Who should serve on health care boards? What should they do and how should they behave? A fresh look at the literature and the evidence

Public boards of directors face challenges in demonstrating effectiveness and return on investment. Health care boards in particular operate in a high risk service and political environment, where both patient safety and financial sustainability are paramount. The motivation in this article is to make sense of the conflicting and competing theories which explain the purpose of boards, and the sometimes weak and contradictory evidence for effective board practices. The main contributions of the study are, first, the use of a realist approach to understand underlying assumptions behind the main theories for health care boards, and, second, practical suggestions in relation to board composition, focus and behaviours, according to circumstances. Amongst its conclusions, this review indicates that board size should be limited, especially for newer organisations, physicians on boards are associated with better quality of clinical care, and choosing to operate diligently with a focus on strategy and on monitoring, a close grip on the business, and strong support for executives are all important. Subjects: Social Sciences; Arts & Humanities; Medicine

ABOUT THE AUTHOR N. Chambers has a wide range of research, teaching and advisory interests. These include board governance in the public and charitable sectors, leadership development, health policy and management, primary care, emergency planning, ehealth and clinical costing. This article is based on a literature synthesis of effective health care boards, which was funded by the Department of Health in England. Her current major research project (also funded by the Department of Health) involves leading a team from the Universities of Manchester and Birmingham, together with the Nuffield Trust, which is investigating changes in board level leadership in hospitals in England following the publication of the Francis Report into the failings at the Mid Staffordshire NHS Hospital Foundation Trust. This study is due to be published in spring 2018.

PUBLIC INTEREST STATEMENT
The Francis public inquiry published in 2013 provided compelling evidence about the extent to which the board of Mid Staffs hospital in the UK failed patients, leading to unnecessary suffering and needless deaths. Public boards of directors, and health care boards in particular, operate in a high risk environment, where both patient safety and financial sustainability are paramount. Meanwhile, the evidence for effective board practices is weak. The main contributions of this study are, to understand the different purposes of boards, and to offer practical suggestions as to who should serve on boards, what they should do, and how they should behave. The review concludes that board members face choices. The composition and focus of effective boards varies according to the specific priorities that organisations face. For safe, high quality and effective care for patients, and to prevent another Mid Staffs, health care boards need to provide close grip and strong support.

Background
Public organisations have to demonstrate optimal performance to their constituencies, and an appropriate return on investment for taxpayers or charitable donors. Nowhere is this more palpable than in the health care sector, where boards hold the ultimate responsibility for the safety, positive experience, effectiveness and efficiency of patient care. Even high performing providers such as Virginia Mason Medical Center in the US which is held up globally as a role model for its high standards of patient safety, has experiences of lapses. In this case, the hospital reported 39 infections and 18 deaths linked to a specific endoscopic procedure between 2012 and 2014, amid accusations that the organisation had been slow to react, or to file reports of adverse effects in a timely fashion (www.seattletimes.com, 2015). Many patient stories in the Independent Inquiry followed by the Public Inquiry into care at the UK Mid Staffordshire NHS Foundation Trust from 2005 to 2009 (The Francis Reports, 2010, 2013 provided compelling and overwhelming evidence about the extent to which the board in this case failed patients. High profile failures in the sector have been financial as well as clinical. The Kids Company, a non-profit organisation with 650 staff who offered psychotherapeutic care to troubled children in various centres in England, abruptly filed for insolvency in 2015 despite receiving government funding over nearly 20 years of around £40 million (US$ 56million) and an emergency grant of £3 million (US$ 4.36 million) paid only 2 weeks earlier. The board of trustees in this case has been accused of failing to impose sufficient control or exercise their proper function (Public Administration & Constitutional Affairs Committee, 2016, p. 24).
These failings call for the gaps in our knowledge about the composition and characteristics of effective boards to be addressed. The functions, features and purposes of health care boards do have much in common with their commercial counterparts (for a review of similarities and differences, see Chambers, Harvey, Mannion, Bond, & Marshall, 2013, p. 35-39). Boards in all sectors share the common endeavour of establishing direction for the organisation, monitoring activities, setting the organisation's values and accounting to shareholders and stakeholders. Generic theories and evidence can therefore provide a helpful guide as to who should serve on health care boards, what they should do and how they should behave. This article will suggest alternative courses of action for members of health care boards, using the learning from different theoretical standpoints on the purpose of boards and sources of evidence about effective boards, and relating it to the health care context. It aims to offer fresh insights into Professor Naomi Chambers, Professor of Healthcare Management at the University of Manchester, talks about the characteristics of effective Boards in the NHS, as part of the forward look seminar on Development and Training. © Mid Staffordshire NHS Foundation Trust. This video is available to view via YouTube at the following link: https://www.youtube.com/watch?v=89EAY-gX6DE effective board composition, structures, processes and behaviours, and to further an understanding of how boards can affect organisational performance. The main question is how can selected corporate governance theories, and the empirical evidence about effective boards, guide choices in health care with regard to composition, structures, focus, and behaviours? Haverland and Yanow (2012) argue that choices of methods and their underlying "ways of knowing" depend on the goal or purpose of the research. This article is based on a realist approach (Wong, Greenhalgh, Westhorp, & Pawson, 2014) to an evidence synthesis of a diffuse literature. This approach emphasises the contingent nature of the evidence and addresses questions about what works in which settings, for whom, in what circumstances, how and why. It draws from the seminal work of Pawson and Tilley (1997) on realistic evaluation which proposes that behaviour choices guiding interventions are embedded in a range of individual, organisational and societal processes and norms. A number of landmark governance texts are characterised by interdisciplinary synthesis, for example, legal and economic (Berle & Means, 1932). A traditional systematic literature review is less able to take account of the multiple and inter-connected variables that influence boards and their performance. A realist angle on the other hand builds on the growing acknowledgement of the importance of contextual factors in board governance (see for example, Bammens, Voordeckers, & Van Gils, 2011). Furthermore, boardroom practices have been described as a black box (Huse, 2007;Selim, Verity, & Brewka, 2009), because of the difficulties in understanding the dynamics and the inner workings of boards. A realist angle therefore seems a sensible approach to take: the study aims to open that black box.

Method for this review
Realist synthesis belongs to the paradigm of theory-driven inquiry. This is an approach rather than a strict technical procedure. It starts with knowledge and theory and ends with more refined knowledge and theory, along the way sifting ideas and empirical evidence. It continuously searches for explanations of programme effectiveness. It draws from Campbell and Russo's (1999) notion of organised distrust and ambition to secure methodological advances and trustworthy reporting. In our case, using evolving standards in realist review (Wong et al., 2014), the synthesis addresses questions about how boards operate, in what circumstances, and why, and the influence that boards may have on organisational performance. These are the context-mechanism-outcome configurations that are the cornerstone of realist methodology. Realist review learns from, rather than controls for, real-world phenomena, thereby providing an acknowledgement, for our study, that no two boards are the same in composition, context or stage of development.
We searched the literature using and linking key terms related to the main research questions. Abstracts were reviewed to test for relevance and to eliminate duplication before selecting a smaller number for closer scrutiny. Recognising the importance of stakeholder involvement in the review process (Jagosh et al., 2012) a combined expert and lay advisory and stakeholder group was convened to support the development, honing and refinement of the research questions and testable propositions and to check emerging findings. The research was initiated in 2011 with subsequent iterations in 2012, 2015 and 2016.

Alternative theories about the purpose of boards
The synthesis uncovered multiple locations of theory and evidence across different disciplinary traditions. In line with a realist view, it seemed to us that underpinning the four main theories about the role of boards were a series of contextual assumptions, mechanisms and intended outcomes. This builds on Lynall, Golden, and Hillman's (2003) view that boards have different needs according to which stage they are in their lifecycle and McNulty, Roberts, and Stiles's (2003) call for theoretical pluralism. We argue that rather than one or other of the theories being, in general, superior or preferred, context and desired outcomes will guide which choice of theory (or combination) and related mechanisms best fits the circumstances. This will be summarised in a realist interpretation framework of selected board theories and choices.
We have focused on four main theories (agency, stewardship, resource dependency and stewardship). These were dominant in our search of the literature, both by volume of articles and also by citation rate (Chambers et al., 2013, pp. 16-17). Other theories -for example, board power or managerial hegemony, public accountability theory and board legitimacy all have merit and relevance. They can be seen as derived, at least in part, from the main four, but detailed discussion of these is beyond the scope of this article.
The overview that follows draws from the generic literature on boards and governance and then is related to the health care sector. For that reason some of the terms, for example, around shareholding and ownership, are not always directly applicable to the health care sector although, as we have signalled above, the principles underpinning the range of purposes and functions of boards cover all sectors.

Agency theory and health care boards
The first and earliest fully developed theory about boards is agency theory, at the heart of which lies questions about the organisation and ownership of assets and the distribution of power that goes along with that. Over the past few centuries, the holding of assets has moved from being an active to a passive affair. When ownership is held by a very large number of individuals and bodies with none holding a significant proportion, control is effectively handed over from owners to managers. Agency theory is predicated on the notion that the shareholders' (or, in the case of the health care sector, the stakeholders') and managers' interests are likely to be different, and that the behaviours of both sets of actors are characterised by self-interested opportunism.
Agency costs are incurred in acting to minimise the gap between the two sets of interests. Jensen and Meckling (1976) elaborate on the three sources of agency costs: monitoring expenditure, bonding costs (to tie agent in) and residual loss (the costs of agents' decisions which diverge from those which are in the best interest of principals). They also emphasise the generality of the agency problem, both at all levels of management and also across different types of organisations, including non-profit organisations, government corporations and cooperatives. This is derived from a view that most organisations (private firms, non-profit organisations, government bodies) serve as a nexus for a set of contracting relationships among individuals (Jensen & Meckling, 1976). Fama and Jensen (1983) also describe the circumstances in which, according to agency theory, a separation of decision management (generation and implementation of proposals) and decision control (ratification and monitoring processes) is indicated. These include large corporations, and also most non-profit organisations and government bodies where there is a degree of complexity or size which means that there is a hierarchy and a diffusion of decision management, and where important decision-makers are not exposed to significant risk by the financial effects of their decisions, which is indeed a distinctive characteristic in public health care sector organisations.
Agency theory carries a set of underlying beliefs about human behaviour, and for health care sector bodies, certain assumptions that governments make about human behaviour. For health care, the assumption behind agency theory is about the need to rein in the self-serving behaviour of managers and clinicians, as well as the need to mitigate against poorly performing managers which has also been termed "honest incompetence" (Hendry, 2005). Critiques of agency theory claim that it downplays the complexity of individual motivations and permutations of organisational life and that it relates to a view about the self-centredness of human behaviour in organisations which is now contested (Perrow 1986). It also diminishes the purpose of the health care board in terms of setting the mission and values for the organisation.

Stewardship theory and health care boards
The second, stewardship theory, was developed as a challenge to beliefs that managers are always self-interested rational maximisers, first by Donaldson (1990) and developed by Davis, Schoorman, and Donaldson (1997) and Cornforth (2003). According to stewardship theory, the goals of board directors and of their managers are aligned, with the latter being intrinsically motivated to act in the best interests of the organisation and to focus on intangible rewards such as opportunities for personal growth and achievement. The emphasis is on the board's role in advising and developing strategy, and on positioning themselves as an additional asset to the organisation, in a common endeavour with managers, rather than on controlling and monitoring performance. Implicit in stewardship theory is the understanding that the owners, or in the case of health care, the stakeholders on the board, are prepared to take risks on how managers will run the organisation, indicating a level of trust that is absent in agency theory.
The main critique of stewardship theory is that it can lead to an oversight vacuum, strategic drift, inertia or a danger of "groupthink" on the board (Davis et al., 1997). This can be mitigated by the rise of institutional investor activism (Anderson, Melanson, & Maly, 2007) or in the case of health care, the influence of patients and the local public.

Resource dependency and health care boards
According to the third, resource dependency theory, the organisation is an amalgam of tangible and intangible assets and capabilities (Barney, 1991). Given that all organisations depend on others in order to survive and thrive, this theory suggests that managing external relationships in order to leverage influence and resources is the prime purpose of the board. Board members are selected for their background, contacts and boundary spanning acumen, facing in to management and at the same time out to shareholders and other stakeholders. The outward-facing board can minimise the uncertainty engendered by external environmental factors and dependencies. Benefits that board directors can bring by using this approach include advice, access to information, preferential access to resources and legitimacy (Pfeffer & Salancik, 2003).
A review of resource dependency theory (Hillman, Withers, & Collins, 2009) confirmed theoretical support and empirical evidence for this lens for understanding boards and its utility early in the life cycle of organisations and in times of stress or decline. The theory can be criticised for an excessive focus on an external focus. It underplays the board role in determining its own future through strategising, and in exercising oversight of internal management actions and performance (Hodgkinson & Sparrow, 2002). On the other hand, for health care boards, productive relationships with government policy actors, local political representatives and the media can play an important part in gaining legitimacy, building reputation and obtaining financial and other kinds of support.

Stakeholder theory and healthcare boards
Stakeholder theory comes from nineteenth century developments of alternative forms of organisation and control in the shape of mutuals and cooperatives. There is a view that an exclusive focus on shareholder interests does not hold the key to good corporate performance and effective accountability. In an age of vocal consumer groups, employee activism, media monitoring and social networking, the assumption that only shareholders are capable of effective monitoring looks increasingly flawed (Clarke, 1998).
According to stakeholder theory, board members work to understand and represent the different interests of individuals and groups who have a "stake" in the organisation. Stakeholders are all whose participation is critical to the survival of the organisation. (Clarkson, 1995). These include managers, employees, customers, suppliers, regulators, government, pressure groups, media and local communities. The argument runs that the inclusion of a range of different stakeholders drives an inclusive approach which represents a wide spectrum of societal opinions, balances competing priorities and avoids dominance by one group with particular interests. Amongst the myriad of stakeholders, it also argues that boards have to identify the critical stakeholders (for example, key staff groups) whose commitment is essential for long-term value creation. In some interpretations of stakeholder theory (Blair, 1995), a hierarchical distinction is made between "taking into account" the views of stakeholders as distinct from "being responsible to" the shareholders.
In practice, given that knowledge is these days the pre-eminent resource, and knowledge is generated by individuals, elements of the stakeholder approach are increasingly utilised. Only by creating great relationships with employees, customers, suppliers, investors and the community will organisations learn and change fast enough (Clarke, 1998) and it therefore makes sense for board membership to include representation from those who add value, assume risk and possess strategic information.
Despite the fact that stakeholder governance models are deeply embedded in some countries in Europe, notably Germany, and in Japan, and claims for these countries' industrial and social success are often based on this model, the empirical evidence for stakeholder theory is weak. Arguments against the stakeholder view include the lack of clarity about stakeholder expectations and complexity of trade-offs if stakeholder interests are to be taken account of. The theory can be criticised for encouraging risk averse, inoffensive but bland and lowest common denominator decision-making. For health care boards, the theory legitimates the purpose and work of the board because between them the stakeholders represent society which the organisation is there to serve. But it can lead to large and unwieldy boards with people recruited for whom they represent rather than for their board-level skills (Greer, Hoggett, & Maile, 2003).
This brief review of the four main and distinct theories about the function of boards demonstrates an absence of consensus, with seemingly competing and conflicting claims for legitimacy. How can this aid boards' understanding of their purpose? Each of the theories highlights important aspects of the role of boards. Agency theory clearly articulates the need to monitor the activities of management. Stewardship theory on the other hand offers the power of the "joint endeavour". Resource dependency theory highlights the impact of external influences and leveraging of expertise. Stakeholder theory is persuasive about the need for organisational and societal inclusivity.
What begins to emerge is that alternative theoretical standpoints offer ways forward in particular circumstances, and depending upon what purpose and outcomes boards are most desirous of achieving. This is highlighted in the table below in relation to health care boards. This suggests that the main purpose or priority of health care boards will be, at any one time, either patient safety, innovation, improved reputation or long-term organisation sustainability and the consequent principal mechanism to achieve these goals will be, respectively, control, advice, external advocacy or stakeholder engagement. These choices of priorities and mechanisms will in turn be driven by differing prevailing contexts, for example, potential or realised failures of care, circumstances conducive to entrepreneurialism, degree of system interdependence, and the importance of collaborative effort. In most cases, health care boards will face a combination of challenges leading to a composite set of priorities and some inevitable trade-offs between them. This therefore now leads us to the proposition that boards do have real choices in relation to composition, processes, focus and behaviours (Table 1).

Choices about board membership
We now review the evidence about different board compositions, which also relate to the main theories about the purpose of boards, before turning to how our findings might shape choices about who and how many people to appoint on to a health care board. In summary, boards face choices about the size of the board, number of independent (non-executive) directors and their background and expertise, the balance between insider and outsider experts, non-experts and lay people, length of tenure of board members and when to pay close attention to diversity issues.
There has been a rise in the proportion of outside or non-executive directors on boards representing the interests of shareholders, owners or taxpayers. In addition, there has been the evolution of different board structures -in particular, the unitary US/Anglo Saxon model where managers and outside directors sit together on a board, in contrast with the two-tier Rhineland model, typified by the German Vorstand (management board) and Aufsichtsrat (supervisory board). To address the perceived problems outlined in agency theory, there has also been a move to split the roles of chair and chief executive, in contrast to a single person holding both positions (CEO chair duality).
The search for the ideal board constituency and size is however far from over. Dalton, Daily, Ellstrand, and Johnson (1998) found no links between board composition, leadership structure and financial performance, and nor did an analysis a decade later (Heracleous & Jacobs, 2008) which examined the separation of the chair and CEO roles and the proportion of managers to independent directors on the board.
The role of the non-executive is not always clear and their impact is highly contingent. There is some weak evidence (Shen & Cannella, 2002) that having a majority of independent outside/ nonexecutive directors on the board is associated with better performance. This is particularly true when the CEO has been with the company for a while and also when pursuing a strategy of cost-efficiency rather than a strategy of innovation. A majority of non-executive directors is not favoured for membership organisations, start-up organisations or in circumstances in which, including in the health care sector, the top team is newly appointed and where a degree of laissez-faire will encourage responsible entrepreneurial behaviour (Ramsay & Fulop, 2011). The work of Perry and Shivdasani (2005) demonstrates that, during crises, boards with a majority of outside directors are more likely to initiate restructuring and secure improvements in operational performance. There is also evidence that longer tenure of independent directors has a beneficial effect, although that eventually diminishes as their critical perspective waxes and then wanes over time. The UK Parliament report on Kids Company (2016) specifically criticised the fact that both the independent chair and the chief executive had been in position for a very long time.
A question that is often asked is: What is the ideal size for a board? Smaller public sector boards modelled on business rather than philanthropic models are associated with swifter decision-making. Bennedsen, Kongsted, and Nielsen (2008) suggest that there may be an inverted U-shaped relationship between board size and performance. An increased number of directors improve monitoring and advising functions but there is a limit (which might be around 19 directors) beyond which coordination, control and decision-making problems outweigh the benefits.
Gender diversity has been examined over a number of years: some authors have found positive and others negative effects on ability to achieve strategic change and on firm performance. A study of 1,500 US public companies found that, on average, the companies managed by a female CEO performed better (Vieito & Khan, 2012) . Others have indicated that the presence of women in firms that otherwise have weak governance has a positive impact on performance, but may have a deleterious effect in other circumstances because of over-monitoring (Adams and Ferreira, 2009). Terjesen, Sealy, and Singh (2009) make the case for increasing the number of women on boards founded on four main lines of argument: the need to tap into the widest possible talent pool; the fact that diverse boards understand their stakeholders better; it prevents "group think": and the increasing (recent) evidence that women on boards are associated with better firm performance. A large-scale survey of non profits in the US demonstrated that ethnic minority diversity contributed to the adoption of stronger accountability practices, for example, around well-functioning audit committees, external audit, conflict-of-interest and whistleblowing policies (Herman, 2009).
With regard to boards in the health care sector, governance arrangements exhibit a hybrid of the corporate and philanthropic models, sometimes dependent on technical or regulatory pressures, and environmental conditions such as urbanisation and degree of competition (Alexander & Lee, 2006). These respective models, with their differences in board size, representation, compensation of board members and committee structure, can be said to relate, in the case of the corporate model, to agency theory, and, in the case of the philanthropic model, to a combination of stewardship and resource dependency theories. Corporate as opposed to philanthropic models of governance were found to be associated with greater efficiency and a larger share of the local market, and these findings were more pronounced in publicly run hospitals than in others (Alexander & Lee, 2006).
In relation to membership of health care boards, Alexander and Lee (2006) in the US, and subsequently Veronesi and colleagues (2014) in the UK, both found that greater physician board participation was associated with better operational performance of a hospital.
In summary, it seems to us that the choices that boards have with regard to membership relate to the circumstances that their organisations face. The case of health care boards illustrates this clearly. Larger boards and a higher proportion of non-executives, with due regard to diversity, may be called for when strong monitoring and robust challenge of executives is indicated. Smaller boards facilitate faster decision-making, enactment of audacious strategic change, and may have a higher risk tolerance, suitable for start-up organisations, when the CEO and their team are finding their feet or when the external environment is conducive to greater risk taking. Women on boards offer a wider set of perspectives. In health care, the paradox of ensuring conformance (patient safety) at the same time as pursuing performance improvement and innovation (clinical effectiveness and efficiency) is ever present. This suggests the need to include physicians on the board both as outsiders (in a monitoring and advising capacity) and as insiders (in an expert capacity). Garratt's (1997) classic and widely cited cycle of activity and four functions of boards (defining organisation goals, making strategic choices, monitoring performance and accounting to shareholders), encompass elements of all the main theories of boards. The emerging evidence is also that high performing boards across all sectors concentrate on shaping strategy, resource identification and use, and talent management. Lorsch and Clark (2008) identified four key areas for boards: defining the long-term, taking the lead in finance discussions, strategy discussions and developing talent. These authors also stress the importance of matching the emphasis attached to different board tasks with the prevailing institutional and external environmental conditions. Boards do face choices with regard, in particular, to how they enact their role in strategy and in monitoring, and in health care specifically how they deliver their role in monitoring the quality of patient care.

Choices about strategy
A number of authors including Useem (2006) and Lorsch and Clark (2008) and also counsel for more attention to the long-term. Too much board time spent in the area of compliance and being too hands-off in the area of strategy brings the danger of unnoticed slow decline. A relatively hands-on approach to strategy formulation and execution, including breaking down large strategic decisions into smaller sequential ones for board-level consideration is urged. The strategic role of the board in Stiles and Taylor's work (2001) is specifically identified by respondents in their empirical study as the primary role. Unlike Useem (2006), Stiles and Taylor pinpointed the board role here as not to formulate strategy but to set the context, answering the question "what business are we in?" and to act as gatekeeper in relation to strategic choices. This view draws from a number of the main board theories, for example, the board is acting as a strategic arbiter in accordance with the agency view, and is executing a boundary spanning role in relation to determining strategic fit with the external environment.
There is some evidence (Stiles & Taylor, 2001) that financial control by the board is more formalised than strategic control, aided by the scrutinising existence of the audit committee. Control systems including external benchmarking can also act as tools for diagnosis but require committee effort to pick out trends, threats and opportunities. The social ties that are built in the course of committee work increase a sense of common purpose and support a stewardship theory of corporate governance. The threat is that they can reduce vigilance and therefore the effectiveness of the board as a mechanism of control (Stiles & Taylor, 2001).
We have already noted that a primary role for boards in general is in the development of strategy which has a strong and close fit with the external environment. The same appears to be true in the health care sector. Over time, a number of authors (for example, Barrett & Windham, 1984;Ford-Eickhoff, Plowman, & McDaniel, 2011;Lee, Alexander, Wang, Margolin, & Combes, 2001) have found a link between the balance of roles taken up by the health care board and the internal organisational and external environmental conditions. Separately, McDonagh and Limbdenstock (2006) and Emslie, Oliver, and Bruce (2006) found two associations in the health care sector using the same board performance measurement tool. Organisations with boards that had a higher score of engagement in all areas of board activity performed better and attention to the area of strategy is strongly linked to good financial performance. Emslie also found that board focus on politics (defined broadly as relationships with internal and external stakeholders) is related to higher levels of staff satisfaction.

Choices about monitoring
We have already noted the importance of patient safety and quality of care for health care organisations. The health care board is moving away from its tradition of deferring to medical staff, synonymous with stewardship theory, and towards a closer monitoring approach akin to agency theory. Jiang, Lockee, Bass, and Fraser (2009) found that a board focus on clinical quality, and in particular, a single board quality committee with physician representation was associated with better processes of care and with lower mortality. Millar, Mannion, Freeman, and Davies (2013) identified the importance of strong and committed, visible and strategic leadership by the board and the presence of well-informed and "quality literate" board members in the effective oversight of quality and patient safety. Mannion, Freeman, Millar, and Davies (2016) found a positive association between staff feeling confident about raising patient safety concerns, and stronger overall self-reported board governance scores, especially in the area of promoting organisation values. This finding connects with the difference that board behaviours can make in relation to shaping organisation culture, which will be elaborated on further in the following section on choices in board dynamics.
In circumstances when the main intended outcome for an organisation is for the minimisation of risk, then it follows that the principal focus or mechanism should be monitoring, predicated on agency theory. The findings in relation to risk management and patient safety in the health care sector suggests that this mechanism may need to be tempered by adding a stakeholder perspective -for example, around the promotion of collectively developed and endorsed organisation values, about which more in the following section -to ensure that staff call out or speak up when they identify concerns. This approach may come at the expense of innovation, and an encouragement of managers to take (responsible) risks. There is also the possibility of crowding out the strategy task, especially for health care boards. Like Selim et al. (2009), Huse (2007 has argued for opening up the black box of boards. On the outside, he argues, there are the internal and external actors, board members themselves, structures, processes and tasks. Inside the box are organisation behavioural concepts such as trust, emotions, politics and expectations.

Choices about board behaviours
Models of board behaviour and exercise of power can be related to the main theories about boards: for example, agency theory is connected to a challenge and compliance set of behaviours, whereas stewardship theory relates to a partnership style of working. Davis et al. (1997) argue that if there is an agreed stewardship relationship on the board, the potential performance of the firm is maximised and if there is an agreed agency relationship on the board, risks and costs are minimised.

Choices about the exercise of power
In a stakeholder model, board members tend to be most vocal when articulating the interests of "their" constituency. Boards motivated by power are those to which influential boundary spanners have been recruited, in a model closely related to a resource dependency view of the board. In his influential piece on the exercise of power by independent directors of the board, Mace (1972) argued that independent directors do give advice, set discipline and provide decision-making in times of crisis, but are otherwise less likely to exercise influence over strategy or to ask discerning questions. This suits CEOs who, according to the managerial hegemony frame, do not want the directors too involved. In a more nuanced contribution, Lorsch and MacIver (1989) chart the rise and fall and rise again of the potential power vested in the independent director. This has to do with the increase in the proportion of outside directors, and the multiple sources of power which include legal authority, stakeholder expectations, personal confidence as well as the power of unity of purpose amongst board members. The negotiation of power sharing, not necessarily equalisation of power, between non-executives and managers offers a way out of either unhelpful board dominance by one or other party. Kosnik (1987) argues for merging agency and managerial hegemony theories to clarify the contingencies that might affect board performance. The argument runs that one is not more valid than the other but that the switching rules need to be identified. Within health care, the power of the CEO, the medical director and the nursing director waxes and wanes and the non-executive members of the unitary board will need to be prepared to step up at times of challenge to the values and long-term success of the organisation.

Choices about the deployment of will and skill of members
A number of authors (for example, Finkelstein & Mooney, 2003;Pye & Pettigrew, 2005) argue that board effectiveness depends on the quality of the individuals who become directors, their discretionary efforts, informal as well as formal working styles and their ability to be more than the sum of their parts collectively to get the work done. This also drives the argument for collective development interventions and programmes to realise the potential of the board. McNulty et al. (2003) characterise the effective non-executive director as "independent but involved", "challenging but supportive" and "engaged but non-executive". These dyadic couplets also illustrate a positive creative tension between agency and stewardship theories of boards. Research carried out by Stiles and Taylor (2001) ranked the qualities brought by non executive directors in descending order as, first, objectivity ("outsideness"), second, advice, and, third, external /expert/ knowledge, This relates to agency, stewardship and resource dependency theories and also maps on to Garratt's (1997) notions in his two sets of board tasks regarding monitoring compliance and contributing to strategy.
In relation to the deployment of discretionary effort by the board to engage with the business of the organisation, the example of the Kids Company given at the start of this article is salutary. The list of the board of trustees indicates strong "outsideness" or objectivity, and it is possible that their purpose (as they saw it) was more to do with lending credibility to the organisation, not least to offer comfort and reassurance to would-be donors, rather than to act as objective challengers of strategy and performance. This is an example of the consequences of having a "trophy" board, with highprofile individuals, and connected with a resource dependency frame that views the board purpose primarily to maximise external political, reputational or financial leverage.

Choices about extent of engagement
The emerging evidence concerning the behaviours of effective boards leans towards a comparatively "hands on" board with able and diligent non-executive directors. The evidence supports the tentative triadic proposition of board dynamics which combines high levels of engagement within a board climate of high trust and high challenge. This connects to a composite theoretical model of boards which combines elements of agency, stewardship and resource dependence and thus represents a development of the dual line proposed by McNulty and colleagues (2003).
The realist lens would suggest that the engaged board, that is both highly trusting and strongly challenging, is likely to be linked to different outcomes depending on circumstances. Some of these dynamics may need to be modified in other conditions, for example, in a start-up phase, or where there is strong competition, an unstable environment or where managerial hegemony threatens the organisation in the longer term.

Board behaviours in health care
In relation to health care, Alexander, Lee, Wang, and Margolin (2009) report two seemingly contradictory findings: hospital boards over time (1989)(1990)(1991)(1992)(1993)(1994)(1995)(1996)(1997)(1998)(1999)(2000)(2001)(2002)(2003)(2004)(2005) are exerting a stronger scrutiny role and at the same time CEOs are more closely involved in board governance. This suggests an intensification of monitoring (agency theory) at the same time as an enhancement of managerialist and stewardship theory like behaviours and has resonance with the dynamics of the engaged board described above.
Endacott and colleagues (2013) report on observations of 24 public board meetings at 8 English NHS trusts and find that 1 mode in their study is generally dominant: chair-led, with an atmosphere of collegiality and variable contribution from non-executive directors. This indicates the dominance of a stewardship model, mixed in with managerial power, and the possibility of board challenge being displaced by a pressure to adhere to a collegial norm.
The range of board behaviours identified in this review, including in the health care setting, indicate their connection with alternative theories about the purpose of the board and the importance of context. The effective board member may therefore wish to be self-conscious in her or his choice of behaviours depending on the circumstances. What we are suggesting here is a board behaviour repertoire which includes reflecting on what kinds of questions to be asking in which circumstances, and when to adopt a robust monitoring stance or to employ an appreciative frame.

Conclusion
We believe that this study is the first use of a realist approach to understand effective health care board governance. It suggests the importance of recognising underlying assumptions when using different theories with respect to effective health care boards.
In brief, we find that it is possible to open the black box, but inside there is no magic formula for an effective board. We conclude that making the most appropriate choices about board composition, focus and behaviours is dependent upon a close reading of the organisation and wider environment context, and upon an agility, as well as an ability, to adapt to changing circumstances. Our research synthesis leads us to propose that there are links between certain board compositions, board actions and organisation outcomes dependent upon a range of contexts.
We find that the appropriate composition of the health care board will depend on the challenges facing the organisation, and organisational maturity. Smaller boards encourage greater risk taking and faster rates of decision-making and innovation. Larger boards, including a greater proportion of independent directors with relevant medical expertise, will have a stronger capacity for more robust monitoring, challenge of the CEO and demand for efficiencies. Beyond the need for a medical presence, there remain gaps in our understanding of the skills and backgrounds needed to be an effective health care board member.
In terms of focus, all health care boards should be mindful of the importance of rigour in developing strategy, managing talent and use of resources, all of which is likely to be connected to improved financial management. Staff satisfaction is found to increase when boards have a strong focus on improving internal relationships. The predominantly outward facing board is important for organisations early in their life cycle or those in the health care sector where there is rebuilding of reputation to be done after a failure of care. There is a range of effective behaviours for boards. Members of health care boards are invited to deploy a broad repertoire to fit different circumstances, always mindful of the need to ensure patient safety, to promote financial sustainability and to create long-term public value.
We additionally propose that a closely engaged board, which exhibits both a strong challenge of, and strong support for management, offers a steer for health care boards; and further empirical research is now required to test these propositions. A study is under way to investigate the actions taken by boards of acute hospitals in England following the publication of the Francis Inquiry report in 2013 mentioned at the start of this paper. This research, due to be published in 2018, aims to unearth what associations exist between hospital performance and the strategic context, board membership, board leadership dynamics and the focus of board activities.
Board members may take some comfort from the fact that the evidence is not there (and unlikely ever to be) that there is one right or wrong way to go about effectively discharging their governance role. This study has shown that we should be wary of prescriptive guidance which carries that underlying proposition. In the research which critiques the assumptions behind agency, stewardship and resource dependency theories, Nicholson and Kiel (2007) found that while each theory can explain a specific case, no single theory explains any general link with organisation performance. As has been argued throughout this article, the board-performance link is likely to be highly dependent on context-specific situations such as stage of organisational life cycle, sector regulation and competitive conditions. This review has a number of limitations. Although it does address the drawbacks of more traditional systematic review methods when dealing with complex social interventions with alternative underlying beliefs, the realist synthesis approach is still a method under development. In addition, the nature of the existing evidence is mostly quantitative analyses of existing large data-sets. Further empirical studies are needed to address the question of composition of health care boards, for example, the appropriate range of skills, backgrounds, expertise and perspectives of board members, the proportion of independent directors and the importance of diversity, which has not yet been tested in relation to organisation performance in health care.
There is, finally, a need for alternative research designs, using mixed methods including case studies, to understand further the black box of board practices, and also the conditions and behaviours which allow for board priorities around patient safety (reducing the risk of harm), improving the experience of care (the collaborative effort between patient and clinician) and enhancing the clinical effectiveness of care (the service improvement capability) to come to the fore.