involved participant observation of prenatal and postnatal referral contexts at Kisumu and Uasin Gishu county hospitals. The initial unstructured nature of our participant observation allowed us to reduce potential researcher biases and assumptions by 'looking around' without a set agenda. In this first stage, a series of issues became salient, namely, issues related to facilities and supplies, training and motivation of medical staff, and application and effectiveness of referral policies. We used these issues as themes to guide a more structured observation in a second phase of our research, together with informal conversations with health-care providers and users. Subsequently, guided by what emerged from the second phase, we proceeded to develop semi-structured questions that guided qualitative interviewing (including in-depth interviews) of key informants. We also made use of our observations to select key respondents, as well as to triangulate their answers not only with themselves, but also with our observations.

We purposively selected 32 key informants for different purposes:

  • To provide us with individual perspectives on the hows and whys of the perceived influence of devolved health services on access to referral maternal health care, we interviewed eight maternal health users receiving care at Jaramogi Oginga Odinga Referral Hospital and at Riley Mother and Baby Hospital having been referred from Kisumu and Uasin Gishu county hospitals respectively (four from each hospital);

  • To provide us with the view from a managerial perspective, we interviewed eight national and county-level medical health officers and four hospital administrators from the two counties;

  • To provide us with vital perspectives (and perceptions) on the effects of devolved health services on access to referral maternity health care, we interviewed 12 health-care providers, namely, two doctors, two nurses, and two midwives/birth attendants from each county.

Data collected from different sources and by different methods were triangulated to strengthen the validity of findings on contextual issues specific to each county. The issues included perceptions on leadership, governance and effectiveness of devolved health services policies, especially those related to access to referral maternal health care. The overall theme was access (using the 4As framework, as detailed in Section 2). Field notes emerging from jottings made over six weeks of observations and recorded interviews were systematically transcribed, cleaned, manually coded, and thematically categorised. We made use of this data to analyse the explanations, interpretations, examples, cases and illustrations of the perceived effects of devolution on availability, accessibility, affordability and acceptability of referral maternal health care. We focused our interpretation of the data on how and why devolved leadership and governance is perceived to affect access to referral maternal health care.

4 Unpacking access: initial findings
This study investigated perceptions on the influence of devolved health services on referral maternal health-care access in Kisumu and Uasin Gishu county hospitals and the origins of those perceptions. Overall, the study indicated mixed perceptions: while nearly all maternal health users agreed that devolved health-care services have a positive impact on access to referral maternal health care, health-care providers expressed various misgivings. Devolution of health services has spawned a widespread perception among users that health services, including referral maternal health care is now closer to mwananchi (citizens) in terms of availability, accessibility, affordability and acceptability. These positive perceptions of improved access to referral maternal health care have led to an influx of patients, an indicator of acceptability of devolved health care by users. But the increase in patients has not been matched with expansion of facilities and human resources, two factors that partly explain the misgivings of health providers. Finally, and similar to what D'Arcy and Cornell (2016) suggest, some health-care providers think devolution has 'devolved' corruption to the county level. This perception is further strengthened by local newspaper stories and television programmes featuring the affluent lives of county government officers, difficult to maintain on their current salaries.

Despite positive perceptions of the effects of devolution on access to referral maternal health care, there are still big challenges including fiscal dependence of county governments on the national government and the delay in fiscal remittances from the national government. Over the period of data collection, remittances were late by eight months. County governments in both Kisumu and Uasin Gishu continued to offer free services using funds from other sources, an advantage of autonomous decision-making by county governments. This finding agrees with Cheeseman et al.'s (2016) argument that Kenya is a unique case in which central government may not manipulate county governments given the dynamism of power relations and political actors, whose presence in the succeeding governments is highly determined by their ability to campaign for new powers and sources of revenue.

A key finding in this study is that users perceive devolved health systems to have positively influenced access to referral maternal health care. This is contrary to some of the studies on effects of devolution on health care, particularly maternal health care (see Bourbonnais 2013), where service providers remain insensitive to the needs of women with respect to cultural and socioeconomic demands. Yet, we found that although users' perceptions strongly associate devolved health services with improved referral maternal health care, there are other factors playing a role in this (real and perceived) improvement. Still, while the policies behind some of these factors originated at central government (for instance, the Free Maternity Services programme and the Beyond Zero Campaign), the implementation happened because of county governments choosing to do so. In other words, county-level decisions had a significant role to play in the perceived positive effects of devolution as regards to access to referral maternal health care.

Finally, users in Uasin Gishu county reported one negative implication of devolution: tribal discrimination of minorities. The perceived discrimination emanates from respondents associating devolved governance with tribal conglomeration. Health-care providers and administrators indicated a difference between this perception and the actual situation, indicating that priority service is provided only on the basis of level of emergency. As Cheeseman et al. (2016) observe, though the introduction of devolution was expected to diffuse Kenya's chronic ethnic conflicts, the findings of this study suggest that this is not the case. We now present the specific findings of this study along the 4As of health-care access: availability, accessibility, affordability, and acceptability.

4.1 Availability
Maternal health patients perceive devolution of health services to have improved availability of referral maternal health care. This availability seems to have increased the number of users seeking services in public health facilities. The situation is still far from satisfactory, but it is a major step from the old centralised health system in which commodities were almost always unavailable. There are no new health facilities in either Kisumu or Uasin Gishu though there were visible repairs and improvements on already existing facilities. None of the study respondents referred to an increase in facilities since devolution. However, they did indicate that with devolved health services, referral maternal health-care services are more readily available than before:

You cannot compare what is happening today with what was happening before. Before the new constitution, it was very difficult to come here. You could die at the district hospital as the doctors tried to transfer you. I say this from experience. My sister died in my arms at the UGC Hospital three days after the doctor told her that they were going to refer her. [She chokes with emotion as tears welled up in her eyes.] The first day they said there was no bed space in the referral hospital. The second day they said there was a bed but there was no ambulance. The third day my sister died. I was so scared when I was admitted in that same hospital for delivery, but I was happily surprised when they brought me here so fast in a brand-new ambulance. (Twenty-five-year old postnatal user, Riley Mother and Baby Hospital4)

Closely associated with improved availability of referral services is availability of ambulance services which was commonly cited by both users and providers. Maternal care patients perceived ambulance services to have greatly improved with devolution. Observations at the four health facilities corroborated this positive perception about the availability of referral maternal health care. During the course of the study, many ambulances in good condition brought or took away patients to the maternal health-care sections of both county and national referral facilities. Follow-up interviews confirmed that the ambulances delivered women who required emergency health care from lower level health facilities and even from homes.

Patient respondents attributed the availability of ambulances to the ability of county governments to access and make decisions over available resources, and so did other respondents. Although other bodies also operate ambulances across the country, and some of them are donated, a health official indicated that the Uasin Gishu county government has bought many ambulances for both county and sub-county health facilities:

Before devolution, there were no funds at the local level to do this so ambulances were rare. And you see we in the county health departments can make decisions on whether to buy ambulances and how to use and maintain them. (In-depth interview, Uasin Gishu County Officer5)

Availability of medicines was also perceived to have improved with devolution. Both interviews and observations revealed that most maternal health-care patients were able to get the prescribed drugs from the hospital pharmacy:

I notice that the pharmacy is well stocked. There are various kinds of medicines. An expectant woman limps to the pharmacy. Her lips are dry and cracked and she appears tired. 'Sema mama [yes ma'am]?' asks the man. 'Wamesema unipatie hii dawa [They said you give me this drug]?', she replies, handing over piece of paper to the man. […] The man gives the woman the drugs and she walks away. Other patients come, give a piece of paper to the pharmacists, pick drugs and go. (Fieldnotes, Uasin Gishu County Hospital)

Some of the patients interviewed noted that initially most hospitals lacked enough and readily available medicine and patients would always be given prescriptions to purchase drugs from private pharmacies. Referral maternal health-care users said that they can now get many, in some cases all, the medicines that they require.

Things have really improved since devolution. Now, you can be sure that you will get the medicines that the doctor prescribes for you. Before, it was like a miracle to get these medicines here in the hospital. That is why many patients only came to this hospital as a last resort. (In-depth interview with male partner of a postnatal user, Uasin Gishu County Hospital6)

While patients overwhelmingly opined that devolution has improved availability of referral maternal health care, health-care providers were not so unanimous:

The perception out there is that devolution has been a magic bullet to all that ails public health services. This perception is mainly spearheaded by the county governments. Devolution has worsened some aspects of health services. For example, the capacity of this health facility has not been expanded to meet the increased referral maternal health-care demand that has been occasioned by free maternal health care policy. (Administrator, Kisumu County Hospital7)

However, the perceptions of health-care providers seem to relate more to their working conditions and environments than to referral maternal health-care access. They noted that, though there is a significant increase in maternal health-care patients being referred from lower public and private health facilities, this is not matched by expansion of the needed facilities and human power. This was clear at both the Riley Mother and Baby Hospital and the Jaramogi Oginga Odinga Referral Hospital where the maternity wards accommodated more than their capacity. Mothers and their children were forced to share beds. This was mainly attributed to both devolution and implementation of policies that encourage maternal patients to seek services in public health facilities:

There is this general widespread notion that services are cheaper, free and better in the health facilities run by the counties. While this may not be necessarily true, this notion encourages users to seek prompt medical attention at various public health facilities. (Hospital superintendent, Kisumu County Hospital8)

In spite of this observation, some health-care providers shared perceptions with users that devolved health services have improved the availability of supplies and ambulance services.

4.2 Accessibility
Devolution in general and devolution of health services in particular has had a visible impact on accessibility to maternal health care. For instance, we observed a remarkable improvement in the means of transportation and condition of roads, both of which directly influence physical and cognitive accessibility to maternal health care. Although the four health facilities under study have always been available, physical accessibility due to poor infrastructure created inaccessibility. Respondents indicated that there is improvement and construction of existing and new road networks by county governments.

The two county governments have also contributed significantly to improved physical accessibility to referral maternal health care through provision of ambulances. As mentioned already in the previous section, both user and service providers averred that ambulances are now more available than before devolution. A patient narrated how she and her baby were quickly ferried by ambulance from the Uasin Gishu County Hospital to the Riley Mother and Baby Hospital on referral. Her baby had developed breathing problems and needed support. This patient's assertion was corroborated by a health official in Kisumu:

These days we have ambulances taking patients from mashinani [grass roots] to the referral health facilities. In terms of referral maternal health, ambulances have become available and mwananchi – hata yule mama wa huko ndani, [the common person, including that woman at the grass roots] is able to access this referral hospital. Devolution has greatly improved health facilities and services. (In-depth interview, Kisumu County Officer9)

Motorcycle taxis locally known as boda boda have had a huge positive impact on physical access to referral maternal health care. The boda bodas have the ability to penetrate to the remotest parts of the countryside along footpaths that even the hardiest of four-wheel-drive vehicles cannot access. Both patients and service providers hailed the presence of these two-wheeled taxis as a game changer in the provision of timely maternal health care. The role of boda bodas surpasses that of ambulances not only because boda bodas can access areas that are impassable by cars but also because many taxi operators have personal relations with their patients which is beneficial to maternal health-care access. Nearly all maternal user respondents indicated that they have personal boda boda operators that they can call on any time, day or night. Although the boda bodas are not part of devolved health services, the conducive environment created by devolution has allowed them to prosper, helping access to maternal health care. According to the World Bank (2016), between July 2015 and May 2016, Kenya improved 21 positions (from 129th to 108th) in terms of ease in doing business. In the same survey, Uasin Gishu and Kisumu were ranked first and second most investor-friendly counties in the country. In both counties, there are boda boda associations registered with the county government, which the county governments use to mobilise the boda boda riders for civic education on road safety.

Access to referral maternal health care has also had a major boost from the explosion of mobile telephony in Kenya. According to the Communications Authority of Kenya (2016), mobile penetration hit 90 per cent in June 2016. Besides the most common use of mobile phones to call taxis to take patients to hospital and to transfer money electronically to people in need of health services, the digital revolution has provided a number of mobile-based health-care services, thanks to information and communications technology (ICT) innovations. Mobile health services and innovations are fast emerging in Kenya such as: mHealth Kenya, a national company overseeing and managing mobile technology health projects; Daktari 1525, which enables over 25 million Safaricom subscribers access to doctors any time to receive expert advice on various health conditions; MedAfrica, a free mobile phone application that allows Kenyans to access health-care-related information (including personalised first aid and reproductive health guidance, as well as access to reputable doctors and hospitals in their area); and the Health Information Technology (HIT) run by the Academic Model Providing Access to Healthcare (AMPATH), at Moi University, Moi Teaching and Referral Hospital and a consortium of North American academic health centres led by Indiana University working in partnership with the Government of Kenya. HIT provides automated patient-specific care suggestions and reminders. Although this digital revolution is independent of devolved health services, the ubiquity of the mobile telephone enables users to access maternal health literally at the click of a button.

4.3 Affordability
Users of maternal health care across Uasin Gishu and Kisumu counties perceive affordability of referral maternal health care to have improved with devolved health services. All respondents were unanimous that the average cost of prenatal and postnatal services was cheaper in the devolved health system than in the old centralised one. While the Beyond Zero Campaign and the policy on free maternal health care in all public health facilities in Kenya is responsible for this, county governments play a key role as devolved decision-making processes allow them to allocate resources to the health facilities even when the National Treasury delays the release of health-care resources to county governments. Both counties grapple with the challenge of inadequate and delayed funding from the national government. At Kisumu County Hospital, we learnt that the last disbursement received was eight months ago. The hospital is forced to use credit, cost-sharing and the Supply Systems Facility Improved Fund. This delay and inadequacy directly affects acceptability of referral maternal health-care services due to low staff motivation, inadequate commodities supply and infrastructure. Cheeseman et al. (2016) argue that Kenya is a unique case in which the central government may not manipulate the county governments given the dynamism of power relations and political actors, whose presence in the succeeding governments is highly determined by their ability to campaign for new powers and sources of revenue. There are efforts, however, for some politicians to be close to the ruling government, and when this happens, the focus is not only on electoral votes but development projects that the central government can afford for the counties.

On administration and management of health facilities, health managers and patients in both Kisumu and Uasin Gishu county hospitals said that devolution has eliminated unnecessary bureaucracies that were bottlenecks to referral maternal health-care access. Some respondents said that the devolved health system had less barriers compared to the old centralised system:

I was referred to this hospital from a lower health facility. At first I thought I would be sent away the way it used to be. However, when I arrived, I was quickly attended to. I will happily return for the next appointment because I know I will be treated well. (Prenatal user, Kisumu County Hospital10)

Hospital management and personnel-related improved health-care access to devolved governance highlights issues such as ease of management of hospital activities without necessarily consulting the national government, easy planning and budgeting, quick decision-making, ability to prioritise key areas of urgent need, time management, better monitoring, supervision, more accountability and public participation. An official said:

In my view, there is improved governance in health-care provision in general because we can make decisions here. This is in spite of financial constraints necessitated by late disbursement of funds from the national government. At the county level we are able to quickly make decisions on what priority health issues to focus on and to allocate these funds without the red tape that characterised the situation before devolution. And of course county governments are more closer [sic] to the people and are therefore able to identify the really [sic] needs of their communities. (In-depth interview, Kisumu County Health Officer11)

Whereas the cost of maternity services is presumed to be free, user respondents mentioned that they have to cater for costs such as transport, cotton wool and drugs. A prenatal user said:

I was once detained at a hospital due to delivery charges. Now I am not worried about delivering in hospital because of high costs. However, I have to have some money for my fare and basic supplies such as cotton wool needed in preparation for delivery. (Referred prenatal user, Jaramogi Oginga Odinga Referral Hospital12)

Use of mobile phones to call taxis and the facility of mobile money transfers which have already been mentioned also help to make services affordable. With extended families and communalism relatively strong in Kenya, individuals can now quickly reach out to their relatives and friends for financial support. Boda boda transport is also cheap compared to taxis, charging about a quarter of what vehicles charge.

4.4 Acceptability
Devolved health services have also had an overall favourable perceptual influence on the acceptability of referral maternal health care. Acceptability is based upon the adequacy of services, infrastructure, equipment, commodities and human resources to local social and cultural expectations. The influx is one of the indicators of acceptability of devolved health care. Yet, the increase in users has not been matched with expansion of facilities and human resources, two factors that partly explain the negative perceptions of health providers regarding the impact of devolved health services on access to referral maternal health care.

While a good number of patients seemed satisfied with the referral maternal care offered in both Kisumu and Uasin Gishu counties, health-care workers at the national referral hospitals are not comfortable with the situation essentially because they are overworked. The proportion of doctors and nurses are important indicators of a county's capacity to provide adequate primary health-care coverage. With the deficits of medical personnel noted in Kisumu and Uasin Gishu counties, the workload remains high yet the number of patients has continued to increase. This breeds mixed reactions from patients and health-care workers on levels of acceptability of referral maternal health-care services. At the Riley Mother and Baby Hospital, for instance, the number of babies born daily has risen from an average of 30 to 60 babies but the staffing remains unchanged. While the nurses are stationed either at the emergency, labour ward, prenatal wards, NICU13 or postnatal ward, at both Uasin Gishu and Kisumu County Hospitals we observed nurses in charge of maternity wards attending to other stations such as emergency labour wards and postnatal wards. Some respondents noted that not only has there been an improvement in community health centres and clinics, but where services are not adequate, access to referral services to county and national hospitals is now easy. A respondent observed:

With devolution, a local dispensary in my village has been improved and maternal health services introduced, hence, I can secure prenatal services within my village on time. This time round I had complications with my pregnancy and I was quickly referred to Kisumu. (Prenatal user, Kisumu County Hospital14)

However, at Kisumu County Hospital focused observation revealed that patients overwhelmed primary care centres in both post-labour and labour wards. There was only one operation theatre, no ambulance dedicated to the maternal wing and the patients outnumbered the beds. We witnessed two or three mothers and their children at the postnatal ward sharing a bed with no mosquito net. One user respondent observed:

Whatever is hanging here as mosquito nets are mere decorations, they are torn and mosquitos have littered every corner of this ward. To make it worse, I have witnessed medical staff attending to us with borrowed equipment. (Postnatal user, Kisumu County Hospital15)

Devolved health services face a huge increase of patient self-referral, putting a lot of strain on staff. We heard that patients who refer themselves to county or national referral hospitals are never turned back because childbirth is always an emergency situation. But it would seem that the strain effect depends on the cadre of staff:

For the doctors, most of us who are teaching with Moi University, we like it when there are many patients because we want our students to learn. And as you know our learning method is problem based (PBL) so the more cases the better for us. Our students are able to encounter all the different kinds of emergencies that a mother or a neonate can have. Actually in this hospital you find more doctors than nurses. You can find that you are doing a ward round with six doctors but there is only one nurse. (Medical doctor, Riley Mother and Baby Hospital16)

According to the doctor, the solution lies in enhancing 'down referral' where expertise from the national hospital is posted to train staff in lower level hospitals and also in improving facilities and equipment and enhancement of supplies in lower level hospitals.

Finally, another negative implication of devolution is a deepening of cultural and religious differences within counties (KPMG International 2013): tribal discrimination of minorities was perceived to negatively affect access to referral maternal health services in Uasin Gishu county. The perceived discrimination emanates from respondents associating devolved governance with tribal conglomeration. As Cheeseman et al. (2016) observe, though the introduction of devolution was expected to diffuse Kenya's chronic ethnic conflicts, the findings of this study suggested that this is not the case. Yet, devolution in Kenya largely followed ethnic lines and as D'Arcy and Cornell (2016) observe elsewhere, some ethnic minorities within some of the counties feel marginalised. This may explain why some maternal health users perceived devolved health services to fan tribalism. Some respondents believed that the dominant tribe in a county receives better services than 'outsiders':

In these hospitals, there is ubaguzi wa ukabila [discrimination based on tribe]. For those of us who do not belong here [are not indigenous], devolution means that we do not get good services if we are not in our home counties. In those other hospitals, all the doctors and the nurses are local people. One feels out of place there. (In-depth interview at Riley Mother and Baby Hospital17)

5 Conclusion
There are clear differences in the general perceptions of the effect of devolved health services on access to referral maternal health care in Kenya. If at one end health workers and other providers seem discontented with devolved health services, at the other end there seems to be considerable satisfaction among users of referral maternal health services and their partners. Users largely perceived county governments to have improved access to referral maternal health care through a series of actions, namely: speeding up referral services, as well as by providing medicines and ambulance services that were previously not available (availability); improving transport to and communication with both county and referral hospitals (accessibility); and supporting the free maternity health-care service even when money is not remitted from the national government promptly (affordability). While the increasing number of referral maternal health patients is perceived to have increased tremendously, there are still some problems of acceptability, including the deepening of tribal discrimination along ethnic lines in provision of services.

Resource mobilisation at the county level and the opportunity for quick decisions have contributed significantly to the perceived positive effects of devolution in referral maternal health-care access in Kisumu and Uasin Gishu counties. However, improved access is partly related to developments that are not directly connected to the devolution of health services, such as free maternity services, the Beyond Zero Campaign, boda boda transport, and mobile telephony. Yet, perceptions matter, and based on that, devolution can be seen to have contributed to improving health systems, specifically in maternal health care. Besides, positive perceptions by users (who are citizens and voters) provide an opportunity for elected officials in Kenya to seek to maintain and even increase levels of user satisfaction.

Devolution is not devoid of challenges. Improved accessibility to referral maternal health care has increased pressure on equipment, commodities, infrastructure and personnel. This renders meaning to the discontent generally found among the health workers involved in this study. Discontent among health workers has been exacerbated by delays in fiscal remittances from national government. There are also perceptions that devolution has increased tribalism. In the study, we did not find any effort to mitigate this even though the perception is not unexpected given the broader Kenyan political context. Therefore, the challenge for policymakers in Kenya at both national and county level is not only to maintain and increase levels of user satisfaction, but also to simultaneously address the misgivings of health workers and mitigate against the perception that corruption and tribalism have increased with devolution. Better governance and accountability structures are critical for success.

Notes
* We wish to thank all respondents who took part in this study. We are grateful to the Partnership for African Social and Governance Research (PASGR) for a grant for this research, and to two anonymous reviewers and the IDS Bulletin team for detailed comments on our article.

1 The Constitution of Kenya (2010), www.kenyalaw.org.

2 The former (enacted in June 2013 through presidential declaration) encourages women to give birth at health facilities under skilled personnel, in keeping with the African Union resolution to exempt user fees for pregnant women and children under the age of five, and the latter (led by the First Lady since 2014) aims to promote maternal, newborn and child health – while controlling the prevalence of HIV/AIDS – through mobile clinics.

3 The research protocol was approved by the Institutional Research and Ethics Committee of Moi University's College of Health Sciences and Moi Teaching and Referral Hospital.

4 Interview, 28 May 2016.

5 Interview, 18 May 2016.

6 Interview, 27 May 2016.

7 Interview, 22 June 2016.

8 Interview, 23 June 2016.

9 Interview, 22 June 2016.

10 Interview, 28 June 2016.

11 Interview, 4 July 2016.

12 Interview, 6 July 2016.

13 Nursing Intensive Care Unit.

14 Interview, 29 June 2016.

15 Interview, 8 July 2016.

16 Interview, 28 May 2016.

17 Interview, 18 May 2016.

References
Baker, Catherine; Mulaki, Aaron; Mwai, Daniel and Dutta, Arin (2014) Devolution of Healthcare in Kenya: Assessing County Health System Readiness in Kenya. A Review of Selected Health Inputs, Washington DC: Health Policy Project, Futures Group

Bossert, T. and Beauvais, J. (2002) 'Decentralization and Health Systems in Ghana, Zambia, Uganda and Philippines: A Comparative Analysis of Decision Space', Health Policy and Planning 17.1: 14–31

Bourbonnais, N. (2013) Implementing Free Maternal Health Care in Kenya: Challenges, Strategies and Recommendations, Nairobi: Kenya National Commission on Human Rights (KNHCR)

Button, G. (2000) 'The Ethnographic Tradition and Design', Design Studies 21: 319–32

Cheeseman, Nic; Lynch, G. and Willis, J. (2016) 'Decentralization in Kenya: The Governance of Governors', Journal of Modern African Studies 54.1: 1–35

Communications Authority of Kenya (2016) Quarterly Sector Statistics Report. Fourth Quarter of the Financial Year 2015–2016 (April– June 2016), (accessed 5 January 2016)

Conyers, D. (2007) 'Decentralisation and Service Delivery: Lessons from Sub-Saharan Africa', IDS Bulletin 38.1: 18–32, (accessed 11 January 2017)

D'Arcy, M. and Cornell, A. (2016) 'Devolution and Corruption in Kenya: Everyone's Turn to Eat?', African Affairs 115.459: 246–73

Inkoom, D. and Gyapong, A. (2016) 'Decentralization in Africa: Local Government and Health Care in Ghana, Malawi and Tanzania', in C. Silva (ed.), Governing Urban Africa, London: Palgrave Macmillan

KPMG International (2013) Devolution of Healthcare Services in Kenya: Lessons Learnt from Other Countries, (accessed 12 July 2016)

Levesque, J.-F.; Harris, M. and Russell, G. (2013) 'Patient-Centred Access to Health Care: Conceptualising Access at the Interface of Health Systems and Populations', International Journal for Equity in Health 12: 18

Maoulidi, Moumié (2011) Health Needs Assessment for Kisumu Kenya, Millennium Cities Initiative Social Sector Working Paper Series 19/2011, New York NY: Columbia University

Masika, Moses (2016) 'Why Kenyan Doctors are on Strike and What Can be Done About It', The Conversation, (accessed 5 January 2016)

McIntyre, D.I.; Thiede, M. and Birch, S. (2009) 'Access as a Policy-Relevant Concept in Low- and Middle-Income Countries', Health Economics, Policy and Law 4: 179–93

Mehrotra, S. (2006) 'Governance and Basic Social Services: Ensuring Accountability in Service Delivery through Deep Democratic Decentralisation', Journal of International Development 18: 263–83

MoH (Ministry of Health) (2006) Reversing the Trends: The Second Rational Health Sector Strategic Plan of Kenya: Annual Operational Plan 2, Nairobi: Government Printers

MoH (Ministry of Health) (2002) Guidelines for District Health Management Boards, Hospital Management Boards and Health Centre Management Committees, Nairobi: Government Printers

Mwamuye, M. and Nyamu, M. (2014) 'Devolution of Health Care System in Kenya: A Strategic Approach and its Implementation in Mombasa County, Kenya', International Journal of Advanced Research 2.4: 263–8

Naeem, I.; Nafees, A.; Zahidie, A.; Fatmi, Z. and Kazi, A. (2012) 'Experience of Devolution in District Health System of Pakistan: Perspectives Regarding Needed Reforms', Journal of Pakistan Medical Association 62.1: 28–32

Oyugi, O. (2015) 'Potential Impact of Devolution on Motivation and Job Satisfaction of Healthcare Workers in Kenya: Lessons from Early Implementation in Kenya and Experiences of Other Sub-Saharan African Countries', Journal of Global Health Care Systems 5.1: 1–30

Penchansky, R. and Thomas, J. (1981) 'The Concept of Access: Definition and Relationship to Consumer Satisfaction', Medical Care 19.2: 127–40

Peters, D.; Garg, A.; Bloom, G.; Walker, D.; Brieger, W. and Rahman, H. (2008) 'Poverty and Access to Health Care in Developing Countries', Annals of the New York Academy of Sciences 1136: 161–71

Robinson, M. (2007) 'Does Decentralisation Improve Equity and Efficiency in Public Service Delivery Provision?', IDS Bulletin 38.1: 7–17, (accessed 11 January 2017)

UNFPA (2014) Counties with the Highest Burden of Maternal Mortality, (accessed 26 October 2016)

World Bank (2016) Doing Business in Kenya, Washington DC: International Bank for Reconstruction and Development

CCBYNC

© 2017 The Authors. IDS Bulletin © Institute of Development Studies | DOI: 10.19088/1968-2017.118

This is an Open Access article distributed under the terms of the Creative Commons Attribution Non Commercial 4.0 International licence, which permits downloading and sharing provided the original authors and source are credited – but the work is not used for commercial purposes. http://creativecommons.org/licenses/by-nc/4.0/legalcode

The IDS Bulletin is published by Institute of Development Studies, Library Road, Brighton BN1 9RE, UK

This article is part of IDS Bulletin Vol. 48 No. 2 March 2017: 'Interrogating Decentralisation in Africa'; the Introduction is also recommended reading.