Institutional analysis of health system governance

Abstract

It is important that researchers who study health system governance have a set of collective understandings of the meanings of governance, which can then inform the methods used in research. We present an institutional framing and definition of health system governance; that is, governance refers to making, changing, monitoring and enforcing the rules that govern the demand and supply of health services. This pervasive, relational view of governance is to be preferred to approaches that focus primarily on structures of governments and health care organizations, because health system governance involves communities and service users, and because governments in many low- and middle-income countries tend to under-govern. Therefore, the study of health system governance requires institutional analysis; an approach that focuses not only on structures, but also on the rules (both formal and informal) governing demand and supply relations. Using this ‘structure–relations’ lens, and based on our field experience, we discuss how this focus could be applied to the three approaches to framing and studying health system governance that we identified in the literature. In order of decreasing focus on structures (‘hardware’) and increasing focus on relations (‘software’), they are: (1) the government-centred approach, which focuses on the role of governments, above or to the exclusion of non-government health system actors; (2) the building-block approach, which focuses on the internal workings of health care organizations, and treats governance as one of the several building blocks of organizations; and (3) the institutional approach, which focuses on how the rules governing social and economic interactions are made, changed, monitored and enforced. Notably, either or both qualitative and quantitative methods may be used by researchers in efforts to incorporate the analysis of how rules determine relations among health system actors into these three approaches to health system governance.

Community, decentralization, economic governance, health systems, institutions

Introduction

The ‘failure of governance’ and the ‘failure of institutions’ are often implicated in many of the problems of poverty, disease and underdevelopment; from the dysfunctionality of governments, the failure of the rule of law, the lack of economic progress within societies, to the inability of health care systems to deliver services to populations (Acemoglu and Robinson 2012). But the ubiquity of the terms—‘governance’ and ‘institutions’—and their ‘strong intuitive appeal’ have led to their use, often without much consideration given to the need for precise definitions (Heinrich and Lynn 2000: 1). In addition, both terms are broadly used and applied concepts in analyses across disciplines in the social sciences. And with this comes the tendency for the terms to mean different things to people in different disciplines. It is therefore important that health system researchers who study governance have clear and collective understandings of governance, which can then inform the methods used in studying health system governance, especially in low- and middle-income countries. In the following ‘methodological musings’, we will highlight a set of approaches to governance based on our field research on primary health care governance in Nigeria, in dynamic exchange with an interrogation of the literature.

The word ‘govern’ was adopted in English and other European languages in the 19th century. It originated as an adaptation of the Greek nautical word kybernan (with a ‘k’ to ‘g’ sound shift) which meant steering, guiding or manoeuvring a ship. The metaphor was originally used in ancient Greece, e.g. by Plato, in the sense of ‘steering the ship of state’, and with the word ‘steersman’ meaning the ‘governor’ of a jurisdiction (Pappas 2013). Governance may therefore be defined as the processes of steering; and according to Bevir (2012: 1) the steering may be done ‘by a government, market or network, whether over a family, tribe, formal or informal organization or territory and whether through laws, norms, power or language.’ Implicit in this definition of governance is the idea that ‘governance’ is not synonymous with ‘government’. Indeed, while governments (i.e. groups of individuals—or organizations—invested with the formal authority to govern a jurisdiction) have the responsibility to perform many governance functions, such functions may be shared with, or performed by, the market, media, lobbies, political parties and community groups. Governments are political organizations, while governance refers to the act of steering, however and wherever they may occur (Bevir 2012).

Indeed, Brinkerhoff and Bossert (2008: 3) described governance in relation to influencing or being influenced by ‘the rules that distribute roles and responsibilities among societal actors and shape the interaction among them’. The word ‘rules’ in this description however requires its own definition. Ostrom (1986) defined rules as prescriptions (i.e. which actions, decisions, relations or outcomes are required, permitted or prohibited); as prescriptions that result from implicit or explicit efforts of individuals and communities to achieve order and predictability within defined situations; and as prescriptions which are commonly known and used by a set of participants to order repetitive, interdependent relationships. The rules-based description of governance by Brinkerhoff and Bossert (2008) and the definition of rules by Ostrom (1986) suggest that governance involves how rules influence the actions, decisions and relations of individuals and organizations, and how they in turn make, change, monitor and enforce rules. The rules may be formal or informal (Ostrom 2010). The word ‘rule’ is often used in relation with the word ‘institution’ (Hodgson 2006). North (1991; 4) for example defined institutions as ‘the rules of the game in society, or… the humanly devised constraints that shape human interaction’.

We acknowledge that there is ongoing debate and discussion on the definition of governance in health policy and systems research which was recently well covered by Barbazza and Tello (2014). However, what we sought to do in selecting these ‘institutional’ definitions was to reach for something elemental about the meaning of governance—i.e. we argue that what is left of governance when stripped to its bare essentials are the rules, both formal and informal, that shape the social and economic interactions among actors. This definition of governance aligns with what Barbazza and Tello (2014) described as ‘tools used to govern’; but in the way we use it in this paper, rules or institutions are not only tools, but also what governance really is essentially. Barbazza and Tello (2014) also identified two additional ways in which the term governance is used; first in a normative sense (with terms such as ‘good’ or ‘democratic’ used to qualify governance); and second in a descriptive sense (with words such as ‘hierarchical’ and ‘networked’ used to describe governance). The emphasis on rules and relations in this article is also driven by our focus on governance in low- and middle-income countries, where the lack of formal rules means that informal rules and relations can be more important in achieving the overall objective of health system governance: to ensure the supply of health services, facilitate the transactions involved in the supply and demand of health services, and protect the rights of people involved in the supply and demand of health services; all to ultimately achieve improved health outcomes (Abimbola et al. 2014).

Notably, health care begins with ‘what happens when someone who is ill (or who thinks he or she is ill or who wants to avoid getting ill) consults a health professional in a community setting for advice, tests, treatment or referral to specialist care’ (Greenhalgh 2007: 2). Health care involves the interactions between demand and supply actors, irrespective of whether the services are provided in public, private for profit or private non-profit health facilities, or whether they are informal or formal health services. Indeed, one of the consequences of ineffective governments in many low- and middle-income countries is that they leave space for unlicensed and unregulated informal providers without formal training (e.g. drug shops and traditional healers and birth attendants) to deliver a large proportion of health services (Abimbola et al. 2016). Beyond the poor regulation of informal providers, the consequences of weak governments also include poor regulation of formal health care providers, poor coordination of referrals and along the continuum of care, and weak institutions for the provision of public goods (such as information, infrastructure and health services). In such settings, economic exchanges (such as in health care markets) are often governed by market forces mediated largely by informal rules such as community, individual and professional norms (Ostrom 2010; Abimbola et al. 2014; Abimbola et al. 2016). What we therefore sought to explore in these ‘musings’ is how researchers might incorporate the analysis of the way in which rules determine the relations among health system actors into the different approaches to framing and studying governance.

Three approaches to health system governance

We identified three different, but overlapping approaches to framing and studying health system governance in the literature (1) the government-centred approach, which focuses on the role of governments, above or to the exclusion of non-government health system actors; (2) the building-block approach, which focuses on the internal workings of health care organizations, and treats governance as one of the several building blocks of organizations and (3) the institutional approach, which focuses on how rules governing social and economic interactions are made, changed, monitored and enforced. By ‘approach’ we mean how governance is framed in order to be studied, and identifying these approaches was informed by the ‘hardware–software’ framework of what constitutes a health system (Sheikh et al. 2011). Hardware refers to ‘concrete and tangible expressions of health systems’ such as finance, medical supplies, information systems, human resources, infrastructure, the organizational structures to provide policies, services, and interventions and their intended targets, users and beneficiaries; while software refers to the less visible and quantifiable components of health systems, such as ‘the ideas and interests, values and norms and affinities and power that guide actions and underpin the relationships among system actors and elements’ (Sheikh et al. 2011: 2).

Each of the three approaches to health system governance discussed in greater detail below includes a mix of focus on both hardware and software, albeit combined to different extent: the government-centred approach (hardware > software), the building-block approach (hardware ≈ software), and the institutional approach (hardware < software). The software aspects of these approaches open up the space for institutional analysis of health system governance as more focus on software allows for greater attention to the rules governing relations. Further, applying the ‘normative–descriptive’ distinction highlighted by Barbazza and Tello (2014) in definitions of governance, the government-centred approach often adopts a normative framing of governance, while the building-block and institutional approaches are descriptive (see Table 1). Notably, either or both qualitative and quantitative methods may be used in any of the three approaches (see Table 2) in efforts to assess whether, how and under what circumstances ‘acts of governance’ contribute (effectively and/or efficiently) towards the goal of facilitating access to appropriate health care and improving health outcomes.

Table 1.

The three approaches to conceptualising health system governance depicted based on how they fit on the hardware-software and normative-descriptive axes

	Hardware	Software	Normative	Descriptive
Government-centred approach	+++	+	+++	–
Building-block approach	++	++	–	+++
Institutional approach	+	+++	–	+++

Table 2.

Examples of research topics amenable to qualitative and quantitative methods for each of the three approaches to health system governance

	Qualitative quantitative
Government-centred approach	Exploring the domains and principles of good governance at national and sub-national levels to explain the decisions or the performance of governments	Quantifying good governance using indices (e.g. the Worldwide governance index, and Ibrahim Index of African Governance) to rank or categorise governments based on their performance
Building-block approach	Exploring organizational or facility governance through the decision space approach to the principal-agent theory or the qualitative evaluation and analysis of governance interventions	Quantifying organizational and facility governance; e.g. using the quantitative component of the Health Systems 20/20 tool, and by quantifying decision space, and the transaction costs of providing health care
Institutional approach	Exploring informal and formal rules governing the demand and supply of services; how rules function, how individuals, groups and governments make, change, monitor and enforce rules, and how these actors are in turn also affected and influenced by rules	Quantifying transaction costs of accessing health care as the costs incurred by individuals in efforts to access and use health services that they would not incur with institutions to ensure optimal information, regulation and coordination of health care transactions

The government-centred approach to governance

The government-centred approach emphasises the central role of governments in governance, often to the exclusion of the role of other actors and organizations. Bell and Hindmoor (2009: 71) define governance as ‘the tools, strategies and relationships used by governments to help govern.’ The sentiment of this definition and its focus on the ‘hardware’ of governance is implicit, e.g. in the approach of global health and development agencies to governance reforms in low- and middle-income countries. This includes their efforts to provide ‘tools and strategies’ and strengthen ‘relationships’ to achieve ‘good governance’, often in reference to a country and its government. The World Bank (1994: xiv) defines governance as ‘the manner in which power is exercised in the management of a country’s economic and social resources’, emphasising ‘the form of ‘political regime’, ‘development’ and ‘capacity …to design, formulate and implement policies and discharge functions.’ The United Nations Development Programme (UNDP 1997: 2–3) also defines governance as ‘the exercise of economic, political and administrative authority to manage a country’s affairs at all levels’.

The government-centred approach is associated with the focus on ‘good governance’ among development agencies, and the principles of transparency and accountability, efficiency and effectiveness, democracy and representation, human rights and the rule of law, and developmental vision and objectives (Gisselquist 2012). This focus on good governance (a normative rather than descriptive focus) has given rise to efforts to assess governance, especially using quantitative measures to summarise or compare the performance of countries whether as a composite measure of good governance, or of issues such as corruption (e.g. the Corruption Perceptions Index created by Transparency International in 1995) or as a guide for international investors (e.g. the Ease of Doing Business Index launched by the World Bank in 2002). The Worldwide Governance Indicators is a widely used, composite measure of factors contributing to good governance. Created by the World Bank in the mid-1990s, it assesses six separate but inter-connected indicators of governance (World Bank 2007). The Ibrahim Index of African Governance, launched in 2007 by the Mo Ibrahim Foundation assesses governance in four dimensions (Mo Ibrahim Foundation 2015). Please see Table 3 for these indicators and dimensions.

Table 3.

Dimensions, principles and domains of good governance according to development organizations

	Dimensions, principles and domains of good governance
World Bank—Worldwide Governance Indicators	Voice and accountability Political stability and absence of violence Government effectiveness	Regulatory Quality Rule of Law Control of Corruption
United Nations	Participation Rule of law Transparency Responsiveness Consensus Orientation	Equity Effectiveness and Efficiency Accountability Strategic vision
Overseas developmentInstitute—World Governance Assessment	Participation Fairness Decency	Accountability Transparency Efficiency
Mo Ibrahim Foundation—Ibrahim Index of African Governance	Safety and rule of law Participation and Human rights	Sustainable economic opportunity Human development

Source: Baez-Carmago and Jacobs (2011).

Like other global health and development agencies, the WHO identifies the principles and domains of good governance in national health systems to include policy guidance, intelligence and oversight, collaboration and coalition building, regulation, system design and accountability (WHO 2007). The WHO describes these roles as the responsibility of governments; and that governance ‘involves overseeing and guiding the whole health system, …requires both political and technical action, …attention to corruption and …human rights based approach’ (WHO 2007: 23). The assessments of health system governance deriving from this premise have focussed on governments, and the principles and domains of good governance as highlighted by WHO, World Bank, UNDP and the Mo Ibrahim Foundation. The Siddiqi et al. (2009) framework for assessing health system governance in low- and middle-income countries takes this approach, using qualitative measures unlike the scoring or ranking system favoured by the World Bank (Kaufmann et al. 1999). Whereas, Kirigia and Kirigia (2011), not satisfied with qualitative measures (because they are not readily aggregated, compared or tracked over time), created a scoring metric based on the qualitative measures used by Siddiqi et al. (2009). Likewise, Olafsdottir et al. (2011) used quantitative measures derived from the Ibrahim Index of African Governance in an ecological analysis to demonstrate the association between governance and health outcomes in 46 African countries. Further, Lewis and Pettersson (2009) and WHO (2010) proposed similar quantitative indices to assess the performance of governments in governing national health systems.

On the other hand, the Health Systems 20/20 (2012) framework combines quantitative measures (i.e. the Worldwide Governance Indicators developed by the World Bank) with the assessment of six qualitative health-specific governance indicators. The qualitative indicators are derived from the frequently used Brinkerhoff and Bossert (2008) framework of accountability in health systems, which in turn is based on a World Bank (2004) framework designed to assess accountability in service delivery. With national governments as the central actor, the framework outlines three two-way relationships among governments, service providers and citizens: first, between governments and citizens; second between citizens and service providers; and third between service providers and governments. With its focus on relationships (and, potentially, the rules governing them) Brinkerhoff and Bossert (2008) added a software aspect to the government-centred approach. Health Systems 20/20 (2012) mixes both quantitative and qualitative measures; and has been widely applied globally, although in these applications, the focus has been on governments, with the role of other stakeholders limited to how their relationships with governments influence accountability. Thus, this approach largely treats governments like a black-box, a singular entity, offers little insight on how to improve governance, leaving out, for example, how power relations influence the decisions of governments. But such insights may arise from more in-depth qualitative explorations and institutional analysis allowed by the Health Systems 20/20 (2012) framework.

The building-block approach to governance

The hardware of organizations is sometimes described as though it includes a building block or a component, or a unit that performs governance functions. Words such as ‘leadership’ and ‘management’ are often used to depict this building block. The building-block approach focuses on the internal workings of government and non-government organizations. Applied to health care organizations and facilities, the building-block approach is exemplified in the inclusion of governance as one of the six WHO health system building blocks (WHO 2007). Like the internal process approach to organizational management, the building-block approach to governance focuses on hierarchical structures, procedures and policies which are used to stimulate routine, regularity, efficiency and performance in relations within and between organizations (Quinn et al. 2003). For example, the governance building-block of a multi-tiered organization such as the district health system coordinates internal operations and external relations through hierarchical lines of command, incentives, descriptions of roles and responsibilities, and technical guidance. In the private sector, the building-block approach may be likened to corporate governance; likewise, clinical governance structures and processes in health care delivery.

In a manner similar to the government-centred approach to governance, UNDP has compiled a catalogue of quantitative and qualitative tools to assess the governance units of lower levels of government, such as district and local governments, based on the principles and domains of good governance (Wilde et al. 2009). Likewise, Mutale et al. (2013) have adapted the quantitative component of the Health Systems 20/20 tool to assess the governance of primary health care facilities in Zambia. However, the building-block approach differs from the government-centred approach in that it is descriptive, and in its focus on both structures (i.e. hardware) and relations (i.e. software). The software aspect of this approach features in studies on decentralization. The range of relationships that may exist between the centre (principal) and the periphery (agent) in decentralised health systems has been extensively explored in the health literature using the principal-agent theory as adapted and applied by Bossert (1998) who proposed the ‘decision space’ approach to the principal-agent theory to capture the range of discretion/choice allowed to agents. Principals engage agents, with the goal of achieving the organizational objectives set by the principal. But while they may share some of the objectives of the principal, agents may also have contrary interests. Principals therefore use rules (with strategies such as incentives and sanctions) to ensure compliance. Thus, the building-block approach is amenable to institutional analysis.

The principal-agent theory has informed the design and institutional analyses of performance-based financing initiatives to improve the governance of health facilities in many low- and middle-income countries (AIDSTAR-Two Project 2011; Bertone and Meessen 2013). The principal-agent theory and the decision space approach have also been used to analyse health system decentralization, especially governance within the district health systems in many low- and middle-income countries. For example, Bossert et al. (2007) in Ghana and Guatemala, Regmi et al. (2010) in Nepal, Frumence et al. (2013) in Tanzania, Kwamie et al. (2015) in Ghana, Seshadri et al. (2015) in India, and others have adopted both qualitative and quantitative methods, singly or in combination, in applying the decision-space approach. But unlike Bertone and Meessen (2013) for example, such studies have not explicitly sought to explore the rules underpinning principal-agent relations. Nonetheless, the studies have explored leadership and management, and indeed, more specifically, the governance of logistics and information systems, human and financial resources and service delivery within district health systems.

Transaction cost economics provides another way to study the internal relations (i.e. software) within health care organizations and facilities. Transactions occur when goods or services (that are standardised, counted, valued and paid for) are transferred between economic agents (Baldwin and Clark 2002). In providing health services, transactions may happen among health care providers, and within a health care organization. (Note that transactions that occur in accessing health care—i.e. between patients and providers—are covered under the institutional approach to governance, rather than within the organizational boundaries that characterise the building-block approach). Transaction costs are costs associated with search, information, negotiation and enforcement of contracts in order to facilitate a transaction (Williamson 1979). They are costs incurred to facilitate a transaction, or the additional costs incurred as a consequence of poorly facilitated transactions, i.e. as a consequence of weak governance. The level of transaction costs incurred in consummating a transfer can therefore provide a means to assess how well governed a system is; i.e. the state of the institutions within a system, such as the health system (Abimbola et al. 2015). For quantitative and qualitative applications of the concept of transaction costs to exchanges among or between health care providers, see Ashton (1998) and Zinn et al. (2003), and within or by health care organizations, see Stiles et al. (2001) and Jan et al. (2008).

Efforts to improve the governance of health care organizations and facilities have also been studied in the form of interventions such as leadership training, performance-based financing, and the privatization of components of the health system. But these interventions tend to be framed as though they are aimed at building-blocks of the health system other than governance; e.g. capacity building for health managers on supportive supervision (at human resources) or the use and analysis of data for clinical audit (at information systems); incentives to increase retention and reduce absenteeism of health workers (at human resources), results- or performance-based financing (at health financing), integrating health service delivery (at service delivery) and privatising supply chain management (at medical supplies). Indeed, while these interventions belong to these building blocks, they are also governance interventions (de Savigny et al. 2009). In a framework designed to help identify areas of weakness and how to intervene in the governance of a health care organization or facility, Mikkelsen-Lopez et al. (2011) reframed governance as a peculiar sort of building block which permeates and determines how other building blocks function; a software framing of governance. The framework is problem-driven and identifies potential areas for governance intervention in each WHO health system building block.

In this article, our position is that in analysing the governance of health care organizations and facilities, it may be more appropriate to frame governance as the mortar holding the building blocks together, rather than as a building block. One drawback of the building-block approach is exemplified in the WHO building blocks, which do not highlight the roles of communities or service users; albeit the place of ‘patient engagement’ has been highlighted in a proposed modification of the WHO health system building blocks (Lazarus and France 2014). In addition, applications of the principal-agent theory in health systems tend to ignore the governance roles of communities or service users. Whereas, within district health systems, community members and service users can be seen as the ‘principals’, with governments and service providers acting as ‘agents’. But the reverse (especially governments as ‘principal’) has been the usually adopted interpretation of the framework. An institutional analysis of governance would however accommodate both interpretations of the theory.

The institutional approach to governance

The institutional approach brings the governance roles of communities into sharper relief. The institutional approach goes beyond a focus on the potential role of communities and service users in accountability, and embraces a broader range of formal and informal roles that communities can play in determining the rules that govern the demand and supply of their health services. For example, Fattore and Tediosi (2013), based on the work of Hood (1998), described how the dominant cultural and societal values may influence the decision to aspire to achieving universal healthcare coverage: fatalist values, hierarchist values, individualist values, and egalitarian values. The institutional approach is also descriptive, rather than normative. For example, arguing that the domains of good governance are much too closely related to be useful as separate entities in empirical research, Baez-Carmago and Jacobs (2011) sought to identify more basic features defining governance that may link together the domains of good governance. They proposed three features, each inextricably combining both hardware and software components, and a focus on institutions: (1) governance input, i.e. ‘how and by whom are the institutions and rules governing the health system constructed’; (2) governance processes, i.e. the ‘basic attributes characterising the implementation of the rules and administrative procedures governing the health sector’ and (3) governance outcomes, i.e. the ‘positive qualities that health system outputs should generate once rules and processes have been designed and implemented’ (Baez-Carmago and Jacobs 2011). This framework therefore highlights health system software as a means of assessing good governance.

In a similar example of the manifestation of software in the analysis of health system governance, Cleary et al. (2013) applied the Brinkerhoff and Bossert (2008) framework to a literature review on accountability in primary health care systems. They identified three factors influencing accountability; the first is a hardware component—resources (including time, space, and especially capacity), and the other two are related to health system software—attitudes and perceptions of health system actors; and the values and culture of the health system. The relationship between these components is that “the values of the system are reflected in resource flows; resources and capacity impact on attitudes and perceptions; and attitudes and perceptions impact on the use of capacity and contribute to enforcing or changing values” (Cleary et al. 2013: 8). Indeed, the institutional approach, with its focus on rules (and especially informal rules), does allow for the institutional analysis of governance. Baez-Carmago and Jacobs (2011; 15) emphasise that informal rules (such as norms and values) are ‘significant phenomena affecting public governance processes, especially in many low-income countries’. They indicate that because the ‘formal rules governing the health system are in many cases not effectively applied in spite of adequate formal regulations, it is critical that a methodology to assess health system performance should include an understanding of informal institutions and actors’ (Baez-Carmago and Jacobs 2011: 10).

Exploring the gaps between formal rules (such as regulations and policies) and actual practice on the ground (governed by informal rules and adaptations of formal rules) in the delivery of health care in low- and middle income countries has been a focus of the work of Olivier de Sardan (2009). This body of work has highlighted for example, the need to understand that the way in which practical norms deviate from formal rules reflects both traditional values, and the weakness of governments in the provision and oversight of public goods and services. Leonard et al. (2013: 71) also applied the institutional lens in a literature review on how to address the information asymmetry problem in health care markets, particularly in ‘under-governed countries, where state regulation is weak’. They found that addressing the information challenge for patients in such settings requires a system in which health providers, whether in the public or private sector, can exercise decentralised control over personnel and finances, but under the close watch of community participation. In addition, their review also highlighted that in these settings, ‘macro-contextual factors such as cultural norms and values matter for service outcomes, particularly on how they determine the performance of community accountability mechanisms, [and] on how they shape provider–recipient relationships (Leonard et al. 2013; 82)’.

Extensive research, especially in economics and political science on the governance of ‘under-governed’ and ‘lawless’ jurisdictions or markets has taken the software approach. These include the analysis of jurisdictions and markets such as medieval trade (Greif 1989, 1993, 2004), prison gangs (Skarbek 2010, 2011, 2012), drug cartels (Hagedorn 1994; Kenney 2007; Jacques 2010), and mafias (Gambetta 1993; Varese 2001; Rafael 2007). What is common within this literature is that in markets where governments provide limited or unreliable protection for property and contract (or in some cases seek to undo whatever protections exist), transacting parties ‘have no choice but to do their best to create their own bilateral or group mechanisms to support otherwise problematic exchange’ (Williamson 2005; 2). Indeed, Olivier de Sardan (2015) likened these situations in which governments are either unable or unwilling to protect and facilitate market transactions, to what happens when governments disappear at local levels. This is the level at which primary health care services are often delivered, such that when government regulation is lacking, social norms fill the gaps. In such situations, local markets for the supply and demand of primary health care may be co-governed (with governments) by community groups. As Dixit (2009: 6) highlighted, governance institutions are provided by non-government actors ‘especially in niches that the government serves poorly, or not at all’.

The institutional approach to governance is referred to in the economics literature as ‘economic governance’ given the focus on the rules underpinning demand and supply relations (Dixit 2009). This is important because health systems are economic systems, and health care interactions are indeed framed in terms of supply and demand relations and the intermediary organizations (e.g. governments and community health committees) that influence supply-demand relations (Janovsky and Cassels 1996). Health services in low- and middle-income countries are often provided in the market, given that in many settings people secure access to health care by paying for services (Leonard et al. 2013). Indeed, such services may be provided in the public or private sector, the payments may be formal or informal, and the services may be formal or informal. As defined by Dixit (2009; 5), economic governance refers to “the structure and functioning of the legal and social institutions that support economic activity and economic transactions by protecting property rights, enforcing contracts, and taking collective action to provide physical and organizational infrastructure.” The three objects of institutions of economic governance outlined in this definition (i.e. providing public goods, protecting property rights and facilitating transactions) are therefore preconditions for the functioning of markets.

The institutional approach to governance has, for example, been applied to better understand how changes in governance institutions impact on the delivery of primary health care in rural communities as China undergoes significant economic transition (Meessen and Bloom 2007). The analysis showed that the economic transition in China creates challenges for the health sector. The emerging market system requires more institutions than were necessary when governance was through authoritarian command and control. The analysis highlights the need for ‘institutional expansion’ in order to adapt to a more hands-off mode of governance via regulation and incentives, especially in the form of hitherto lacking informal and professional norms, ethics, code of conduct, and checks and balances by civil society (Meessen and Bloom 2007). The institutional approach to governance has also been applied to analyse the design of performance-based financing initiatives in Rwanda (Meessen et al. 2006) and Burundi (Bertone and Meessen 2013); and the role of community groups in performance-based financing initiatives (Falisse et al. 2012) in Burundi. These studies suggest that beyond aligning institutions with incentives, governance interventions to improve organizational performance in under-governed settings should also ensure that non-government actors within the community have the decision rights and the capacity to monitor and enforce the newly made or changed rules governing the supply of health services.

More recently, we adapted for primary health care governance (Abimbola et al. 2014) what Kiser and Ostrom (1982; 179) described as ‘the three worlds of action’ or three levels at which to conduct institutional analysis of governance—operational level (where individuals make direct economic and social choices); collective level (where community or group choices influence those operational choices) and constitutional level (where the rules governing collective and operational choices are determined). Weakness at one level is compensated for by governance at another level. Each level has bidirectional relations with the other two, as in the Brinkerhoff and Bossert (2008) framework; the difference is that actors at each level are deemed as able to make, change, monitor and enforce rules (Abimbola et al. 2014). We also applied the theoretical work of Williamson (2010) on transaction costs to primary health care governance (Abimbola et al. 2015). For Williamson (2005) transactions are the basic unit of institutional analysis of governance such that when institutions are not in place to facilitate transactions, partners to an economic exchange incur transaction costs. Navigating health care markets with high levels of information asymmetry and where regulation is weak and coordination is poor implies that, prior to making appropriate health care transactions, patients incur transaction costs while obtaining inappropriate care from formal and informal providers. The “transaction costs of accessing health care” can be used to quantify the effects of weak governance, and to determine at which level of governance it is more cost-efficient to intervene in order to reduce transaction costs (Abimbola et al. 2015; Abimbola et al. 2016).

Conclusion

In summary, we have presented three approaches to health system governance, each of which can adopt a quantitative, qualitative or mixed approach to research. One approach is the ‘government centred approach in which the focus is on a key actor; the government. The other one is to take governance as ‘pervasive’ entity with the focus on institutional arrangements. The building-block approach is a mix of the two; a focus on the internal workings (including institutional arrangements) within typically government-run organizations or health care facilities. However, we argue that because the institutional approach to governance is at the same time encompassing as it is granular, it is much better suited to the analysis of governance in settings where governments under-govern, such as health systems in many low- and middle-income countries. We argue for a bottom-up perspective on what happens on the ground; what Olivier de Sardan (2009) referred to as ‘real’ governance, encompassing both the formal and informal rules governing the demand and supply of health care, and both the formal and informal ways in which the rules are made, changed, monitored and enforced whether in the government-centred, building-block or institutional approach to health system governance. We hope that our reflections on the approaches to health system governance in the literature will contribute to and support much needed methodological discussion among health policy and system researchers.

Funding

During the completion of this work, Seye Abimbola was supported by the Rotary Foundation through a Global Grant Scholarship (grant number GG1412096) and by the Sydney Medical School Foundation through a University of Sydney International Scholarship. No additional external funding was received for the work that led to this paper. The funders had no role in the preparation of this manuscript.

Conflict of interest statement. None declared.