Should GPs be identity relative paternalists?

Richard Armitage is a GP and Public Health Specialty Registrar, and Honorary Assistant Professor at the University of Nottingham’s Academic Unit of Population and Lifespan Sciences. He is on twitter: @drricharmitage


When our patients make choices that will predictably risk or cause them harm – such as smoking, not exercising, or regularly over-eating – GPs have a professional duty of beneficence to advise against such decisions.1  But what if patients choose – as they frequently do – to ignore our advice?  In these scenarios, the GP’s professional responsibility to beneficence comes into tension with their duty to respect patient autonomy.1  This clash of ethical obligations is usually resolved by appealing to John Stuart Mill’s ‘Harm principle,’ which holds that “the only purpose for which power can be rightfully exercised over any member of a civilized community, against his will, is to prevent harm to others. His own good is not a sufficient warrant.”2  In other words, GPs allow patients to make harmful choices – thereby affording primacy to autonomy at the expense of beneficence – when that harm accrues only to the individual making such choices.  As such, the Millian harm principle strongly defends against medical paternalism – against the GP’s well-meaning interference in the lives of their patients.

…rather than considering oneself as a single self who exists from birth until death, it may be useful to regard oneself as a number of “successive selves” over time.

This approach to choice-making – which draws a distinct line between choices that harm the self and those that harm third parties – has been recently subjected to intense debate by medical ethicists.  A contemporary argument by Dominic Wilkinson3 utilises the ‘reductionist view’ about personal identity defended by the Oxford philosopher Derek Parfit, which holds that the continued existence of an individual over time can be reduced to particular physical or psychological continuities.4  On this view, the extent to which someone at point t1 is physically and/or psychologically continuous to and connected with someone at the later t2 can be determined by asking whether, for example, they share the same body, the same memories, the same personality traits and patterns of thought?  For Parfit, the answers to such questions reveal how the later person is related to the earlier person, while asking whether the person at t1 is the same as that at t2 is often an empty question.  As an intuitive example, when an individual sustains a profound brain injury and is rendered permanently unconscious, families often report that the person who they were has “gone” despite the individual’s body remaining alive,5 suggesting that the discontinuity in the memories, personality traits and patterns of thought make the person at t1 (pre-brain injury) a meaningfully different person to the individual at t2 (post-brain injury).

Parfit’s reductionist view about personal identity holds various striking implications.  One is that, rather than considering oneself as a single self who exists from birth until death, it may be useful to regard oneself as a number of “successive selves” over time.4  Another is that, rather than asking whether I am the same person at t2 as I was at t1 (which generates a binary, yes on no, answer), it may be more useful to ask how closely related I am at t2 to myself at t1 depending on the extent of psychological connectedness between the persons at t1 and t2 (which generates an answer of degrees).4  From this, a moral implication of this view about personal identity emerges: that the distinction between the self and others is less clear than conventionally viewed and, therefore, we should regard decisions about our future selves as we would about decisions that affect another person.4

Paternalism in medicine is generally regarded as ethically impermissible when patients have capacity to make autonomous decisions about their lives, and receives two standard objections: firstly, that it is good for individuals to be able to learn from their mistakes (the consequentialist objection); secondly, that individuals are in stronger positions than others to know what is best for themselves (the epistemic objection).2,4  However, the recent argument by Wilkinson applies the Parfitian ‘reductionist view’ about personal identity to medical paternalism.  By regarding future selves as other people – third parties who are meaningfully different from current selves – Wilkinson generates the ‘identity relative paternalistic intervention’ (IRPI): “Individuals should be prevented from doing to future selves… what it would be justified to prevent them from doing to others.”  While his argument is complex, Wilkinson uses the IRPI to undermine the harm principle and suggest that doctors and states may be justified in being more paternalistic.

What does this mean for, and how might it apply to, contemporary general practice?  To support his argument, Wilkinson uses three cases – vaccine refusal, home birth refusal, and advance resuscitation refusal – each of which contains different degrees of connection between the individuals making decisions and their future selves that are later harmed by their decision, to demonstrate the varying degree to which such medical paternalism may be justified.  I shall now describe three hypothetical cases relevant to general practice, then explain the potential implications of the IRPI in each.

Treatment refusal

Maggie, a 65-year-old woman with a new-onset headache, tenderness over the temporal artery, jaw claudication, and scalp tenderness is suspected of having giant cell arteritis.  Her GP explains the need for early treatment with high doses of prednisolone to reduce the chance of serious complications including visual loss.  Despite understanding the potential consequences, Maggie refuses this treatment as she is aware that steroids may cause weight gain.  Shortly afterwards her vision deteriorates and is permanently lost.

Screening refusal

Donna, a 31-year-old heavy smoker who uses the combined oral contraceptive pill and was too old to receive the HPV vaccine when it was introduced into the UK routine childhood vaccination schedule, refuses to attend cervical screening because she is worried that the procedure would be uncomfortable.  18 months later she is diagnosed with advanced cervical cancer.

Smoking cessation refusal

Alan, 23-year-old man who smokes 40 cigarettes each day refuses an invitation to a smoking cessation programme because he enjoys the immediate effects of smoking.  35 years later he develops and dies from metastatic adenocarcinoma of the lung.

In Treatment refusal, there is only a short period of time between Maggie’s decision to refuse steroid treatment at t1 and the harmful consequences of that decision at t2.  This means the psychological connection between the two Maggies is strong.  In contrast, in Smoking cessation refusal, many years pass between Alan’s initial decision to continue smoking heavily (t1) and the metastatic cancer that it causes (t2), meaning the psychological connections between Alan at t1 and the harmful consequences of that decision at t2 are likely to be weak.  Between these two extremes is Screening refusal, in which an intermediate amount of time has elapsed between Donna’s decision at t1 to refuse cervical screening and the harmful consequences of that decision at t2, meaning the psychological connection between the two Donnas is moderately strong.

This raises the provocative claim that GPs should once again act in a paternalistic manner…

On the reductionist view of identity, the psychological closeness of the two Maggies means that they are fundamentally the same person, while the large psychological distance between the two Alans makes them third parties to one another, and the intermediate psychological connection between the two Donnas makes them more different than the two Maggies and less different than the two Alans.  As such, according to the identity relative paternalism account, the reasons for paternalism are strongest in smoking cessation refusal, weakest in treatment refusal, and moderate in screening refusal.  Accordingly, since doctors apply the harm principle and thus do not permit harm to third parties, GPs should potentially not permit Alan to refuse smoking cessation, and potentially not permit Donna to refuse cervical screening, whilst they should allow Maggie to refuse steroid treatment.

These implications of identity relative paternalism in general practice contexts raise various problems.6,7,8,9,10  One is a potential paradox in which identity relative paternalism appears to allow Maggie to make a decision that will almost certainty and imminently induce lifelong suffering (blindness) whilst prohibiting Alan from making a decision that has much lower certainty of causing harmful consequences at some indeterminate future point (adenocarcinoma) – in acting for the patient’s own good, ‘true’ paternalism would sooner overrule Maggie’s decision than that of Alan, but not on the identity relative account of paternalism.  Another problem is that identity relative paternalism may prohibit altruistic self-sacrifice, such as that inherent to living organ donation.11  Another is the impact of ‘transformational experiences’ (such as becoming a parent) on the psychological connection between a person between times t1 and t2, which may render these selves meaningfully different from one another despite only a short amount of time having passed between them.12 Yet another is that identity relative paternalism cannot obviously be both permissible and paternalistic.  This is because, if present and future Alans are different people, then interference with t1 Alan’s decision is potentially permissible (by applying the Harm Principle) but not paternalistic (because paternalistic action is for the own good of the patient, not that of a different person).13

The identity relative paternalism argument has brought the ethical permissiveness of paternalism back into the medical ethics limelight.  Since primary care is largely concerned with facilitating improvements in lifestyle and controlling modifiable risk factors (such as those in Smoking cessation and Screening refusal), identity relative paternalism may be mostly justified in the context of primary care.  This raises the provocative claim that GPs should once again act in a paternalistic manner, albeit in the contemporary interpretation as identity relative paternalists.

Deputy Editor’s note: It’s worth reading this carefully and considering the implications. Should GPs do a harder sell on decisions with implications in the far future? Does the lack of certainty about the distant future undermine these examples? Are these examples of moral luck, where the rightness of an action is judged on an actual rather than predicted outcome? We would love to see some debate in the comments below or a fuller reply article submitted to


  1. Tom Beauchamp and James Childress. Principles of Biomedical Ethics. 8th edn, Oxford University Press, 2019
  2. JS Mill. On liberty. London: Longman, Roberts, and Green, 1864
  3. D Wilkinson. The harm principle, personal identity and identity- relative paternalism. Journal of Medical Ethics 2023; 49: 393–402. DOI: 10.1136/medethics-2022-108418
  4. D Parfit. Reasons and persons. Oxford University Press: Oxford, 1984
  5. C Kitzinger and K Kitzinger. ‘This in-between’: How families talk about death in relation to severe brain injury and disorders of consciousness. In: L van Brussel and N Carpentier, editors. The Social Construction of Death: Interdisciplinary Perspectives. Basingstoke UK, Palgrave MacMillan, 2014. Chapter 12.
  6. E Braun E. Identity-relative paternalism fails to achieve its apparent goal. Journal of Medical Ethics 2023; 49: 413–414. DOI: 10.1136/jme-2023-109119
  7. R Dresser. Medical choices and changing selves. Journal of Medical Ethics 2023; 49: 403. DOI: 10.1136/jme-2023-109120
  8. EG Schantz. Identity-relative paternalism is internally incoherent. Journal of Medical Ethics 2023; 49: 404–405. DOI: 10.1136/jme-2023-109009
  9. D Birks. Identity-relative paternalism and allowing harm to others. Journal of Medical Ethics 2023; 49: 411–412. DOI:10.1136/jme-2023-109118
  10.  L Sheahan and L Campbell. On Wilkinson: unpacking Parfit, paternalism and the primacy of autonomy in contemporary bioethics. Journal of Medical Ethics 2023; 49: 415–416. DOI:10.1136/jme-2023-109122
  11.  C Garstman, S de Jong, J Bernstein. Does identity-relative paternalism prohibit (future) self-sacrifice? A reply to Wilkinson. Journal of Medical Ethics 2023; 49: 406–408. DOI: 10.1136/jme-2023-109028
  12. LA Paul. Transformative experience. Oxford University Press, Oxford, 2014.
  13. B Saunders. Journal of Medical Ethics 2023; 49: 409–410. DOI: 10.1136/jme-2023-109121

Featured Photo by Kyle Glenn on Unsplash

Notify of

This site uses Akismet to reduce spam. Learn how your comment data is processed.

Newest Most Voted
Inline Feedbacks
View all comments
Ben Hoban
Ben Hoban
1 year ago

I believe that I’m the best custodian of my future self’s welfare on the basis that we’re already acquainted, and if I mess things up, future me can live with it more easily than if someone else does. Viewing health outcomes as the measure of sensible decision making undermines the idea of capacity and hands unlimited power to the state.

Andrew N Papanikitas
Andrew N Papanikitas
1 year ago
Reply to  Ben Hoban

Future people can be quite unforgiving of healthcare professionals for failing to stop them from unwise choices… is this an either or are there tipping points?

Elke Hausmann
Elke Hausmann
11 months ago

Why is medical paternalism always wrong? Just like our own unique selves change over time (partly due to the decisions we have made in the past), so does medical knowledge, and even the best advice I can give as a doctor today on the basis of my current knowledge may be wrong tomorrow (plenty of examples of that in the history of medicine). I can advise a patient, but the patient has to make their decision, and if he or she does not OWN that decision, that will create more problems in the future if there is a negative outcome, which may lead to profound unhappiness for the patient constantly ruminating about what might have been, or grief for the doctor, if the patient feels that the doctor had an undue role in influencing that decision and decides to complain. And obviously it feeds into what kind of world we want to live in – one that allows us to make our own decisions over our own lives or one where we are told how we should live and deviation is punished?

Previous Story

The impatience of convenience: a reflection on ‘Four Thousand Weeks’

Next Story

Episode 125: Results from two national cancer audits – what’s changed in referrals and early diagnosis for cancer between 2014 and 2018?

Latest from BJGP Long Read

Cautionary tales

Stories entertain us, but on a deeper level, they also help us make sense of our

Would love your thoughts, please comment.x
Skip to toolbar