A compassion-Based Restorative Approach in Healthcare GNM Cases Involving Systemic Failure

Abstract

Prosecutions for healthcare gross negligence manslaughter (GNM) raise a persistent tension. A strict focus on individual culpability may obscure systemic failures that shape clinical work, while an exclusive focus on systemic failure may appear to weaken personal accountability. Restorative justice has been proposed as a complementary mechanism in England, yet the existing proposals emphasise process design and risk management without offering clear normative criteria for distinguishing individual fault from systemic contribution. This article develops a compassion-based restorative approach for healthcare GNM cases involving systemic failure. Drawing on Martha Nussbaum’s account of compassion as structured by three judgements (seriousness, nondesert, and the eudaimonistic judgement), it treats compassion as a normative discipline for a restorative approach that runs alongside England’s criminal process. The approach does not revisit the legal causation and culpability assessments that underpin liability under the Adomako framework. Instead, it focuses on normative criteria and shows how compassion can constrain what may be forgiven or negotiated, and how it can guide the allocation of accountability and remedial duties between individuals and healthcare systems. The article specifically applies the judgement of nondesert through two practical questions: whether the clinician has continued this behaviour as a matter of habit or choice, despite being educated, counselled, and informed by a designed system, and whether the clinician is a victim of the poorly managed healthcare system or poor working conditions. In this way, the framework resists demonisation, avoids impunity, and redirects attention towards prevention through actionable institutional responsibility.

Introduction

The prosecution of gross negligence manslaughter (GNM) cases involving healthcare professionals with profound evidence of systemic failure has drawn criticism for their punitive focus on individuals, often overlooking systemic failures that contribute to fatal errors. This approach risks undermining fairness and fails to address the root causes of harm. Restorative justice, which prioritises repairing relationships and addressing harm holistically, offers a complementary response. ¹ However, critics warn that restorative processes may exploit victims’ goodwill or diminish accountability, ² and that highlighting systemic failure may hide or excuse personal fault. ³

To better enable restorative justice to operate in healthcare GNM cases involving systemic failure, this article proposes a compassion-based restorative approach inspired by Martha Nussbaum’s theory of compassion that systemises the normative role of compassion in restorative practice, as a complement to the current criminal response.

Compassion has long been discussed in philosophy and ethics, yet its role in law remains contested. For example, apart from Aristotle, whose work will be discussed later in detail, David Hume argued that our sentiment of virtue depends on extensive compassion. ⁴ Adam Smith held that compassion universally exists in everyone such that even ‘the greatest ruffian, the most hardened violator of the laws of society, is not altogether without it’. ⁵ Jean-Jacques Rousseau also emphasised compassion’s positive role in public affairs. ⁶ However, in the legal field, compassion’s relevance has been questioned. ⁷ For example, Dan Markel argues that ‘compassion to wrongdoers has a home, but it is not in the courthouse or at clemency hearings’, ⁸ since he contends that a legal basis of compassion is a legal basis of bias. ⁹

Although compassion has been viewed as subjective and incompatible with legal reasoning, scholars such as Martha Nussbaum argue that compassion, when grounded in cognitive judgements, can enhance justice. ¹⁰ Nussbaum identifies three key judgements – seriousness, nondesert, and eudaimonistic – that structure compassion and tether it to rational evaluation. ¹¹ The judgement of seriousness is a belief or appraisal that the suffering is serious rather than trivial. ¹² The judgement of nondesert is a judgement that this person did not bring the suffering on himself or herself. ¹³ Eudaimonistic judgement refers to the belief that this person, or creature, is a significant element within one’s scheme of goals and projects, an end whose good ought to be promoted. ¹⁴ This article explores how these judgements can inform legal decision-making.

This article argues that integrating Nussbaum’s three judgements into restorative justice frameworks can mitigate the concerns about restorative justice mentioned above. The eudaimonistic judgement confirms the value of a compassion-based restorative approach, requiring concern for all people, seeing their humanity, resisting demonising defendants, and revealing systemic failures in cases. The judgement of seriousness functions as a reality check, presenting the dual attributes of victim and offender of the clinicians in healthcare GNM cases involving systemic failure. Finally, the judgement of nondesert sets the boundary of compassionate responses, providing a normative basis for differentiating responses to personal fault and to systemic failure, thereby complementing current legal practice. This restorative approach may operate before and after a criminal trial, but it must be kept procedurally separate from the determination of accountability under Adomako. Its focus is on dialogue, acknowledgement, and commitments made by the system. What it offers is a set of compassion-based criteria to guide how a restorative process should respond to both individual faults and systemic failure alongside the criminal process.

This article makes two contributions. First, it develops a compassion-based restorative approach as a normative framework for a restorative track and shows how Nussbaum’s three judgements set clear limits and tasks: seriousness sets a minimum standard for recognising harm, nondesert guides how burdens are assigned, and eudaimonistic judgement helps prevent dehumanisation. Second, it puts the judgement of nondesert into practice by setting out a clear way to separate individual fault from systemic contribution and to assign accountability and remedial duties between clinicians and healthcare organisations.

This article is structured as follows: it first explores the background, covering healthcare GNM cases involving systemic failure, restorative justice, and compassion. Then, it discusses what compassion is in the field of law, laying a theoretical foundation for the further development of the compassion-based restorative approach. The last three sections before the conclusion carefully examine the three judgements in the context of healthcare GNM cases involving systemic failure to demonstrate how a compassion-based restorative approach works.

Background

The law on GNM has been heavily criticised, particularly when it involves healthcare practices. In such cases, medical professionals will face a criminal charge if they make a serious mistake amounting to gross negligence that leads to a patient’s death. In England and Wales, the leading authority on GNM is R v Adomako; ¹⁵ the Adomako test was later confirmed in R v Misra. ¹⁶ The cases of Adomako and Misra form the basis of the five-part test for GNM:

1. Did D owe V a duty of care?

2. Did D breach that duty of care?

3. Did the breach cause V’s death?

4. Was there a serious and obvious risk of death?

5. Was the breach grossly negligent? ¹⁷

The test is largely objective, rather than subjective, which means that liability does not require proof, as an explicit element, that the defendant was subjectively aware of the risk in order for a conviction to follow. The test also requires the result of death. The combination of these two points presents a disadvantage to a specific profession, namely medical professionals, in the criminal justice system. In an empirical study on the prosecution of GNM involving medical professionals, Oliver Quick found a sense of discomfort among prosecutors regarding the severity of subjecting individuals, whose errors have catastrophic consequences, to criminal litigation. ¹⁸ This unease is partly because the offence is occupationally targeted, and individuals working in the medical field are more likely to face criminal penalties because of their profession. Some prosecutors believe that it would be inappropriate for individuals to face prosecution solely on the basis of their occupation, particularly when the nature of their work inherently carries a higher risk of fatal outcomes due to potential errors. ¹⁹

Moreover, because the test for GNM does not require defendants to be subjectively aware of the risk, the absence of a mens rea requirement lowers the threshold for medical professionals to meet the elements of the offence. This heightens concerns about a punitive approach in legal practice when the criminal justice system deals with GNM cases in healthcare, especially the cases with clear evidence of systemic failure.

Scholars such as Anne-Maree Farrell, Amel Alghrani, and Melinee Kazarian find that England and Wales stands out for its escalating severity in penalising incidents where healthcare professionals’ errors result in fatalities compared with other common law jurisdictions. ²⁰ Farrell et al. argue that the punitive approach is not helpful in addressing the problem or preventing medical accidents. This is because, first, there is very limited empirical evidence indicating that the application of criminal law serves as an effective deterrent to such behaviour, especially within the context of intricate and progressively strained healthcare settings. ²¹

More importantly, the causes of the fatal errors are not solely attributable to individuals’ mistake or negligence. ²² They are the results of systemic failure. ²³ Amel Alghrani et al. argue that failures in healthcare delivery are rarely the result of individual errors. Instead, they stem from systemic issues, including organisational factors such as trust targets, pressure to deliver results, unsafe supervision, and inadequate staffing. ²⁴ These systemic failures create conditions for unsafe practices, such as clinicians working while fatigued or while juggling multiple roles simultaneously. ²⁵ The case of R v Bawa-Garba ²⁶ provides an example.

Dr Hadiza Bawa-Garba, a junior paediatrician, was convicted of GNM following the death of a 6-year-old patient, despite her defence emphasising systemic failures within the healthcare setting. ²⁷ These failures included a malfunctioning IT system that delayed critical test results, severe understaffing of consultants, doctors and nurses, and operational breakdowns in nursing care. The criminal court, however, focused narrowly on her individual omissions, applying the Adomako ²⁸ test.

In contrast, the subsequent professional regulatory proceedings acknowledged these systemic dimensions. ²⁹ The Medical Practitioners Tribunal (MPT), after reviewing the same facts, concluded that erasure from the medical register would be disproportionate and instead imposed a suspension. ³⁰ The Court of Appeal (Civil Division) upheld this decision, explicitly recognising the contextual pressures and systemic deficiencies that contributed to the tragic outcome. ³¹

This divergence between the criminal and regulatory responses exposes a tension at the heart of the current legal response to healthcare GNM for which systemic failure is an element. The criminal justice response to such cases is punitive and targets individuals, ³² marginalising systemic issues. ³³ While the criminal trial treated Bawa-Garba as an isolated wrongdoer, the appellate judgement in the regulatory sphere implicitly accepted that culpability cannot be assessed in a vacuum when systemic failings are deeply implicated. This inconsistency underscores the inadequacy of a purely retributive model in cases of healthcare GNM involving systemic failure.

Scholars argue that reform to the legal response to GNM is necessary to reflect the systemic nature of such grave errors and to prevent them from happening again. Thus, Farrell et al. promote restorative justice as a complement to the current retributive model. ³⁴ Howard Zehr first mentioned restorative justice in his book, Changing lenses: a new focus for crime and justice. ³⁵ This approach asserts that crime is a violation of people and relationships, and justice should focus on rebuilding the relationship, involving the victim, the offender, and the community. ³⁶

Farrell et al. argue that restorative justice in healthcare GNM cases involving systemic failure will be helpful in offering a much more holistic approach in the sense of healing and righting wrongs. ³⁷ The restorative approach in these cases should have two parts: redress to victims (bereaved families) and the rehabilitation of offenders (healthcare professionals). ³⁸ Redress includes payment of financial compensation, explanations, apologies, professional accountability, systems learning to prevent harm from happening again, and an acceptance of institutional accountability. ³⁹ Rehabilitation includes preparedness to admit error on the part of healthcare professionals and a statement of regret/apology to the bereaved family, determination of fitness to practice, and review of systemic failure and poor performance that led to the fatal errors. ⁴⁰

However, there are numerous critiques of restorative justice. Scholars, such as Annalise E. Acorn, emphasise that restorative justice may be exploited by offenders. In her critique of restorative justice in criminal law, Acorn describes it as a procedure that seeks to elicit compassion for the victim from the offender and compassion from the victim for the offender. ⁴¹ In reply, Sam Garkawe has argued that restorative justice is beneficial to both victims and offenders because it can achieve a balanced restoration of the interests of both parties by allowing victims to participate directly in decision-making and obtain compensation and psychological healing, while encouraging offenders to take responsibility and promote their social reintegration, with a framework that ensures procedural fairness and provide necessary support for all participants. ⁴² However, this response does not answer Acorn’s concerns. This is because even Garkawe admits that the effectiveness of restorative justice is limited by the risks of power imbalances, inadequate mediator training, and cultural applicability issues. ⁴³ Acorn believes that our own vicarious discomfort will lead us to forgive offenders when they perform compassion and remorse, taking advantage of the good nature of the victim and community. ⁴⁴ Furthermore, Garkawe’s concerns about power imbalances can also be seen in other studies. These studies explore the limitations of restorative justice in addressing racism, ⁴⁵ gendered power imbalances, ⁴⁶ juvenile delinquency, ⁴⁷ and the legacies of colonialism. ⁴⁸ This is because restorative justice emphasises informality and operates outside formal procedures, relying on personal communication. In this informal and personal form, which focuses on individual cases, it is difficult for restorative justice to address the systemic issues underlying the offences and matters mentioned above.

In discussions of restorative justice in healthcare GNM involving systemic failure, scholars have expressed their concerns that the emphasis on systemic failure may obscure personal responsibility that should be addressed. In an article on the aftermath of COVID-19, Roberto Catello notes that systemic factors are important, but they should not automatically absolve individuals of their duty of care and decision-making responsibilities. ⁴⁹ Robert M. Wachter and Peter J. Pronovost argue that systemic explanations are important tools, but when certain safety practices have matured and systemic obstacles have been largely eliminated, continuing to use the ‘systemic failure’ narrative may become a ‘cover-up’, allowing repeated or deliberate wrongdoers to evade personal responsibility. ⁵⁰

Those concerns focus on how to determine the boundary between individual responsibility and systemic failure. However, Stephen Buetow, citing John Kekes, argues that without such a clear distinction, forgiveness towards offenders – a core component of restorative justice – will be difficult to implement because forgiveness and blame are incompatible:

The two responses to wrongdoing cannot coexist since blame sustains the moral weight of harm, while forgiveness releases it. Blame maintains moral disapproval that forgiveness inherently negates. ⁵¹

Farrell et al. reply to the concerns mentioned above. First, their approach is a complement to current legal practice. On this view, the restorative process is not designed to determine criminal responsibility, including the legal causation and culpability assessments that underpin GNM liability under the Adomako framework. Those questions are still to be determined by the formal criminal process. What the restorative process provides is greater scope to consider how systemic failures contribute to substandard care and patient deaths, while also responding to the concerns of clinicians, institutions, and families in cases that give rise to GNM investigations or prosecutions under English law.

What Farrell et al. emphasise is how to use system design to prevent the abuse of compassion and power imbalances between victims and the healthcare system. They note that the implementation of restorative approaches must be based on the consent and willingness of the parties involved. ⁵² They also specifically envisage the restorative process in a criminal context in a manner that addresses the risks that ‘informality, power asymmetry, and case context’ undermine the protection of rights and respect for the severity of harm. ⁵³ To this end, Farrell et al. suggest establishing pilot projects first and using case referral eligibility assessments and a monitoring panel composed of representatives from professional regulators, prosecutors, the criminal justice system, and the NHS to ensure oversight. ⁵⁴ They also suggest continuously refining relevant agreements and referral standards through ongoing evaluation. ⁵⁵

Farrell et al. offer a procedural risk control scheme centred on pilot programmes and oversight panels to address issues such as the informality and power asymmetries in restorative justice in the criminal context. However, they do not elaborate further on the key normative issues behind these risks such as under what circumstances a demand for ‘compassion’ would become a denial individual responsibility, and what normative framework should be used to determine which oversight and procedural safeguards are sufficient and do not allow restorative justice to rely on untested goodwill. Therefore, this article introduces Nussbaum’s structured three-judgement framework of compassion to provide a set of normative constraints for restorative justice, thereby responding to criticisms that restorative justice may ‘take advantage of the good nature of the victims or communities’. ⁵⁶

Overall, this section explored why healthcare GNM is contested in systemic failure cases and why a purely retributive response can leave the system invisible. Farrell et al.’s restorative proposal offers a parallel, process-focused track with safeguards to manage risks such as informality and power imbalances. This article builds on that proposal and argues that a restorative approach, including a compassion-based restorative approach, should adopt a parallel and process-focused track. Detailed questions of implementation, such as timing, governance, oversight, and the precise relationship to sentencing, fall outside the scope of this article. However, the compassion-based restorative approach proposed uses Nussbaum’s three judgements of compassion to clarify how to separate personal fault from systemic contribution and how to allocate remedial duties in a way that is fair, accountable, and prevention-focused. It, therefore, focuses on normative criteria and does not displace the Adomako test.

What Is Compassion in the Field of Law?

Compassion is an issue of long-standing concern in ethics and philosophy. Aristotle defined a compassion-like emotion: eleos. He described it as ‘a feeling of pain at an apparent evil, destructive or painful, which befalls one who does not deserve it, and which we might expect to befall ourselves or some friend of ours, and moreover to befall us soon’. ⁵⁷ In modern times, some scholars translate eleos into English as ‘pity’. ⁵⁸ However, Nussbaum argues that it should be translated as ‘compassion’ because Aristotle’s definition of it is closely related to modern usage of ‘compassion’, whereas the word ‘pity’, in a modern context, has developed nuances of condescension and superiority towards the suffering. ⁵⁹

Aristotle’s understanding of compassion is similar to the cognitive judgement approach in the current fields of law-and-emotion and the philosophy of emotion. Terry A. Maroney notes that some scholars study emotion through an approach that is cognitive appraisal and deeply Aristotelian. ⁶⁰ Maroney explains the approach as a set of perceptions and evaluations, distinctive sets of which underlie any given emotion. ⁶¹ For example, Robert Solomon holds that emotions are constituted by normative and moral judgements. ⁶² Solomon argues that emotions follow a logical form, which he expresses as ‘A be xxx (the emotion) about yyy’. ⁶³ For instance, ‘I feel happy about watching the movie’ is a complete expression of an emotion. On the other hand, merely ‘I feel happy’ is not complete because, without the object phrase ‘watching the movie’, it is uncertain and unverifiable whether ‘feeling happy’ is true or not. ⁶⁴ The original expression involves at least two judgements: I have watched the movie, and the movie is greatly beneficial to me. Without either statement, ‘I’ would not feel happy.

Furthermore, in Hiding from Humanity, ⁶⁵ Nussbaum argues that emotions are bound up with thought. It is hard to describe an emotion without thought about the object of that emotion. When we feel angry, we mean that we are angry about something. When we are afraid, we are afraid of something. In her own words, ‘If I attack a person who has just raped my child, my anger, again, is not just a mindless impulse. It involves thoughts about the terrible damage my child has just suffered and the wrongfulness of the offender’s act’. ⁶⁶

Therefore, scholars such as Nussbaum hold that emotion is part of people’s cognition, and, as in the example of anger, emotion is full of cognitive judgements about specific objects. Most importantly, this realisation means that if a decision-maker has compassion towards others, it indicates that the decision-maker has made or is making cognitive judgements. Based on this, scholars have put forward several theories of the cognitive judgements of compassion regarding which kinds of cognitive judgements constitute compassion, or, in other words, which cognitive judgements decision-makers should employ to have appropriate compassion.

Maksymilian Del Mar sees compassion as a ‘cognitively evaluative relational feeling’. ⁶⁷ He understands compassion as a cognitive and evaluative process with different layers and stages. Del Mar describes five stages of the process of having compassion, including awareness of another person and the interest or concern for that other person, imagining the other’s situation and perspective, feeling the pain that one imagines the other person to be suffering, returning to imagining the situation of the other and their perspective on it after feeling the pain, and finally, feeling sadness for their suffering. ⁶⁸ As for evaluation, he employs a thin perspective: ‘compassion is necessarily thinly evaluative, meaning that the other person is salient to us and that we have some concern for them’. ⁶⁹ Hooria Jazaieri et al. hold a similar idea about the cognitive structure of compassion to Del Mar’s cognitive structures, except that Jazaieri et al.’s cognitive structure includes the element of a wish to see the relief of that suffering. ⁷⁰

Nussbaum gives a different answer to the question of the evaluative feature of compassion. Nussbaum’s account, as mentioned before, holds that compassion’s three judgements – the judgement of seriousness, the judgement of nondesert, and the eudaimonistic judgement – embody significant evaluation of the event and constitute assessments of values. In the judgement of seriousness, one decides which losses are serious and what is essential to human flourishing. To provide a solid foundation for assessing loss, she develops the Central Human Capabilities List. ⁷¹ The judgement of nondesert involves assessing the sufferer’s actions and determining the proper response to the sufferer. The eudaimonistic judgement is similar to Del Mar’s first process of compassion, which concerns the awareness of another person and interests or concern for that person.

To be clear, this article does not reduce compassion from an emotion to a pure judgement; instead, it adopts Nussbaum’s approach: compassion is an emotion with a feeling dimension, but for legal operability, it is structured into reviewable judgements. This article explores those judgements through healthcare GNM cases involving systemic failure in the later sections. The first one discussed in the next section is the eudaimonistic judgement and the value of compassion in legal affairs, specifically resisting demonisation while maintaining accountability.

Eudaimonistic Judgement and Why Compassion?

As mentioned before, compassion’s positive role has been explored by Enlightenment scholars, such as Rousseau’s pity in Émile, ⁷² Hume, ⁷³ and Smith’s sympathy. ⁷⁴ Based on the historical and philosophical discussion regarding compassion’s role in law, contemporary theorists have offered important insights. First, the early literature on law and compassion originated from feminist legal scholars and judges in the United States in the late 1980s, seeking to challenge the orthodox view that compassion was an emotion negatively associated with women, contrasted with dispassion, which was considered within men’s sphere. ⁷⁵ In 1996, Nussbaum suggested that compassion is the basic social emotion, as it contains a powerful, if partial, vision of just distribution and provides imperfect citizens with an essential bridge from self-interest to just conduct. ⁷⁶ Later, inspired by Smith, Nussbaum provided an account of how to involve such an emotion in judicial thought and writing: we want emotion that is tethered to the evidence, that has been carefully filtered for bias, and that is informed by legal reasoning of the best sort. ⁷⁷ In the most recent literature, scholars such as Jonathan Herring, ⁷⁸ Del Mar, ⁷⁹ and Lorana Bartels and Anthony Hopkins ⁸⁰ have provided their respective approaches of introducing compassion in law.

Herring, ⁸¹ Del Mar, ⁸² and Bartels and Hopkins ⁸³ hold that introducing compassion into the legal decision-making process will greatly improve the fairness of legal practice. Herring believes that ‘the idea of being cared for by someone compelled to be compassionate under fear of legal sanction sounds horrific’. ⁸⁴ However, he argues that a society with compassionate caring relationships is desirable; to achieve this goal, he believes that, with appropriate modification, liberalism’s concepts of rights and best interests can be used to promote the exercise of compassionate relational care. ⁸⁵ Del Mar proposes using compassion as a motivator to prompt judges to exercise appropriate imagination in delivering justice. ⁸⁶ Bartels and Hopkins propose using compassion to reform the Australian criminal law system, and they argue that compassion can be the foundation for cultivating understanding towards offenders, and that it is possible to be compassionate towards both victims and offenders. ⁸⁷ Furthermore, in GNM cases involving healthcare professionals, Andrew Graeme Rowland and Keerthi Mohan promote compassionate leadership and just cultures when investigating and addressing incidents of gross negligence in healthcare. ⁸⁸

Compassion focuses on the inevitable relationships between individuals and society or relationships among individuals. This emphasis on relationships is in line with theories, such as communitarianism, ⁸⁹ care ethics, ⁹⁰ and vulnerability theory. ⁹¹ It is also in line with the restorative approach to rebuilding trust and relationships. Its value is best represented by the eudaimonistic judgement.

Eudaimonistic judgement means that this person, or creature, is a significant element in one’s scheme of goals and projects, an end whose good is to be promoted. ⁹² The eudaimonistic judgement provides the basis for the value of healing in restorative justice. The first value of this judgement is that it requires caring for all the people in a given society, which is different from Aristotle’s account and from the common understanding of ‘concern circles’.

Aristotle defined pity or compassion as ‘a feeling of pain at an apparent evil, destructive or painful, which befalls one who does not deserve it, and which we might expect to befall ourselves or some friend of ours, and moreover to befall us soon’. ⁹³ According to this definition, there are three requirements for compassion: first, the evil must be destructive or painful, not something meaningless; second, the sufferer does not deserve it; third, observers might expect similar suffering to befall themselves or their families and friends.

In the last requirement, Aristotle states that individuals feel compassion towards another person because they or their loved ones could be in the sufferer’s position. They see the similarity between the sufferer and themselves or their loved ones. This judgement is that of similar possibility: the person suffering is similar to my loved ones or me, and because of the similarity, the suffering may befall me or my family and friends; therefore, I want to help.

Aristotle has modern allies in care ethics. Nel Noddings expresses a similar idea in her book Caring. ⁹⁴ She states that we are in the centre of concentric circles of care. The innermost circle is our loved ones, while the outer circle is those for whom we have personal regard, and so on until the circle of strangers. She believes that the core feature of a care relationship is that the one-caring and the cared-for (two terms she uses to describe people who provide care and who receive care) are reciprocally dependent. Therefore, we cannot form a reciprocal relationship with a distant person. Thus, our obligation of care towards distant others is less strong than the one we have in the inner circle. ⁹⁵ After being criticised, she revised her position in her book Starting at Home and applied the principles of justice to guide the action we would take for distant others. However, the hierarchy of care implicitly exists, as the domestic arena is the most significant for being the originator and nurturer of justice. Without the ‘best homes/families’, it would be hard to develop the best care policies for distant others. ⁹⁶

The circles of concern are problematic, even though they might be natural. We tend to focus more intensely on people who are similar to us, but this can open the door to racism, sexism, and other discriminatory or exclusionary ideologies. We may find that some people are more similar to us than others. These are people who look like us, share the same religion, have the same nationality or gender, live in the same society, and so on. To address this problem, Nussbaum modifies the judgement of similar possibility into the eudaimonistic judgement to transcend the circles.

As noted before, eudaimonistic judgement means that this person, or creature, is a significant element in one’s scheme of goals and projects, an end whose good is to be promoted. This judgement concerns all people, even distant others, and recognises them as an important part of one’s own plans. Thus, in light of this definition and Nussbaum’s emphasis, Aristotle’s judgement of similar possibility is a significant but not necessary part of forming the eudaimonistic judgement to bridge the gap between individual’s goals and the eudaimonistic judgement. ⁹⁷

The eudaimonistic judgement says that we should show concern for everyone, including people who may have committed offences. This does not mean excusing wrongdoing or allowing people to avoid responsibility. It means recognising the humanity of everyone involved, so that we can pursue accountability without demonising anyone or reducing a person to their worst act. ‘Seeing humanity in everyone’ is the second and most important value of compassion and of its eudaimonistic judgement.

For example, Bartels and Hopkins argue that compassion can help build relationships. ⁹⁸ This compassion does not mean condoning criminal behaviour. It means holding a person responsible in a way that still recognises their humanity. ⁹⁹ They cite the case of Adrian Bayley, who raped and murdered a female journalist while on parole for multiple sexual offences. They argue that if we treat Bayley as a monster rather than as a human being, ‘we will fail to see the prevalence of male violence against women and the way in which society shapes, enables, permits and condones that violence.’ ¹⁰⁰ If society can respond with compassion, it may be easier to identify and address the deeper causes of Bayley’s offending at social and structural levels, ¹⁰¹ which may include the role of toxic masculinity in sexual violence. ¹⁰²

Bartels and Hopkins also discuss the decision in Cardinal George Pell’s case, in which Pell was accused of sexually assaulting two 13-year-old choirboys in 1996 when he was Archbishop of Melbourne. They note that the judge condemned the offences while also acknowledging other important factors, including the wider context of child sexual abuse in the institution, the victims’ trauma, and their need for justice. ¹⁰³ The judge also took account of factors relevant to Pell as a member of the community, including the likely hardship of imprisonment in his late seventies. ¹⁰⁴ Bartels and Hopkins regard this as a just decision because the judge avoided demonising the defendant and showed compassion towards Pell, the victims, and the community.

These two examples show that having compassion for everyone does not mean commit compassion to impunity. Concern for everyone means seeing the humanity in everyone, understanding how individuals and their behaviours are shaped by given contexts and the underlying systemic issues that the offence represents, to prevent it from happening again. This is where restorative justice is especially useful. A restorative process does not ask us to excuse the individual or to minimise the wrong. It asks us not to treat the individual as a ‘monster’, stopping the analysis there. Instead, a restorative approach uses a structured dialogue to make clear how the harm happened, what pressures and organisational gaps shaped the conduct, and what risks remain if the system does not change.

In healthcare GNM cases involving systemic failure, this shift matters. If the case is framed only as the moral failure of one clinician, the system that enabled the death may remain invisible. By contrast, a compassion-based restorative approach holds the person to account while also directing attention to the wider conditions, such as staffing shortages, unsafe workloads, poor supervision, weak escalation routes, and blame-oriented cultures, which often lie behind grossly negligent outcomes. In this way, restorative justice helps replace ‘monster-making’ narratives with a clearer focus on systemic responsibility and prevention. For example, in Bawa-Garba, a compassion-based restorative approach would require us not to deny Dr Bawa-Garba’s humanity but to inquire into how the fatal error occurred and which systemic issues should be highlighted and addressed to prevent similar harm from recurring.

In addition, the eudaimonistic judgement requires concern for everyone, including victims and bereaved families, providing a normative basis for a restorative approach. It can create a structured space where the family can ask questions and articulate the impact of the death, and where the clinician and relevant institutional actors can respond with acknowledgement and concrete commitments. In a healthcare GNM case involving systemic failure, such as that of Bawa-Garba, restorative outcomes should be designed to make the system visible: not only apologies or explanations but also agreed actions aimed at reducing future risk (e.g. changes to escalation protocols, supervision arrangements, staffing safeguards, reporting processes, and learning mechanisms). The eudaimonistic judgement supplies the normative reason for this design: if we take everyone’s faring well and flourishing seriously, then justice must include steps that reduce the conditions that give rise to similar harm.

However, the value of compassion within a restorative approach in this context lies not only in the eudaimonistic judgement’s requirements to resist the demonisation of individuals that can hide systemic issues but also in the fact that healthcare professionals are indeed harmed by the poor working conditions and problematic management, which make these professionals rightful objects of compassion and create space for compassion to play an appropriate role in such restorative approaches. Recognising them as objects of compassion does not negate accountability for the death but helps situate responsibility within the wider systemic context. This leads to the next judgement of compassion, the judgement of seriousness.

The Judgement of Seriousness

Nussbaum notes that the judgement of seriousness is the belief or appraisal that suffering is serious rather than trivial. ¹⁰⁵ That is to say, when decision-makers have compassion for others, they make a judgement that other people’s suffering is serious enough for them to reach out and help. In analysing healthcare GNM cases involving systemic failure through the lens of compassion, there are two groups of people who are the objects of compassion, bereaved families (victims) and healthcare professionals. The judgement of seriousness requires decision-makers to consider the seriousness of the suffering experienced by both groups.

To begin with, the judgement of seriousness is self-evident to the bereaved families. Their loved ones have died because of others’ failures. This imposes serious emotional and potentially financial burdens on the bereaved families. The salient question is whether the punitive treatment of a GNM conviction is disproportionately serious for healthcare professionals whose fatal errors amount to gross negligence within a context of systemic failure.

As for the individual healthcare professionals facing criminal charges within a context of systemic failure, as previously discussed before, a growing body of literature and many scholars argue that solely focusing on individual fault without considering systemic failure is unfair in many healthcare GNM cases. ¹⁰⁶ This is not only because this offence does not require subjective awareness of the risk of death but also because such fatal errors in healthcare delivery often are tied to collective institutional processes. Thus, these healthcare professionals often have two identities: offenders in the eyes of the criminal law and victims of a failing system.

The intersection of the offenders’ dual attributes of offender and someone who is also harmed by their social and cultural conditions has been widely documented. ¹⁰⁷ In a study of female offenders, Susan L. Miller finds evidence that female offenders have a clear history of victimhood before they offend. ¹⁰⁸ Miller, citing empirical evidence, finds that female offenders are four to five times more likely to experience physical or sexual violence during childhood, which puts women at greater ‘risk’ of offending. ¹⁰⁹ In more general terms, Bartels and Hopkins argue that it is a false dichotomy between ‘offenders’ and ‘victims’, and that many offenders are products of their own victimisation experiences. ¹¹⁰

Healthcare professionals in GNM cases may also have dual attributes. On the one hand, they directly commit fatal errors. On the other hand, they may be harmed as well. They may suffer from a problematic working environment or face significant work pressure. For example, in Bawa-Garba, Leicester Royal Infirmary Hospital had a staffing shortage, which made staff work excessive hours. ¹¹¹ In more general terms, the Royal College of Physicians (RCP) published a report, ‘Being a Junior Doctor’. ¹¹² In the report, RCP interviewed 498 junior doctors. The report showed that 90.5% of doctors said they had stayed at work beyond their scheduled shift at least once in the past month; additionally, 46.6% reported working at least 2 hours past the end of their shift at least once during the same period. ¹¹³ The heavy workload led 60.6% of the doctors to think the workforce pressures were having a serious or extremely serious impact on patient safety, ¹¹⁴ and nearly all junior doctors in the survey (96%) reported that gaps in nursing rotas were having a negative impact on patient safety. ¹¹⁵ Poor working arrangements make healthcare professionals both exhausted workers and victims of the system. Most importantly, from their perspective, they are working in a compromised emotional and physical state, which will ultimately impact patient safety.

Then, a restorative approach should consider the dual identity of healthcare professionals and recognise the seriousness of their suffering arising from systemic issues. As to how a restorative approach should consider this point and respond to it, the section on the judgement of nondesert will unpack this further as it addresses what treatment individuals deserve.

The Judgement of Nondesert

Before exploring the judgement of nondesert, it is worth noting that the concept of fault is closely related to compassion, especially to the judgement of nondesert. As mentioned before, a compassion-based restorative approach, in terms of eudaimonistic judgement, which treats everyone as a significant element in one’s scheme of goals and projects, does not mean excusing the individual or minimising fault. In this sense, a compassion-based restorative approach requires us to distinguish between fault and what the individual does not deserve, or to put another way, to find the true target of compassion in a given situation. Guiding decision-makers to find the true target of compassion by distinguishing fault from nondesert is the unique value of a compassion-based restorative approach in healthcare GNM involving systemic failure.

One more point that needs to be clear is that as stated before, the Adomako test is largely objective: liability does not require proof, as an explicit element, that the defendant was subjectively aware of the risk as many fault-based offences require. This leads to an ongoing debate in healthcare GNM cases involving systemic failure about whether these criteria of ‘grossness’ should be understood as including a dimension of personal fault (subjective awareness of the risk) or not (an objective departure from the standard of care). ¹¹⁶ This article does not aim to settle the academic dispute. Instead, under both situations, there might be a risk of creating a gap between individual accountability and how offenders should be treated morally. Under the current objective model, the issue is clear and has been discussed above. Under the subjective model with an emphasis on individual fault, it is still necessary to consider whether the fatal individual fault was constructed under high pressure, poor working conditions, and problematic management. The compassion-based approach then serves as a normative limit instead of reopening the debate about liability under Adomako. It guides how responsibility and prevention duties should be allocated between individuals and the system, preserving accountability while avoiding excuse for both individual fault and systemic issues.

Conceptions of the Judgement of Nondesert

The judgement of nondesert is a judgement that this person did not bring the suffering on himself or herself. ¹¹⁷ That is to say, when decision-makers have compassion towards others, they assess whether the consequences were proportionate to the person’s conduct.

This judgement is more controversial than the other two judgements. However, this article argues that the judgement of nondesert is necessary for applying compassion in law, as is particularly evident in the healthcare GNM cases involving systemic failure. It can be a useful theoretical basis for distinguishing fault from nondesert to avoid misusing compassion and goodwill. This is because the judgement of nondesert can help decision-makers understand the true cause of the misfortune. With the help of the judgement of nondesert, decision-makers can reach decisions that fit not only with the purpose of punishment in criminal law but also with the purpose of educating and rehabilitating offenders, so that they can successfully reintegrate into society and reduce re-offending.

To begin with, scholars are divided on whether the judgement of nondesert is necessary in engendering compassion. Nussbaum, like Aristotle, explicitly argues that the judgement of nondesert is necessary for compassion. By citing ancient Greek tragedy and empirical research, Nussbaum insists that people will blame a person rather than feel compassion when that person causes suffering through their own fault; therefore, compassion requires an assessment of responsibility and blame. ¹¹⁸

Del Mar disagrees with Nussbaum and argues that it is possible to have compassion even where the sufferer is thought to deserve their suffering, provided the observer has received proper training. ¹¹⁹ He calls this skill ‘the skill of overcoming strongly evaluative compassionate resistance’. ¹²⁰ In another study, Heather Keating and Jo Bridgeman focus on compassionate killing, a category of criminal case in which a parent or spouse kills, or assists in the death of, a child or spouse with severe disabilities, debilitating injury, or chronic or terminal illness. ¹²¹ Keating and Bridgement claim that a relative can feel compassion for a family member who uses a wheelchair and wishes to end their suffering even though it may be the relative’s own fault for drink-driving or for playing rugby. ¹²²

If the permissive understanding of compassion, proposed by Del Mar and Keating and Bridgeman, namely that compassion can arise without the judgement of nondesert, is adopted, in healthcare GNM cases involving systemic failure, compassion towards defendants may be readily available without a careful normative separation between personal fault and systemic failure. However, in such contexts, this permissive approach risks allowing individual fault to be hidden behind the emphasis on systemic failure, meaning that the danger of exploiting goodwill remains. Nussbaum’s judgement of nondesert does not deny compassion in cases involving fault. Rather, it insists on distinguishing fault from suffering and on evaluating whether the suffering exceeds what is deserved. Therefore, it offers a normative boundary against neglect of fault.

Nussbaum gives an imaginary example of a teenager who is caught drink-driving and is charged with an offence. ¹²³ It is the teenager’s fault, but as an adolescent, they are experiencing an uncomfortable and confusing process, which can make them moody and cause poor judgement. Therefore, the teenager will face punishment, but the punishment will not be as severe as that of an adult in the same position. Thus, the scope of Nussbaum’s judgement of nondesert is broad enough to encompass both innocent situations and situations of wrongdoing or fault.

The judgement of nondesert is necessary for compassion, not only because it encompasses situations involving fault but also because it can guide compassion towards the key aspect of the misfortune. In this sense, the scholars mentioned above, who believe that the judgement of nondesert is not necessary for compassion, misunderstand the object of compassion or the role of the judgement of nondesert in the example in which they claim that compassion exists without the judgement of nondesert.

In Keating and Bridgeman’s case, the relative can feel compassion for a family member who uses a wheelchair and wishes to end their suffering even though it may be the family member’s own fault for drink-driving. There is a question as to what exactly the relative has compassion for. If someone drink-drives and is arrested and justly punished by law, but is not injured, the relative may not feel the same level of compassion because, in Keating and Bridgeman’s argument, the suffering that the relative wishes to end is the fact that their family member uses a wheelchair. That is, the compassion in Keating and Bridgeman’s example is directed towards the disability, not the consequence of wrongdoing. After all, the reasonable direct consequence of wrongdoing is moral or legal condemnation, but no one deserves to be disabled in such a way.

This difference in what people are truly compassionate about is important, as it distinguishes between the fault for which the individual should be held accountable and nondesert, in virtue of which the individual can be treated with compassion. This distinction can help determine the means to alleviate their misery. In the legal decision-making process, this analysis of the object of compassion or of the precise suffering that the person does not deserve can help decision-makers target the root of the suffering and determine the appropriate means within a restorative approach, as well as any punishment.

For example, in the study of female offenders mentioned above, Miller finds that female offenders have a clear history of victimisation before they committed the crimes. ¹²⁴ Miller finds that female offenders are four to five times more likely to experience physical or sexual violence during childhood, which puts women at greater ‘risk’ of offending. ¹²⁵ The judgement of nondesert can have a twofold assessment in this example. First, the female offenders have done something criminally wrong, and they should face the consequences as a result. Second, their misfortune arises from their victimhood in their childhood. Therefore, they deserve the specific help, psychologically or in other professional ways, to help them heal from their childhood trauma. Merely realising them or sentencing them more lightly because of their victimhood will not be helpful to prevent re-offending. The root of their misfortune should be addressed at the social level. Therefore, there should be a systematic measure for protecting children from multiple perspectives, which will lead to more nuanced consideration of distributing public resources to education, infrastructure, the social worker system, and so on.

Furthermore, because the judgement of nondesert can help us identify the real issue behind suffering, it can help us address the problem ‘should we or can we have compassion for offenders while considering victims?’ Bartels and Hopkins believe that we can because it is a false dichotomy between ‘offenders’ and ‘victims’, and many offenders are victims of their own victimisation experiences. ¹²⁶ However, even though theoretically speaking, we can punish offenders for their fault and address the systemic failures that make them suffered, how can we do so especially in healthcare GNM cases involving systemic failure? The next section will explore how the judgement of nondesert can be applied in those cases.

The Judgement of Nondesert and the Healthcare GNM Cases Involving Systemic Failure

As stated before, the reasons scholars support a restorative approach in healthcare GNM cases involving systemic failure are that fatal errors may be conditioned by systemic failure and that healthcare professionals may neither intend the outcome nor be aware of the risk. That is to say, such scholars believe that merely punishing these healthcare professionals is excessively harsh for individuals lacking malice, especially when systemic factors contribute significantly to the error.

Thus, to lift the healthcare professionals out of this nondesert situation with concern for the faring well and flourishing of everyone, including healthcare professionals, victims, and others within the healthcare system, decision-makers should consider systemic failure because, without systemic change, the errors will recur and professionals will be in the same situation again. ‘Decision-makers’, in this regard, are not only the criminal court judges, as a criminal court has the focused mission on determining whether a crime has been committed and what punishment to apply, but also other decision-makers, including facilitators and the independent office of the Health and Disability Commissioner mentioned by Farrell et al., ¹²⁷ as well as decision-makers at the MPT. Bawa-Garba shows that it is important to have a mechanism to determine the fitness of practice of the healthcare professionals in question, to help them return to the work for which they have trained and practised for years. This is exactly what the MPT did in the case of Bawa-Garba, where the MPT recognised these systemic problems and decided to suspend her registration rather than erase her from the medical register. ¹²⁸

Apart from emphasising systemic failure, a compassion-based restorative approach would not hide or deny personal fault. Unlike Farrell and her colleagues’ proposal, which is a process-focused approach, the judgement of nondesert is a normative constraint on the restorative track: it sets limits on what may be forgiven and structures how accountability and remedial duties are allocated between individuals and systems. The key method is to ask two questions: whether the clinician has continued this behaviour as a matter of habit or choice, despite being educated, counselled, and informed by a designed system, and whether the clinician is a victim of the poorly managed healthcare system or poor working conditions.

On an ordinary reading of the first question, it seems to clarify when the breach is more plausibly attributed to individual fault in a way that may support the grossness evaluation under the Adomako test. For example, Wachter and Pronovost propose that if the clinician refuses or repeatedly forgets to perform safety practices that are ‘recommended by most safety authorities and required by accreditors, and are associated with relatively low cost, complexity, and risk’, the fault leans towards personal accountability, such as refusing or constantly forgetting to perform hand hygiene practice. ¹²⁹

Wachter and Pronovost’s account is intuitively consistent with the grossness limb under the Adomako test: where a safety rule is clear, persistent non-compliance is more plausibly attributed to individual fault and may therefore support a finding that the breach was ‘gross’ in context. However, the judgement of nondesert and the compassion-based restorative approach do not revisit legal causation or criminal liability, which will be determined within the criminal justice process. The judgement of nondesert puts the spotlight on which burdens should be borne by the healthcare system, especially where problematic management and structural pressures shape the conditions in which the offence occurs, even if individual fault is established.

Asking the first question is to establish the nature of the individual breach before moving to the systemic context. It enquires whether the clinician’s failure occurred despite a well-designed, functional system that clearly communicates safety expectations. If a clear safety rule was in place but ignored, then, as Wachter and Pronovost’s account suggests, the fault leans towards personal accountability. However, if the system fails to consider and address these safety issues in its design and practice, the foundation for attributing sole personal fault is immediately weakened.

Most importantly, regardless of whether individual fault is established under the first question (and subsequently evaluated under the Adomako test in criminal proceedings), the compassion-based restorative justice approach does not stop at individual accountability. Unlike Wachter and Pronovost’s approach, the judgement of nondesert then asks the second question: the systemic contexts in which the fault was shaped. This is an essential step to identify which burdens the healthcare system should bear and to determine whether the individual clinician, even if at fault, deserves a restorative arrangement rather than purely retributive treatment.

The second question shifts the focus directly to the systemic context. Building on the findings of the first question, the approach does not deny or hide personal fault; rather, it asks whether the personal fault was shaped by an unjust environment. It is crucial to avoid demonising the individual and to prevent organisations from hiding the systemic failures behind individual accountability. Ultimately, it requires decision-makers outside the current criminal courts to address systemic failures and to recognise the shared humanity of this individual and others who are also victims of the system, through specific institutional commitments capable of implementation and oversight.

For example, in the case of Bawa-Garba, the clinician was a victim of the poorly managed healthcare working environment, which, as mentioned before, included a malfunctioning IT system that delayed critical test results, severe understaffing of consultants, doctors, and nurses, and operational breakdowns in nursing care. Applying the first question reveals that the system itself failed to provide safe conditions. While these systemic failures weaken the basis for attributing sole personal fault, they do not entirely remove individual responsibility. It remains the role of the current criminal process and laws to determine whether the breach was ‘gross’.

On the second question, these conditions show why a nondesert analysis is needed in the restorative track. Even when individual fault is found, the case shows burdens that should not fall on the clinician alone and risks that cannot be addressed by punishing one person. This approach resists ‘monster-making’ while keeping accountability and makes systemic responsibility clear and practical. The MPT and the Court of Appeal’s decision to suspend her practice licence, rather than erase her from the medical register, represents an appropriate step. However, a compassion-based restorative approach would not stop there. It would put the spotlight on the systemic failures and require other decision-makers to make systemic changes to protect future healthcare professionals as well as their patients from similar systemic risks. These changes could involve requiring hospital management and, when needed, regulators to recognise these systemic failures and to make clear and verifiable commitments to reform, such as fixing the IT reporting pathway, improving escalation and supervision, and ensuring safe staffing levels.

Conclusion

This article argues for a compassion-based restorative approach to healthcare GNM cases involving systemic failure. It follows Farrell et al.’s restorative approach and proposes that this compassion-based restorative approach should complement current legal practice in English law. It treats compassion as the guiding standard for a restorative approach and holds that Nussbaum’s three judgements of compassion can help such an approach to avoid the abuse of goodwill to excuse individual accountability.

The central contribution lies in how the judgement of nondesert can provide a structured way to allocate burdens and remedial duties between individuals and systems. It operates through two main questions: whether unsafe behaviour reflects an individual choice despite clear rules and adequate support, and whether the clinician act in poorly managed conditions that increase the risk of error. This approach helps avoid merely focusing on, or excusing, individual accountability, and it also helps ensure that systemic failures are identified and addressed. A compassion-based restorative approach therefore offers a principled way to pursue accountability alongside prevention.

More work is needed to identify suitable institutional settings and safeguards for this restorative track, but the normative framework set out here provides a clear starting point for this task.