Abstract
Gross negligence manslaughter is a well-established common law offence in English jurisdiction. Nevertheless, it remains a troubled area of law. Academic commentators, as well as the Law Commission, have frequently called for reform. We explore this offence through the prism of medical manslaughter and argue that the offence is insufficiently nuanced in certain situations of death caused by gross negligence. This applies particularly to complex professional circumstances where a coalescence of inter-linking factors may converge with disastrous results for the parties concerned. Using the clinical care model as an exemplar, we argue that relentless focus on the offender can result in a disproportionately high level of blame being foisted on an individual. On the basis of decision theory, we suggest that insufficient regard may be given to confounding contextual matters. Failure to address underlying systemic factors may impede wider patient safety initiatives such as professional medical candour and strategies designed for institutional improvements directed at broader societal benefit. We offer an alternative approach based upon principles of fair attribution of blame and transparency. A new statutory offence is suggested for a fault-based standard that is set by law that gives greater prominence to contextual matters. We argue for wider disposal options that include ancillary court orders for rectification of systems failures and more proactive engagement of the coronial process.
Introduction
The gross negligence manslaughter (GNM) action brought against Dr Bawa-Garba in 2016 involved a junior doctor specialising in paediatrics who had recently returned to work following 14 months of maternity leave. She was charged, convicted and given a suspended sentence following the death of a critically ill 6-year-old child who had been in her care. 1 Subsequently, a Medical Practitioners Tribunal found that the doctor’s fitness to practise had been impaired and she was suspended from practice for a year. The General Medical Council appealed the tribunal’s decision as being too lenient and called for her removal from the medical register. The doctor’s subsequent appeal against removal was allowed by the Court of Appeal. 2 The case and series of events that followed perturbed the medical profession both nationally 3 and globally 4 and prompted two systematic reviews of the offence in the context of clinical care. 5 Both reviews pointed to the need for accountability in systems that support medical practice. The review led by Professor Sir Norman Williams strongly recommended working towards a just and learning culture in healthcare, whereby professionals feel able to raise concerns and reflect openly on their mistakes and where only those responsible for providing unacceptable standards of care are held to account. This is expected to improve patient safety and develop a clearer understanding of where the (legal) bar ought to be set for gross negligence in cases of medical manslaughter. 6 The strength of response from the medical profession against Dr Bawa-Garba being charged with GNM, as well as from respected members of the public 7 prompted us to express our thoughts in reconsidering the role of GNM where death has been caused by (grossly) negligent medical care.
GNM is a common law offence that has been refined gradually over the years. Its contemporary basis was enunciated first in Bateman, 8 almost a century ago. The current law is defined by Adomako 9 and piecemeal development has ensued subsequently, such as the need for foreseeability of a serious and obvious risk of death, 10 guidance as to how juries ought to be directed, 11 further precision on causation in cases of omission and an increase in severity of sentencing to bring the offence more in line with other forms of manslaughter. 12 Nevertheless, GNM may not be the optimal response to death caused by gross negligence in certain circumstances. In particular, we consider situations where death is caused by a defendant’s act or omission during the course of conducting legitimate work within a complex managed environment. In cases such as these there may be a range of suboptimal interrelated factors or failures which, at the material time, might have impacted on or influenced that conduct which is later deemed to be erroneous or careless.
Our principal objection to GNM is that the offence in its current form focuses too narrowly on the defendant as an individual. Insufficient consideration is given to contextual factors such as the challenges and pressures caused by inadequate systems and processes in the work environment and the cumulative adverse impact this may have upon professional judgement with potentially tragic consequences. We explore this issue through the prism of clinical care and use ‘medical manslaughter’ as an exemplar while recognising that cases of GNM also occur in settings other than healthcare. 13
We do not propose to rehearse well-trodden arguments about GNM other than as background. In this article, we argue that context is highly relevant, particularly where fatalities are caused by medical negligence. There are three main reasons for this. First is the centrality of interdependency between context and decision-making in clinical care. Second, there may be a putative place for institutional liability for inadequate systems and processes where grossly negligent conduct occurs in the context of managed clinical care. Third, context-driven acknowledgement and learning from mistakes requires transparency for future avoidance and improvement.
This article is set out as follows. We begin by considering some of the shortcomings of GNM and argue that the offence is insufficiently nuanced to account for contextual factors in medical manslaughter such as the inherent risk of death in patients who are seriously ill. Next, we consider the interdependency of an individual’s decision-making process in the clinical environment. Following this, we examine the role of healthcare institutional liability for death of patients in the context of wider failures of systems and processes. Then we focus on the medical professional duty of candour that may be inhibited by fear of recriminations through prosecution for manslaughter. Opportunities for transparency, and learning from errors, may be lost thereby negating the potential to improve patient safety and ultimately benefit the wider healthcare economy. We rethink GNM and propose a model for a new offence of ‘death caused by grave professional negligence’ a label specific to this context that better reflects the wrong concerned more accurately, alongside wider ancillary orders for quality improvement and involvement of the coronial system. Finally, we offer our concluding thoughts.
Problematic issues with GNM
Background
The ingredients of the offence can be stated simply: the defendant owed a duty of care to the deceased; there was breach of that duty; the breach caused (or significantly contributed to) death of the victim; and the breach should be characterised as sufficiently ‘gross’ such that it deserves criminalisation. 14 More recently in R v. Rose, Sir Brian Leveson added that it must be reasonably foreseeable that the breach of duty gave rise to a serious and obvious risk of death and that the circumstances of the breach were truly exceptionally bad and so reprehensible as to justify the conclusion that it amounted to gross negligence and required criminal sanction. 15
The concept of ‘gross negligence’ has developed incrementally since its initial formulation almost a century ago. 16 Disquiet has long been expressed about the vagueness of its definition as articulated in R v. Adomako: ‘whether having regard to the risk of death involved, the conduct of the defendant was so bad in all the circumstances as to amount in their judgment to a criminal act or omission’. 17 It is for the jury to decide whether or not the conduct of the defendant is criminal. In effect, therefore, juries are told to convict if they think that a crime had been committed. 18 The circularity of the offence is highly troublesome. Nevertheless, although GNM has been criticised frequently, it has been affirmed consistently by the courts. 19 In Rose, Sir Brian Leveson indicated that five elements must be proved to find a person guilty of GNM. 20
The direction to be given to the jury has been refined recently and standardised. When a jury is faced with several issues to consider in a case, experience suggests that jurors can be assisted by written sequential lists of questions or ‘routes to verdict.’ This can focus deliberations and provide a logical route to reaching a decision. In more complex cases, judges may provide the jury with a chart detailing permissible combinations of verdicts. A route to verdict should relate to the actual evidence in the trial and confined to the matters in issue. 21 In R v. Sellu, for example, the Court of Appeal was highly critical of the trial judge for not providing the jury with a ‘route to verdict’ document, which is now required routinely in all complex criminal cases. 22
In relation to the risk of death, the direction in R v. Broughton should be followed: ‘At the time of the breach there [must be] a serious and obvious risk of death. Serious, in this context, qualifies the nature of the risk of death as something much more than minimal or remote… An obvious risk is one that is present, clear, and unambiguous. It is immediately apparent, striking and glaring…’. 23 It is therefore an objective question. As to the potentially difficult question as to whether the risk was obvious, a subjective standard is to be used. The risk must have been obvious to the reasonable person in the defendant’s position who demonstrates the same level of negligence as the defendant. 24 The circumstances of the breach must have been ‘truly exceptionally bad’ and so reprehensible as to justify a criminal sanction and constitute manslaughter. A high threshold is therefore required.
In Sellu, the court emphasised that to repeat the word ‘gross’ was insufficient. The jury must be satisfied that the failure was not just serious, or very serious, but ‘truly exceptionally bad’. 25 According to Leveson, the jury must be assisted sufficiently to understand the line that separates ‘even serious or very serious mistakes or lapses, from conduct which was “truly exceptionally bad” and was such a departure from that standard [of a reasonably competent doctor] that it consequently amounted to being criminal’. 26 The jury must then be directed on principles of causation to prove that the defendant’s breach of duty caused, or significantly contributed to, the death. What is less clear, however, is how the jury ought to be directed to make a judgment about causation against a background of underlying illness. This can be a complex decision since even at the beginning of treatment the inherent risk of death, due to the condition itself, may be very wide ranging. 27
Guidance can be found in recent case law. As Langley J stated in Misra: If the prosecution has made you sure that [the doctor] did fail so grossly in their duty of care, then you must consider whether it has also made you sure that the failure or failures were a substantial cause of death.
The key point that emerges from the above judgments is that the defendant’s negligence should have been a ‘substantial cause’ or made a ‘significant contribution’ to the death of the victim. Yet how can this be evaluated by the jury other than through their subjective opinions? There is no clear judicial direction on what is meant by ‘substantial’ or ‘significant’. Furthermore, computation of these terms is even less clear when measured against a background of an existing inherent risk of death due to an underlying medical condition which might well be the reason for the hospitalisation.
Consider a situation where the inherent risk of death (r) is nil or very low. In these circumstances, it might be easier for a jury to assess whether the defendant’s negligence (n) made a substantial or significant contribution to the death, because of the absence or limitation of any other significant contributory cause other than the negligence itself. Matters are less straightforward if r is high because then the significance of the negligence (in causing death) needs to be evaluated against a background where the risk of death is considerable even at the start and before the occurrence of any negligence.
The matter becomes even more troublesome when the inherent risk is expressed as a range. Consider r within a range of a% and b% (where a is less than b). The contribution of negligence would be greater at a than at b (because a being less than b would confer a lower risk compared to any contribution from the negligence). Assuming n is constant, the significance of the contribution of the negligence to the death would be less, depending upon a higher level of risk against which that contribution is evaluated. Stated differently, the confidence with which it might be said that the contribution to death by the negligence was significant, or substantial, would decrease as the base or inherent risk of death increases. This would be highly germane to determining a significant or substantial contribution to death caused by the negligence itself. In Bawa-Garba for example, the inherent risk of death of the patient was given by experts as lying somewhere between 4% and 20.8% (a fivefold higher top end compared with the lower end). Depending upon the comparator (top, middle or bottom of the range), different outcomes may be obtained regarding the substantial or significant nature of the contribution to death from negligence, thereby leading to inconsistency for defendants. There is little judicial clarity as to how this range should be addressed in determining significant or substantial causation by negligence.
In broad terms, it appears that causation ‘can be established by showing…that death, although inevitable, was hastened by the conduct in question’. 30 Although it must be proved, beyond reasonable doubt, that non-negligent care ‘would have saved or prolonged life’, 31 this fails to distinguish between the concepts of prolonging life and hastening death. There seems to be little difference (in law) between living ‘appreciably longer’ and to life being ‘appreciably shorter’. 32 The court appears to regard shortening life and failing to prolong life as a corollary of one another and holds the two concepts as not being inverse but merely representing different sides of the same coin. 33
It is also unclear as to what ‘significantly sooner’ means (in respect of time). The court has persistently shied away from this question, employing terms such as ‘significantly’ and ‘substantial’ interchangeably in similar contexts. 34 Nevertheless, what may be significant will not always the same as substantial, as the latter has a higher threshold. The practical application of these concepts may be problematic, particularly with patients who are severely ill. Such patients may have a high chance of imminent, or near imminent, death in any event as part of the natural history of their condition.
Not only is it difficult to direct a jury in circumstances such as these, but even medical experts may not be able to pinpoint the expected outcome accurately. Prognostic figures quoted from the medical literature are based generally on aggregate or extrapolated data and may not necessarily mirror exactly the circumstances for a particular patient. By definition, an expert opinion is just one individual’s opinion and not a precise prediction. Opinions are based on best available medical evidence, as well as personal belief and experience. Some experts may spend more time on research than in clinical practice and there may be a gulf between the theory of what might be done, or expected, and what is actually done in practice.
There is undoubtedly a need for medical experts to assist the court with its understanding of clinical issues. However, in line with any complex criminal litigation, there is a risk of the jury being swayed by expert opinion. The more technical and scientific the opinion, and the stronger the personality of the expert, the more likely that influence will be. A concern is where experts for the prosecution provide their opinion in a way that potentially over-influences the role of the jury, in determining ‘gross negligence’. 35 There is a danger that the evidence given by experts, if not accompanied by detailed (non-technical) explanation, may result in the jury having to trawl through evidence adduced by leading questions asking whether the behaviour of the defendant was (or was not) gross negligence and then merely accept this as such. Thus, the jury’s role as a decision maker could be supplanted or even usurped. 36
What does ‘gross negligence’ mean?
Exactly what transforms negligence into gross negligence remains unclear, but the negligence must show such disregard for the life and safety of others as to amount to a crime and consequently be deserving of punishment. 37 It is a question for the jury to decide. 38 While there is no mandatory formulation, the task of the trial judge is to assist the jury in understanding what separates a very serious breach from (ordinary or less serious) negligent conduct. A very serious breach would be one which was ‘truly exceptionally bad and was such a departure from that standard [of a reasonably competent doctor] that it consequently amounted to being criminal’. 39
Negligence is measured by the traditional Bolam standard. It is not necessary to show disregard for the patient, or recklessness, or even that the accused was doing his or her best while working in under-resourced conditions. 40 It would, however, be fallacious to state that circumstances do not matter if there is to be further assessment of that doctor’s negligent conduct as being ‘truly exceptionally bad’. Making a determination on a very serious departure from standard clinical practice cannot be undertaken reasonably without considering the circumstances or working environment in which that conduct occurred, as both may be relevant to the outcome for the patient, as well as the actions of the practitioner.
In Bateman, the court held that …in order to establish criminal liability the facts must be such that in the opinion of the jury, the negligence of the accused…showed such disregard for the life and safety of others as to amount to a crime against the State and conduct deserving punishment.
41
…the jury must go on to consider whether that breach of duty should be characterised as gross negligence and therefore a crime. This will depend on the seriousness of the breach of duty committed by the defendant in all the circumstances in which the defendant was placed when it occurred.
42
Luck, or bad luck?
The moral justification for GNM remains problematic as the alleged perpetrator may not have had a morally blameworthy state of mind. 45 However, intention or recklessness is not a requirement for the offence. Prosecutors, judges and juries all struggle with the ill-defined concept of gross negligence. 46 In fact, whether or not one is prosecuted or convicted may depend very much upon luck. 47
In Adomako, the defendant doctor failed to notice disconnection of the ventilator during anaesthetic. As a result, the patient suffered catastrophic and ultimately fatal brain damage. The doctor’s subsequent conviction and sentence for GNM is generally regarded as appropriate in the circumstances. However, was Dr Adomako singularly unlucky? Lyons 48 provides evidence to show that similar cases have occurred where patients have not been ventilated properly and then have similarly suffered hypoxic brain damage, due to negligent placing of the endotracheal tube into the oesophagus rather than the trachea. The argument here is that such mistakes do happen, and no charges or convictions were brought in these other cases, albeit with similar outcomes.
It is interesting to speculate about what might have happened in Adomako if the patient had not died but had continued to survive in a state of severe brain damage. In the absence of death, a charge of manslaughter cannot be brought. 49 In fact, there are only limited circumstances for a criminal sanction if an individual is injured or incapacitated by mistreatment. 50 If a patient survives, but has serious injury as a result of negligence that is gross, the outcome for the defendant might be quite different.
In considering the requirement for gross negligence as stated in Rose and Broughton (foreseeability of a serious and obvious risk of death), it is possible to envisage a counterintuitive outcome. A doctor who assesses a patient negligently, and thereby fails to ascertain a foreseeable and obvious risk of death (and as a consequence acts incorrectly), could be convicted of GNM. On the other hand, if the doctor fails to carry out the required assessment (because of ignorance or lack of knowledge), then a serious and obvious risk of death would remain speculative as such a risk would only crystallise if the doctor had performed the appropriate checks. Thus, an egregious failure might actually escape conviction for criminal manslaughter.
51
The rationale behind this paradox may lie in the words of Sir Brian Leveson: The implications for medical and other professions would be serious because people would be guilty of gross negligence manslaughter by reason of negligent omissions to carry out routine eye, blood and other tests which in fact would have revealed fatal conditions notwithstanding that the circumstances were such that it was not reasonably foreseeable that failure to carry out such tests would carry an obvious and serious risk of death.
52
there is a perverse incentive for those who owe a duty of care to another to do as little as possible to discharge it and in doing so avoid potential criminal liability. Whilst this may be unlikely to impact on the high standard of care that doctors provide to their patients, it is not inconceivable that a landlord might decide not to provide his tenants with a carbon monoxide detector so that he remains ignorant should gas ever leak from the boiler.
53
The context of clinical care: Complex interdependent decision-making
From the perspective of decision theorists, clinical decision-making is influenced by at least three main factors. The first is that of ‘pattern recognition’. Medical students and clinicians are primed to recognise certain patterns of symptoms and signs that point them towards a most likely diagnosis. In the acute clinical setting, a diagnosis has to be reached fairly swiftly. There is a continuum in the complex process of human cognition, which at one end involves responses that are effortless and also rapid. Kahneman refers to this as ‘system 1 thinking’, which is the automatic ‘fast thinking’ that is used for familiar situations and is also used by default. In contrast, there is a slower process of thinking, which is more active and effortful, and is used typically for problem solving. ‘Slow thinking’, which is ‘system 2 thinking’, is more logical and conscious. 54 In reality, individuals switch from fast to slow thinking and then back again. The ‘oversight’ of automatic or intuitive thinking is linked to the prevention of error by feedback from system 2 into system 1 processes. 55 The more experienced the clinician, then (usually) the faster an accurate initial ‘most likely’ diagnosis will be made through the ‘fast process’ of system 1 thinking. Recognition-primed cognitive process is in contrast to a model where all alternatives may be compared within a complete information framework with the luxury of time and provides some insight as to why ‘error’, at least to some degree, is inevitable.
Second, clinicians (particularly in hospitals) tend to work within environments of ‘uncertainty’. There are several reasons for this. The information provided to them at any one time may be incomplete. This is not necessarily due to inadequacy, either on the part of the doctor (or patient), but because some, potentially highly relevant, information might not be available at that particular time. For example, the results of investigations might be delayed or not transmitted to the doctor for whatever reason. What is underestimated frequently is that clinical diagnosis is not always a precise computation. A clinician is often required to make the best decision at any material time that best fits with the particular clinical profile and circumstances of the patient. Even evidence-based medicine or ‘gold standard’ randomised controlled trials are not a ‘perfect science’, as the composition of subjects in these studies may not mirror the exact profile of the specific patient being treated at any particular time.
Third, clinicians tend to work in teams, rather than isolation. Clinical decision-making will often depend upon information, observations and responses provided by other members of that team. Salas and colleagues have defined a team as a ‘distinguishable set of two or more people who interact dynamically, interdependently, and adaptively towards a common and valued goal’. 56 Effective outcomes are more likely to ensue from effective teamwork, which is based upon agreements of belonging, trust, leadership and communication, as well as a shared commonality of thinking within that team. Optimal results can be achieved only by stable teams with shared ownership, objectives and values. For this reason, stable teams are likely to perform better than groups of professionals gathered hastily together to provide a clinical service. In Bawa-Garba, for example, an agency nurse failed to properly observe and communicate the child’s deterioration. She failed to take timely action that would have been commensurate with the seriousness and urgency of the child’s clinical condition. She had also turned off the oxygen saturation monitoring equipment prematurely (because the child was ‘looking better’). 57 Permanent nurses who were part of an established team would have been more familiar with working practices compared to agency nurses and may not have acted or taken decisions without further consultation with medical staff. Clinical decision-making goes beyond the skills and behaviours of single individuals to the culture and collective nature of human performance.
Justice and patient safety would be better served by a more sophisticated contextualisation of the relationship between individuals and human cognition in decision-making. 58 Clear distinctions need to be made between an act (or omission) that is an error, and one that is careless or negligent. Reason has defined error as a planned sequence of mental and physical activities that fails to achieve the intended outcome when that failure cannot be attributed entirely to chance. 59 Error is not intentional. If moral culpability is dependent upon making a free choice (either intentionally or recklessly), then an act that is not the result of such a choice should not be blameworthy. 60 What might be a fairer approach is to consider an act that results in error as being evaluated on the underlying thinking and reasoning upon which the decision is based, rather than being influenced by outcome alone. 61 It is questionable as to why outcome bias 62 should predicate a harsher appraisal of error.
The context of institutional liability
It is not uncommon for several adverse systemic factors to be operative at the time of a medical mishap. In Bawa-Garba, there was a catastrophic failure of the hospital’s electronic computer system which delayed the transmission of blood test results. There were agency nurses on duty, a shortage of supporting middle grade medical staff and a series of communication failures between nursing and medical team members. In addition, there was a range of work load pressures on the doctor herself. 63 Given such a miasma the question that arises is who is responsible for the welfare of a patient in hospital? Is it the doctor or nurse, other ancillary staff, managerial staff or the Secretary of State for Health? As Wells states aptly, ‘The truth is that none of these would function without the other’ and therefore all are responsible for the final outcome in a patient. 64 Inefficiency in systems for clinical care as well as resulting ‘near misses’ (from death or serious injury to a patient) have long been recognised. 65 Yet manslaughter charges within clinical care focus squarely on individuals with little or no regard to those attendant systemic issues. Although systemic failings have not escaped judicial notice, they are deemed to lie outside the purview of a trial. 66 The effectiveness of alternative legal vehicles (aimed at capturing organisational failings) through the Health and Safety at Work Act 1974 (HSWA) and Corporate Manslaughter and Corporate Homicide Act 2007 (CMCH) for enforcing institutional liability manslaughter through gross negligence appear to be limited.
Healthcare institutions are liable to prosecution under the HSWA if they fail to ensure, as far as reasonably practicable, the health and safety of their employees, clients, visitors and the general public. 67 In many prosecutions for corporate manslaughter, health and safety violations are also present. In fact, section 19 of the CMCH permits juries to return verdicts on health and safety as well as manslaughter charges. In R v. Misra and Srivastava, 68 for example, an action was brought under the HSWA as well as GNM. Four successful prosecutions were brought under the HSWA against the Mid Staffordshire NHS Foundation Trust on account of significant health and safety breaches causing the death of four patients. 69 More recently, Southern Health was fined £2 million after admitting having breached health and safety laws that led to the death of Connor Sparrowhawk who drowned in a bath following an epileptic seizure at Slade House in Oxford in 2013 and Teresa Colvin who died at Woodhaven Adult Mental Health Hospital in 2012. 70
Imposing fines on NHS organisations following breach of health and safety legislation creates inevitable dilemmas. In R v. Southern Health NHS Trust, 71 the court acknowledged the inevitable tensions between the need to hold the Trust to account for its actions while recognising its duty as a provider of healthcare in that every pound paid as a fine is a pound less that is available for clinical care. In similar vein, in R v. Shrewsbury and Telford NHS Trust, 72 the court reduced the fine imposed by 50% to reflect the constrained financial circumstances of the Trust and its role as a provider of public healthcare.
Following its enactment, the CMCH was widely perceived to be a landmark development in creating a new statutory offence by which companies and organisations could be found guilty of corporate manslaughter following serious managerial failures that caused death due to a gross breach of duty. A main driver behind the reforms that culminated in the CMCH was the acknowledged difficulty in securing a conviction for manslaughter against large corporate organisations and particularly those with diffuse management structures. 73 The Act abolished the common law offence of corporate GNM and the need to find the ‘directing mind’ of the company to be at fault. Under the CMCH, NHS Trusts could (in theory) be liable to prosecution in that all operational managerial decisions can be taken into account when ascertaining whether systemic failures in health and safety have caused death. However, the Act also requires that senior management has made a substantial contribution to the gross breach of duty in the way that they (mis)managed or (mis)organised organisational activities. 74 While this might be satisfied in small or medium enterprises, where senior management assume some operational tasks, it is less likely within large organisations. To compound matters further, the CMCH is acknowledged to be complex with interpretive ambiguity and multiple exceptions, such as situations of medical emergency. In fact, there are exceptions within exceptions regarding medical treatment which would seem to be superfluous since the offence is only committed when an organisation’s senior management has played a substantial role in the gross breach of care which causes death. 75
Although responsibility may lie under the CMCH, there are further hurdles to assigning organisational liability. Proof of causation requires that death was caused by the way that the organisation was managed to organise its activities. Clearly, organisations act through personnel who carry out their functions at the front line. It is often easier to define one person’s actions rather than tease out the contribution to culpability that might occur within multilayered managerial activity. Even if a breach of duty can be proved, that breach must be ‘gross’, meaning that it falls far below what can reasonably be expected of the organisation in the circumstances. 76 In addition, the jury must consider whether the organisation failed to comply with any Health and Safety legislation that related to the alleged breach 77 as well as the attitudes, policies, systems or accepted practices within that organisation which would have encouraged or produced tolerance of failure.
The barriers to prosecuting large organisations under the CMCH are clearly substantial. The first and only NHS Trust to be prosecuted was the Maidstone and Tunbridge Wells NHS Trust. 78 A woman aged 30 failed to recover from an anaesthetic and died of a cardiac arrest following surgery to remove residual placental tissue. The prosecution alleged that the Trust was liable for appointing doctors that the Trust knew, or ought to have known, lacked the training, qualifications or competence for their roles, and for failing to ensure the doctor with primary responsibility for care was supervised by a consultant. Having heard the prosecution, the judge ruled that while there was evidence that the Trust had breached its duty of care, there was no evidence by which a jury could properly conclude that the breach was gross or that the breach had caused the death. 79
A company that owned and operated Autumn Grange care home was convicted of corporate manslaughter under the CMCH in December 2015 following the death of a resident due to appalling care and systemic neglect. 80 In the circumstances, the organisation was classified as a ‘micro company’. Since the director who pleaded guilty to GNM was also an acting manager with operational duties, the statutory requirement that senior management made a substantial contribution to the gross breach of duty could be satisfied. The successful prosecution of Autumn Grange is a rare occurrence that succeeded only because of the unusual position that a director with strategic accountability also had added operational duties.
In the context of institutional liability, it appears that the HSWA and CMCH offer little comfort for the hope of achieving recourse following the death of a resident in care where this was caused by, or contributed to, systemic organisational failure. The hurdles against successful prosecution are considerable. Tracing direct lines of responsibility (for causing death) through opaque managerial structures is difficult, if not impossible. Only time will tell whether there is a realistic prospect that prosecutions of NHS Trusts can be achieved using these provisions, 81 although at present, in our view, this seems unlikely.
The context of candour for advancing patient safety
The statutory duty of candour requires organisations to ‘act in an open and transparent way with relevant persons in relation to care and treatment provided to service users in carrying out a regulated activity’. 82 Although this applies to organisations and Health Service bodies, it is (inevitably) clinicians, while working as representatives of that organisation, who are required to discharge the duty. In addition, doctors are bound by a professional duty of candour to be ‘open and honest with patients if things go wrong…to put matters right (if that is possible)…offer an apology [and] explain fully and promptly what has happened’ as well as the long and short term effects. 83 This duty includes informing patients when medical treatment has caused them harm, empathising and saying sorry, as well as reflecting and learning positively for the future. Although apologies may contribute to the well-being of recipients (even if not sincere), conventional wisdom suggests that these serve no purpose if they lack sincerity. On this basis, van Dijck suggests that mandated apology should be seen as fulfilment of a legal obligation rather than a statement of genuinely held feeling. 84 Proposals that are made to enact rules that try to force compassion are likely to be counterproductive and result in cynicism.
There may be a certain reluctance to disclose openly following death of a patient for fear of reprisal in the form of legal action. Openness and improving healthcare proactively by learning from adverse events will be undermined if doctors are fearful of a culture that looks to apportion blame. Potential prosecution is a very real hazard that will diminish critical reflection following patient safety incidents. It is a profound irony that for an ethos that supports a ‘blame free’ environment that fears such as this can impede frank disclosure and open discussion about mistakes. 85
A potential way towards fostering openness about adverse clinical incidents and encouraging learning has been through appraisal and revalidation. 86 Part of medical appraisal requires honest reflection that can be included within the appraisal portfolio and then discussed in terms of learning achieved for the future. Reflection is a personal and confidential matter between the doctor and the appraiser that is discussed at appraisal and premised on the tenet that genuine reflection can promote deep learning, reduce clinical error and improve patient safety. It is axiomatic to state that openness and honesty in reporting mistakes is integral to the professional duty of a doctor. It is essential to sophisticate a system that is proactive and well informed to enhance patient safety. Anxiety and fear about backlash can stifle any desire to undertake such reporting and can seriously erode the trust between doctors and the public if reflective practice is undermined by its potential to be used in criminal proceedings.
The possibility of educational reflective documents being used within court proceedings caused considerable concern and debate following Bawa-Garba and is a matter not encountered before in the context of GNM. As part of learning from an adverse event, Dr Bawa-Garba had been asked to reflect on the circumstances that had led to the patient’s death and to sign a trainee form setting out what she should have done differently with the benefit of hindsight. 87 The Medical Protection Society confirmed that her e-portfolio had not formed part of the evidence for the court, 88 and her personal reflections are not detailed in the judgment. 89 However, it would seem that elements from those reflections ‘was fed into the trial…[this] was a reflective document from [the] e-portfolio, which she filled in seven days after the incident’. 90 Although these e-portfolio reflections may not have formed part of the evidence directly, it appears that these may have had an indirect effect. It is unclear as to what weight, if any, the jury might have put on this. If, however, educational reflection that flows from the duty of candour might be used to apportion blame, then this could have a detrimental effect on openness following a medical error.
Independent studies from the United States 91 and Australia 92 reveal that fear of litigation and the threat of punishment is a major barrier to open disclosure. It seems unlikely that assurance, legal or otherwise, would serve to redress the loss of confidence that one might feel in terms of openness and frankness being the most appropriate way to reflect and learn. This may impair strategies for improving healthcare following medical errors. Recorded reflections, such as those used in e-portfolios, are not subject to legal privilege and may be requested by the court if considered relevant. The possibility that educational reflections may form part of evidence of culpability has led some to state that unless reflections are completely protected, the only sensible recourse for any doctor would be to avoid direct acknowledgment of one’s own shortcomings, but instead to reflect in the abstract. 93 It has further been suggested that verbal coaching might be more appropriately cautious than written reflection when appraising performance. 94
Studdert and Richardson have argued proactively for a wall between disciplinary action and systems for learning, with few limited exceptions. 95 Countries, such as Australia and Canada, have introduced laws of qualified privilege that act as a limited shield against liability. 96 In Canada, there is a compromise position of disclosing facts to patients but keeping private opinions for professional review. This has been criticised as being difficult to implement and unjustifiable. 97 The University of Michigan Health System has an open disclosure system with an offer for responding swiftly to compensate patients who have sustained injury due to unreasonable medical care. This also aims to support patients and promote learning from untoward events. Early results indicate positive outcomes in terms of reducing new claims, possibly reducing costs and promoting a culture of honesty and shared learning. 98 In England, a Health Investigation Safety Branch 99 has been created to investigate patient safety incidents and disclosures, although the findings will not normally be used in disciplinary or legal proceedings unless required by the court. It remains to be seen how implementation will work in practice and whether it will support or potentially impair the duty of candour. 100
Reconsidering death in the context of clinical negligence
In our view, for the reasons above, GNM is not optimally suited for all circumstances. It is an offence of ‘individualisation’ through its narrow focus upon the microcosm of the actions (or omissions) of an individual, with scant regard paid to the wider contextual factors in which those actions took place. By using a clinical context, we argue that the wider context is relevant because of fundamental interdependencies of cognitive decision-making and necessary reliance on optimal functioning of systems and processes. An adverse outcome may be the result of a ‘domino effect’ of failures wider than that of the individual alone.
We are not alone in thinking that GNM as an offence is not wholly satisfactory. Quick has argued for change through raising the culpability threshold to that of subjective recklessness. 101 Laird, by comparison, calls for change through an authoritative Supreme Court judgment, rather than the piecemeal approach of the Court of Appeal. 102 We believe that it is time to reconsider GNM. 103 Although the focus of this article has been on clinical negligence, many of the issues raised may apply equally to other areas where persons work within environments that require complex decision-making and where decisions of individuals are shaped by contextual factors.
We propose a new statutory offence of ‘death caused by grave professional negligence’ for reconsidering GNM in circumstances such as these. The proposal is based on twin pillars of rethinking the current formulation of the offence alongside wider additional options. Our reconsideration is based on an approach premised upon a model of a ‘just culture’ built upon fair attribution of blame and transparency for improvement. 104 Our proposal incorporates more nuanced assessment of negligence that encompasses proportionate assessment of the failings of the individual practitioner alongside disposal options of ancillary court orders directed towards systems improvement and wider involvement of coronial processes.
A new offence: Death caused by grave professional negligence
We strongly support the view of the Law Commission that the criminal law should be properly concerned with questions of moral culpability. 105 As a general premise, a person should not be held responsible for an unforeseeable consequence of death. Against this is the view that personal liability for causing death is warranted in the public interest and for public protection and confidence. Moral culpability would carry force particularly in circumstances of medical or (other) situations of death caused by professional negligence. In our view, and as a matter of principle, a person acting lawfully and within a complex managed professional environment should be held responsible for (inadvertently) causing death by negligence, only if they fail to advert to an obviously foreseeable risk of death to the victim, and if that failure is a very serious departure from expected practice in the circumstanes. It remains unclear, however, whether such liability should be addressed through an offence badged as manslaughter.
In citing the example of Adomako, Tadros states that the defendant arguably did everything he believed to be appropriate to ensure that the victim did not come to harm and that the main criticism directed against him could only be for his lack of ability as an anaesthetist.
106
In accepting that death caused by such lack of professional ability should attract criminal liability, two points remain troublesome. First, is it appropriate that the threshold for criminalisation is left to the jury? If the test for criminality is intended to reflect personal decisions by the jury about the proper scope of the criminal law…it will not achieve the desired coherence [expected of the criminal law], and will not be subject to the appropriate constraints of a public system of law.
107
Our proposed option would be enactment of a new offence: death caused by grave professional negligence. 108 The offence and its elements need to be defined with a focus on the ingredients of the offence, culpability and relevance to circumstances. The offence would apply in those circumstances where a professional duty of care arises. We envision a two-stage process. In the first, negligence would need to be established in accordance with the existence of a duty of care, breach of that duty, the breach having caused death of the patient, together with the test of foreseeability requiring proof of a serious and obvious risk of death at the time of the breach.
For the second stage, we suggest that the culpability element of ‘grave professional negligence’ required (for the criminal offence) should be defined as a substantial failure by the defendant to take reasonable care in the totality of circumstances of the breach. 109 The conduct of the defendant should be one that reflects a very great departure from the standard of professional practice expected of a reasonable practitioner of equivalence in the defendant’s position, at the material time. Assessment of the seriousness of the negligence should take into account all the circumstances of the negligence and the impact this may have had on the defendant’s decision-making. This would be particularly relevant in institutional settings (including hospitals) where there is ‘organisational responsibility’ for the efficient and smooth running of systems. Doctors (and other professionals) mostly do their best in the circumstances that are available. In reality, however, information systems, real-time data analysis, diagnostic tests and other support systems may be lagging disastrously behind the growth in healthcare needs. Such factors need to be taken into account explicitly.
The offence that we propose differs from GNM in two aspects. First, the fault standard for the offence would be set by legislation. We suggest setting that standard as ‘grave professional negligence’, which is substantial failure by the defendant to take that level of care that ought to have been taken by a reasonable professional of equivalence. Currently, in GNM, the decision for criminalisation of the defendant is left to the jury. In Adomako, Lord Mackay stated that the essence of GNM is whether the conduct of the defendant was so bad ‘as to amount in their judgment to a criminal act or omission’. 110 Therefore, the conduct of the defendant is criminal if it is grossly negligent, and it is grossly negligent if it deserves to be criminal. Notwithstanding the circularity of this statement (which has achieved much eristic academic attention), it is the jury that decides whether or not to criminalise. Although Adomako has been refined and clarified over the years, the position of the jury in respect of deciding on criminalisation has not changed. Even recently, Sir Brian Leveson in his statement in Rose, in defining the elements of GNM said that a key element of the offence is that the defendant’s breach should be so reprehensible that it ‘amounted to gross negligence and required criminal sanction’, 111 a matter that is still left in the hands of the jury. Our proposal shifts this position to one where the fault standard for the criminal offence is set by law. This is the preferred position to maintain the expected consistency of the criminal law.
Second, our proposal gives weight to wider circumstances and context of the case. This requires assessment of the seriousness of the breach in the totality of the circumstances. In Adomako, Lord Mackay explained that making a decision about gross negligence and whether this amounts to a crime, ‘will depend on the seriousness of the breach of duty committed by the defendant in all the circumstances in which the defendant was placed when it occurred’. 112 However, this has generally been taken to mean merely those circumstances related directly to the facts of the case and does not include those circumstances that may be directly related to the context of the defendant’s decision-making. The totality of circumstances and context are distinctly relevant as they may have led, or impacted upon, those facts that are now scrutinised as potential acts or omissions that are indicative of a very serious degree of negligence.
The approach taken to ‘context’ in medical cases of GNM is narrowly focused on evidence of facts, and not on evidence of circumstances leading to those facts: The problem is one which arises acutely in medical cases where the evidence of the ‘context’ will often be disputed and be based on what the deceased said. For example, in Rudling, the issue may turn on whether the mother described the boy’s genitals as black (would have triggered a realisation in a reasonable doctor of a risk of death) or discoloured (not immediately obvious that that suggests life-threatening illness).
113
Ancillary court orders and coronial processes
Following conviction, current disposal options for GNM serve only to punish the offender. 116 While punishment may be rightly seen as just retribution, from a consequentialist perspective, it achieves little in terms of potential benefits such as ‘closure’ for the bereaved, systems improvements or enhancements to patient safety. Further justification for punishment is based on the premise that this may deter. 117 Exactly how this may act as a deterrent in cases of error, or negligence, is not entirely clear because in these circumstances the action of the defendant is not wilful. The defendant’s conduct was either ‘innocent’ (a mistake) or inadvertent (careless). Modification of future behaviour would seemingly be better addressed by learning and remediation in such situations.
Even if one were to accept the approach of punishment as ‘caution for the future’, this does not appear to translate into practice. In commenting on the case of doctors Prentice and Sullman, 118 (who made a terrible mistake by injecting vincristine into a child’s spine, eventually causing death), Dyer states that bringing the full weight of the criminal law on two young doctors does ‘little to remedy a system which lets juniors loose on patients with too little training, too little support and too little sleep’. 119 It also fails lamentably in strengthening wider accountability or patient benefit. In our proposal, we suggest that disposal options should be more creative to include ancillary court orders addressed at failings in systems and wider use of coronial processes.
Ancillary orders made by the court may have potential as a powerful mechanism for strengthening accountability of hospitals following death due to ‘gross negligence’, as well as for developing more robust mechanisms for future learning and prevention of death caused by negligence or systems failures. Where an incident has resulted in a doctor being charged with GNM the systems within the department and the environment within which the event took place should be subjected to independent scrutiny. According to the Hamilton Review, this is a particularly important consideration for trainees to ensure that the environment for work is safe and supportive. 120 In fact, external scrutiny does not usually happen and it is left to the organisation to conduct an internal investigation. Following Bawa-Garba, 121 for example, the hospital undertook its own investigation which revealed a catalogue of failures. 122
Perhaps, some lessons can be learnt from the processes of coronial investigations and inquests. 123 Article 2 of the European Convention on Human Rights plays a growing role and obliges states to have robust mechanisms for investigating deaths where the state may have had some responsibility. This is well recognised, for example, in cases of death during police custody or mental healthcare settings. In Middleton, the House of Lords ruled that some coronial inquests can take a wider approach to explore how death came about to include ‘by what means and (emphasis added) in what circumstances’ the death occurred. 124 The so-called Article 2 inquest has a wider remit and is at the discretion of the Coroner. As public authorities and emanations of the state, NHS organisations must comply with the Human Rights Act 1998 and the European Convention on Human Rights. Article 2 imposes procedural and substantive obligations on the state to carry out investigations in circumstances where the NHS owed a duty to take reasonable steps to protect patients under NHS care where it knew (or ought to have known) there was a real and immediate risk to the person’s life. Investigations carried out for these purposes are intended to identify circumstances, correct mistakes, identify good practice and learn lessons for the future to prevent recurrence of similar incidents. To satisfy the procedural obligation, the State must carry out an investigation that is reasonably prompt, effective and independent of those implicated. A coroner’s inquest is the means by which the state ordinarily discharges its procedural obligation. During an inquest, four questions must be answered: who has died, when, where and how they came by their death. Article 2 inquests are more inquisitorial and seek to ascertain in greater detail how a person came about by their death, by what means and in what circumstances. 125 The Coroners and Justice Act 2009 permits coroners to issue Regulation 28 Reports to individuals, organisations, local authorities or government departments and their agencies where the coroner believes that action is necessary to prevent further deaths.
Griffiths and Saunders
126
suggest that there is greater scope for Article 2 inquests. Although we agree in principle, in practice this might offer only limited potential. In Parkinson
127
an inquest was convened, the Coroner determined that it was not an Article 2 inquest, and found that Mrs Parkinson had died of natural causes.
128
Following judicial review of the Coroner’s decision regarding the non-engagement of Article 2, the claim was dismissed by the High Court. The court confirmed that where there was adequate provision for securing high professional standards and protection of patient’s lives, errors of judgment by doctors or negligent coordination of care during the treatment of a patient are not sufficient to invoke article 2. The Court stated: Where a state has made provision for securing high professional standards among health professionals and the protection of the lives of patients, matters such as errors of professional judgment or negligent coordination among health professionals in the treatment of a particular patient will not be sufficient to engage article 2.
129
The outcome in Parkinson is unsurprising as it aligns broadly with jurisprudence from the European Court of Human Rights (ECtHR) that has set a high threshold for the engagement of Article 2 in cases of medical negligence. In Fernandes, 132 the ECtHR reaffirmed that in the context of alleged medical negligence, the states’ substantive positive obligations relating to medical treatment are limited to a duty to regulate and put in place an effective regulatory framework compelling hospitals (private or public) to adopt appropriate measures to protect patients’ lives. Even if medical negligence were to be established, the Court would normally find a substantive violation of Article 2 only if the relevant regulatory framework had failed to ensure proper protection of the patient’s life.
In very exceptional circumstances, the responsibility of the state may be engaged under Article 2 in respect of the negligence of healthcare practitioners: (a) where a person’s life is knowingly put at risk by denial of life-saving emergency treatment (not where a patient is considered to have received deficient, incorrect or delayed treatment); (b) where a systematic or structural dysfunction in hospital services results in a person being denied access to lifesaving emergency treatment and the authorities knew or ought to have known about that risk and failed to prevent that risk from materialising and therefore putting the patient’s life at risk.
On the facts, it may not be easy to distinguish between cases of mere medical negligence and those where there is a denial of access to life-saving emergency treatment (particularly as a combination of factors may contribute to death). For a case to fall into the latter category, the following factors must be met: first, the acts and omissions of the health care providers must go beyond mere error or negligence; second, the dysfunction must be objectively and genuinely identifiable as systemic or structural to be attributable to the state authorities; third, there must be a link between the dysfunction complained of and the harm which the patient sustained; and fourth, the dysfunction at issue must have resulted from the failure of the state to meet its obligation to provide a regulatory framework in the broader sense.
Although there is no absolute bar to an Article 2 investigative inquest in cases of very serious clinical negligence, in our view, based on the above rulings, it is likely that only a few cases will cross this threshold. Wider scrutiny of failings that may have contributed to the death, or the circumstances, will undoubtedly be of benefit for taking future preventative action. We propose that this may be achieved through ancillary orders similar to those undertaken in a Coroner’s Court.
The (longstanding) ‘Rule 43 order’ of the 1984 Coroners Rules has been incorporated into the Coroners Act 2009 as ‘a regulation 28 prevention of future death report’. 133 Coroners can therefore impose a duty upon organisations with a view to making necessary improvements to prevent future deaths from similar failings. This could improve patient safety where systemic factors may have contributed to the death by gross negligence. These ‘learning and prevention of future deaths reports’ offer considerable potential for implementing changes to make the clinical environment safer. In an analysis of 30 cases, Quick identified a range of causative factors that include lack of guidance, failure to implement guidelines, communication errors, failure to follow care plan, medication errors, poor medical records, safe levels of staffing, lack of a learning environment, which could include inadequate supervision, and errors leading to a wrong diagnosis. 134 These interesting findings indicate a potentially large source that is yet untapped and which may provide a basis for directing future improvement strategies to prevent death by gross negligence.
We propose that the court adopts a similar approach to that of the coronial system whereby learning for prevention future deaths orders may be issued. If obvious systemic deficiencies are apparent during a trial for GNM, the court may make an ancillary order for the hospital to investigate with a view to implementing learning and improvement. Such an order (with the force of law behind it) could be monitored by an independent healthcare enforcement organisation. 135 In fact, it seems hard to imagine why such an order for learning from unanticipated death may not be issued by the criminal court (when it may already be issued by a coroner’s court). This might go some way towards improving the system and would be preferable to no orders at all, in attempting to address potentially contributory systemic failings in the circumstances of death alleged to being caused by gross negligence in clinical care.
Conclusion
The deplorable outcome of death of a person in the circumstances of undergoing medical care is usually not solely due to the failings, errors or negligence of a single health professional but instead may represent a coalescence of individual and systemic factors. The current ‘crime and punishment’ model of GNM is therefore unsatisfactory. It may result in a disproportionately high level of blame being assigned to an individual.
GNM is essentially an offence of ‘individualisation’ in its focus upon the person with scant regard to the environment where the misconduct occurred. Context is an essential consideration since individual decision-making in complex professional environments is the product of mutuality between interdependent cognition and systems and processes that are institutionally driven. Giving insufficient weight to context may fail to provide opportunities for future learning and systems improvement. A tragic death due to error or carelessness is often the end result of a cascade of failures that include contextual factors and suboptimal institutional processes.
We espouse an alternative approach to death caused by gross negligence during the course of professional duties within complex environments. Our proposal is based on a model of ‘just culture’ for fair attribution of blame and future learning for wider benefit. We suggest a statutory offence of ‘death caused by grave professional negligence’. The standard of proof would require substantial failure by the defendant to exercise reasonable care expected of a professional of equivalence, when considered in the totality of the circumstances of the breach. Our new offence differs from GNM in two substantive ways: first, the standard for fault is to be set by law rather than being a matter that is left to the jury; second, the totality of the circumstances includes all relevant matters including the context of the work environment at that particular time. We suggest more creative disposal options through ancillary orders by the court for addressing systemic failures and proactive involvement of coronial processes. We hope that our discourse may stimulate further debate and help to refine the law in this troubled area of death caused by gross negligence. ‘Crimine ab uno disce omnis’. 136
Footnotes
Acknowledgement
The authors are grateful for the comments of Dr James Slater and two anonymous reviewers on an earlier draft of this article.
Declaration of conflicting interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The author(s) received no financial support for the research, authorship, and/or publication of this article.
