Describing failures of healthcare: a study in the sociology of knowledge

Abstract

In 2008, five ‘serious untoward incidents’ occurred on a small maternity unit in a hospital in the UK. The prevailing view, held by clinical staff, hospital managers, and executives, was that these events were unconnected and did not signal systemic failures in care. This view was maintained by the testimony of staff and governance procedures which prevented the incidents from being considered together. Drawing on the inquiry report of the Morecambe Bay Investigation (2015), I examine how the prevailing view was built and dismantled, eventually being replaced with a very different description of events. Overturning this view required affected parents to engage with governing bodies and legal processes, challenge clinical staff, lobby for inquests, and mobilise social media and the national press. Tracing how different descriptions of events weaken or gather force as they travel through different forums, processes, and are presented to different audiences, I explore the sociology of knowledge around establishing failures of care.

In the context of concerns around healthcare safety, questions of knowledge – what is known, when, by whom, and how they know it – are critical. In this article, I explore these questions from a sociology of knowledge perspective, which relates the content of knowledge to the context in which it is developed and employed, analysing the report of an inquiry into healthcare failures at a maternity unit in the UK. In my reading, the inquiry report details a struggle for understanding around how practice should be viewed; from the outside, the incidents that occurred represented systemic failures in safety, but from the inside (for a considerable period of time), they were considered an unfortunate, but coincidental, series of events. The positions from where these perspectives emerge are important in shaping the content of knowledge and can explain much about how understanding develops.

Such ‘situated knowledge’ is the subject of much scholarship across many fields of qualitative social science. I draw insights from science and technology studies, anthropology, sociology of health and illness, sociological studies of disasters, and from ethnographic methodological literature to develop a discussion around the located character of knowledge, perception, and interpretation that explains how different perspectives on events can develop and to illustrate their safety implications. I then focus on events that formed the subject of the Morecambe Bay Investigation, an independent inquiry chaired by Dr Bill Kirkup into failures of care in the maternity unit of Furness General Hospital. My focus is not only to explain the co-existence of contrasting views on the sufficiency of care but also to explore how one version comes to predominate and what it takes to overturn that view. I draw on principles of narrative analysis to analyse the inquiry report, tracing how descriptions of events weaken or gather force as they travel through different forums, processes, and are presented to various audiences, with shifts in context changing the way events are viewed. Thus, I examine the sociology of knowledge in establishing failures in care; the claims that are made – their positionality, weighting, authority, mobility, and audience – as different descriptions of events circulate.

The located character of knowledge and its relevance for risk and safety in healthcare

As Goodwin (1994: 606) states ‘All vision is perspectival’. We do not all see the same thing. What is seen, understood, how it is described, and from what or whose perspective are issues central to ethnography and there has long been methodological debate around the virtues of insider and outsider perspectives. The outsider has the advantage of critical distance, taking nothing for granted, resisting assumptions (Hirshauer, 2006; Naaeke et al., 2011). The insider’s view is rich in local knowledge and interpretation. As May and Pattilo-McCoy (2000) explain, their different degrees of familiarity with the setting led to perceptual inconsistencies reflecting their personal and intellectual backgrounds. For example, the native Chicagoan’s knowledge of local gangs allowed for meticulously detailed descriptions, rich in meaning, whereas the non-native’s fieldnotes of the same scenes demonstrated curiosity but contained little detail and interpretation.

Perception is also structured through activities and work, which give rise to varying interpretations. Goodwin’s (1994: 606) concept of ‘professional vision’ sheds light on how professionals learn to see and understand phenomena in professionally relevant ways. Analysing the 1992 Rodney King trial, in which a recording of four white police officers assaulting an African-American motorist stopped for speeding was used by both sides to contest the opposing side’s account, Goodwin (1994: 606) argues ‘the ability to see a meaningful event is not a transparent, psychological process but instead a socially situated activity accomplished through the deployment of a range of historically constituted discursive practices’. These practices include ‘coding’, using a classification scheme to structure perception and transform the event into an object of knowledge salient to the particular profession, and ‘highlighting’ which makes specific features of a complex perceptual field notable by marking them out. Using these techniques, the defence described the events, not as a man on the ground being beaten, but as movements by the man recognisable as aggression. Goodwin therefore situates perception, not in the individual mind but in disciplinary activities designed to make significant features of a phenomena recognisable to others. Focusing on how work allocates different information to professionals depending on their role, Anspach (1987) also takes up the issue of how professions develop their particular perspectives. She analysed decision-making in neonatal intensive care and argued that the organisation acts as an ‘ecology of knowledge’. She found that doctors had limited contact with the infants and relied heavily on diagnostic technology, whereas nurses’ perspectives were informed by perceptual cues developed through continuous contact with the infants. Consequently, physicians and nurses differed systematically in their view of babies’ prognoses (Anspach, 1987). The point to note is the way one’s position in the organisation serves as an interpretative lens. Developing these ideas about how the organisation acts on the content of knowledge, Waring (2009) describes the transformations that occur as knowledge travels through organisational processes. Clinicians’ initial verbal accounts of safety incidents reflect the complexity of clinical work, and understanding of events develops through reflection and collegial discussion with experienced colleagues. In formal written reports, descriptions were supplemented with further actions to justify actions or deflect blame. Once received by risk managers, descriptions were stripped of technical and contextual detail to conform to database classification systems.

The sociology of knowledge holds important implications for risk and safety, and for the case examined below, Vaughan’s (2016 [1996]) concept of ‘normalisation of deviance’ is particularly illuminating. She traced the factors that produced a cultural disposition among NASA engineers and managers towards the rationalisation of risk, and how a pattern of decision-making that repeatedly led to the normalisation of deviant findings gradually moved practice away from safety standards and towards disaster. Vaughan (2016: 409) says:

the Challenger launch is a story of how people who worked together developed patterns that blinded them to the consequences of their actions. It is not only about the development of norms but about the incremental expansion of normative boundaries; how small changes – new behaviours that were slight deviations from the normal course of events – gradually became the norm, providing a basis for accepting additional deviance. No rules were violated, there was no intent to do harm.

By ‘normalisation of deviance’, Vaughan does not mean becoming accustomed to wrongful behaviour, rather, she points to the ‘prerational forces’ such as organisational priorities, which shape the options of a decision. She further explains how the uncertainty of engineering knowledge, a belief in technological redundancy, and the endorsement of decisions through official channels normalised signals of potential danger, resulting in mistakes with catastrophic consequences. Emphasising the locality of such perspectives, Dekker (2011: 39) further suggests that from the inside, drift away from safety standards is invisible:

From the outside, such fine-tuning constitutes incremental experimentation in uncontrolled settings. On the inside, incremental nonconformity is an adaptive response to scarce resources and production goals. This means that departures from the norm become the norm. Seen from inside people’s own work, deviations become compliant behaviour.

The message is that if work is routinely accomplished in ways that are different to safety policies, this is not deviance, it is normal.

The literature above all emphasise the located and embodied nature of knowledge, and that as knowledge is inescapably positioned and structured through activities and work, this gives rise to varying interpretations and multiple perspectives. How events are viewed, interpreted, and described are intrinsically connected to the positioning of individuals and the communities to which they belong. Different professions develop knowledge of varying characters, and the organisation acts to distribute different forms of knowledge to different groups of staff and to mediate and transform accounts of events. The sociology of knowledge holds important implications for the safety of healthcare, in particular how parameters of safe practice can shift, producing a cultural disposition toward accommodating anomalies and rationalising them as ‘normal’. These arguments raise concerns about how serious incidents are interpreted, how failures of care are established, and how the prevailing interpretation of events may change.

Methodological approach and case-study selection

The analysis below stems from an interest in sociological approaches to safety in healthcare and from a familiarity with the location. Being a social scientist at the local medical school, one who teaches and researches patient safety, provided a good reason for reading The Report of the Morecambe Bay Investigation (2015). On reading the report, it became clear that events documented within the report related to the concerns I pose above – how incidents with serious human consequences are understood when elevated levels of risk become accepted, the practices that produce these levels of risk become normalised, and how failure or substandard care is established.

The events I examine relate to a ‘cluster’ of five ‘serious untoward incidents’ (SUIs) involving patients on the maternity unit at Furness General Hospital (part of Morecambe Bay NHS Trust) during 2008. These events and others became the subject of an inquiry in 2013, and the inquiry report (Kirkup, 2015) is my primary source of data for the analysis below. The report covers more than this cluster of incidents, however, how these incidents should be viewed – whether they were related – was pivotal in establishing whether there were systemic failures in care.

The five SUIs were:

• (1) A baby damaged by the effects of perinatal hypoxia after preterm delivery

• (2) A maternal death due to high blood pressure, thought to have been unavoidable.

• (3) A maternal death due to an amniotic fluid embolism, thought to have been unavoidable, and the death of the baby due to shortage of oxygen.

• (4) An intrapartum stillbirth due to shortage of oxygen in labour.

• (5) The death of a baby from sepsis following prolonged rupture of membranes and maternal illness.

These events all happened within 1 year and the inquiry found in each case care was seriously deficient. Prior to the inquiry, numerous investigations were conducted and although some criticisms were levelled at the Trust, a conclusion of systemic failures in care was resisted. In my reading, the inquiry report depicts an intense struggle as to how standards of care should be viewed, and it is revealing (and sometimes surprising) to see which description of events holds in which circumstances. As the report documents, from the inside:

staff considered that there had not been failures of care and that they were being unfairly criticised. This was most graphically illustrated by the comment made to us by an interviewee as she left the room that “sometimes bad things happen in maternity – people just have to accept it”. (Kirkup, 2015: 17)

Yet, from the outside:

deaths that occur during labour to a baby that had developed normally, described as intrapartum stillbirths: these should not happen, and their uncommon occurrence must be regarded as a serious incident requiring investigation. We were distressed to find that not only were intrapartum stillbirths a too-regular occurrence at FGH, they seem to us to have been treated with far less concern that we expected, and as a result opportunities were missed to identify substandard practice. (Kirkup, 2015: 174)

These quotes represent the polarised views from the inside and outside. To explore how the prevailing interpretation of events shifted from the inside to the outside, I have drawn extracts from The Report of the Morecambe Bay Investigation that explain the different positions on the adequacy of care and the processes by which those positions were strengthened or weakened. In line with the principles of narrative analysis, I constructed a chronology of events, preserving the sequences of action and the way participants negotiate language and narrative genres (Riessman, 2016). Narrative analysis maintains that it is useful to question (1) In what context was this account generated, (2) why was the story told that way, and (3) what did the story accomplish for the speaker (Riessman, 2016). A narrative approach considers both the author and audience, includes analysis of the local context – setting, questioner, listener – and positions stories, talk, and descriptions within a broader dialogue (Riessman, 2016). These issues are central to the analysis below and sit comfortably within the sociology of knowledge perspective. In particular, how perception is structured by one’s position, what that position enables one to see, and what it blinds one to, offer an entry point to thinking about how a belief in the adequacy of care can be maintained in the face of growing evidence to the contrary. Moreover, examining descriptions as artefacts that are shaped and circulate through many forums and processes, allows insight into the organisational practices that promote and obstruct acknowledgement of failures in care.

Given the emphasis on how events are described and the primary source of data – an inquiry report – there is need to reflect on inquiries as processes of knowledge construction and inquiry reports as texts that deliver a definitive version of events. First, in circumstances of public discontent, inquiries are said to serve a political agenda, conveying a sense of decisive action, providing reassurance, and rebuilding public confidence (Greer and McLaughlin, 2017; Timmins, 2019; Walshe, 2019). Consequently, Boudes and Laroche (2009: 79) indicate the need to avoid positioning the reports as ‘unbiased windows’ on events. They highlight how a seemingly objective chronology of events entails choices that suggest a storyline: other start and end points, the inclusion of some occurrences and not others, may cast the events in a different light. Further, the insertion of comments and the way data is presented, for example, with mitigation or challenge, distributes accountabilities and informs how evidence should be interpreted (Boudes and Laroche, 2009; Goodwin, 2018).

For these reasons, I searched the inquiry’s open interview records for context on the events and interview excerpts detailed below and for further relevant details not included in the inquiry report. A major limitation of this source of data is that, for reasons of confidentiality, discussion of specific cases of patient care occurred private sessions and are not publicly available. I have also drawn on a book recounting personal experiences of the events and some media reports. These additional sources were used to challenge and enrich the developing arguments (Squire, 2011). The point to note is that while all sources claim some authoritative knowledge of ‘what happened’, they are based on different forms of knowledge and access the events in different ways. None can be taken as a straightforward account of ‘what happened’. Multiple versions of ‘what happened’, however, does not mean they are all equal, and I am interested in how the shifts in consensus occurred and what that means for differentially positioned individuals in acknowledging failures of care.

I will now discuss, first, how the view from the inside – that the 2008 SUIs were unconnected – was constructed and sustained before moving on to explore the work it took to overturn this view and to have an inquiry commissioned; a move which cemented the view that the events were connected and did represent systemic failures in care.

Constructing the view from the inside: descriptive practices of internal investigations

The occurrence of an SUI automatically triggers an internal investigation. In theory, this process at Morecambe Bay involved a review by the clinical director and the head of midwifery. In practice, the review was mostly undertaken by the maternity risk manager, and the clinical director was not routinely involved (Kirkup, 2015). The maternity risk manager was a senior midwife, a supervisor of midwives ¹ , and also a staff representative (formerly a Royal College of Midwifery union official) – a combination of positions that prioritises support and advocacy for midwives. The inquiry report suggests the conflicts of interest this combination of roles produced informed the maternity risk manager’s sympathetic perspective and shaped the views of those around her:

We believe that this [conflict of interest] was significant in the events that developed, not only in encouraging the group think among midwives that all was well but also in promoting a view at more senior levels that there were no systemic problems in the unit (Kirkup, 2015: 180).

The maternity risk manager is positioned as central to maintaining the status quo, reassuring those above and below that all is well. Hutter (2005: 72) points out that organisational barriers to seeing exist, such as a reluctance to move bad news upwards through the hierarchy. But, referring to the rules, roles, and authority relations that inform perceptions of ‘appropriateness’, Vaughan (2016: 197) highlights the ‘prerational dynamic’ by which organisational arrangements determine the range of choices people see as rational in a given situation. Vaughan’s emphasis on the ‘prerational’ is a reminder that reluctance to convey bad news may not operate on a conscious level.

The inquiry report goes on to criticise the quality of internal investigations, and the quote below indicates the importance of the position from which these descriptions emerge:

We were distressed to hear and see evidence that the investigation of maternal deaths was also sometimes superficial and rudimentary, and failed to identify clear examples of substandard care. In some cases, this reflected an over-reliance on poorly completed records, when it would have been evident from a conversation with the relatives of the deceased that warning signs were missed some time in advance of the subsequent acute deterioration of the patient’s condition. (Kirkup, 2015: 175)

Instead of identifying failures in care, internal investigations were dominated by the comparatively benign criticism of the need to keep better records. The dependence on such records, noted by the inquiry, again points to the limited view these investigations took. Patients and families had no input at this stage and there appears to be only a limited role for other healthcare professionals. So, the first move in knowledge-making about these incidents was the internal investigations led by an individual deeply situated within the midwifery community whose commitment to supporting midwives is evidenced by her combination of roles. The findings of the internal investigations reflect these commitments in their refusal to attribute significant fault to midwifery practices. Internal investigations, it seems, were superficial and decidedly partisan.

However, constructing a view that the SUIs did not represent systemic failures in care required a further move – that they should not be formally examined together, a move primarily located at Trust Board. Although the Chief Executive insisted the SUIs had repeatedly been discussed together, the inquiry found ‘no documentary evidence of any systematic review of the cluster of SUIs in 2008, and no record that they were notified to or discussed by the Board’ (Kirkup, 2015: 20). Nevertheless, the Chief Executive defended his view of events as unconnected by stating that:

a connection could be characterised by clinical similarity of presentation or complications, or involvement of the same clinicians, or because they all demonstrated a similar pattern of deficit in clinical quality and standards. (Kirkup, 2015: 86)

Superficially, there was no pattern to the 2008 incidents; the events involved different clinicians and different things went wrong. It was considered an unfortunate, but unrelated, cluster. The inquiry panel, on the other hand, were astonished that (with one exception discussed below) ‘none of the unit clinicians, clinical director, or executive directors appeared to have considered that there may have been a pattern to the occurrence of these extremely rare events in a small unit’ (Kirkup, 2015: 175). This lack of curiosity could be explained by the Chief Executive’s comments:

“I had no reason to believe that anything had happened or changed there that changed the overall safety of the Unit. I was assured by the Head of Midwifery and by the – and by the Associate Medical Director that the Unit was safe. We hadn’t changed anything in terms of number of midwives or doctor’s rotas or anything that would have. . . destabilised it in that sense. So I guess we were as confident as we could be that, you know, that we were dealing with something that we thought we understood.” (Kirkup, 2015: 87)

There is a circularity in perception here. Like the ‘acceptable risk’ procedures Vaughan describes, organisational procedures for investigating incidents had been followed and had satisfied questions about safety, resulting in the knowledge that endorsed the status quo. By limiting the frame of reference (each SUI being considered individually) and the sources of information drawn upon (clinical records and staff statements), internal investigations produced a perspective where failures in care were held outside the field of view. The Chief Executive’s comments demonstrate Anspach’s (1987) ‘ecology of knowledge’, wherein what is known depends on what sources of information are seen, and what is seen depends on one’s position within the organisation. Even at the head of the organisation, one’s position shapes what one sees and understands, and this understanding circles back in the form of resistance to further investigation discussed below.

Descriptions under scrutiny: sustaining the ‘unconnected’ view

The view that events were unconnected did not go unchallenged. Although there was a widespread lack of curiosity about a connection between the 2008 SUIs, it was not a perspective held by all. In 2008, an obstetric consultant attempted to show connections between the fourth 2008 SUIs – an intrapartum stillbirth – and an earlier neonatal death in 2004. He:

wrote a letter identifying some of the deficiencies that had contributed to the disastrous outcome, drawing a parallel with the early neonatal death in 2004 . . . and warning that in his view further tragedies would ensue unless action followed. His letter was addressed to the Clinical Director [of Obstetrics and Gynaecology], and copied to the Trust’s Medical Director…. He did not receive a reply, and we could find no evidence that his concerns were taken seriously, acted upon or investigated. (Kirkup, 2015: 16)

It is puzzling why this account had so little effect; it carried the hallmarks of authoritative knowledge being informed by local and professional knowledge. The lack of response shifts attention from the author to its audience. As Hirshauer (2006) points out, the fate of a description is settled by the reader. Perhaps, the lack of effect was because the letter was circulated to a limited and, crucially, internal audience. When questioned by the inquiry team, the Medical Director had no recollection of receiving the letter, despite stating he always took a formal approach to handling concerns (Morecambe Bay Investigation: Open Interview Records, p. 31), and the Chief Executive, who was later alerted to the existence of the letter, could not recall any response to the concerns raised (Open Interview Records, p. 50–51). This account was also in the context of exceedingly ‘dysfunctional working relationships’ between obstetricians and midwives. In the Open Interview Records, the Chief Executive alludes to the letter covering ‘relationship issues’, and elsewhere in the inquiry report, the dysfunctional relationships receive considerable attention; midwives are portrayed as fiercely protective of their ‘normal childbirth’ boundaries (and the autonomy it confers), and the obstetricians are cast as complacent, content to be called and dismissed as the midwives see fit (see Goodwin, 2018). In this context, it may be that safety concerns were interpreted as a product of persistent interprofessional differences.

A more persistent challenge came from outside the organisation. Following the death of a baby from sepsis, the father (Mr James Titcombe) made a complaint to the Chief Executive ² . It resulted in an external investigation by a midwife, an obstetrician, and a paediatrician (the Chandler, Hopps, and Farrier report) into the care of Joshua Titcombe, and a review by the supervisor of midwives. The supervisor of midwives investigation was organised through a professional midwifery structure and intended to be an external process, however, in this case, the investigating supervisor of midwives was internal to the Trust. The inquiry report identifies conflicting outcomes of these two investigations; the external report concluded that opportunities to detect and treat the infection before it became life threatening were missed, highlighting deficient monitoring of the baby by midwives, whereas the supervisor of midwives’ investigation played down the likelihood that better monitoring would have resulted in a different outcome. These two investigations are both descriptions forged through the lens of professional knowledge, but place inside and outside views in contest. The value of an external, professional, and multidisciplinary viewpoint is that it allows for comparison between maternity practices here and elsewhere, to see professionally salient details but, unaffected by local cultural knowledge, to see them differently, in the awareness that things could be otherwise.

However, while there were external reviews and scrutiny from regulatory bodies, there was still no analysis of all five cases together. Furthermore, action in response to the SUIs supported the view inside the Trust that they understood the situation and had taken appropriate measures. As the Chief Executive recalls,

I was convinced the circumstances were different, and there was a different reason for them . . . at the time, I definitely believed that we had worked out what had gone wrong, or hadn’t gone wrong . . . So I think we had convinced ourselves that we were doing the right things to try to put right the issues that had come up from those cases. (Open Interview Records, p. 8)

In February 2009, when the Trust declared the 2008 cluster of SUIs in their application for Foundation Trust status (a status that confers more autonomy), one of the regulatory bodies (North West Strategic Health Authority) questioned if there was a gap in understanding regarding whether the five 2008 incidents were connected and suggested further investigation. The Trust commissioned another review; however, in setting the terms of reference, they steered the review away from examining the cases together and onto governance arrangements. The Chief Executive explains:

“. . .I commissioned the report from Pauline Fielding. . . that was around trying to look at clinical governance and – not to review the individual cases, but to come back and say, ‘Look, in terms of the governance across this patch, what could we do? What structure could we put in place. . . that would get us past this?” (Kirkup, 2015: 87)

Resistance to an external review of all five cases was based on the conviction that the incidents had been investigated, understood, and acted upon. The perspective of the review, therefore, should be on orientated towards the future. As the Strategic Health Authority officer tells it in the Open Interview Records (p. 29), the Chief Executive explained this framing of the Fielding Report as being to examine issues arising from all of the incidents, which she assumed would necessitate a re-examination of the incidents altogether. It did not, and this move prevented any connection between the incidents from becoming visible.

How the ‘unconnected’ view began to change: relatives’ activism

In the face of such resistance, it required complaints and other forms of activism by relatives, as well as media attention, inquests, and a police investigation to successfully challenge the ‘unconnected’ view. However, it is in relatives’ activism where the bulk of the descriptive work necessary to overturn the received view took place. On receipt of the Chandler, Hopps, and Farrier report, a contest of views began between Mr Titcombe and the midwives:

the report was shared with Mr Titcombe, who challenged several aspects, and it became clear that there were significant discrepancies between the accounts given by midwives and the record made by the Titcombe family shortly before Joshua died. (Kirkup, 2015: 170)

Crucially, the Titcombes made their own record, documenting their first-hand experiences, whereas the external review was based on medical records and the midwives’ accounts given to the internal investigation. Here, the lines of comparison are between the lay and professional and available sources of knowledge; the Titcombe’s account rests on what Hirshauer (2006) calls a naturalistic rhetoric of authentication, ‘I was there’, based on direct and close observation, whereas the external report can claim the epistemological authority of ‘professional vision’ wherein perception is honed to emphasise certain details over others through professional education and training (Goodwin, 1994). But given these asymmetries of clinical knowledge, the outcome is not what one might expect, reflecting May’s (2007) contention that lay/professional analyses of knowledge contests should take greater account of the societal context in which shifts in organisational structures and the repositioning of patients and healthcare professionals have reshaped the dynamics of knowledge.

The Titcombes’ successful challenge of the midwives’ account bears testimony to the changing dynamics of knowledge but there were further inadequacies in the midwives’ account. Initially, they had argued that Joshua had not been hypothermic.

Only when Mr and Mrs Titcombe presented a convincing account that Joshua had been significantly hypothermic on two occasions, an account accepted by the midwives, did their version of events change to a universal lack of awareness of the significance of neonatal hypothermia. (Kirkup, 2015: 18)

What the Titcombes’ description contained is not documented in the inquiry report; however, in Joshua’s Story, a book subsequently written by James Titcombe, he describes how his wife recalled two distinct temperature measurements recorded on the observation chart that were hypothermic. He also describes the conversations he had with the midwives, and the actions taken to warm Joshua up. These recollections are detailed and precise; there is a persuasive account of Joshua being moved into a heated cot with a heater directly over him only for it to be removed when, following Mrs Titcombe’s concern, the midwife touched Joshua’s skin and recognised he was too hot (Titcombe, 2015).

On the midwives’ subsequent account that they were all unaware hypothermia in a neonate was a cardinal sign of infection necessitating urgent action, the casting vote is with the audience. The inquiry report, however, gives a steer on how to interpret this account, describing it as ‘extraordinary’ and noting that ‘many experienced interviewees expressed varying degrees of surprise and disbelief’ on hearing it (Kirkup, 2015: 18). The shifting of accounts when challenged, and apparent implausibility of the subsequent version of events, did not make for a strong epistemological standpoint. The hospital admitted liability but the Titcombes were not satisfied that the hospital had been honest. Moreover, in relation to the argument being developed here, admitting liability in one case did not represent a pattern and evidence of systemic failures in care. Relatives’ activism, then, did not end with letters of complaint; in the Titcombes’ case, it also involved liaison with Strategic Health Authority, lobbying for an inquest, taking their complaints to Parliamentary and Health Service Ombudsman who are responsible for making the final decision on unresolved disputes, and setting up a campaign group.

Inquests and police investigations: opening up and closing down descriptive spaces

Of all the activist strategies, the Titcombes’ pursuit of an inquest and their willingness to engage with local, national, and social media were particularly consequential. Taking the inquest first, Joshua’s death occurred in Newcastle where he had been transferred for highly specialised treatment. However, due to his extremely poor condition on arrival, the coroner regarded the death as expected, and thus an inquest unnecessary. Only after considerable efforts by the Titcombes to inform the coroner of events preceding Joshua’s transfer, did he refer the case to the coroner in Cumbria.

The inquest finally took place in June 2011 during which the coroner criticised Trust staff for ‘collusion in preparation for the inquest and possible destruction of evidence’ (Kirkup, 2015: 37). In addition to the collective change in midwives’ accounts over whether Joshua was hypothermic, the ‘collusion’ refers to some model answers circulated by maternity risk manager in response to ‘difficult questions’ posed by the Trust solicitor. The inquiry report notes:

This distortion of the process underlying an inquest was picked up by the coroner, who commented on the similarity of the accounts that he heard from different witnesses and the concern that this caused him. (Kirkup, 2015: 19)

‘Distortion’ and ‘collusion’, however, is not how the maternity risk manager saw it. When questioned on why she circulated the model answers to staff participating in the inquest, she explained that it was to obtain answers for the Trust solicitor, and to give staff the opportunity to make amendments and correct inaccuracies, not to coach staff:

It was never the intention to coach staff to do anything, we never did that. We just told them to tell the truth. That they would have statements in front of them and they would answer questions that the Coroner gave them. That was certainly not my intention when that document was completed and sent back to the solicitor. (Open Interview Records, p. 51)

Traditionally, decisions to violate rules have been explained as calculated, amoral choices (Vaughan, 1999: 291) yet Vaughan (2016: 408) urges us to ‘wonder about processes that normalise deviance, possibly allowing organizational members honestly to view their actions as normal rather than deviant.’ Given the maternity risk manager’s unfamiliarity with inquests, her roles within the Trust, and her supportive stance towards the midwives, it is possible that she did not see this preparation as violating procedures. Yet, further doubt was cast on the veracity of the midwives’ accounts by the loss of a key observation chart. So concerned was the coroner at this ‘possible destruction of evidence’ (Kirkup, 2015: 37) that he issued a rule 43 letter (a report on action necessary to prevent further deaths) and this initiated a police investigation.

Before discussing the police investigation, I would like to briefly consider another inquest as it illustrates how they are a process through which descriptions are constructed, circulate and gather force. Regarding the fourth 2008 incident – a stillbirth – the Trust argued that an inquest should not take place as a coroner’s legal powers extend only to deaths that follow live births (Kirkup, 2015: 173). This argument can be read as an attempt to close down space for potential criticism for it is here the view from the inside can be publicly scrutinised and descriptions can gather force or be dismantled. Moreover, as coroner’s courts are public proceedings and frequently attended by journalists, their conclusions are a matter of public record and may be widely disseminated. This is a space that allows for external reporting.

Legal processes both open up and close down descriptions. Above, the preparation for inquests solidified the midwives’ account, but its paucity was publicly exposed giving space to alternative views. The police investigation, which was characterised as ‘thorough’ and ‘persistent’ in the inquiry report but ended without prosecution, also shaped accounts but in ways that closed down subsequent descriptions:

Police interviews of FGH staff almost all ended with solicitors advising their clients to say nothing in addition to their written statement and answer no questions. . . . Although the majority of present and former staff were helpful and informative when interviewed [in the inquiry], some appeared to us much more constrained in answering and stuck doggedly to previous lines even when these were difficult to sustain under challenge. We could not help but detect echoes of both the inquests and the police investigation in this approach. (Kirkup, 2015: 37)

While the police investigations and the inquests were counterproductive to producing a fuller description, importantly, they are external processes that cast a view from the outside, exposing concerns and disseminating them widely.

Increasing external scrutiny: strengthening the view from the outside

In replacing the dominant ‘unconnected’ view, a tipping point had been reached and events picked up the pace at this point: the inquests and police investigation prompted further external investigations by governance agencies (Monitor and CQC) as well as widespread adverse publicity, two further SUIs occurred in 2011 which bore ‘unmistakable similarities to earlier incidents’ (Kirkup, 2015: 38), in February 2012 Chief Executive resigned, further Board changes followed, and in September 2013, the inquiry was commissioned.

Throughout all this, James Titcombe kept up a regular media presence using social media and engaging with the national press. In Joshua’s Story, he describes collaborating with other affected parents to establish the campaign ‘Morecambe Bay Inquiry Action’ which involved setting up a Facebook group, a website, a launch event filmed by national news media, liaison with members of parliament, and ultimately being instrumental in having the inquiry commissioned. He is quoted in the Independent (March 2014) national newspaper as saying:

For me, social media made the difference. The spread of people that you can contact, the coming together, the working together. Twitter is an awesome, powerful tool for a campaigner. Anybody can ignore one person – you’re just a grieving dad. If there’s a group of you saying the same thing, that’s harder to ignore. Without social media I don’t think this movement would have happened.

The growing importance of social media in campaigning has been noted elsewhere (see, e.g. Segerberg and Bennet, 2011); here, James Titcombe highlights the collectivity of a description as being critical to its effectiveness. However, while the consensus around failures in care was growing, vociferous public debate persisted on both sides. Many local people joined the campaign; others vocally resisted it, arguing that intense negative publicity would close a much-needed maternity unit. Moreover, public debate and publicity came at a cost for all involved; James Titcombe (2015) was subject to the anger of local residents and, referring to the constant, conspicuous presence of reporters at the hospital, the Chief Executive commented, ‘there were days I’m not sure why the midwives turned in for work in Barrow; they were under that much pressure’ (Open Interview Records, p. 47).

The inquest, the police investigation, the campaign, and media coverage all contributed to increasing public suspicion of systemic failures in care, and the inquiry cemented this view. Over 18 months, the inquiry examined hospital statistics, case notes, emails, minutes of meetings, internal investigation reports, coroner’s reports, external reviews, as well as conducting interviews with staff, members of regulatory bodies and other professionals associated with the hospital (for further discussion of inquiry methods, see Goodwin, 2018). It concluded that ‘these events represent a major failure at almost every level’ (Kirkup, 2015: 11). These were no longer individual cases.

Regarding the decisiveness of inquiries’ versions of events, there are a number of points to be made. First, it has been noted that inquiries construct a ‘univocal and coherent view’ of complex and uncertain events (Brown, 2003: 96) and that inquiries are structured around discovering the objective truth of matters under investigation (Walshe, 2019). However, this is what inquiries are charged to do; one version must win, given the terms of the investigation. Taking the above analysis into account, what can be said is that considerable descriptive work had already gone into constructing the view that there were systemic failures in care, and an investigation taking place years after the events, when the outcomes are known, provides a vantage point not available to those inside the Trust at the time of the events.

Second, the authority of the inquiry’s account has a number of bases; inquiries are forensic in their approach, they adhere to rules governing scientific knowledge production and representation, such as making provenance claims around the commissioning of the inquiry, and comprehensively detailing the evidence reviewed (Brown, 2003). The report must narrate an engaging story by using direct quotations from witnesses, include large amounts of contextual and micro-situational detail, and conform to generally held notions of how people behave in crisis situations (Brown, 2003) being informed by current knowledge of safety science and human factors principles (Goodwin, 2018, 2019).

Third, credibility is established in large part by the expertise of the inquiry team (Goodwin, 2018), in this case being chaired by Dr Bill Kirkup, a medical doctor whose career has included being an obstetrician, serving as the Associate Chief Medical Officer for England, and participating in a previous inquiry. He was assisted by experts in Obstetrics, Nursing and Midwifery, Paediatrics, and Healthcare Law. But, beyond credentials, the analysis here shows that the inquiry team brings a mix of perspectives, being an outsider to local knowledge which facilitates questioning of local rationalities, but insider to professional knowledge thus bringing a view rich in interpretation on the adequacy of professional knowledge and practice.

Finally, the public nature of the inquiry, and breadth of the intended audience, shapes inquiries in important ways. Inquiries are generally commissioned in response to public outcry, and the degree to which affected families are involved and kept informed makes them distinct from other forms of regulation (Penhale and Manthorpe, 2004). Moreover, the availability of the report, the methods of investigation, and sources of information to interested members of the public and national press potentially allow the weight of public judgement to play into the consensus shifting from inside to outside views.

Conclusion: descriptions, sociology of knowledge, and establishing failures in healthcare

Social science scholarship has often made a distinction between description and explanation with description being comparatively undervalued (for a discussion of this, see Marcus et al., 2016), yet for Latour, a description’s ability to explain is the hallmark of quality:

the opposition between description and explanation is another of these false dichotomies that should be put to rest . . . If a description remains in need of an explanation, it means that it is a bad description. (Latour, 2005: 137)

Descriptions matter, they are a key tool in the construction of knowledge; they confer weight, involve interpretation, and bestow meaning. The analysis above shows the import of descriptions where there is a struggle for understanding and the processes by which their content is shaped. For the author, this involves the relationship between knowledge, observation, and interpretation, the author’s positionality and how this frames perception. But further, descriptions are shaped by where and to whom they are presented, and their strength derives in part from their circulation and collectivity.

Analysing the Morecambe Bay Investigation report has shown the contest of views at play between individuals and collectives, those on the inside and the outside, and lay and professional. Looking at how descriptions were put together, and attempts to stop them being put together, the Morecambe Bay Investigation illustrates some important features of descriptions when establishing failures of healthcare. Taking their epistemic origin – lay or professional – first, the interplay was more complex than one pitted against the other. Professional knowledge is still authoritative knowledge, but it is not beyond challenge; the direct experience of Titcombe’s account was far more persuasive than the midwives’ which drew considerable scepticism. Nor were professional and lay views always in contest; there were criticisms from professionals as well as relatives: the obstetrician who raised concerns, the Strategic Health Authority officer who questioned whether the five SUIs were fully understood, the coroner, and the inquiry panel. The significance in each case is how individuals were positioned and the audience they addressed. The obstetrician and the Strategic Health Authority officer addressed a limited audience of senior management within the Trust resulting in no change to the predominant perspective, whereas the coroner and the inquiry panel addressed a wider public audience and were more consequential in shifting the predominant perspective.

Next, the position from where the descriptions emerged was largely (although not exclusively) organised around internal and external perspectives. In short, external perspectives were more likely to be critical, but terms of reference and sources of information fundamentally limited their vision. The external review by a midwife, obstetrician, and paediatrician identified some failures of care, but criticism was limited by only having access to medical records and staff statements. The Fielding Review’s potential for criticism was circumvented by shifting the terms of reference away from exploring a connection between the five SUIs. Externality of perspective is not always enough; external control over the investigation’s remit or terms of reference is essential, as with the inquests and inquiry. Again, it seems, audience is key; the descriptions most effective at establishing failures in care were those that allowed public scrutiny – inquests, media reporting and social media campaigns, and inquiries.

Lastly, if descriptions are to hold, they must create a shared perspective. Collectivity defends against the charge that it is one view against another. The case above illustrates the importance of making detailed knowledge of events accessible to the public, and the processes that allowed a collective viewpoint to develop – inquests, social media campaigns, and the inquiry. These were critical in shifting the received view that the SUIs were unconnected, but public debate and the exposure it involves come at a personal cost to all concerned.