Article Text
PDF
Abstract
In Nigeria’s federal government system, national policies assign concurrent healthcare responsibilities across constitutionally arranged government levels. Hence, national policies, formulated for adoption by states for implementation, require collaboration. This study examines collaboration across government levels, tracing implementation of three maternal, neonatal and child health (MNCH) programmes, developed from a parent integrated MNCH strategy, with intergovernmental collaborative designs, to identify transferable principles to other multilevel governance contexts, especially low-income countries.
National-level setting was Abuja, where policymaking is domiciled, while two subnational implementation settings (Anambra and Ebonyi states) were selected based on their MNCH contexts. A qualitative case study triangulated information from 69 documents and 44 in-depth interviews with national and subnational policymakers, technocrats, academics and implementers. Emerson’s integrated collaborative governance framework was applied thematically to examine how governance arrangements across the national and subnational levels impacted policy processes.
The results showed that misaligned governance structures constrained implementation. Specific governance characteristics (subnational executive powers, fiscal centralisation, nationally designed policies, among others) did not adequately generate collaboration dynamics for collaborative actions. Collaborative signing of memoranda of understanding happened passively, but the contents were not implemented. Neither state adhered to programme goals, despite contextual variations, because of an underlying disconnect in the national governance structure.
Collaboration across government levels could be better facilitated via full devolution of responsibilities by national authorities to subnational governments, with the national level providing independent evaluation and guidance only. Given the existing fiscal structure, innovative reforms which hold government levels accountable should be linked to fiscal transfers. Sustained advocacy and context-specific models of achieving distributed leadership across government levels are required across similar resource-limited countries. Stakeholders should be aware of what drivers are available to them for collaboration and what needs to be built within the system context.
- Maternal health
- Health policies and all other topics
- Qualitative study
Data availability statement
Data are available upon reasonable request.
This is an open access article distributed in accordance with the Creative Commons Attribution Non Commercial (CC BY-NC 4.0) license, which permits others to distribute, remix, adapt, build upon this work non-commercially, and license their derivative works on different terms, provided the original work is properly cited, appropriate credit is given, any changes made indicated, and the use is non-commercial. See: http://creativecommons.org/licenses/by-nc/4.0/.
Statistics from Altmetric.com
WHAT IS ALREADY KNOWN ON THIS TOPIC
Healthcare is a concurrent responsibility of three levels of government in Nigeria. However, policies and programmes have predominantly been designed and implemented with a top-down approach. These designs have not fully fostered cross-jurisdictional collaboration.
WHAT THIS STUDY ADDS
Misaligned governance structures in Nigeria do not firmly support its concurrent healthcare responsibilities. In multilevel governance contexts, cross-jurisdictional collaboration for health policy processes requires adequately distributed leadership and explicit policy and fiscal mandates. Mismatched administrative and fiscal decentralisation constrain policy processes where there are inadequate intergovermental accountability measures to foster collaboration.
HOW THIS STUDY MIGHT AFFECT RESEARCH, PRACTICE OR POLICY
Adequate measures that facilitate cooperative federalism need to be instituted in multigovernance contexts to enhance intergovernmental policy processes, especially in weak health systems of some low-income and middle-income countries.
Introduction
The WHO conceptualises health system governance as involving the combination of strategic policy frameworks with effective oversight, coalition-building, regulation, system-design, and accountability.1 There is much debate about administrative decentralisation of governance and how this affects health systems in low-income and middle-income countries (LMICs), particularly in Africa.2 Given the colonial history, many countries in Africa have struggled with adapting suitable governance systems acceptable to their diverse populations, as well as able to optimise governance.3 Thus, many countries have found themselves evolving through authoritarianism (often unitary) to democracy, with varying degrees of application of the principles of decentralisation.4 These systems of governance have direct implications for health service delivery.
Administrative decentralisation has been classified into three forms: deconcentration, delegation and devolution.5 Deconcentration redistributes decision-making power by appointing representatives of the central authority to administer subnational regions. With delegation, decision making is transferred from central authority to semiautonomous institutions like the ministry of health. Devolution occurs when central government cedes decision-making powers to autonomous subnational or local authorities, which typically have legally defined geographical boundaries and independent fiscal administrations.5 6 In this article, except where clarified, our reference to decentralisation is with respect to devolution.
Collaborative governance (CG) broadly describes cross-sector (or intersector) public policy decision making and execution across boundaries of governments and public agencies where this cannot be achieved otherwise.7 8 CG has long gained ground in Europe and the USA, predominantly in the environmental sector,9 but is now being increasingly applied to address challenging and complex public health issues10 and also used to explore different levels of governance within countries.11 12 CG is further elucidated in the Methods section. In this study, we refer to intersectoral collaboration across the national and subnational governments as intergovernmental collaboration.
In most LMICs, maternal, neonatal and child health (MNCH) outcomes are partly determined by how the health system is structured and governed.13–18 Brazil’s progress in improving MNCH in the last three decades was partly attributed to decentralised governance, which allowed policies to be tailored to the MNCH needs of states.18 19 In India, variations in outcomes of subnational MNCH policy implementation were attributed to decentralisation reforms.13 Other examples from other countries, including Nigeria, show that decentralisation impacts policy implementation,20–22 although we are still learning about the mechanisms by which this happens.
Nigeria is administratively decentralised into three levels: national level, mid-level (36 states and federal capaital territory and 774 local government areas (LGAs), and grouped into six non-administrative zones. The mid-level (states), presided over by elected state executive governors, are the federating units. State governments have significant autonomy to adopt, reshape or reject national policies, particularly with respect to healthcare; they are also not constitutionally prevented from making state-level policies. Thus, states are often the units of analysis in evaluation of national policies.23–26 However, fiscal administration in Nigeria is significantly centralised, such that revenues from taxes, and key sectors such as oil and gas, are received centrally, with a percentage (approximately 43%) redistributed subnationally according to formulae developed by an agency of the central government.25 The LGAs receive the lowest revenue allocation from the central government, channelled through a joint account with states and disbursed to LGAs at the discretion of the states. Although this state influence over LGAs varies from one state to another, with some states more accountable than others,27 the resultant lack of fiscal autonomy has, over time, impacted the administrative capacity of LGAs28–30 who are assigned responsibility for primary healthcare (PHC) services, where the bulk of MNCH services are accessed in the public sector. The three national-level programmes under study were interventions designed partly to mitigate these shortcomings.
The National Council on Health (NCH), which includes state commissioners of health and is presided over by the federal minister for health, is the highest national health policy-making body in Nigeria. However, state governors are the ultimate policy decision makers at the state level. Beyond policies, state laws are made by the state legislatures and assented to by state governors. This subnational political decision-making structure overrides any national-level decisions taken by the NCH.31–33 Thus, state governments have powers to determine their healthcare priorities irrespective of the NCH. States are also not accountable to the federal government regarding healthcare spending. This limits the national government’s power over subnational implementation of the national policy agenda, resulting in a misalignment that impacts health policy processes.32 34–36
Nigeria registers persistently high rates of maternal and child morbidity and mortality37–39 and significantly contributes to the global burden.40 In 2007, the Federal Ministry of Health (FMoH), with key development partners, collaboratively integrated existing fragmented MNCH policies41–44 into a single policy strategy to reduce maternal, newborn and child morbidity and mortality within the framework of the National Health Sector Reform Programme (2003–2007).45 This integrated maternal, neonatal and child health (IMNCH) strategy was ratified by the NCH for subnational implementation across the country. Three interventions, implemented between 2009 and 2019, were among the primary mechanisms by which the strategy was meant to be executed: the Midwives Service Scheme (MSS), the Subsidy Reinvestment and Empowerment Programme for Maternal and Child Health (SURE-P MCH) and Saving One Million Lives Programme for Results (SOML PfR). These three programmes were selected for the present study because they were consecutively implemented in every state, all year round between 2009 and 2019. There are other MNCH programmes based on the strategy, notably the National Health Insurance Scheme–Millennium Development Goals (MDG)–Free Maternal and Child Health Programme, which was implemented in only 12 states, and the MNCH Weeks, implemented twice a year.46 47 While we may draw some comparisons, these other programmes do not form part of this study.
To be effective, the overarching strategy and all three study programmes required collaboration across government levels, and indeed all three clearly outlined collaboration aspirations and intent in their design.48–51 Specific roles and responsibilities, within the collaboration, were outlined for each government level (online supplemental file 1), and the strategy and programmes had collaborative aspirations in their goal, funding sources and other policy components (online supplemental file 2).
Supplemental material
Drawing insights from these policy processes, this paper examines Nigeria’s experiences of intergovernmental collaboration with respect to MNCH, which remains a priority on the national health agenda. This paper aims to elucidate the processes and mechanisms of translating national policies for subnational implementation and unpack the impact of the governance arrangements in Nigeria’s federal decentralised system on health policy processes. It is an overview paper, the first in a series of reflective pieces based on the first author’s research thesis findings.52 A subsequent paper will critically analyse and compare the role of states in bringing about the mandates of the IMNCH strategy through the study programmes. A third paper will specifically focus on actors and roles and relationships in these processes.
Methods
Theoretical and analytical framework
Initially guided by the policy stages heuristic in data collection and analysis,53 this study frames the Nigerian health system as a multilevel governance (MLG) system29 and applies the MLG lens to situate the challenges of federalism and decentralisation. Within this system, policy processes straddle government boundaries. Therefore, degrees of participation and inclusivity during various stages of the policy process impact adoption and implementation. Emerson et al propose that one or more drivers—leadership, consequential incentives, interdependence and uncertainty—are necessary to initiate a collaborative governance regime (CGR), in interaction with the system context. A CGR comprises the collaboration dynamics and the collaborative actions brought about by these dynamics. Collaboration dynamics are made up of the iterative relationship between three concepts: principled engagement (wide and inclusive stakeholder engagement) and shared motivation (mutual understanding and trust building), which generate joint capacity for action. The more drivers present, the more likely a CGR will be initiated. The form and direction of the CGR is shaped initially by drivers that emerge from the system context; however, development of the CGR, as well as the degree to which it is effective, is influenced over time by collaboration dynamics and collaborative actions.7 These are crucial especially in federal entities where healthcare is not restricted within one jurisdictional level or within one organisation because of its complexities, which usually include multi-level challenges.54 For our analysis and discussion, we apply the Emerson framework as shown in figure 1.
Analytical framework adapted from Emerson et al.7
Study design and setting
The study was conducted in Abuja, federal capital territory (national level), and in Anambra and Ebonyi states, in Nigeria’s southeast zone. These states were purposefully selected to compare subnational implementation experiences of national programmes based on variations in their socioeconomic and MNCH contexts, but with similar subnational governance and political structures. Populations of both states are predominantly Christian and of Igbo ethnicity, and both are largely dependent on the national level for fiscal transfers.33 55 Anambra state has a population of 4.2 million, including 1.1 million women of childbearing age, while Ebonyi state has a population of 2.2 million, of which 567 757 are women of childbearing age. Anambra consistently maintained a higher socioeconomic and MNCH index through the study period, 2009–2019.37 39 Each state defines its health priorities in a 5 yearly state strategic health development plan, guided by a national strategic framework.33 55 Detailed and critical comparative analysis of contexts and MNCH policy implementation across both states and the role of actors, as explored in the larger thesis,52 are the foci of other papers.
Sampling and data collection
This study was guided by the Consolidated Criteria for Reporting Qualitative Research.56 Study respondents were drawn from national level, zonal level and subnational level (table 1). Sampling was purposive of people who had been or still are involved with various stages of the MNCH policy process to include policymakers, legislature, programme managers, development partners, civil society organisations, non-governmental organisations and media.
Summary of data sources
Documents reviewed (online supplemental file 3) include national-level and state-level policies, programme evaluation reports and relevant global documents from Google Search engines and PubMed searches, websites of the FMoH and National Primary Health Care Development Agency (NPHCDA), peer reviewed publications and grey literature from government organisations and snowballing from document references. Document reviews were conducted by EE,1 supervised by UL throughout the process.
National-level data were gathered in Abuja. At the zonal level, respondents included those who had oversight of both study states. In the study states, sampling included health commissioners, programme managers, key informants (KIs) in the State Ministry of health (SMoH) and parastatals (State Primary Health Care Development Agency, State Health Insurance Agency, Ministry of Finance, local government (LG) MNCH coordinators and state-level development partners. All respondents were either involved in policy development, implementation or advocacy for MNCH.
In-depth interviews (n=19 national, n=25 subnational) were conducted by EE1 and two research assistants. Initial pilot exploratory interviews (n=3) between October and November 2018, which, with relevant documents, guided the development of the key interview guides. Guides were continuously adapted to new information and to the respondent category.
The initial round of data collection also helped build a picture of relevant themes for analysis. Further documents and respondents were then identified in an iterative manner. Other interviews were concluded between May and August 2019. All interviews were conducted in English language and lasted between 40 min and 70 min.
Data analysis
The document review involved tracing and mapping how MNCH policies evolved over the past decade, with a focus on background, contextual factors, changes over time, actors, relationships, policy design and characteristics, policy implementation outcomes, collaboration intents and collaborative processes, and how these shaped the policy process. Interviews were transcribed verbatim, and transcripts were organised with NVivo qualitative analysis software V.11. Analysis of transcripts were carried out by EE1 and UL. Familiarisation with the data was followed by organising the data sets into themes. The Emerson’s CG framework (figure 1) was applied to categorise and organise the themes (contextual drivers, stakeholder engagement, shared motivation, capacity for joint action, policy implementation actions and outcomes) at national and subnational levels, iteratively triangulating information from interviews and documents.
Patient and public involvement
Patients were not involved in this study. Reports of the study were presented at a national seminar to study participants and stakeholders.
Results
In this section, we present our findings of the experiences from the different stages of the policy process, in line with the CG framework. First, we identify the macro and micro system context and drivers, then specific programme context and drivers and describe observed collaboration dynamics.
Macrosystem context and drivers
Key international contexts were the then MDGs and the push for Nigeria and other high MNCH burden countries to achieve MDGs 4 and 5.40 The debt relief (2005) and World Bank’s provision of financial support for the SOML PfR in 2015 were key international economic drivers.
At national level, the key broad national contexts are the governance, political and economic arrangements in the country. Full administrative decentralisation constitutionally gives states autonomy over national-level policies. Politically, national and subnational elections are held every 4 years, partially or fully renewing the politically elected legislature and politically appointed health sector actors. Economically, fiscal resources (mainly from oil) are centralised. In addition to these macro-contextual factors, specific programme contexts and drivers influenced the MNCH policy process during the study period, as described further. Table 2 outlines findings of intergovernmental collaboration dynamics, collaborative actions, and outcomes in the IMNCH strategy and programmes.
Intergovernmental collaboration dynamics,collaborative actions and outcomes in the IMNCH strategy and programmes
Programme contexts and drivers
A positive driver for CG was the ongoing health sector reform,45 which formed a framework for developing the parent IMNCH strategy.
Financial arrangements for MNCH programming were also an important contextual factor. Dedicated funds for MSS, SURE-P MCH and SOML PfR were held separately from national-level and state-level health budgets. MSS was largely funded from MDG resources, pooled at the national MDG office at the presidency. MDG funds allocated to states were also pooled at MDG offices of the state government house, for all sectors.57 For SURE-P MCH, funds from oil subsidy removal were domiciled at the SURE-P secretariat at the presidency. Similarly, SURE-P funds allocated to states were held at the state SURE-P secretariat for all sectors.57 In contrast, SOML PfR funds from the World Bank were clearly earmarked, held at FMoH and disbursed directly to SMoH according to programme design, and it was hoped that this would also strengthen federal-state fiscal federalism.51 These funds, however, were disbursed at the discretion of the state executive governors.
The overarching mismatch of fiscal and administrative decentralisation, where states have administrative autonomy but fiscally dependent on the national level, also constantly interacted with 4-year election cycles, which come with new actors and new political dynamics. Across these cycles, MNCH remained high on Nigeria’s health agenda, which some respondents reported was more in response to pressures from global context:
…advocacy efforts around maternal mortality rates in Nigeria which got O’s (then president) attention …there was also a global push…and a sort of a global awareness of Nigeria’s weakness, so it became one of the factors in negotiations around the Paris Club Debt relief…. At the WHA [World Health Assembly], UN General Assembly when Nigeria was mentioned repeatedly and got O. uncomfortable…
Collaborative governance regime
Collaboration dynamics (principled engagement, shared motivation and joint capacity) and collaborative actions make up the CGR. This section presents identified collaboration dynamics (or absence of them) and what collaborative actions or inactions were observed as a result.
Collaboration dynamics (IMNCH strategy, MSS and SURE-P MCH)
Shared motivation
In the early stages, during development of the parent IMNCH strategy in 2007, there was a shared motivation among policymakers to meet the MDGs, which was further enabled by the ongoing health sector reform initiative. The strategy included a structure for collaboration across government levels and partnerships with other sectors. A clear structure for proposed roles and responsibilities of all government levels, private sector and development partners was outlined.48 58
The perceived inclusive leadership of the then health minister who led the development of the IMNCH strategy was also key in initiating collaboration dynamics:
He had that outlook…you have to do very broad consultation to both primary and secondary stakeholders, because any stakeholder that will either benefit or be injured by your policy ought to be heard during policy formulation, and that is what we were doing…! (NL06_national-level policymaker/technocrat).
Principled engagement
Deliberate wide stakeholder engagement and consultation facilitated the adoption of the framework strategy by subnational levels, which was then used to develop guidelines and plans for subnational implementation, as noted by this KI:
…for [the] state to implement, they didn’t need to change because states were also part and parcel of development. When we were developing it, all state people came, commissioners, directors of the state health care board were there with all members, everything that happened within that strategy, they were all part of it, so even when we did state visits, they were there, that momentum was everywhere, at the state level (NL11_national development partner).
This level of principled engagement observed in the early stages were not sustained when policy goals were translated to subnational levels. Following the launch of the Strategy, subnational levels were to take ownership of the programme and roll-out interventions in their respective states and LGAs. As noted in the strategy document:
When states indicate readiness to commence implementation of the Strategy, National Technical Team will provide them support for formation and training of a Technical Team which will be responsible for rolling out the process to LGAs, wards and communities. Each State Technical Team will guide the development of LGA implementation plans…. (Federal Ministry of Health, p65)48
Since the national level does not have jurisdictional authority over states, states rolled out implementation at different times. An annual evaluation survey of the strategy in 2009 showed that only 23 states had requested implementation support from FMoH, and at the ime of evaluation, FMoH had visited 18 of these.58 At this point, initial cracks in the collaborative intent started showing up.
Designs of subsequent programmes (MSS and SURE-P MCH) were reportedly not as widely consultative as the strategy. There was no explicit subnational stakeholder engagement in the early stages, as respondents reported at both national and subnational levels,
I can't remember being involved in the design of MSS, and the fact that some of us at the state level were not involved in the planning introduced some kind of disconnect in execution … (SLA03_subnational policymaker/technocrat)
while a national-level stakeholder reflected:
Looking back now, I will say that the state was not totally involved. If it is now, I would involve them more. It was later, and it was almost like a done deal. We did go for advocacy visits and all that, but it wasn’t that intensified in MSS… (NL03_national-level implementer)
Joint capacity
States and LGAs adopted programmes by passively signing memoranda of understanding (MOUs), without adequate intent to implement:
…whatever you do at the national level and send across, no state will reject it. There is one document I have where all 36 state governors signed including the President and Vice-President. I treasure that document because I use it to teach how much you can get chief executives of states to do something and then at the end of the day, it doesn’t still work…we all know. So, signing of MOU or going into agreement doesn’t translate into desired action because we don’t have accountability mechanisms in place in Nigeria, where people are held accountable for anything they sign… (SLE04_subnational academic/policy broker).
Subnational stakeholders conceded that programme goals of MSS and SURE-P MCH were important to implement to improve MNCH outcomes. However, they also pointed out that programme logistics designs (in MSS and SURE-P) were difficult to implement subnationally due to the costs associated with remunerating health workers by state and LGs, providing accommodation for midwives and community health extension workers, supporting PHCs, and retaining health workers after initial phases of MSS and SURE-P MCH, as outlined in the programme MOUs. So, while there was agreement with the policy intent, inadequate stakeholder consultation, as well as policy content and design, undermined chances for implementation success.
States and LGAs were required to pay programme staff a proportion of their monthly salaries, but this was not backed by any dedicated programme funds and not provided for by states in their health budgets. These factors undermined subnational adoption and implementation of these policies.
National level also intended that after the 2-year pilot phase, MSS would be taken over, owned and sustained by states and LGAs. However, there was no explicit design path in the document as to how this was going to happen, nor how it was going to be funded:
The plan was for most part wishful thinking on the part of federal government. Part of the founding idea of MSS was that federal government would pull out and state government would take control, that was the rhetoric, and it continued until SURE-P materialized. But if SURE-P hadn't happened, it [MSS] would have died a natural death in the name of handing over to state government (NL01_national implementer/academic).
In SURE-P MCH, which also intended a collaborative approach, state and LGA stakeholders were, again, not adequately engaged in the hastily done design, as acknowledged in the concept document (FMoH, NPHCDA, p.5)50 SURE-P MCH was an unexpected opportunity to carry on with MSS activities, following the oil subsidy removal in January 2012.
So, transiting to SURE-P (MCH) was like a soft landing but it was supposed to learn some lessons from MSS so that it will be tidier… (NL05_national implementer)
Perhaps, lessons learnt in MSS were not adequately incorporated into SURE-P MCH. For example, similar MOUs, which did not adequately hold the subnational levels accountable,49 50 were drawn up for both programmes, although states and LGAs had not honoured the MOU for MSS. These unintended consequences resulted in poor initiation and sustainability of collaboration dynamics, which were insufficient to bring about intended collaborative actions.
Collaboration dynamics SOML PfR
Shared motivation
The third programme the SOML PfR was developed and implemented after the MDGs. The original SOML (not the SOML PfR) was conceptualised around the same time as the SURE-P MCH, as an umbrella concept (rather than a programme/policy) to house the MSS, SURE-P MCH and other MNCH activities/programmes, which it was hoped would collectively save one million lives by the end of the 4-year tenure of the government at the time (2015) and, hopefully, meet MDGs 4 and 5.59–61
At the end of SURE-P MCH, which was also the end of the MDGs, through the efforts of the Minister of State at the time, the SOML concept evolved into the SOML PfR, a results-based programme with funding from the World Bank.51 By the time the SOML PfR was designed (2015-post MDGs), it had become clear that the wide policy-implementation gaps were partly due to the design of the MSS and SURE-P MCH, in the existing governance structure, which did not foster adequate collaboration as intended. These lessons are reflected in SOML PfR’s emphasis on results, as evident in the design document:
…availability of many of the needed inputs…suggest that governance… is the binding constraint, As a response, FMoH introduced an innovative financing mechanism which we hope will address observed challenges … the PfR will help strengthen fiscal federalism and encourage the Federal-State relationship….51
With this intent, 82% of SOML PfR funds were devolved to states, in recognition that implementation occurs at state level and through health facilities where services are delivered to patients, although via federal programmes.51
Principled engagement
The design of SOML PfR further facilitated a collaborative approach, given that activity of states developing their own specific programme guidelines was one of the programme output indicators, with an attached incentive.51 As a subnational KI reflected,
We had to start at state level to get the work done. We were asked to meet severally, we put down a work plan and sent to national and there were corrections, they even sent us program support unit. They were here to make us know how they really want SOML, that is performance for result… and finally it was approved, that was when we qualified to receive the money (SLA01_subnational implementer).
Joint capacity
Implementation of SOML PfR was, however, constrained by subnational governance structures, which give governors executive powers over all sectors, hence programme funds, and by inadequate accountability mechanisms regarding these funds. So, ring-fenced programme funds still required governor’s consent for expenditure. Programme funds were sometimes extended to other health sector projects, at the insistence of the executive governors:
To make the Saving One Million Lives functional…, we develop a work plan which if national approves, it is returned, then you start implementing. But you find out that occasionally you might send a program, and it may not go down well with the governor, and he will slot in his own program, which was what happened to us (SLE02, subnational policymaker/technocrat).
As a result, there were misprocurements of equipment like tricycle ambulances and generators, by states, which, although essential, were not direct SOML PfR components:
When we got our own (seed money), instead of applying them in these indicators, we went and bought keke na pepe (tricycle). At the end of the day, our indicators did not change, we lost the money because we decided to boil our seed and eat… if we are honest, that we are not paying a priority attention to maternal and child health (SLE_09, senior executive/technocrat).
Collaborative actions (IMNCH strategy, MSS and SURE-P MCH)
The collaboratively formed national-level and state-level core technical groups facilitated smooth translation to states that where ready to roll out implementation. However, not all states rolled.58 Collaborative aspects built into the IMNCH strategy did not account for leadership changes that necessarily occur as part of election cycles. When development, approval and launch of the strategy was happening in 2007, Nigeria was involved in national elections, which ushered in new actors—ministers, governors, commissioners, etc. The transition may have contributed to the delay in (or abandoning of) roll-out and pre-implementation activities of the strategy in some states:
There were visits to most states and meetings with governors, but the problem was that there was a new administration. So when new people came on board, that advocacy and momentum was not there… they felt it wasn't their mandate, because they had their own arrangement of what they planned to do… Some states bought into it, but some did not… (NL11_national development partner).
In the MSS and SURE-P MCH, MOUs were passively signed by all levels. Other intended collabotive actions (staff recruitment and deployment, selection of intervention facilities and staff remuneration) were either partially or not collaboratively implemented at all (table 2).
States were not involved in the recruitment and deployment of programme staff and selection of intervention facilities. As some respondents reflected,
…we didn’t have the same involvement. What I know is that the MSS was like a scheme trying to bridge the gap in the manpower of midwives in the state.]hey told us that they have posted the midwives, and we were wondering, how can you post midwives to the states without going from the ministry but direct to the facilities and then we will be required to seek to know them? We started asking those questions, at every presentation we asked these same question,…where are they, how are they paid, how do we know where they are so that we can monitor them? (SLE01, subnational policymaker/technocrat).
…The recruitment procedure was a bit faulty. They just get the list, fly in, at times they will even post them from Abuja to facilities where they don't know if they exist (SLE_03, senior executive/technocrat)
There was a similar experience with the SURE-P MCH programme:
…in one of our monitoring, integrated supportive supervision (ISS) we went at U (facility), I saw people who introduced themselves as SURE-P staff. I asked him what he does, he said he is a health educator… So, just like the MSS, I never really knew the impact of that. MSS was even more specific because all the people there were midwives. But for SURE-P, we don’t know who they were or what their professional qualifications were and the work they were supposed to do (SLE01_subnational policymaker/technocrat)
States also resisted criteria set out by national level for identification of intervention facilities, hence constraining collaboration. According to a national-level KI,
…when they gave us the list of facilities, we sent people to go and check. Those that did not meet the required criteria/population were identified and the states were advised/convinced to change it. We recommended that they change those ones; but you find out that in some few places where we asked them to change the facilities, they did not. Instead, they will tell you that the powers that be want it to be there. In that case we don’t fight. We try to convince them and some of them will change it at the end of the day, some will not. (NL04_national implementer)
The proportion of the staff remuneration that was the responsibility of the states and LGAs, as per the MOU, was either haphazardly paid or in most cases, not paid at all:
I cannot vouch to what extent either party kept to these agreements especially at the state level, I think they were getting the national part of their remuneration, but that of the state was not very regular, or maybe they were not paid at all. (SLA_03, state executive/technocrat)
A national-level implementer also buttressed this:
None of them (Anambra and Ebonyi states) paid. It is only the local government that paid N10,000 from the Ministry of Local Government and Chieftaincy Affairs. The local government no doubt is more responsive to PHC than the state, but the problem is, currently their (LG) money is now getting stuck at the state and I don’t know what they are doing in that area to get their money back… (NL04, national implementer)
States and their LGAs did not adequately provide accommodation for programme staff, as per the MOU:
…We tried to give them accommodation but not necessarily the ideal accommodation that was envisaged because at times we had to pair two in a room… A few will accept and others will say no, so we were not able as the local government to provide accommodation (SLA_10_LG executive/technocrat)
Programme staff were not explicitly retained at the end:
…the state government in Anambra I’m aware absorbed some of them but I don’t think it was direct absorption because they still had to go through the interviews other people who applied for the job went through then. Maybe they were successful because they were already in the state, and they seemed to know their way better. I never saw anyone that said, ‘I was taken because I was doing MSS’…so, beautiful programme, poor continuity…you sign an MOU, but you are not enabled. (SLA_10_LG executive/technocrat)
In Ebonyi state, the opinion was
…from the government perspective, you cannot have many people from your state that are unemployed, and you retain people from all over the place… (SLE_03, senior executive/technocrat)
Collaborative actions: SOML PfR
In addition to signing of MOUs, a number of other activities on the SOML PfR were fully collaboratively implemented, although programme fidelity was constrained by the subnational government decisions. During implementations, both states deviated from the collaboratively designed activities and goals. Anambra state initially started implementing the programme disbursement linked indicators, but because they were already highly performiing in some of these indicators prior to the programme, they feared they would not meet the required upward change in the indicators:
…we selected a few facilities, gave them seed money but we didn't really continue because it wasn’t like we expected…We were not getting the outcomes we needed from them. So, somehow, we stopped that… (SLA01_subnational implementer).
As explicitly put by another KI,
…now if somebody improves by 0–10%, he wins the extra money but if you are already 93%, like in Anambra state immunization you get nothing… some other states like for instance in immunization, Kebbi state, the other time improved from 5% to 55% and they were awarded the best in the country. Of course, you know that to move from zero to the first 50 is very easy but when you begin to score excellently…it becomes very difficult and that was what we were not very happy about from the beginning. We complained… (SLA07_state executive/technocrat).
Ebonyi state also explicitly deviated from implementing activities for programme indicators, because of government decisions:
To make the Saving One Million Lives functional…we (the PMU) …sit down together, we develop a work plan which will be shown to the honourable commissioner, if it is approved, then you now send it to the national; if national approves then it is returned, then you start implementing. But then you’d find out that occasionally you might now send a program like that, and it may not go down well with the governor, and he will slot in his own program which was what happened to us. (SLE02, subnational policymaker/technocrat).
Hence, Ebonyi state embarked on misprocurements (tricycle ambulances and electricity generators) that were not direct components in the programme design.
Outcomes
As a result of inadequate accommodation in some intervention facilities, the staff deployed to rural areas took turns to work, rather than work together as per programme design (two midwives were supposed to run a shift together) but worked alternate days because they had to commute long distances to the health facilities. In some LGAs, staff did not report to duty or exited before the end of the programme. Dissatisfaction with accommodation facilities and poor remuneration contributed to staff attrition62 and, as also reflected by a KI programme goal, were not sustained nor scaled up in the MSS and SURE-P MCH. MDG goals 4 and 5 were not achieved.
In the SOML programme, there was explicit deviation from programme targets and activities, resulting in less than expected outcomes,63 and reflected by a KI:
When we got our own (seed money), instead of applying them in these indicators, we went and bought keke na pepe (tricycle). At the end of the day, our indicators did not change, we lost the money because we decided to boil our own seed and eat…if we are honest, we are not paying a priority attention to maternal and child health… (SLE_09, senior executive/technocrat).
In summary, idenitified contextual factors were not sufficient to generate adequate collaboration dynamics for aspired collaborative actions. The initial robust stakeholder engagement during the parent strategy development in 2007 waned thereafter. Persisting top-down institutional designs and power imbalances, which had led to prior conflicts and mistrust, constrained shared motivation. Hence, capacity for joint action was not adequate to collaboratively achieve programme goals.
Discussion
Despite a strong overarching policy framework (the IMNCH strategy), substantial resourcing and strong national leadership, policy implementation and outcomes fell short of expectations for the three programmes examined in this study.57 63 64 Aspirations for a collaborative MNCH policy process across levels of governance have not worked optimally. Early stages of the strategy development showcased significant collaboration across government levels but were not sustained. However, existing system context and drivers, misaligned governance structures, and mismatched fiscal and administrative capacity and responsibilities combined to undermine collaboration dynamics, which were insufficient to drive collaborative actions between government levels. Given Nigeria’s constitutional governance structure, collaborative actions did not have adequate checks and balances to overcome existing contextual barriers to collaboration, notably at the level of subnational executive powers. Consequently, policy outcomes were far from intended goals: at the process level, staff were not adequately integrated and remunerated and not retained at the end of programmes. In terms of policy impact, there were no appreciable improvements in MNCH outcomes (figure 2).
Summary of collaborative governance in the MNCH strategy and programmes. LG, local government; MNCH, maternal, neonatal and child health; MSS, Midwives Service Scheme; SOML PfR, Saving One Million Lives Programme for Results; SURE-P MCH, Subsidy Reinvestment and Empowerment Programme for Maternal and Child Health; SDG, Sustainable Development Goals.
Fiscal/administrative mismatch is common in other postcolonial LMICs, where federal authorities can retain paternalistic attitudes and be unwilling to share power with subnational entities.3 24 65–68 Identified system contexts in Nigeria share similarities with India, which also has poor MNCH outcomes (with wide interstate variations) and a three-tier (federal govermnent, state govermnent and LG) government system. Health is also a constitutional concurrent responsibility of the three tiers in India. However, given persistent poor performance in health outcomes, in the last decade, a national-level policy sought amendments to empower the LGs to be more actively involved in public health administration. States like Kerala and others, which have complied and taken steps to empower LGs, devolved funds and responsibilities to LGs, have registered positive gains, although there has also been resistance from several states to devolving funds to their LGs.13 69 Experiences from Kenya show that devolution of healthcare policies (MNCH and others) led to increased county-level decision space after initial challenges of political interference and lack of clarity in roles of national and county level actors with respect to health system functions.70 71
However, in Nigeria, when national initiatives and policies fail to translate directly to subnational level,32 this is less a result of inadequate fiscal capacity than of a lack of the will of states as federating units to shift their positions and willingly collaborate with the national level. Health policies are perceived as top-down on the one hand and there is also inadequate support for state level health policies, as reflected in the number of other interventions with state level ownership and support that have been implemented in one of the study states72 and in other states of the country. These include MNCH73 74 and non-MNCH interventions.27 These programmes, although successfully and elaborately initiated, were also not sustained due to poor funding by successive state governors.
The reverse was originally the case in Brazil, where fiscal decentralisation to subnational government left the national government with funding deficits for its responsibilities and consequent inflation. This situation persisted until the disjuncture was addressed through a constitutional amendment.75 Neither Brazil nor Nigeria has achieved optimal allocation of authority to their tiers of government that could foster adequate intergovernmental collaborative actions. Brazil still aspires to a cooperative federalism, but there remain challenges of inadequate fiscal and administrative capacities at the lowest tier. Hence, 30 years after their unified health policy, although there were regional improvements in health indices in the north-east,19 policy outcomes remained short of intended goals, and inadequate fiscal capacity at lower levels was a contributory factor.76 77 However, Brazil has done relatively well in MNCH outcomes when compared with Nigeria, partly attributable to social movements challenging traditional beliefs in Brazil.78–80
Intended intergovernmental collaborative actions (development and enactment of policies, recruitment and deployment of programme staff and other resources, selection of programme facilities, staff remuneration and monitoring of programme activities) were not adequately implemented in Nigeria. Signing of MOUs by multilevel stakeholders were the first collaborative actions, but constituted passive collaboration, underpinned by a prior history of mistrust and domineering top-down power practices at government boundaries.32 64 81 In contrast, under the Whole of Society Approach programme of the Western Cape province in South Africa to address the social determinants of health, for example, the provincial (state) government quickly recognised the need to address the ‘long-standing mistrust between government and communities’ through shared learning and distributed leaderships, among other strategies.82 In the short-term these strategies became important collaborative actions which laid foundations for future positive collaborative outcomes.
The inadequately collaborative designs of MSS and SURE-P MCH placed the national level in the forefront of implementation, an arrangement which was incompatible with the attributed responsibilities for the different levels. Therefore, this arrangement was not sustainable and encountered difficulties at the point of handing over ownership to states, since policy initiatives did not have adequate accountability levers to foster adherence and lacked constitutional powers to do so. A similar experience occurred with Uganda’s Environment Policy, which was participatory and user-focused in design but in practice was mediated by legal and administrative structures and procedures established for implementation of nationally determined programmes.83 84 However, where there is adequate subnational political will to implement national initiatives, successes have been recorded, irrespective of design flaws.85
Feedback of lessons learnt in policy processes is a key component of evolving and complex adaptive health systems.86 Analysis and learnings from the policy processes and outcomes of the first two programmes contributed to adaptation to a certain level of collaboration across government levels in development of SOML PfR, a process aided by the unifying incentive of leveraging new funds from World Bank for a performance-based programme. However, subnational governance executive powers and possible mistrust arising from prior contestations also costrained this programme, despite these adaptations. Subnational executive powers were not adequately mobilised to commit to the programme. The non-hierarchical structure of CG challenges the existing historical top-down structure in place in Nigeria. The subnational executive powers further complicate this structure, leaving governance in permanent contestation.
Intergovernmental CG is even more challenging in LMICs as a result of weak social and political systems, hierarchical leadership and entrenched political patronage and corruption,10 87 and high transaction costs of collaboration.88–91 In northern Ghana, there have been attempts for adaptive CG for climate change management, but these have been constrained by mistrust and diverse interests.89 These challenges partly contribute to the lack of incentives to work across jurisdictional boundaries. Within the vertical programme context, prior to the strategy development, there had not been a prior history of cross-level stakeholder dialogues and deliberation, as was also identified as a constraint in Ethiopia’s water management collaboration.92
This study has some limitations. While the study demonstrates that distributed leadership is critical for collaboration to successfully happen, it does not explore avenues for institutionalising this or ways of mitigating existing collaboration barriers like transaction costs. Methodologically, the long time span of the study—a decade long—and continuing reforms in the overall health (and MNCH) policy environment made researching this phenomenom like chasing a moving target. Capturing various events and how they changed over the long duration of the study was a challenge, in part due to recall bias, creating the risk of taking a reductionist approach. However, adequate saturation was reached in the interviews, providing confidence that the phenomenon under study was thoroughly captured and understod.
The CG framework has proved a very useful tool in this study in understanding the nuances around the Nigerian governance structure and its impact on collaborative policy aspirations. Reflecting on the use of the CG framework for analysis in the Nigeria context and for that matter, other resource-limited settings or low-income countries, we would isolate fiscal capacity (funding sources, control) as a key driver, alongside leadership, rather than being subsumed under incentives, especially in countries with quasi-federal structures.
Conclusion and recommendations
The level of collaboration initiated by the FMoH at the beginning of the IMNCH strategy development, which was new, required iterations and was not achieved, despite implementation of three consecutive programmes. The aspirational collaborative intent of the national level across jurisdictional boundaries was constrained by governance arrangements and power imbalances across the three tiers, partly due to a lack of will to collaborate, and partly to challenges of collaboration across government levels and between multiple actors. Post MDGs, despite adaptations, these malalignments continue to plague MNCH policy implementation in particular and healthcare delivery as a concurrent responsiblity of the government tiers in general.
To achieve sustainable improvements on MNCH policy, the FMoH, with primary responsibility for policymaking, should refocus on the governance structure. Collaboration could be facilitated via full devolution of healthcare responsibilities by national authorities to subnational governments, and national level providing independent evaluation and guidance, using available intergovernmental collaboration dynamics, to be agreed between both levels. These actions would bring the policy space down from national to subnational level, invariably affecting subnational commitment to policy implementation decisions. This recommendation, however, transcends MNCH programmes and the health sector and has implications of higher-level political and economic reforms, including the protracted debates about constitutional amendments and fiscal restructuring.93 A more health sector-related recommendation is that leadership for MNCH services needs to be distributed both vertically and horizontally across states and LGAs, and LGA actors empowered with resources and knowledge. Stakeholders need to be aware of what contextual drivers are available to them for collaboration, such as wide engagement platform of the NCH, and these should be purposively replicated subnationally. With the existing fiscal structure, innovative reforms which hold all government levels accountable should be linked to cross-level fiscal transfers. Since advocacy has worked,94 sustained advocacy for MNCH and other context-specific innovative ways of achieving distributed leadership across government levels are required in Nigeria and similar resource-limited LMICs.
Data availability statement
Data are available upon reasonable request.
Ethics statements
Patient consent for publication
Ethics approval
This study involves human participants and ethical clearances to conduct this study were obtained from the biomedical research ethics committee of the University of Western Cape (BM 18/7/9) and from the research ethics committee of the University of Nigeria Teaching Hospital, Ituku-Ozalla, Enugu State (NHRC/05/01/2008B-FWA00002458-IRB00002323). The participants gave informed consent to participate in the study before taking part. Respondents were initially sensitised to the study, and then appointments were secured for interviews. The participants were assured of confidentiality and were made aware they could withdraw from the study/interview at any stage. With written, informed consent, interviews were recorded, and recordings were only accessible to the researcher and assistant who transcribed interviews. Anonymised transcripts were stored with identifier codes in a passworded computer.
Acknowledgments
The authors acknowledge the contributions from Professor Helen Schneider in developing the original study proposal. We thank health authorities at the Federal Ministry of Health, Abuja, and the Ministry of Health of the Anambra State and Ministry of Health, Ebonyi State, in granting permission for the study and engaging in provision of information and documents, all our interviewers and study participants.
References
Supplementary materials
Supplementary Data
This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.
Footnotes
Handling editor Stephanie M Topp
Contributors EE and UL jointly conceived the study. EE conducted the data collection with support and guidance from UL and led the writing of this paper with contributions and revisions of several versions from EAE, SLD and UL. All authors read and approved the final version of the manuscript. EE acts as the guarantor for this study.
Funding The authors have not declared a specific grant for this research from any funding agency in the public, commercial or not-for-profit sectors.
Competing interests None declared.
Patient and public involvement Patients and/or the public were not involved in the design, conduct, reporting or dissemination plans of this research.
Provenance and peer review Not commissioned; externally peer reviewed.
Supplemental material This content has been supplied by the author(s). It has not been vetted by BMJ Publishing Group Limited (BMJ) and may not have been peer-reviewed. Any opinions or recommendations discussed are solely those of the author(s) and are not endorsed by BMJ. BMJ disclaims all liability and responsibility arising from any reliance placed on the content. Where the content includes any translated material, BMJ does not warrant the accuracy and reliability of the translations (including but not limited to local regulations, clinical guidelines, terminology, drug names and drug dosages), and is not responsible for any error and/or omissions arising from translation and adaptation or otherwise.