The Health Professions Council of South Africa (HPCSA) has introduced significant updates to its ethical guidelines on two fronts that affect South Africans far more directly than many may realise: end-of-life care and the use of artificial intelligence in healthcare. Although these issues may feel abstract or uncomfortable, the revised rules shape what could happen to you or a loved one when clear decisions matter most.
The first set of changes, contained in the updated booklet 7 on withholding or withdrawing treatment, formalises a new safeguard for patients who lose the capacity to speak for themselves.
According to an analysis by Bowmans senior associate Jay Page, the guidelines now expressly allow individuals to nominate a representative in writing to make healthcare decisions on their behalf if they become unable to do so. This introduces a practical form of advance planning that can give families clarity when emotions are running high and decisions are time sensitive.
Page notes that end-of-life decisions must always be made in the patient’s best interests and wherever possible, in consultation with the patient, their nominated representative or close relatives. The guidelines recognise that not every patient will want to confront the details of life-prolonging treatment. In such cases, the practitioner is now permitted to brief the patient’s chosen representative so that an informed decision can be reached.
The revisions also codify the hierarchy of consent when a patient has not appointed a representative. If a patient lacks capacity, and no written proxy or court order exists, authority flows in a strict order: spouse or partner first, then parent, grandparent, adult child or sibling. The aim is to reduce uncertainty and prevent disputes when families are already under strain.
Another important clarification affects scenarios where treatment is considered futile. The longstanding principle remains that clinicians may decide to withhold or withdraw treatment even when the patient or family requests its continuation. The updated guidelines now explicitly state that such decisions must be grounded in the patient’s best interests, reinforcing the centrality of dignity and clinical judgement.
Guidelines on the use of Artificial Intelligence
The second set of revisions to booklet 20 addresses a fast-moving area: the role of AI in healthcare. Here, Page describes an approach that embraces innovation but insists on patient safety and accountability. AI may support and speed up clinical decision-making, but it cannot replace a clinician’s own expertise. The guidelines state that the practitioner must always make the final call, and Page emphasises that “there is limited room for AI that serves only to make practitioners’ lives easier”.
The revisions also highlight risks that matter to the public. Patient confidentiality is not guaranteed when algorithms can re-identify individuals from supposedly anonymised data. Minority and disadvantaged groups may be further marginalised because AI systems often lack robust data about their specific characteristics. And there is still no comprehensive regulatory framework governing the safety and accuracy of AI tools.
To address these gaps, the HPCSA outlines three “pillars of AI”: ethics, legality, and technical standards. Practitioners must ensure that any AI tool respects patient autonomy and privacy, complies with South African law, and is technically sound, reliable, and safe. Only AI systems that have been validated for reliability and cultural appropriateness may be used, and SAHPRA has now detailed what counts as AI-enabled medical devices, spanning imaging software, predictive algorithms, diagnostic support tools, and certain wearables.
Crucially, the standard of care must remain the same whether or not AI is used. Patients who decline AI cannot be disadvantaged or denied access to services. And practitioners are cautioned to avoid informal or unrecognised tools, with Page drawing a parallel: clinicians would not use an untested medicine and should treat unvalidated AI tools with the same caution.
For the public, these updates highlight two realities. First, planning for end-of-life scenarios is no longer an abstract legal exercise but a practical step that can ensure your wishes are respected when you cannot voice them yourself. Second, as AI becomes more embedded in healthcare, the regulatory guardrails are tightening to protect patients from both over-reliance and under-regulation.
These are sensitive areas, but they touch every household at some point. The HPCSA’s revised guidelines set clearer expectations, giving South Africans firmer footing in moments where clarity, trust and sound judgement are essential.
