Discretion without oversight: the general practitioner´s unsupervised power to decide against a forensic autopsy in South Africa

¹Department of Accounting and Law, Mangosuthu University of Technology, Durban, South Africa

^&Corresponding author
Anathi Phela, Department of Accounting and Law, Mangosuthu University of Technology, Durban, South Africa

Commentary    

When a patient dies in a general practitioner' care, South African law places the first, and often the only, decision on whether that death will be independently examined in the hands of the very practitioner who was treating the patient at the time of death. The general practitioner (GP) completes the Notification/Register of death form, records a cause of death, and, in doing so, decides whether the death was 'natural', in which case the body may be released for burial or cremation without further investigation, or 'unnatural', in which case the law requires that it be reported to the South African Police Service and referred for a medico-legal, forensic autopsy [1,2]. South Africa's death investigation system is, in this respect, a hybrid: it borrows features of both the coronial and medical examiner models used elsewhere, but the initial classification decision, on which everything else depends, rests with the certifying doctor rather than with any independent judicial or medical officer [3]. This commentary argues that this arrangement contains a structural weakness that has received too little sustained attention: the same practitioner who may have made a clinical error is generally the one deciding whether that possibility is significant enough to trigger the one process, the forensic autopsy, capable of exposing it. This is not a hypothetical concern. The definition of 'unnatural' in South African law is drafted precisely to capture this scenario, yet the practical mechanism for enforcing it depends almost entirely on the individual practitioner's honesty, insight, and willingness to implicate themselves. A related, already well-documented, version of this problem arises when a doctor's classification is not compromised by self-interest but by pressure from the deceased's relatives, who may ask a doctor to record a death as natural and release the body, precisely to avoid a medico-legal autopsy. McQuoid-Mason has set out clearly why South African doctors have no legal discretion to accede to such requests once a death appears unnatural [4]. That literature addresses external pressure on the doctor's classification. This commentary is concerned with a different, and in some ways harder, version of the same structural gap: what happens when there is no external pressure at all, only the doctor's own unsupervised judgement about their possible error.

The legal framework: a duty that depends on self-assessment

The Inquests Act 58 of 1959, as amended by the Inquests Amendment Act 8 of 1991, requires that any person who has reason to believe that a death was due To causes other than natural causes must report it, and that where such a death is alleged, a medico-legal post-mortem examination must be conducted [1,5]. Section 2(2) of the Act makes failure to report an unnatural death a criminal offence, although the maximum penalty on conviction, a fine not exceeding R1 000, is a striking indication of how lightly the obligation is treated in practice [1,5]. The operative definition of an 'unnatural' is set out in the Regulations Regarding the Rendering of Forensic Pathology Service, made under the National Health Act 61 of 2003 [2,6]. That definition covers four categories of death: those resulting, directly or indirectly, from physical or chemical influence; those that are sudden, unexpected, or otherwise unexplained; those related to a therapeutic, diagnostic, or palliative procedure; and, critically for the argument here, any death that would normally be considered due to natural causes but, in the opinion of a medical practitioner, may have resulted from an act or omission that is criminal in nature [5]. That fourth category is the legislature's recognition that a death can look entirely natural on its face, as in a patient with known heart disease found dead at home or an elderly patient who deteriorates and dies in a GP's rooms, yet still legally requires referral if there is reason to think that negligent diagnosis or management contributed to it. The difficulty is that the Regulations place that judgement squarely `in the opinion of a medical practitioner' [5], and in the great majority of ordinary deaths in private or primary care, the practitioner best placed, and often the only person present, to form that opinion is the GP who was treating the patient. Where the negligence in question might be the GP`s own, such as missing a diagnosis, mismanaging a known condition, or delaying referral, the law is, in effect, asking the practitioner to identify and act against their own potential interest, with no independent second check built into the certification process itself.

Uncertainty, not just incentive, is part of the problem

It would be too simplistic to frame this purely as a question of dishonesty. Even a conscientious GP acting in good faith must interpret a legal test, whether an act or omission 'may be criminal in nature', which lies well outside their normal clinical training and which legal commentators have described as difficult to apply consistently in practice [7,8]. Jansen van Vuuren, writing specifically for a general-practice and family-medicine audience, notes that the current, broadly drafted legislation was intended to remove some of this discretion for procedure-related deaths by including them all, but acknowledges that important aspects of the framework remain difficult for practitioners to interpret, and that consultation with the local forensic pathologist is often the only practical way to resolve genuine uncertainty [7]. Madiba, reviewing the 2008 amendment to the equivalent provision in the Health Professions Act, raised similar concerns about whether the amended, deliberately all-inclusive wording gives practitioners any greater real clarity, or simply a wider net drawn around a decision they are still left to make largely alone [8]. This means the discretion at issue is not only an incentive problem, a doctor's reluctance to implicate themselves, but also a competence and system-design problem: doctors who are genuinely unsure whether a given death meets the threshold and who have no routine, structured second opinion to consult at the point of certification, however willing they may be to do the right thing.

What happens when the discretion fails: the life esidimeni illustration

The consequences of relying on individual practitioners' discretion, with no independent oversight at the point of certification, were demonstrated on a devastating scale in the Life Esidimeni tragedy. Between 2016 and 2017, more than 1300 mentally ill patients were moved from Life Esidimeni facilities in Gauteng to a network of non-governmental organisations, many of which were later found to be unlicensed and grossly under-resourced; 144 patients died, and an official Health Ombud investigation, led by Professor Malegapuru Makgoba, found that these deaths were overwhelmingly attributable to neglect, starvation, and dehydration rather than the natural progression of any underlying illness [9,10]. Investigative reporting found that almost half of the death certificates reviewed were incorrectly completed, with the cause of death left blank or simply recorded as `natural causes', and that the great majority of these deaths never reached a forensic pathologist at all, precisely because the certifying medical officers did not classify them as unnatural [9]. The Health Ombud's own report concluded that nearly every one of these deaths should, in fact, have been treated as unnatural [10]. Life Esidimeni is an extreme case of institutional and systemic neglect rather than a single GP's clinical error, but it illustrates the exact structural mechanism this commentary is concerned with: a legal framework that already defines these deaths as reportable, operating through a certification process that depends on the very practitioners whose conduct is in question to recognise and act on that fact. Where that recognition failed at Esidimeni, it did not fail because the law lacked a definition broad enough to capture what was happening; it failed because nothing in the certification process itself required, or even prompted, an independent check on the treating practitioner's own classification.

Towards a structural rather than a purely ethical answer

The response to this problem should not rest solely on reminding individual doctors of their existing legal and ethical obligations, important as that is, because the Esidimeni case shows that repeated, large-scale failure of exactly that kind is possible under the current framework. Three structural changes merit consideration. First, the death notification process could incorporate a mandatory, brief, structured checklist that directly tracks the four regulatory categories of unnatural death, rather than leaving the practitioner to recall and apply an unfamiliar legal test from memory, an approach consistent with Jansen van Vuuren's observation that practitioners are generally undertrained in this area of the law governing their practice [7]. Second, deaths occurring in, or shortly after treatment by, a solo practitioner or small practice, where no colleague is naturally positioned to notice an anomalous pattern, could be flagged for a low-burden, proportionate second review, whether by a peer, a local health authority, or the district forensic pathology service, rather than resting entirely on self-report. Third, given that the current maximum fine for failing to report an unnatural death has stood at R1 000 since 1991 and is a wholly negligible deterrent in current terms, the penalty structure in the Inquests Act itself warrants legislative reconsideration, alongside consideration of professional consequences through the Health Professions Council of South Africa.

Conclusion     

South African law already recognizes, in its statutory definition of unnatural death, that a death that appears natural on its face may warrant forensic investigation precisely because of the possibility of negligence in the deceased's diagnosis or management. What the law has not adequately addressed is the practical reality that, in the ordinary run of cases, the decision to act on that possibility is left entirely to the discretion of the practitioner whose own conduct may be the very issue in question, operating with minimal training in how to apply the legal test, no routine second check, and a statutory penalty for getting it wrong that has not kept pace with the seriousness of what may be at stake. The life Esidimeni tragedy shows what that gap can permit when it fails on a large scale; there is no reason to assume it does not also fail, quietly and one death at a time, in ordinary general practice.

Competing interests     

The author declares no competing interests.

Authors' contributions     

Anathi Phela is the sole author of this article, takes full responsibility for drafting and researching it, and has agreed to the publication of this final manuscript.

Acknowledgement     

The author would like to thank Dr Kevin Ramsarghey Dean (Acting): Faculty of Management Sciences for the motivation to publish this manuscript, having presented it at a research colloquium in Scotburg, KZN South Coast, in September 2025.

References