Enacting corporate governance of healthcare safety and quality: a dramaturgy of hospital boards in England.

The governance of patient safety is a challenging concern for all health systems. Yet, while the role of executive boards receives increased scrutiny, the area remains theoretically and methodologically underdeveloped. Specifically, we lack a detailed understanding of the performative aspects at play: what board members say and do to discharge their accountabilities for patient safety. This article draws on qualitative data from overt non-participant observation of four NHS hospital Foundation Trust boards in England. Applying a dramaturgical framework to explore scripting, setting, staging and performance, we found important differences between case study sites in the performative dimensions of processing and interpretation of infection control data. We detail the practices associated with these differences--the legitimation of current performance, the querying of data classification, and the naming and shaming of executives--to consider their implications.

Introduction

Patient safety remains a high‐profile health policy issue, traceable internationally since landmark publications in the USA (Institute of Medicine 1999) and UK (Department of Health 2000) highlighted the scale of medical error and harm to patients. Errors were framed as conditioned, precipitated and exacerbated by systemic and latent organisational factors ‐ and thus amenable to prevention (Waring et al. 2010). Empirical research informed by organisational psychology located clinical failures within organisational contexts (‘clinical micro‐systems’), seeing errors as the result of embedded unsafe practices rather than individual failings (Nelson et al. 2008). The solution proposed for this framing of the problem was implementation of standardised processes and ‘designing out’ errors, both at the level of specific interventions and whole‐organisation safety systems. While this perspective informs much empirical patient safety research, Lamont and Waring (2015) offer a subtle reading of a tension evident within the literature: is patient safety a ‘thing’ that may be enhanced through technical solutions; or a more nebulous, contested phenomenon requiring attendance to the socio‐cultural context of proposed changes to practice (Rowley and Waring 2011)? From a socio‐cultural perspective, transfer of technical ‘solutions’ between industries risks subversion by existing professional hierarchies, as observed by Currie et al. (2009) in an evaluation of incident reporting techniques developed in the aviation industry and subsequently implemented within a hospital. Crucially, it is feared that embedded social and organisational practices which make technical systems work within their original contexts may be overlooked (Macrae 2014); and that this explains the limited impact of patient safety interventions (Sheldon et al. 2004). The implication is that greater attention is required to ambiguities at play, and greater insight into the lived experience of patient safety work requires combination of theory with empirical data.

The failure of the Executive board to ensure safe clinical practice was implicated in events at Mid Staffordshire NHS Trust in the UK (Francis 2013) where many patients received sub‐standard care in a context of chaotic management systems. The detailed response to these events is outlined in A promise to learn – a commitment to act (Department of Health 2013) which informed development of a National Patient Safety Alert System, publication of ‘never events’ data, and a Patient Safety Collaboratives programme to support improvements. As entities with statutory responsibilities for oversight, boards have ultimate responsibility for upholding the quality and safety of care delivered within their organisation, and are charged with a fundamental role in the governance of patient safety through defining and managing objectives, strategy, priorities, culture and systems of organisational control (NHS Leadership Academy 2013).

Empirical literature identifies governance practices as potentially important in safeguarding patient safety, including routine feedback and monitoring of statistical data, strategic involvement of clinicians in quality improvement, and attention to external governance systems (Jha and Epstein 2010, Jiang et al. 2009, 2011, Vaughn et al. 2006). However, as with the broader patient safety research literature above, considerable weaknesses in study design and theoretical orientation remain (Millar et al. 2013). While qualitative and case‐study research is emerging (Baker et al. 2010, Ramsay et al. 2010), significant gaps remain in our understanding of the processes of organisation associated with board governance of patient safety in hospital settings (Chambers and Cornforth 2012). Specifically, we lack detailed understanding of what board members do, and the manner in which they do so, as they seek to discharge obligations with regard to patient safety (Millar et al. 2013, Nicolini et al. 2011, Waring 2007). The concept performativity may prove helpful in exploring board practices (Freeman and Peck 2007, 2010), the purpose of this article being to explore its application empirically.

Below, we introduce performativity and consider its influence within the study of organisational life; trace its foundations to the work of Austin (1962) and Goffman (1974); and note its empirical application in the context of participatory governance (Hajer 2004). We then employ Hajer's framework of scripting, setting, staging and performance to hospital executive board meetings, and explore implications for patient safety governance.

The foundations of performativity: Austin and Goffman

Austin (1962) coined the neologism ‘performativity’ to describe instances in which the utterance of a phrase constitutes an action which changes reality rather than describes it; a simultaneous ‘saying’ and ‘doing’ which requires others to act in accordance with its implications. Austin's paradigmatic case is the phrase ‘I do’ when spoken within the context of a marriage ceremony; a phrase requiring those exchanging vows to act, and be acted upon by others, as a married couple from that point forward. Austin additionally stipulates that performative utterances are meaningful actions that are neither true nor false but generative, that is, they create a social reality.

While Austin defines performativity, Goffman's (1974) dramaturgy of social interactions applies a theatrical metaphor to indirectly explore its operation, principally through framing. For Goffman, the framing of the stage – the separation of front of stage, backstage and audience, and reciprocal acceptance of the roles of audience and players – shapes the performativity of utterances onstage. In an extension of earlier work on the presentation of self (Goffman 1959), Goffman notes that performativities are made possible through framing. Frames are essentially classification systems which actors use to order and make sense of diverse social phenomena, so that the performative potential of an utterance depends upon the frames available.

The rise of ‘performativity’

From its inception within linguistic philosophy (Austin 1962), the reach of performativity has ballooned. Early areas of influence include theoretical development in the emergence of order in complex interactive systems (Bateson 1972); the framing, staging and (re)creation of social life (Goffman 1974); and the ‘language games’ structuring performative utterances (Lyotard 1979). Later applications include the constitution of gendered identity through interactions (Butler 1993, 2010); the continuous (re)construction of society (Latour 1986, 1987); and the effects of economic theory upon action (Callon 1998). Organisationally, performativity has informed analysis of continuous change present in the enactment of organisational routines (Feldman and Pentland 2003); the enactment of technology within social settings (Law and Singleton 2000, Orlikowski and Scott 2008); the role of storytelling in coordinating within and between organisations (Diedrich et al. 2011); and organisational change as active translation rather than passive diffusion (Czarniawska and Sevon 1996).

The performativity of governance: Hajer

Informed by Goffman's earlier work, Hajer (2005, Hajer and Versteeg 2005) provides a dramaturgical framework for analysing the performative dimensions of public governance through consideration of the setting(s) in which deliberation takes place and the enactment of organisational frames, operationalised through consideration of the scripting, setting, staging and performance of governance. Scripting refers to the determination of actors involved in the decision‐making forum. Consistent with the generative potential of performativity it considers how participatory practices construct participants as either active or passive; collaborators or protesters; competent or incompetent. In contrast, setting concerns the physical environment of interaction, including artefacts (e.g. minutes of previous meetings) that participants bring to the physical environment and which shape performance. Deliberate attempts to organise interaction between participants is identified as staging. It is achieved by drawing on existing symbols, the invention of new ones, and conventions governing distinctions between active players and passive audiences; what might be termed ‘unwritten rules of engagement’. The final category, performance, concerns interactions which (re)construct knowledge / power relationships that shape future interactions, providing opportunities for change over time.

If Goffman is correct that reality is mediated through the application of frames to make sense of available information then all accounts of reality are shaped. The implication, in the context of patient safety governance, is that the examination of safety could be framed in many ways with radically different consequences for action. For example, in presenting data that indicates a breach of an infection control performance target, is this: an instance of unreliable data requiring a defence of organisational practice? Or a worrying event requiring detailed diagnostic work to uncover systemic problems and institute quality improvement activity? Or an impossibly harsh target imposed by regulators requiring robust challenge? We explore empirically the operation of such performativities in relation to the board governance of patient safety, applying Hajer's analytic framework to qualitative data collected through overt non‐participant observation at four case‐study sites within the English NHS.

Methods

Design

The overall study design is a comparative case study across multiple (n = 4) study sites (Stake 1995). Each case had at its centre an acute Foundation Trust (FT) hospital – an independent public benefit organisation based on mutual traditions subject to independent external regulation (NHS Leadership Academy 2013). Membership of FT management boards comprises a chair, chief executive, executive directors and independent (non‐executive) directors (NEDs) who are, with the chair, in the majority. All board members bear responsibility, individually and collectively, for performance and the quality of services (NHS Leadership Academy 2013).

Sample selection

Cases reflect the diversity of English NHS hospital FT trusts: a district general; a teaching hospital with a global reputation for specialist services and innovation; a regional centre offering specialist services; and a trust undertaking large‐scale service redesign. Further specific details, including cultural characteristics of board meetings, is provided in Table 1.

Table 1

Summary of each case study site

Arran	‘World‐class provider’, Global aspirations built on research & development Strategic focus; external horizon‐scanning to secure compliance with policy directives and safeguard the self‐image of the trust as a pre‐eminent provider. Strong ‘shaping’ steer by the CEO: low challenge by non‐execs and a strong medical executive team Needing to address emerging performance issues while minimising damage to its ‘world class’ self‐image
Skye	‘Local service under pressure’ District general hospital trust (3 sites) Rotating board membership for over 3 years. A myriad of problems Focus on internal problem solving, limited wider strategy High internal challenge by non‐execs, a strong chair, technocratic CEO finding his feet
Lewis	‘Embattled regional powerhouse’ A large teaching hospital Seen as the main regional provider Dominant CEO acting as a political antagonist; defending local interests from competing regional and national forces Estate, finance, and legal disputes
Islay	‘Faith in quality improvement methodologies’ A district general hospital trust An ‘intelligent’ board 1 – Reasoned and assured questioning by non‐execs; 2 – CQI culture; 3 – Emphasis on patient experience Strategic focus 1 – reconfiguration and integrated care strategies 2 – Clinical oversight 3 – divisions ‘invited’ to provide updates

Strategic focus

Cases were additionally selected on the basis of their performance trajectory over the last three years on a range of safety and quality indicators selected from the Dr. Foster database for 2011. Every year Dr. Foster publishes a Hospital Guide that uses publicly available statistics to measure and assess what is happening in hospitals in England to increase transparency in relation to variations in performance. The data focuses on indicators in three domains of quality obtained from the available statistics, combined with information from self‐reporting of safety aspects from an annual questionnaire. We used these data to select two case study sites indicated as getting worse (in the light of overall improvement of other hospitals) and two that were improving. Trust sites have been anonymised by being renamed after Scottish islands; the two improving trusts are Islay and Arran, the two getting worse Lewis and Skye.

Data collection and analysis

Acting in pairs, the authors undertook overt non‐participant observation of four sequential management board meetings at each case study site, totalling approximately 50 hours of observation. Of the major topic areas discussed, issues relating to service quality, patient safety, performance measurement, and risk formed a substantial part of board meeting time at each site (Table 2). The authors are all experienced social policy researchers with established interests in the governance of healthcare quality, and data collection was designed to facilitate an analysis of Hajer's analytic framework. Descriptive free‐text field notes were taken by observers at each meeting, supplemented with documentary data including the agenda, supporting papers and (retrospectively upon completion) minutes of each meeting. Data from field notes were compiled across multiple board meetings within each trust to detail board operations and identify the manner in which board governance of patient safety was enacted. We used the analytic dimensions of scripting, setting, staging and performance identified by Hajer (2005) as a deductive a priori template (code book) against which qualitative material were organised. Thematic analysis revealed patterns (similarities and differences) within the data, showing the divergent dramaturgies at play within each specific site. While these enactments were found to be stable across multiple meetings within sites, they proved substantially different between sites.

Table 2

Summary of board time by topic area across all four observed meetings

Topic areas	Arran		Skye		Lewis		Islay
	Mins	(%)	Mins	(%)	Mins	(%)	Mins	(%)
Service quality, patient safety, performance measurement, risk	252	(60)	379	(42)	236	(28)	448	(54)
Strategy and capacity	30	(7)	185	(21)	158	(19)	315	(38)
HR	56	(13)	79	(9)	13	(2)	31	(4)
Finance	83	(20)	55	(6)	105	(12)	28	(3)
Estates	–	–	–	–	198	(23)	8	(1)
Other	–	–	194	(22)	130	(17)	–	–
Total	421	(100)	892	(100)	845	(100)	830	(100)

The table summarises the amount of time spent in minutes (percentage of total time) on specific, frequently occurring, topic areas within each case study site, totalled across all of the four board meetings at each site. All board meetings devoted considerable time to issues of service quality, patient safety, performance measurement and risk, and this composite measure took the longest amount of time at each of the sites.

Findings

We first outline the operation of the board at each site to retain the integrity of each organisational setting, then offer a comparative analysis of Hajer's dramaturgical categories ‘scripting’, ‘setting’ and ‘staging’ in sequence across sites, indicating the scope and scale of differences between sites (table 3). We then explore the final dramaturgical category, ‘performance’, in relation to board deliberation of indicators associated with infection control, a key component of the external performance management regime within the English NHS.

Table 3

Summary of Hajer's dramaturgical categories by case study site

	Arran	Skye	Lewis	Islay
Scripting	CEO shaped events through ‘CEO Report’ agenda item; Low levels of Non‐exec challenge; Opportunities to challenge ‘managed’	chair dominant; Non‐execs forthright challenges; adversarial (agency) model with execs lauded or shamed	CEO a dominant personality; Much deferral to his experience, non‐Execs able to question; ‘embattled’ narrative of CEO the main coda	Non‐execs robust yet respectful challenge; endemic, framed and legitimated as ‘improving patient experience’
Setting I: physical	Large, airy meeting room in Trust HQ Education centre; Non‐adversarial ‘Horse‐shoe’ arrangement of tables at the front of the room, space for many observers as required; high quality projection facilities routinely used	Rotating venue, typically cramped, dated and poorly equipped; no consistent seating arrangement; ‘audience’ very close to board tables	Board room at main site – imposingly furnished (dark wood, large tables); a place ‘where important decisions are made’. Separate table for those invited to present to the board – adversarial; calls to mind the layout of an ‘inquiry’	Rotating venue, always in well‐furnished, low‐key ‘office’ environment. Spacious, with room for a wide range of attendees to observe as required
Setting II: artefacts	Presentations modelled on medical lectures; routinely used to summarise main points of reports in board papers and frame discussion. Successfully limited the scope for alternative challenges	Detailed information presented using software tools; typically ‘dry’ delivery; used by non‐execs to facilitate challenge	Not used other than in ‘special’ presentations by external speakers – and then not supported by software or projection. Relied on orations from presenters, and challenges made with reference to supporting information in board papers	Presentations routinely used in clinical updates – supported by presentation software and projection facilities
Staging	Highly structured and formal; CEO's report placed early in agenda affording scope to craft an over‐arching narrative which reinforced the self‐image of the organisation as a high performer even where there were difficulties (e.g. Infection control); Performance data summarised to 1 page of A4 and used to identify ‘red’ areas for development work – assurance mode	Highly structured and formal; chair steered focus on breach of targets in summary indicators to facilitate robust challenge	CEO dominance of ‘matters arising’, placed early in the agenda, set the tone. Most of the time spent on these items. The ‘CEO show’	Highly structured, yet ‘permission to speak’. Opens with a narrative patient story ‘to concentrate our minds’ – patient experience a guiding principle, invoked routinely. Performance data reported within a ‘Quality improvement strategy’ section – improvement mode
Performing: infection control	Despite existence of summary A4 ‘traffic light’ indicators showing breach on infection control, CEO used their report to frame perceptions (successful reclassification with commissioners and Monitor) to remove the breach. There was no non‐exec challenge.	chair adopts tone of ‘disappointment’ at data presented on C. diff target breach to demand action by the Nurse Director. No attempt to ‘explain away’ or contextualise the breach	C. diff targets presented as unobtainable by CEO; Legal challenge threatened to commissioners as part of ‘embattled’ narrative	Infection control target (reported in QI strategy section) is met but CEO warns against complacency. Data is disaggregated to identify potential ‘hot spots’ to further drive performance

Site 1: Arran – a world‐class provider

Arran has a strong research culture and global reputation for service innovation. The board acts strategically and is externally oriented, seeking to secure compliance with central policy directives while further developing its global reach and status as a pre‐eminent provider and academic research centre. In operation, board meetings are highly formal and structured to enable the CEO to ‘steer’ interpretations of events. The executive team are experienced with low turnover and a strong medical presence. Non‐executives exhibit low levels of challenge even when performance difficulties are under discussion; indeed, the dynamic of board meetings could be summarised as maintenance of the narrative of ‘world class’ status while engaging with potential performance difficulties evident in summary indicator data.

Site 2: Skye – a local service under pressure

In marked contrast, Skye exhibits high levels of executive turnover and overt recognition of multiple problems in service provision, focusing on internal problems rather than strategy. board meetings are steered strongly by the chair who dominated a newly appointed CEO; strong challenges from Non‐executives are common including expressions of disappointment at poor service performance and support for executives considered high performing. During the period of observation one executive stepped down, replaced by a deputy who received similarly high levels of challenge. Board dynamics could be summarised as routine non‐executive challenge with limited long‐term strategic direction.

Site 3: Lewis – an embattled regional powerhouse

Protective of its reputation as a regional centre, board meetings are structured to ensure that issues considered to be of strategic importance receive due consideration, events considered as they unfold over successive meetings. The image revealed is that of a guardian of local and wider regional interests, requiring continued political influence. The executive team is experienced and longstanding and the CEO personifies the organisation, dominating board meetings and acting as a political antagonist to defend local interests. During observation considerable time was spent on a small number of financial and estate issues and related legal disputes considered strategically important. Non‐executives were typically involved in clarifications related to their specific areas of interest rather than challenges over strategy or performance. Overall, the board dynamic could be characterised as one of defending ‘regional champion’ status.

Site 4: Islay – faith in quality improvement methodologies

The overwhelming impression of this site is one in which board structures and processes focus attention on reconfiguration and integrated care strategies. Central to this approach is the consideration of performance issues related to external regulatory requirements (infection control; pressure sores) directly linked to improvement work to support strategic plans. Patient experience is routinely invoked and normalised by the presentation of a patient experience narrative at the start of each board meeting. Non‐executives are encouraged to question and their contributions are actively sought by the executive team. This board could be characterised as drawing on a wide range of soft and hard intelligence, seeking to maintain strategic focus while discharging external regulatory accountabilities.

A dramaturgy of board governance of patient safety

Scripting

While often commenting on information presented to the board, at Arran NEDs were limited in their challenges to executives and avoided overt conflict. In marked contrast the chair and non‐executive directors at Skye confidently challenged executives, particularly in relation to poor performance in infection control and capacity issues, demanding that the CEO improve standards. There was clear differentiation in the way that different executives were seen and treated, being either lauded or shamed. The medical director (MD) was widely seen as competent and dependable, and his pathway development work was formally acknowledged:

(Field note extract)

Can we please congratulate [Chief Operating Officer name] and [MD name] they're doing a great job.

Yes, the Medical Director is bringing it down to local levels which is great to see…

In contrast, the Nurse Director received tougher questioning and her hesitancy was drawn to attention (considered in detail below in the section on ‘performance’).

At Lewis, board meetings could be summarised as the ‘CEO show’; the dominant narrative was of an embattled trust, fighting for local interests – a trust under fire enmeshed in structural, legal and financial disputes. This strongly adversarial framing is typical and exemplified in the CEOs threat to resign in the face of fierce opposition from external opponents:

(Field note extract)

It's insulting quite frankly. We're dealing with animals… I'm willing to resign if this drags on and isn't resolved.

The CEO's views were rarely challenged internally, with board members supporting his position and accepting his performance as part of the ritual and crafted with humour, emotion, and political point‐scoring. The chair had a supportive presence, summarising issues as appropriate.

At Islay the board maintained a clear focus on strategy and service improvement. The CEO projected a calm influence; a moderniser with a similarly oriented executive team who had received quality improvement training which clearly informed board practice. Robust yet respectful challenge was normalised among non‐executives, framed and legitimated as ‘improving patient experience’. Board dynamics were driven by this central narrative, which granted permission for depersonalised challenge, exemplified in the following field note extract concerning theatre safety, in which improvement activity is explored dispassionately in relation to standardisation and cultural change, rather than personal challenge:

The human factors training is important but has the culture changed? I need to push you that we want to take it forward.

We do adhere to the protocols but we constantly want to be individuals as well…

(Field note extract)

This is the most challenging area, it should be processed, your personal leadership is needed… but there needs to be some kind of standardisation.

[name] has been amending the checklist to accommodate different interpretations…. If you go round now, as a result of these changes, you'll see the difference… The problem was the theatre culture; we need a change of attitude related to infection control. The drive for quality and safety's the most important

Setting I: physical

Arran board meetings were held within the education centre at Trust HQ in a large room with high‐quality presentation facilities. The location reinforced a sense of ‘information transmission’ (passive) rather than deliberation (active). The room was large, bright and airy, the audience typically consisting of five governors and / or members of the public, ordered into rows of (sparsely populated) chairs facing presentation screens and set back from board members, themselves seated in a ‘horseshoe’ arrangement of three tables at the front of the room; a non‐adversarial setting consistent with a ‘common purpose’.

Skye meetings were held at different locations across the trust. With one exception, meeting rooms were cramped, hot, and lacking air circulation. Because of the poor conditions the audience (3–5 governors) tended to split into two groups at opposite sides of the room, very close to the board table. The chair, chief Executive, trust secretary and trust minute‐taker sat together, with other seats haphazardly assigned with no consistency across meetings.

Board meetings at Lewis were held at the board room at the main trust site whose décor consisted of dark wood tables and chairs with red leather covers; an imposing environment reflecting civic pride. The seating arrangement was stable across meetings. The chair, chief executive and observers sat together, the business development director and non‐executives together on an adjacent table and the nurse and finance director with some spaces for guest speakers and observers. The trust secretary, medical director, and more non‐executives sat at an adjacent table. Those invited to present to the board faced an experience akin to an inquiry; waiting in an anteroom for the appointed time on the agenda, invited in when required, positioned on a table in front of the board to face questions, and escorted out once their item concluded.

While the venue for Islay board meetings rotated, all were held in well‐furnished, low‐key ‘office’ environments. Rooms were spacious and accommodated observers. Seating was arranged around a large single table on an ad hoc basis. Attendees included members of the quality improvement team, and clinical leads who attended to report updates regarding specific clinical reviews.

Setting II: artefacts

Board papers

All sites provided a briefing pack of information ahead of the meeting, consisting of a formal, standardised agenda and appended reports for agenda items in sequence. The format and contents were similar across sites. Two frames of reference underpinned board papers: RAG (red‐amber‐green) ratings and ‘narrative’ executive summaries. Board papers presented information in the form of an executive summary of the ‘key issues’ over the past month and the options for moving forward, and additional quantitative material included line graphs, run charts and bar charts related to performance.

Presentations

At Arran presentations were undertaken by medical staff in response to specific issues highlighted within prior board meetings. The format broadly adhered to conventions associated with academic medical lectures, presenters using available electronic facilities to display detailed presentations, summarising the main points of associated reports appended within the compendium of board papers, and responding to questions from the floor. While affording Non‐executive board members the opportunity to raise detailed questions, presenters framed material to anticipate questions and structure the following discussion.

While similar conventions operated at Skye presentations were more informal in tone, lacked vibrancy and afforded Non‐executive board members the opportunity to raise questions in relation to detail.

Staging

Formal in operation, the ordering of the agenda at Arran facilitated the crafting of an over‐arching narrative by the CEO presenting the organisation as a high performer – even where more detailed reporting suggested potential difficulties.

The detailed ‘performance report’ delivered by the director of strategy provided a high‐level summary and narrative overview of key performance indicator (KPI) data related to activity, efficiency, access, cancer, infection, quality and safety, workforce, finance, and the monitor compliance framework. Performance data were presented for the current month and year to date summarised on a single A4 page, and disaggregated by sub‐division (medicine, surgery and cancer, specialist) with additional data in a supporting document. While the structure (KPIs) and form (‘traffic‐light’ RAG indication of performance against targets) of reporting used to inform the board potentially supported the adversarial discharge of public governance in the form of assurance (‘conformance’), Non‐executives did not use the data to publicly challenge executives at any of the observed board meetings.

The CEO avoided such overt challenges by contextualising shortfalls detailed in the performance report in his preceding ‘CEO report’, offering an interpretation of figures and outlining on‐going work to diagnose and / or address shortfall. We consider this in further detail in the ‘performance’ section below. The CEO's report was wide‐ranging, provided opportunities for clarification and discussion, and consequently took considerable time. The CEO used this composite section, delivered very early in the board meeting, to consider multiple items related to trust activity / performance and their implications, in the light of external policy directives and / or the strategic opportunities they provided.

In contrast to the close executive control above, at Skye the staging was typified by the chair's dominance. His steering of the agenda, and contributions from others, focused principally on the breach of targets (particularly C. difficile (C. diff)) and issues related to coding and validity of statistics, with very limited external horizon scanning or evidence of wider strategy. The chair was active throughout, particularly so in relation to quality and patient safety, evidenced in steering challenges to the nurse director about infection control and the failure to achieve the C. diff target. There was a sense (picked up by an audience member in an aside to researchers) that he dominated the meeting. By controlling the staging and encouraging NED interjections the chair prevented other narratives from unfolding.

Staging at Lewis was shaped by an extensive ‘matters arising’ section placed immediately after the opening item of ‘apologies’. This was by far the longest section and invariably consisted of the following issues considered most important including a major capital investment project; qualified provider procurements; an academic health science network; and commissioner penalties in relation to C. difficile.

In contrast, board meetings at Islay were tightly structured around the agenda. The tone of each meeting was set by an opening ‘patient story’ narrative, introduced by the chair as ‘something to focus our minds’. These narratives, collected by the improvement team, concerned issues such as the fears, anxieties and negative experiences following diagnosis, processes of care and treatment outcomes. They were emotive and used by the chair to legitimate challenge and concern with service improvement; the language of ‘patient experience’ was available to, and used by, board members to depersonalise challenging questions. Strategic focus on service improvement was supported by the agenda structure, ‘service improvement’ placed as a standing agenda item immediately after the ‘patient story’, and led by members of the quality improvement team presenting data and updating progress on strategically important projects including intensive care, community services, falls and infections. The incorporation of indicators for infections into this section is important, as it ties external regulatory requirements to internal strategy in an agenda item specifically oriented to service improvement.

Performance(s): board governance of infection control

Board meetings at Arran were dominated by the CEO and typified by a need to present trust performance in a favourable light, consistent with a ‘world‐class’ reputation. This dominance was not exercised through explicit force of personality, but principally through careful structuring of the agenda and the use of framing in the ‘CEO Report’ agenda item to shape perceptions of, and thus actions in relation to, items appearing later in the agenda.

As outlined above, the CEO's report drew attention to shortfalls in performance. Notable examples in relation to patient safety included failure to achieve reduction in C. diff targets for three quarters in succession, and a potential breach of A&E waiting time target in the fourth quarter (and thus for the whole year). While both of these topics could have raised vociferous challenges from non‐executives, no such challenges were made. Rather, in relation to the former, the CEO announced that commissioners had waived penalties for eight cases, and on this basis he had conferred with the external regulatory body for reclassification as ‘green’ (healthy) under their governance performance indicator, previously flagged as amber and placing the trust under external scrutiny. This was presented as evidence of over‐sensitivity to borderline non‐compliance:

(Field note extract)

Monitor reports C. diff in A&E is now [six over target] but there are eight cases that we have appealed. These shouldn't have been in the figures in the first place; it should have been identified before coming in (…) we have missed (the target) for three quarters (…) So the problems with A&E combined with C. diff means our governance rating's likely to go down the fourth quarter. It means the organisation hasn't deteriorated but it can give a negative impression to the outside world.

In the case of the A&E issue, the seriousness of the potential breach was minimised by reference to the ubiquity of such problems (‘everyone faces such difficulties’), addressed with reference to current plans to expand A&E provision (‘we are already dealing with this’), and attention focused on the role of limited out‐of‐hours cover by GPs in increasing demand on A&E services (‘the problem lies elsewhere’).

In both instances, the CEO's report identified poor performance against indicators that could have been addressed as ‘breaches’, pre‐interpreting the data as misleading. That these analyses were convincing is clear from the fact that neither of these issues were picked up by non‐executives for further exploration during the subsequent detailed performance report. Given the important work involved in framing these issues the CEO report was typically one of the longer, and on one occasion the longest, agenda item(s) at 46 minutes out of a total of 110 (42% of the total time).

In contrast, board performances at Skye were typified by high levels of NED activity. In relation to breach of a C. diff target, the chair adopted a tone of disappointment at failure to reach the target – presented as a serious issue which needed to be tackled, requiring explanation by the director of nursing. There was no serious attempt to reframe the target breach; the ND is held responsible:

(Field note extract)

We've not got continuous quality improvement for C. diff. This has increased from last year which in my opinion is unacceptable… So, is the C‐Diff policy in place [name]?

The policy and plan is good, but [she hesitates]

I feel there was a struggle to answer, I'm disappointed with the reaction from the new Nursing Director.

Detailed discussion followed on the implications for the trust:

(Field note extract)

What about costs?

Seven over [and] it's a £3.5 million fine, so we need to talk to Commissioners, plus the cost of additional cleaning.

How likely will the surplus be wiped out?

CCG won't want us bankrupt but there are higher powers than the CCG.

High levels of challenge were also evident in relation to other performance issues, including a ‘red’ rating against A&E performance and a lower than anticipated level of emergency admissions.

At Lewis, CEO dominance and use of humour, emotion and politics is exemplified in discussions of a breach of infection control targets. Discussion focused on the financial penalties associated with breach of the target, the trust facing a fine of over £1 million per excess case. Disputes with commissioners ensued and legal proceedings developed, board discussions centred on ‘proportionality’ and perceived unfair treatment in comparison with regional neighbours:

(Field note extract)

There's no drive for improvement. [local rival] has had the same target as last year. Why have we been given such a challenging target? [Local rival] isn't in breach of contract even if they miss target! I mean we could use the money and reinvest in community if we are in surplus…. I need a steer if we're not budging on £1.3 million then court mediation is the answer … It's ridiculous, it's a disincentive to improve, make improvements. We're also going to discuss this with the FT network. The penalties are completely disproportionate… .

Yeah, they've got the process wrong here.

Yes, I know, I've heard [local rival #2] is at <70 more than trust>. We accept but it's not proportionate.

The CEO suggested they had no choice but to take commissioners to court as an act of safeguarding local services:

(Field note extract)

Someone has to stand up for the NHS and it's going to be . We are going to… .

London will be imposing further penalties; I'm pretty sure of that based on my intelligence working with London,

This isn't about London, this is about the NHS. It's about people delivering the services, people receiving services…

While board consideration of infection control at Islay was also framed in relation to external targets, its service improvement orientation was very evident. The MD presented data showing that the trust was meeting its target. This was warmly welcomed by a non‐executive who praised the effort, and then quickly considers the interventions required to ensure continuous compliance, prompting a discussion of the educational interventions being made to embed infection control in staff behaviour:

(Field note extract)

We are under trajectory, we are making good progress. MRSA is an area of anxiety. We had an outbreak on , fogging, deep cleaning all as you would expect. A root cause analysis of all cases couldn't find anything. No cause found. It was just patient specific. For E‐Coli … we are trying to distinguish the ones that we can prevent and the ones that are unpreventable.

These are fantastic results to be under the targets but work to be done to assure and provide more details. We need to work within the Divisions with the high turnover of junior doctors in order to educate them about infections; otherwise it's never going to end.

It's getting permanent staff to communicate with junior doctors and it's also about getting Advanced Nurse Practitioners involved.

These very different emphases and approaches across sites are summarised in Table 2.

Discussion

Obtaining, processing and interpreting performance information are acknowledged as important aspects of boards’ oversight role (Department of Health 2013). While there is limited empirical support internationally that those hospital boards which prioritise the collection and analysis of performance data tend to have improved quality outcomes (Jiang et al. 2009), our analysis draws attention to the mediating role played by processing and interpreting performance data and to the performativities at play during deliberation. Consistent with a socio‐cultural perspective (Lamont and Waring 2015), performative analysis enables empirically grounded exploration of the refraction of external requirements for patient safety governance through local practices. Specifically, we highlighted the practices involved in realising the potential of patient safety performance indicators (Freeman 2002), and more importantly how the ‘warning signals’ provided by such data may be muted, deflected or silenced. We consider further the implications of our analysis for recent international empirical research findings, specifically time spent on patient safety; availability of patient safety indicators; and the training required by board members. While cautious of the reification involved in lauding specific interventions as ‘solutions’ to patient safety problems, we maintain that close analysis of the understandings and conventions in board interactions may be helpful in informing action.

A very high proportion of English NHS hospital Trust boards carry out processes that international research indicates may be associated with higher performance (Jha and Epstein 2013, Mannion et al. 2015). All report quality sub‐committees (Jha and Epstein 2010) and almost all have explicit objectives related to improving patient safety (Jiang et al. 2008, 2009). All of our case study sites sought to provide strategic assurance by establishing organisational structures and processes for reporting safety‐related information throughout the organisation and to the board (Botje et al. 2014, Jiang et al. 2009); making patient safety a strategic priority (Jiang et al. 2008); developing and nurturing an ‘open and fair culture’ (Vaughn et al. 2006); and using high level information to ensure compliance with safe practices / standards and external targets (Jha and Epstein 2010). Yet, the degree to which aspirations were fully met was moot, and governance activities remain contingent on board dynamics. Our case study sites exhibited governance behaviours variously related to: agency theory (Chambers and Cornforth 2010), in seeking to measure performance to ensure compliance and hold staff accountable for their actions; stewardship theory (Cornforth and Edwards 1998), in attempting to implement a framework of shared values built on trust; stakeholder theory (Chambers and Cornforth 2010), in managing complex trade‐offs between stakeholders, including staff, patients and the public; and resource dependency theory (Zahra and Pearce 1989) in managing internal and external relationships to leverage influence.

Our analysis highlights the role of, and differences in, local processes of organising in relation to board governance of patient safety. Thus while the amount of board time devoted to discussing patient safety has been identified as potentially important (Jha and Epstein 2010), we draw attention to the fact that boards used this time differently. Similarly, while the availability of summary data has been indicated as important (Jiang et al. 2008, 2009), and similar levels of performance indicator data relating to infection control were available at each of our case study sites, differences in use were significant and related to the practices legitimated within each setting. At Skye, Arran and Lewis attention focused on data indicating shortfalls – a quality assurance oriented approach. Operationalising the governance of patient safety largely in terms of assurance through retrospective use of performance data to alert the board of poor performance encourages under‐reporting and does not indicate how to address deficiencies. Specific responses noted at our sites included challenge and blame by NEDs at Skye; interpretive work by the CEO to forestall challenge at Arran; and framing targets as unrealistic and requiring challenge of regulators at Lewis.

Our results also have implications for the nature of training required by boards. Findings highlighted the challenges board members face in terms of scripting and staging, especially when decisions in these areas often pass unchallenged, unremarked or even unnoticed. A better understanding of these issues may feed into revised training and induction processes for board members. In the USA, a number of states have introduced initiatives in the form of mandatory induction and orientation programmes for board members (Jha and Epstein 2012) and this may also be an area that requires greater attention in the UK. Crucially, while the availability of information may be a necessary precursor for improvement, it is not of itself sufficient. While development of technical ‘competencies’ in analysis of performance indicators may be helpful, the important role played in setting, scripting and staging in the use and interpretation of information and performance data suggests that a broader curriculum is required directly attending to these issues. While summary reporting of quality indicators is important (Jha and Epstein 2010), local processes of organising that make it possible for non‐executive board members to use such information to hold executives to account sensitively are required – most evident at Islay and rather less so elsewhere. Thus the dramaturgical arrangements of boards require attention if they are to become fora for effective debate, challenge and the instigation of action.

Only at Islay did the board seek to provide a strategic focus on quality improvement (to the extent of structuring the board agenda around strategic quality improvement initiatives). This approach seemingly offers a way of managing tensions between assurance and improvement by combining the requirements of external assurance with ongoing quality improvement through development of national and local level datasets, and required additional qualitative information from reports by teams from clinical service areas.

It is important to note that we do not claim privileged status for our own analysis, or suggest that our perspective somehow reflects an underlying ‘truth’ to which we had special access. We do not claim that the performance indicator data presented at these board meetings showed deficiencies in patient safety that were inappropriately interpreted away. Indeed, the very value of a performative approach is that it traces the way in which words both ‘say’ and ‘do’; to consider meaningful actions that are neither true nor false but create a social reality acted upon by others. Our analysis shows the use of data to create a social reality in relation to infection control rates, which then guided future actions. We would assert however that a set of behaviours / practices that use information to support cycles of improvement activity are required.

Conclusions

Studies of statistical associations between organisational structures and patient safety outcomes are not able to explore the dynamic and messy lived reality of board governance practices related to patient safety (Millar et al. 2013). Our analysis of such practices at four Hospital boards indicates the importance of local processes of organising in relation to governance of patient safety. While the availability of summaries of quality indicators to board members is undoubtedly important, so equally are the operation of processes of organising that make it possible for local actors to use such information to make interventions to sensitively hold executives to account with regard to patient safety processes and outcomes. In this regard, our findings indicate the challenges faced by board members in terms of the artefacts at their disposal and the limitations of the scripts and staging associated with board practices. Encouragingly, in drawing attention to practices associated with the enactment of safety and quality they also indicate possibilities and opportunities for enhanced deliberation of information with which to improve the corporate governance of safety and quality, and we encourage additional applications of performativity‐based approaches.