10 best resources on power in health policy and systems in low- and middle-income countries

Abstract

Power is a critical concept to understand and transform health policy and systems. Power manifests implicitly or explicitly at multiple levels—local, national and global—and is present at each actor interface, therefore shaping all actions, processes and outcomes. Analysing and engaging with power has important potential for improving our understanding of the underlying causes of inequity, and our ability to promote transparency, accountability and fairness. However, the study and analysis of the role of power in health policy and systems, particularly in the context of low- and middle-income countries, has been lacking. In order to facilitate greater engagement with the concept of power among researchers and practitioners in the health systems and policy realm, we share a broad overview of the concept of power, and list 10 excellent resources on power in health policy and systems in low- and middle-income countries, covering exemplary frameworks, commentaries and empirical work. We undertook a two-stage process to identify these resources. First, we conducted a collaborative exercise involving crowdsourcing and participatory validation, resulting in 24 proposed articles. Second, we conducted a structured literature review in four phases, resulting in 38 articles reviewed. We present the 10 selected resources in the following categories to bring out key facets of the literature on power and health policy and systems—(1) Resources that provide an overarching conceptual exploration into how power shapes health policy and systems, and how to investigate it; and (2) examples of strong empirical work on power and health policy and systems research representing various levels of analyses, geographic regions and conceptual understandings of power. We conclude with a brief discussion of key gaps in the literature, and suggestions for additional methodological approaches to study power.

Introduction

Power is central to understand and transform health systems. Health systems refer to those ‘organizations, people and actions whose primary intent is to promote, restore, or maintain health’ (World Health Organization, 2007), and like any social system, function with people at their core. Health systems are therefore influenced by the power dynamics that underlie societal interactions (Nichter, 1986; Walt, 1994; Freedman, 2005; Sheikh et al., 2014). Power manifests implicitly and explicitly in diverse ways in the interactions of health system actors at the local, national and global level. These interactions span the dynamics between patients and providers at primary health facilities to the negotiation between national and global level actors regarding resource distribution and health policy priorities. Indeed, power shapes health policy and practice, including community collaboration, participation and ownership; access, affordability and quality of health services; and the prioritization and development of health policy (Erasmus and Gilson, 2008; Buse et al., 2009; Kickbusch and Szabo, 2014; Shiffman, 2014; George et al., 2015). Power also shapes social inequalities experienced by individuals and communities by interacting with their ‘social locations’—race, ethnicity, gender, sexuality, religion, class—to influence healthcare needs and experiences (Iyer et al., 2008; Hankivsky, 2014; Larson et al., 2016). Put differently, power is present at each actor interface in the health system, and influences when and which individuals require health intervention. It therefore shapes all actions, processes and outcomes.

Analysing and engaging with power has the potential to promote more transparent, equitable and fair health systems in low- and middle-income countries (LMICs) (Kesby, 2005; Sheikh et al., 2014; Freedman, 2016). In explaining the reasons behind the stubbornly high inequities observed across and within populations in relation to health outcomes, access to health services, and the broader social determinants of health, scholars and practitioners have pointed toward the central role of power imbalances (Samuelsen et al., 2013; Friedman and Gostin, 2017). Grasping and addressing the foundational causes for inequity in health therefore require us to engage directly with power at multiple levels—from individual interactions to societal structures (Kesby, 2005). Further, power privileges certain forms of knowledge and discourse in global health, influencing the types of interventions used to address health systems challenges (Shiffman, 2014; Benatar, 2016). Understanding how and why certain policies and interventions gain more attention than others, and the implicit value judgments underlying those processes and decisions requires an exploration of power dynamics (Walt, 1994).

Power analysis has long been a discrete area of work within the social sciences, which explore both theoretical foundations and practical applications of the concept (Haugaard, 2011). In the area of health policy and systems research in LMICs—defined further in Box 1—a small, but growing, body of empirical work has examined the impact of power at multiple levels of health systems (Gilson, 2012). In addition to power analysis, complementary disciplinary approaches, often from the social sciences, have been applied to describe how power affects health policymaking and health service delivery in LMICs. For example, anthropologists have used critical ethnography to uncover and scrutinize the power imbalances that shape the interplay between local realities (for communities and for health providers) and national and global forces (Pfeiffer, 2003; Janes and Corbett, 2009; Feierman et al., 2010; Storeng and Mishra, 2014). Political scientists and economists have applied the tools of case study research, econometrics and other approaches to trace and analyse how national political systems, state-society relations, international relations and macroeconomic shifts impact services and policy at multiple levels (Ottersen et al. 2014). Historians have used the explanatory potential of historical research to provide critical insight into discourse and policy trajectories at global and national levels (Birn, 2009; Packard, 2016); and medical sociologists and social epidemiologists have examined power flowing through, and emerging from, the social positioning of individuals and groups, the structures and patterns of institutions and organizations, and the networks connecting actors (Krieger, 2001; Williams, 2003; Bradley et al., 2011; Berkman et al., 2014). We discuss some of the relevant theoretical basis for this broader field of work in the section on theories and frameworks of power below.

In practice, civil society networks, such as the Community of Practitioners on Accountability and Social Action in Health (COPASAH) actively work to make power visible and unlock power amongst communities, through accountability efforts such as citizen monitoring of health services (Shukla et al., 2011; Flores and Ruano, 2015). The role of civil society and citizen participation in health policy and systems through embedded social accountability efforts has been noted as crucial for achieving meaningful gains in health (Gaventa, 2005; Flores et al., 2009; Freedman, 2016). A few international development organizations, such as regional offices of SIDA and Oxfam, have explicitly adopted power analyses as a mechanism to determine the structural roots of poverty, and to strengthen participatory approaches to development (Pettit, 2013). Power is also sometimes discussed, although often implicitly, in the context of research and action-oriented approaches that seek to foster equity in health policy and systems, such as rights-based programming, social justice interventions, epidemiology and broader work on the social determinants of health, critical gender, race and feminist and queer theory.

Notwithstanding the examples of action and scholarship above, power analyses in health policy and systems research generally, and in LMICs in particular, remain scarce. One likely reason for this scarcity is that engaging with power in research and practice can be challenging. Conceptually, power is multi-faceted and contested, shaped by theory from a range of disciplines (Haugaard and Clegg, 2009). It is also ‘in principle unobservable, and can only be studied indirectly’ (Murphy, 2011). In LMICs, furthermore, a large proportion of health systems and policy research is driven by (donor-funded) demand for outcome evaluations with a focus on measurable indicators and far less attention is paid to exploring relational underpinnings (Topp and Chipukuma, 2016).

A key lacuna in health policy and systems research thus appears to be the application of existing theories and frameworks around power in the context of LMICs (Buse et al., 2009). Specifically there is a need to: (1) examine key health policy and systems challenges in LMICs through the lenses of existing theories on power; (2) build on these theories and cultivate new theory/frameworks given the particular context of LMICs when compared to high-income countries; (3) expand the application of research to transforming power relations at multiple levels of the system, not just at organizational or grassroots levels; and (4) encourage discussion, analysis and action amongst all actors in the system—for example researchers, practitioners and communities.

This article seeks to facilitate greater engagement with the concept of power by serving as an entry point and guide to researchers and practitioners focused on health policy and systems in LMICs. Our focus is narrowly on the role of power in shaping health policy and systems functioning, rather than exploring power as a determinant of health outcomes (e.g. in the context of social position or macroeconomic structures). The article is divided into four sections. First, we provide a broad overview of the concept of power, in order to familiarize readers to major theorists, ideas and frameworks; this overview is intended to act as a high level summary for those unfamiliar with power research rather than a comprehensive review of the power theory and debates, which lie beyond the scope of this paper. Second, we describe our methodological approach in selecting 10 key resources. Third, we present 10 key resources—frameworks, commentaries and peer-reviewed articles—that exemplify thinking and research on power in health policy and systems in LMICs. We conclude with suggestions for future research, action and commentary in this area.

Theories and frameworks of power

In this section, we provide a broad overview of select key theories and frameworks pertaining to power. Power is defined in the Oxford Dictionary as the ability or capacity to ‘do something or act in a particular way’ and ‘direct or influence the behaviour of others or the course of events’. However, numerous definitions of power exist, many bearing what has been termed a ‘family resemblance’ to one another (Wittgenstein, 1958; Haugaard and Clegg, 2009). English-language discourse on power reflects the ‘Western’ or North American/Western European tradition. Over the course of the twentieth century, theorists from these regions working in the social science disciplines of political science, sociology, anthropology and philosophy have engaged in rich debates over the meaning of power. Debates have included is power best understood as power over (one’s influence over others) or power to (one’s capacity to achieve our goals)? And, how do we reconcile the power of underlying social structures in shaping individual behaviour, with human agency in determining behaviour, thought and action? These discussions are ongoing in social science scholarship, and reflect the pluralistic and contested nature of power, as well as the abundance of ideas that researchers and practitioners may draw upon.

Power is conceptually fluid, and is discussed in the context of closely related constructs, such as authority, accountability, intersectionality and hierarchy. Power also flows through different levels of the system—global, national, provincial and frontline/local. Finally, power is shaped by distinct historical, political, social and cultural forces, reflected in scholarship on rights-based approaches, critical feminist theory, post-colonialism and critical race theory, to name a few.

We have summarized some of the theories and concepts most pertinent to health policy and systems in the following categories—dimensions (how power is channelled) and sources (how power emerges) —recognizing that our synthesis touches on only some facets of the discourse around power, and that there is significant overlap among the categorizations.

Dimensions of power

Power operates through different dimensions, resulting in varied theories regarding the mechanisms for generating influence and controlling decision-making. Historical debates saw cultural anthropologists such as Lévi-Strauss (1968) assert the power of ‘structure’—including the structure of symbols, organizations and language while others, drawing on enlightenment traditions, focused on power held by an individual - or ‘agency’. Still others argued that separation of agency and structure is a false dichotomy. Giddens (1984) for example posits that power, agency and structure are interwoven, noting a duality of structure that allows for power to be both voluntarily utilized by actors via agency, and involuntarily shaped by our surrounding social structures.

Foucault (1994) describes a more ubiquitous and diffuse dimension of power, which is power is neither held by actors to coerce others nor operationalized by structures (Gaventa, 2003). Instead, power is inseparable from the development of knowledge systems (‘power-knowledge’) and is manifested in the creation and acceptance of ‘truths’ within society, which shape and limit discourse and behaviour. Gramsci built on Foucault’s work by describing ‘hegemonic power’, the often unacknowledged or unrecognized dominance of particular viewpoints which generate consent among groups by making specific cultural and social discourses normal, invisible and thus unquestioned (Gramsci, 1999).

One of the most commonly discussed dimensions of power is its ‘relational, zero-sum nature’. Power is sometimes tightly concentrated amongst particular actors, such as policy elites (Wright Mills, 2000) or dispersed more broadly among certain individuals and groups, who then compete for influence (Dahl, 1957). Bachrach and Baratz (1962) expanded on these explicit dimensions of power by positing that another ‘face’ of power was non-decision-making; in other words the underlying social and political norms that block or suppress actors from raising issues or taking action. Lukes (1986) added a further dimension, noting that underlying perspectives on issues or actions are often deliberately influenced by the covert use of power through ‘socialization, education, media, secrecy, information control, and the shaping of political beliefs and ideologies’ (Gaventa and Cornwall, 2001). This concept from Lukes bears similarity to the Gramscian idea of hegemony, but is distinct due to its wilful and deliberative nature.

From a frontline, implementation perspective, the concept of ‘discretionary power’ is often invoked. Lipsky (1980) put forward the notion that the way in which decisions or actions are taken by frontline actors, for example, health providers, becomes the policy itself—as this is the way that consumers and communities experience the policy. Due to the difficulties involved in implementing their tasks at the frontlines, ‘street-level bureaucrats’ use their discretionary power to reinterpret or make policy in such a way that redirects and sometimes contradicts the goals of their organization, but ultimately allows them to address complexity, manage stress and gain more control (Erasmus, 2014).

Finally, a taxonomy of power developed by Barnett and Duvall (2005) for the field of international relations, is increasingly being applied to health policy and systems research, particularly in the context of global health policy (Shiffman, 2014). In this taxonomy, the authors develop the ideas of ‘compulsory power’ (‘the direct control of one actor over the conditions of existence and/or the actions of another’), ‘institutional power’ (‘actors’ indirect control over the conditions of action of socially distant other’), ‘structural power’ (‘the co-constitutive, internal relations of structural positions’) and ‘productive power’ (working through ‘diffuse constitutive relations to produce the situated social capacities of actors’). This last category can be linked to concepts such as discursive power or ideological power.

Sources of power

Among theorists who engage with actor agency, an important domain of power analysis focuses on how actors may draw power from particular sources such as resources, skills, knowledge, access or individual attributes (French and Raven, 1959). Recognizing how individuals, organizations and networks derive their power is important for sharpening our understanding of how and why power flows in particular directions or accumulates with certain groups. Such an understanding also facilitates an awareness of how those sources of power are distributed unevenly as well as how they may be used to improve equity in health policy and systems. Several sources of power have been identified in social science literature, including (but not limited to):

• Technical expertise—Power derived from knowledge, skills and information held by individuals, and the individuals’ authoritative claim to that knowledge (Haas, 1992); power and knowledge are inextricable, and shape discursive norms in society (Foucault, 1994).

• Political power—Power derived from political authority—legitimate, traditional or charismatic (Weber, 1946)—may then be channelled toward achieving certain objectives. Political power can also be understood in terms of state power and its impact on social life (Mann, 1986).

• Bureaucratic power—Power derived from the knowledge and authority of bureaucracies and the administrative machinery through which formal policies are often designed, implemented coordinated (James, 2011).

• Financial power—Power derived from accessibility to financial resources, such as money, assets and property, and the use of that power in influencing decision-making (Bourdieu, 2008).

• Networks and access—Networks in their many structures and forms, such as issue networks or epistemic communities, often serve as a key source of power, derived from sources including collective knowledge, action and homophily (Rogers, 1983; Rhodes and Marsh, 1992), and social spaces (Bourdieu, 1989), which can be described as ‘interactive moments … [that] … create contexts for new social representations … to emerge’ (Scott et al., 2017, p. 2).

• Personal attributes—Weber (1946) defined charismatic authority as emerging from ‘exceptional powers or qualities’; such attributes are tightly wound up with other individual factors such as gender, race, sexuality and religion.

Bourdieu further describes sources of power as having their roots in different types of ‘capital’ such as social capital (the actual or potential resources one gains from membership in a group), cultural capital (the skills, educational qualifications etc. acquired through membership in a group), economic capital (money, property and other assets) and through ‘habitus’ (the socialization of actors to their social disposition). As noted by Erasmus and Gilson (2008), policy processes, and certainly most health systems processes, are dynamic, and therefore, ‘it is important to see sources of power as relational and context-dependent, rather than as fixed possessions or properties of actors’ (p. 364).

Frameworks for practitioners

Directly applying concepts of power to facilitate real-world change is a key goal for many researchers and practitioners focused on health policy and systems (Flores et al., 2009; Sheikh et al., 2014). Several resources and tool-kits have been developed in international development field. A key framework informing several of these resources is the ‘power cube’ (Gaventa, 2005), which operationalizes concepts from many theorists, including Lukes (1974), and consists of three intersecting dimensions—levels, spaces and forms of power. This framework is meant to enable a range of stakeholders—grassroots and civil society organizations, governments, development agencies and researchers—to situate themselves in a multi-dimensional and dynamic conceptualization of power relations, and to then consider the entry points for addressing imbalances of power. Similarly, VeneKlasen and Miller (2002) shared a categorization of power—power over, power with, power to, power within—and provide participatory exercises to help facilitators explain these concepts to grassroots organizations and practitioners, alongside reflections and examples from different settings. An action-oriented handbook from Hunjan and Pettit (2011) also provides approaches for analysis and strategy for grassroots organizations, such as detailed workshop and activity plans. Beyond the resources mentioned here, however, conceptual and applied research can support effective use of power theory and frameworks.

Methodology

We undertook a two-phased approach to identifying and selecting the 10 best resources. The first phase was a collaborative crowd-sourcing initiative to solicit suggestions for excellent resources on power in health policy and systems research. We started by reaching out to experts and inviting contributions from members of the SHAPES (Social Science Approaches for Research and Engagement in Health Policy and Systems) Thematic Working Group of Health Systems Global, via email. The multi-disciplinary nature of our authorship team also facilitated the incorporation of resources from various disciplines. Twenty-four suggestions were received, and following a screening process, shortlisted articles were shared and comments invited via a participatory exercise at the Fourth Global Symposium of Health Systems Research in Vancouver (November 2016).

In the second phase, following critical feedback, we undertook a structured literature review to identify further resources, including from a range of social science disciplines (review table available in the online Supplementary data). We used the search terms ‘power’ AND ‘health policy’ OR ‘health policies’ OR ‘health system*’ in PubMed and Scopus, and identified a total of 2138 articles. We removed 732 duplicates, which resulted in 1405 unique references. Pairs of co-authors screened the title, and where necessary, abstracts, against an inclusion criteria, which specified that articles needed to explicitly engage with power to explore health systems issues. In 242 cases, the reviewers disagreed, and the titles and abstracts were assigned to a third, blinded reviewer. Twenty-nine titles did not have abstracts and were removed from consideration based on title alone. Following the additional blinded screening, a total of 171 articles were identified for detailed abstract review. Of these, 12 articles had already been identified as potential resources in our first phase, and were automatically included for consideration, while the remaining 159 were each ranked by two reviewers on a scale of 1-5, 1 being ‘Article should not be considered for full review’ and 5 being ‘Article should definitely be considered for full review’. Selected articles received a high score from both reviewers.

Thirty-eight new articles were selected for full review—37 from the search process, and one on recommendation from the research team. Two reviewers read each article, and scored them on a scale of 1-5, using the criteria outlined in Box 2, and a similar ranking system as described above. Finally, following discussion amongst the group, we selected 10 resources that in our view exemplified research in health policy and systems in LMICs. Our selections were meant to capture conceptual/methodological and empirical pieces on power in health policy and systems research, and were also meant to represent various levels of analyses, methodologies, geographic regions and conceptual understandings of power.

Some resources in our review included those authored by research team members. To avoid conflict of interest in our selection, research team members (1) did not recommend their own articles; (2) did not review their own articles and (3) did not participate in decision-making on the inclusion of their articles.

Results

Top 10 resources

Our selected resources, summarized in Table 1, are divided as follows—(1) conceptual pieces that explore the way power shapes health policy and systems, or provide methodological guidance on researching power in health policy and systems research; (2) strong empirical examples that showcase various levels of analyses (global, national, frontline/local), methodologies, geographic regions and conceptual understandings of power.

Conceptual

Rushton and Williams (2012) explore factors driving global health governance. By showing how policy debates are shaped by competing ‘frames’ or understandings of global health, the authors draw attention to the way different forms and distributions of power—including the perceived authority of those attempting to ‘frame’ the issue—can result in the material or discursive promotion of certain ideas over others (linking with productive power). The authors suggest that analysis of global health governance often ignores the role and impact of policy-making processes. The authors adopt an inter-disciplinary approach and offer a framework for incorporating such analysis into our understanding of global health governance, by focusing attention on four pillars of the policy process, namely: framing, paradigms, power and the ‘deep core’ of neo-liberalism. This framework is particularly useful for those engaged in research and dialogue pertaining to global health governance and policymaking, as it encourages readers to acknowledge deeply embedded assumptions about the global health ‘playing field’ and outlines an approach to categorizing drivers of global health policy (theoretically and empirically) and to exploring mediating relationships. This resource was selected because the framework may guide the design and analysis of research studies, and can also equip practitioners with concepts to deepen their understanding of how some issues become prioritized over others on the global policy agenda, and the ways that different powerful actors intervene to exercise their influence.

The next resource is a pair of editorials by Shiffman, the first of which stimulated nine commentaries. In his initial editorial, Shiffman (2014) proposes that power is ‘exercised everywhere’ in global health, but that some expressions of power are more evident than others. Shiffman too engages the notion of ‘productive power’ by describing two types of less apparent power—epistemic and normative—and highlights the questions they raise regarding legitimacy and accountability. Epistemic and normative power stem from actors’ claims to expertise and moral authority. Critically, given their bases in science and humanitarian intentions, these types of power are rarely problematized in global health. But Shiffman argues that they should be. Shiffman deploys specific examples; he describes the power exercised by global philanthropies, international scientific journals and multilateral agencies, noting that while global processes undertaken by these groups may appear to be rational, deliberative and neutral, they inevitably result in the elevation of particular individuals and/or agendas. Thus, there is a ‘critical need to investigate how epistemic and normative power get exercised in the global health field’ (p. 299).

In his subsequent response letter, Shiffman (2015) synthesizes key elements of some of the nine commentaries to make three suggestions about how we might better understand the exercise of epistemic and normative power in global health. Shiffman suggests we should acknowledge that: (1) global health is not and cannot be entirely rational; power struggles and political processes to resolve normative questions, such as those related to equity, are intrinsic to the endeavour; (2) global health may be considered a ‘field’ (a concept from Bourdieu), whereby actors employ various tactics and resources to advance their preferences and interests and, (3) inclusive, transparent and fair processes that engage those whose lives are affected by policies are an important way to address the global health legitimacy and knowledge ‘deficit’ (Shiffman, 2015). These two resources (Shiffman, 2014, 2015) utilize theory to describe how power is exercised by global health actors, to question the public accountability of those with power, and to advocate for more inclusive agenda setting. Together, the Shiffman articles—and responses to them—question the distribution and legitimacy of discursive power that has heretofore largely remained implicit and unexamined. We selected this pair of editorials due to their introducing new arguments in the debate on power and global health governance, their effective linking of concrete examples with theoretical concepts, and their generation of robust debate within the field.

Erasmus and Gilson (2008) make a major contribution to health policy and systems research in this paper by outlining the methodological tools that researchers can utilize to investigate power in organizational settings of policy implementation. First, the authors provide an overview of implementation theory from public policy, such as forward and backward mapping (Elmore, 1980). Next, they describe cases that elucidate these concepts. For example, they outline empirical exemplars including organized resistance by frontline health workers (implementers of policy) and the discretionary power these same workers have over patients. The authors go on to provide a robust analysis of methodological techniques to investigate power in various organizational settings, such as clinics and administrative offices. The article includes a detailed discussion of the process of utilizing interviews, observation and document review to illuminate issues of power. The authors also draw attention to the different sites for investigating power within the organizational setting, such as physical objects (e.g. posters outlining policy issues), internal and external discourse (e.g. the use of humour to subtly resist or criticize), and organizational routines (e.g. the order and manner in which consultations with patients are conducted). Finally, the authors walk readers through the difficult terrain of interpreting data and making judgments about power. We selected this piece for its valuable discussion of methodological approaches that enable rigorous research on health policy implementation and outcomes, using theories of power.

Empirical

Front-line/local level

Lehmann and Gilson (2013) provide an example of exploring the influence of power at the frontline of the health system, by investigating the actor dynamics at the heart of the implementation of a community health worker programme in rural South Africa. The authors utilize VeneKlasen and Miller (2002) categorization of power to draw out the expressions and sources of observed power. Furthermore, they utilize ‘actor interface analysis’ to explore the ways in which the ‘interests, relationships, modes of rationality and power of policy actors intersect’ (p. 360). The authors first present an overview of the policy and its implementation process, followed by a description of the policy outcomes observed, in this case, a ‘thinning’ of the stipulated policy to simplify the disbursement of stipends to a few community health workers. Lehmann and Gilson then use actor interface analysis to explore how actors utilized and expressed power, power over others (authoritative power), power with others (discretionary power), power within and power to act. The analysis illustrates how the interaction of various streams of power resulted in a thinning down of the policy at multiple levels in the system. This study was selected as an important empirical contribution to the literature in this field, which links power to policy outcomes through concepts such as contestation, negotiation and tension amongst stakeholders.

Drawing upon social science concepts, including Foucault’s notion of ‘power-knowledge’, and social spaces (Lefebvre, 1991; Kesby, 2005), Scott et al. (2017) describe the role of village health committees (VHCs) in four rural North Indian village in renegotiating and reproducing power dynamics—both within the community, and between community members and external stakeholders. Their findings help to address a key gap in health policy and systems research regarding the gender, caste and other power dynamics of health system structures that straddle the state/society divide, such as VHCs; and the power dynamics between community members and the health system. The authors discuss how VHCs served as a space for some women to test out new boundaries for gender norms, even though these ‘micro-transgressions’ came up against the stark reality of their ‘everyday’ power relations. On the other hand, they also describe how the discourse framing the rationale for community participation committees reinforced existing power relations. Despite the hegemonic discourse of local responsibility, the VHCs were often unable to engender government responsiveness. This reinforced a negative collective identity in the community. We selected this study for its clear and robust methodology (longitudinal qualitative data involving repeat interviews, focus group discussions and non-participant observation), strong theoretical foundations and its contributions to under-examined topics of power, community-level social and gender dynamics, and structural linkages between communities and the state.

Velloso et al. (2013) also draw upon Foucault in their power analysis of an emergency care system in a large city in Brazil. Specifically, this paper is a useful example of researching power relations amongst cadres of health workers, and the impact of these dynamics on service delivery. In describing how ambulance drivers, nurses, nursing assistants and doctors have the power to report or not report on one another, being potentially watched by others at all times, the authors illustrate how health workers find themselves formally accountable to managers but also informally accountable to one another, the community and even the media. Foucault’s concepts of hierarchical surveillance, normalizing judgment and ‘the examination’ (usually discussed in the context of patient examination, but here relating to public examination of providers via the media) are used to explore how power operates in a fluid and anonymous manner among emergency care teams, creating unanticipated and fluctuating structures of accountability, standardized behaviour and values against which individuals can be judged. This article was selected because it engages theories of power to interpret observed behaviour among the healthcare providers, making visible the subtle and often unanticipated ways in which systems of surveillance and judgment become normalized in social settings.

National level

Dalglish et al. (2015) apply a power analysis to examine the policy development and implementation of integrated community case management (iCCM) for childhood illnesses in Niger. The authors utilize an iterative approach, allowing for their initial data analysis and theory review to inform their framework of power, and then reapplying that framework back to the full data set. The authors identify three main sources of power—political authority, technical expertise and financial resources—and in their analysis, identify both positive and negative uses of these forms of power in the promotion of iCCM in Niger. A key strength of this article is the richly detailed contextual analysis in which the case study is situated. For example, the authors provide a nuanced historical and political analysis, enabling readers to understand better how sources of power were able to find expression. From a policy analysis perspective, this article also highlights how different forms and expressions of power interact to shape political priorities and policy trajectories. We selected this article due to its contribution to our understanding of power and health policy analysis in LMICs; the analytic approach incorporating deductive and inductive theory on the forms of power; and its strong grounding of the policy process in historical, political, and social context.

Parkhurst et al. (2015) demonstrate the importance of broader contextual factors in understanding power in their analysis of political resistance to male circumcision in Malawi. Drawing on interviews, document review and historical analysis, the authors examine how and why a ‘narrative of defiance’ and a ‘narrative of doubt’ towards male circumcision emerged in Malawi, situating the findings in the country’s ethnic, religious and identity politics. For example, the authors highlight the historical context of Malawian resistance against European colonialism, and how this shaped local policy actors’ sense of needing to defend their country against outside influence, in this case, perceived donor hegemony, to promote male circumcision. Although this article does not use an explicit power theory, it adopts an explicit power frame, and the authors illuminate several key issues pertaining to power—resistance, pressure and hegemony. In doing so they inform and complicate our understanding of why some countries are considered ‘leaders’ and others ‘laggards’ by the global health community in their adaptation and implementation of global recommendations. This article was selected as an example of applying a power analysis to national-level health policy, drawing out sharp connections among post-colonialism, local socio-cultural dynamics, global health governance and national health policy.

Kapilashrami and McPake (2013) explore the dynamics between global and national level in their study on the Global Fund and the governance issues associated with its HIV programme in India. Using the concept of public and hidden transcripts, advanced by Scott (1992), the authors contrast the Global Fund’s public and official descriptions of its organizing principles and operations, with concealed resistances and discourses of domestic actors that are not reflected in the dominant narrative of ‘successful inclusiveness’ on the global stage. The study utilized critical ethnography, with fieldwork taking place in five Indian states over a two-year period, and involving interviews, consultancy by the first author with a grant recipient, observation and site visits. The authors find that within the Global Fund’s ‘mixed form of governance’ involving a range of Indian and transnational networks and organizations, the public and hidden transcripts differ on a number of fronts. Differences include the publically unacknowledged role of social, economic, political and network power in enabling certain organizations to more successfully secure grants; the reality of insufficient oversight of the Country Coordinating Mechanism by partnering national and sub-national organizations; and the fact of competition among partnering organizations, brought on by the intensive reporting demands. This richly detailed and analytically robust study was selected as a valuable example of utilizing ethnographic research in studying power at the interface of global and national health policy and systems.

Global level

In this book chapter, McNeill et al. (2013) discuss the role of power in shaping international dynamics in the area of childhood immunizations, through an exploration of the functioning of key actors at the global level. McNeill et al. use prior research and published accounts to describe the establishment of the global vaccine alliance, GAVI, its impact on institutional relationships at the global level, and its influence on national-level decision making. The authors build their narrative by examining three sources of power and authority—financial, expertise and mandate—and highlighting how these sources of power enabled global-level players such as governments, international organizations, foundations, transnational private firms and international NGOs to shape outcomes. The authors also indicate the consequences of the power asymmetry between ‘strong’ and ‘weak’ actors in the global sphere, noting particular threats to the autonomy of LMIC governments. Although this chapter does not present a detailed methodology, the authors effectively use three key sources of information in power analyses—primary interviews, secondary evidence and observation. They present a compelling example of power analysis by drawing upon multiple sources of evidence (not just their own primary data), and interpreting that evidence using a strong theoretical base. We selected this resource due to its success in weaving together multiple sources of evidence to interpret and put forward an empirically sound power analysis, and to analyse power at the interface between global and national actors.

Conclusion

The exercise of gathering and reviewing these materials highlights the growing attention to this topic within health policy and systems research, but also that more research, theory building and reflection are required. We highlight three areas requiring more attention below.

First, more work is needed to understand how sources, dimensions and expressions of power differ at different levels of the health system, how these may be linked across levels of the health system and what theories and frameworks are most suitable to understand power at these various levels. Second, few studies explicitly engage with issues of power at the frontline of healthcare provision in LMICs (e.g. among providers, between providers and community members, and through management and organizational culture), with notable exceptions primarily from southern Africa and India (Behague et al., 2008; Grimen, 2009; Sheikh and Porter, 2011; Lehmann and Gilson, 2013; Oliver et al., 2015; Saprii et al., 2015). This is a major gap given that frontline workers, and the power dynamics that influence their performance, sit at the juncture between clients/communities and the broader health system and have great influence over equitable delivery of care. Furthermore, we observed few conceptual frameworks specific to power and frontline health workers. Frameworks that have been used include street-level bureaucracy theory from Lipsky (1980), an adaptation of the hardware/software model to explore power, trust and accountability in frontline health services (Topp et al., 2015), and the concept of social spaces. The emerging field of respectful maternity care is one potentially fruitful avenue of research and theory development on frontline workers and power (Bowser and Hill, 2010).

Third, we noticed a lack of empirical work on action-oriented research addressing power dynamics in health policy and systems. This perhaps reflects the broader imbalance between power analyses seeking to contribute to knowledge generation, and those seeking to directly address power asymmetries through dialogue, organizing and activism (Haugaard, 2011). It is also likely that practitioners communicate their experiences of power through channels other than academic journals. Recognizing this, a review and curation of audiovisuals and training materials (such as those available at www.equinetafrica.org and www.copasah.net) specific to power may be a more appropriate mode of synthesis and communication for this type of work.

Although researchers and practitioners sometimes work toward distinct goals, both groups have much to gain from sharing their respective insights. For example, the concepts of power to and power over are well grounded in theory, and have been used to formulate action strategies for grassroots groups (VeneKlasen and Miller, 2002; Kaim, 2013). Similarly, in development studies, the ‘power cube’ has been previously used in research to unpack power dynamics (Hossain and Akhter, 2011). To harness the capabilities of both sets of stakeholders, methods such as participatory action research (Loewenson et al., 2014) and research that is co-produced by researchers and implementers that ‘increase the empowerment and social change potential of their research’, should be widely encouraged, to ensure that power analyses are democratized (Baum, 2016).

There is also a need for additional methodological tools to explore power. Although historically quantitative health policy and systems research has not fully engaged with issues of power (in either design or analysis), social epidemiologists conducting research in the United States and other high-income settings are building an extensive evidence base of the social gradient in health, and social network analysis is increasingly being used to visualize and quantify how power is distributed in policy networks (Wang, 2013). We also support recent calls for more integration of social science theory into health policy and systems research, allowing us to build a more robust evidence base and theories about how power and other dynamics affect health policymaking and service delivery (Daniels et al., 2017; Van Belle et al., 2017). Further, as is the case for peer-reviewed research in general, research on power authored by individuals based in high-income settings is over-represented in the published literature, and efforts should be taken to spotlight voices from LMICs. Finally, researchers and practitioners have a responsibility to reflect on their own position and role in propagating the status quo, and attempt to use their positions to ‘co-create the conditions at every level in the system that can make that locally driven transformation possible’ (Freedman, 2016).

Our methodology presents certain limitations. First, our search terms and review process was designed to identify resources that explicitly engage with power; resources that only discussed power in broader constructs, such as hierarchy, competition, resistance and human rights, were not included. Second, our review was limited to English language publications, thus potentially resulting in the omission of important and useful non-English resources. Third, although we used Scopus to capture the social science literature on power in health policy and systems, we recognize that a broader search (e.g. different databases and books) might have resulted in more resources directly from the social sciences.

In conclusion, power analysis is a key pillar of health policy and systems research. However, more methodologically and theoretically diverse studies are needed, and more insight and analysis from a range of stakeholders, particularly those from LMICs and those at the frontlines. In this list of resources, we have attempted to provide a broad overview of power, and to share exemplary scholarship pertaining to power, and seek to encourage researchers and practitioners in health systems to weave analysis of power into their research. Further efforts in this space will contribute to expanding and deepening our collective understanding of the underlying power dynamics that shape health systems and health inequity around the world.

Acknowledgements

Our thanks to Anna Abelson for her contributions to the literature review. We thank the members of the sub-cluster on Power and Health Systems of SHAPES (Social Science Approaches for Research and Engagement in Health Policy and Systems) Thematic Working Group of Health Systems Global for their early contributions, and the individuals who provided feedback during and after the Health Systems Global Symposium in 2016. We also thank the anonymous peer reviewers of this manuscript whose feedback substantially strengthened the final product.

Supplementary data

Supplementary data are available at Health Policy and Planning online.

Conflict of interest statement. None declared.

Funding

VS is currently supported by the Agency for Healthcare Research and Quality under grant award T32 HS000087 (PI: Jane Holl, MD, MPH). Its contents are solely the responsibility of the authors and do not necessarily represent the official views of AHRQ.

References