Break my bones, but don’t let me die! Should doctors provide ‘futile’ CPR?

by Dominic Wilkinson

Two recent cases in a Toronto hospital illustrate a dilemma that hospital doctors face all too frequently. What should they do if patients or their representatives insist on treatment that the doctor believes would be futile? Should they just go along with the patient despite their misgivings? Alternatively, should they unilaterally withhold treatment if they feel it would be inappropriate to provide it?

In the most recent case, 41 year old Mann Kee Li, with spreading cancer despite chemotherapy, left detailed specific instructions when he entered hospital. He wanted his life continued even if he were in a coma, never to regain consciousness. If his heart stopped he wanted doctors to attempt to provide cardiopulmonary resuscitation (CPR) even if this might be painful, might break his ribs and puncture his lungs.

However, the doctors caring for Mann Kee decided recently not to follow his instructions. They made a note in his chart that he was not for resuscitation in the event of a cardiac arrest. Mann Kee’s family sought a court order to compel the hospital to follow his wishes. The hospital’s executive expressed the medical point of view:

“When clinical teams determine that further interventions would have no benefit to the patient . . . ethically and legally, health-care providers are not obliged to provide interventions that lie outside the standard of care and would be of no benefit, and indeed may well cause harm to a patient”

This case comes only a month after a family launched a lawsuit against the same hospital for failing to provide resuscitation for an 88 year old patient, Douglas DeGuerre, who died in hospital following surgery to amputate his legs.

Disputes over resuscitation decisions like these have given rise to dozens of court cases, and hundreds of academic articles on medical futility. Often, what is at stake is the definition of ‘futile’ treatment, and whether treatment for the individual patient would or would not be futile. But debates about the definition of futility miss the crucial ethical point. There are only two ultimate reasons for doctors to decline to provide treatment that a patient believes would benefit them. The first is a paternalistic judgement that treatment would be contrary to the patient’s interests, it would harm them. The second is on the basis of insufficient resources.

The statement by Sunnybrook hospital executive appears to be based on the first reason. What doctors are saying in effect is, “we know that you want this treatment, and you think that it would benefit you, but we know better”. A judgement of this sort, that resuscitation should not be provided if it will only allow a patient to survive for a very short period, or has a very low chance of success, is necessarily based on a judgement about the value of survival, and often about what constitutes a good death. But the problem is that this prioritises the physicians’ values over those of the patient. Many people might believe that life would not be worth continuing in a state of permanent unconsciousness, but others would hold a quite different view. I personally would not choose to have CPR in my last moments if it had minimal or no chance of restoring me to a life I consider worth living. But others appear to regard cardiac massage almost as a form of secular last rite – evidence that doctors have done everything possible. Is one of these right and the other wrong? Why is the doctor’s view of the world given precedence? One reason that the patient’s wishes might be disregarded is that doctors may believe that the patient didn’t really know what he was requesting at the time that he asked for “full resuscitation”. The current case is distinctive because Mann Kee appeared to know exactly what his request for CPR might involve.

Paternalism may sometimes be justified. Some choices might be sufficiently obviously irrational that they should not be respected. Imagine for example, that a patient wanted treatment continued even if that meant that they were in prolonged constant unrelieved agony? It would be clear in that case that there could be no rational reason for such a choice, and that acquiescing to it would harm the patient. But providing CPR in circumstances like Mr Li, or Mr DeGuerre is not like that. The most likely outcome of resuscitating them would be that they would die despite all the efforts of doctors to forestall death. If we were really worried about them suffering pain from attempted chest compressions, we could always provide intravenous or intramuscular analgesia at the start of resuscitation.

In fact, the second justification for withholding futile treatment is far stronger – though it is much less often discussed. If providing a treatment would harm other patients there is a much clearer and less ambiguous justification for refusing to provide it. Providing CPR to a patient where there is a 0.5% chance of it being successful may or may not be enough to warrant the indignity, and potential pain of resuscitation. But providing CPR to this patient could mean that the resuscitation team is not available for another patient (with a much higher chance of recovery). More significantly, if resuscitation did result in a return of circulation for a patient like Mr Li or Mr DeGuerre, they would require admission to intensive care to be able to survive. And intensive care beds are a highly expensive and scarce medical resource. When the intensive care unit is full (which is not uncommon) other patients who could benefit from admission must be transferred to another hospital, or managed without intensive care. Patients who are declined admission have a significantly higher mortality than those who are admitted.

One final justification that is sometimes given for declining to provide futile treatment is that this may violate doctors’ professional integrity, or ‘abrogate their professional responsibility’. But it is not clear what this means. If a treatment were neither harmful to the patient, nor harmful to any other patient (through denying them access to limited resources), why should the professional’s integrity be at threat? If it is one of those reasons that leads to the integrity claim, then there is no need to talk about professional integrity.

In the case of Mann Kee Li, mentioned at the start of this post, family members withdrew their request for CPR only hours after the start of the court proceedings. His physical deterioration had led them to believe that he would no longer wish for resuscitation measures to be provided. Although cases of intractable disagreement (cf Golubchuk, Betancourt) often are the ones that receive public attention, most futility disputes are resolved with time and good communication. However, sometimes doctors should provide treatment that they regard as 'futile' – where treatment would cause little or no pain to the patient, and would not prevent other patients from accessing treatment. At other times, doctors are right to refuse to provide treatment that they judge to be inappropriate. However, the strongest reason for doing so is on the basis of distributive justice and the impact of providing treatment on others.

* In one survey of patient’s views about resuscitation 10% of patients indicated that they would want resuscitation even if they were permanently unconscious,

  • Facebook
  • Twitter
  • Reddit

15 Responses to Break my bones, but don’t let me die! Should doctors provide ‘futile’ CPR?

  • Not an ethical comment, but I am wondering to what extent the portrayal of CPR and other extraordinary measures in medical TV dramas and in other programmes has influenced peoples conception about the efficacy and benefits of those treatments? None of us wants to watch doom and gloom stories when we watch TV for escapism, but the general public has no other avenue for information regarding these sorts of interventions. As such, on TV we are presented with ‘miracle’ cases of recovery which in no way mirror the actual medical reality.

    On a medical note providing intravenous or intramuscular analgesia at the start of resuscitation is unlikely to have any effect on the pain of chest compressions. Firstly, the chest compressions themselves may noit be effective enough to get the drugs around the body, and secondly, due to the mode of action of these drugs they would not have enough time to take effect (but even if they did there would still be a painful period at the start of the resuscitation). Therefore, I don’t think the option of painless resus simply to be seen to be doing something is an option.

  • Dominic Wilkinson says:

    Thanks Muireann for both of these excellent points.

    For the first – you are right that television dramas give a very distorted view about the likelihood of recovery from CPR. In a paper in the NEJM in the mid 1990s ( ), John Lantos and colleagues found that patients having CPR on television had a 75% chance of recovery. In another recent study ( ) of patients with chronic lung problems, watching “educational” medical TV programs was associated with an exaggerated perception of the likelihood of CPR being successful.
    But, the paper that I linked to above ( ), suggests that even when patients are given realistic information about CPR and its likelihood of success, they still choose CPR far more often than doctors would choose it for themselves.

    As for your second point. You are probably right that analgesia would not be fully effective in a patient mid-arrest. However, other medications are sufficiently circulated to have an effect fairly quickly when CPR is being provided (for example adrenaline or suxamethonium), so an analgesic might reasonably be anticipated to be at least partly effective. I take your point that this wouldn’t prevent pain at the start of resuscitation.
    There is no evidence that I am aware of about perceptions of pain from CPR in patients who do survive. However, ( ) only 10% of survivors have any memory of the event. Pain from the resuscitation itself is not prominent in my experience of looking after patients who have received CPR (though some have ongoing discomfort from bruising or rib fractures). (Pain might be experienced but not remembered – which raises interesting separate questions about whether non-remembered pain is bad, or as bad as remembered pain)
    In any event, as pointed out in a recent article in JME ( ), even if CPR is painful, the question would be whether the disvalue of that pain outweighs the value of a chance of survival. Then we would be led to ask – who should do that weighing up of values – the doctor or the patient?

  • Thanks for the links to those studies Dominic.

    I must look up the Lantos et al one. I am intrigued as to the research methodology involved and am currently wondering which research council will fund me to watch years worth of ER re-runs in order to collect data!

  • Dominic Wilkinson says:

    For more TV CPR studies see – things are a bit bleaker on Flemish TV

    and – more recent study including HOlby city, Casualty, ER and Grey’s anatomy – survival rate not unrealistic, but few poor outcomes shown

  • I do not think that refusals to provide non-beneficial treatment are paternalistic. It is almost never the patient herself contemporaneously asking for the aggressive interventions. It is the patient’s surrogate. When providers resist the surrogate’s demands, they usually doubt that the surrogate is really doing what the patient wanted. Here, in the Li case, the patient had an extensive written advance directive. But he did not anticipate the precise circumstances that obtained this week. Therefore, it is not the case that the patient AUTONOMOUSLY chose X and the providers chose Y. Rather, the providers (and ultimately the family too) was skeptical that the patient actually chose or wanted X.

  • Dominic Wilkinson says:

    Hi Thaddeus,

    thanks for your thoughtful comment. There are three questions here
    a. Do doctors refuse to provide futile treatment because they do not think that the patient would actually want it, or do they refuse to provide it because of some other reason (perhaps hard paternalistic judgements about what would be best for the patient, or insufficient resources)
    b. If doctors refuse to provide treatment because they think that the patient doesn’t actually want it (though the surrogates claim otherwise) is this paternalism?
    c. Whether or not it is called paternalism, is it justified for doctors to refuse to provide treatment on the grounds that they don’t think that the patient actually would want it

    a. I take your point that doctors’ motivation for refusing to provide futile treatment is sometimes that surrogates appear to be misrepresenting the patient’s wishes. If that were the case, then, as Robert Truog has argued, the most appropriate response to futility disputes would be to challenge the legitimacy of the surrogate ( ) rather than to focus on whether or not the treatment is futile.
    However, as you have argued elsewhere ( ), the majority of intractable futility disputes are religiously motivated – as was the case for Samuel Golubchuk for example. In those instances it appears that doctors have good reason to think that advance directives, or surrogates’ requests are genuine expressions of the patient’s wishes. What doctors are saying in those cases is (if they are relying on the first reason highlighted in the post) a form of paternalism.
    (More generally, my sense is that among doctors or even in written guidelines, there is a sense that denoting treatments as ‘futile’ or ‘medically inappropriate’ is to imply that they should not, or at least *need not* be provided even if the patient autonomously requests them, or the surrogates are genuinely representing the patient’s wishes).

    b. Is this paternalism, if doctors are disputing whether a choice is really what the patient would autonomously choose (if they could)? You are right to highlight that this is not hard paternalism. Soft paternalism is sometimes interpreted as referring to attempts to cast doubt on autonomous nature of choices ( ), at other times it refers to forms of persuasion or coercion. But it certainly appears to be a type of paternalistic judgement (the doctor claims superior knowledge of the patient’s actual or hypothetical wishes), and is almost certainly based on a deeper type of paternalism (the doctor believes that he or she knows what would be best for the patient, and this is the reason why he/she thinks that the patient would want this).

    c. But of course the important question is whether it is justified for doctors to overrule surrogates on the basis that they know better than the family what the patient would have wanted. The problem with this type of (soft) paternalism, is that although the doctor may have superior knowledge about the medical facts, the likelihood of survival with/without treatment, the possible side effects (including pain) of providing treatment, a patient’s wishes about treatment are going to be based on both medical facts and the patient’s values. And, a priori, there is no good reason to think that the doctor would have better knowledge than family members of the patient’s values – quite the opposite in fact. The survey by Ian Kerridge pointed to above highlights that doctors have quite strikingly different value judgements than patients about treatment in the face of small probability of success, or likely significant impairment.
    That is not to say that surrogates always get it right. Their own grief, or guilt, or inability to let go, may lead them to mistakenly believe that the patient would have wanted treatment to continue, when in fact it is they who want treatment to continue. Where there is good evidence that that is what is going on, doctors should try to convince surrogates that the patient would not really want the disputed treatment. If necessary, perhaps they should go down the Truog line of having surrogates replaced. But the problem remains that if doctors want to overrule surrogates because they believe that they know better (than the family) what the patient would have wanted, it requires a difficult to substantiate claim about the values and preferences of the patient. Resource limits may provide a much stronger and clearer (although I accept more controversial) justification for refusing to provide treatment.

  • Peter Wicks says:


    At the end of your post you state that doctors should sometimes provide treatment that they regard as futile, while at other times they are right to refuse to provide treatment that they judge to be inappropriate. You then state that the strongest reason for doing so is on the basis of distributive justice and the impact of providing treatment on others. Personally I find these statements convincing, in the sense that they seem to accord with my own moral intuition, but given that they are normative statements I do not regard them as being true. It has also occurred to me that the argumentation contained in your post may be circular in the sense that it also contains several other normative statements.

    I’d be interested in your views on this. The way I would argue this would be to start from the utilitarian premise that we should be trying to maximise overall well-being, and then work up from there using empirical arguments that at least appear to be plausible, even if they lack a genuinely robust evidential basis. But it’s not clear to me that you are actually taking a utilitarian position here.

  • Dominic Wilkinson says:


    the short answer is that I believe that at least some normative statements are truth-apt, and hence my meta-ethical viewpoint is cognitivist. I am not sure where you find my arguments circular, but I can understand that as a non-cognitivist you find the assertion of normative claims perplexing.
    My understanding of normativity is that, when enough information is available, and they are appropriately specified, some normative statements are false and some are true. Other normative statements are neither unequivocally true nor false, but, we may hold them to be true or false where are good reasons to believe this.
    The approach that I take in this issue, as in most, is not to work from the ground up, from meta-ethics to ethical theory, to specific application. Rather, I prefer to start from the arguments and reasons that we have in favour of different alternatives, to look at the values and presumptions that appear to be widely held, and subject them to critical scrutiny. Using something like a Rawlsian reflective equilibrium I try to subject our intuitions to rational analysis, and to test the results of ethical analysis against shared intuitions.

  • Peter Wicks says:

    Many thanks for this Dominic.

    It’s not quite that I find normative claims perplexing. After all I make them myself, frequently. I’m also not sure to what extent I am really “non-cognitivist”. I think my position is rather that normative statements are declarations of the speaker’s values, where I am assuming that values are something one decides rather that something one discovers. Linguistically, of course, they take the form of propositions, but I don’t think one should assume that this is how they are usually intended. What is clear, though, is that I am not a moral realist, so where normative claims are made in the context of moral realism (and are thus clearly intended to be understood as propositions) I would regard them as false (and for that reason somewhat annoying). Unlike you I don’t believe that this depends on the amount of information available.

    That being said, I think the approach that you favour (critical analysis of pre-existing arguments and reasons) is perfectly consistent with my own understanding, in fact in a sense it’s all we can do, since there is no “ground” from which to work up. But it’s true that my interest is especially in seeing how far one can go in constructing an overall moral framework that accords reasonably well with our intuitions and then deducing ethical positions from that framework, rather than inducing general principles from my intuitions on specific issues. I guess this explains my affinity for utilitarianism…

  • I find it problematic to state that it is unclear what is meant by violating the doctor’s integrity. (Not to mention the rest of the medical team–nurses, respiratory therapists, etc.– who are affected by the medical decision.) It sounds as if the ethical position here is that it doesn’t matter if the procedure the patient wants harms the providers who have to deliver it. We don’t ask physicians who oppose abortion to perform them, out of consideration of their values. It is certainly repugnant and damaging to many in the medical field to perform resuscitation and deliver complex and prolonged ICU care to patients in the last minutes, hours, and days of their lives. It can have the effect of desensitizing medical providers to our own feelings and limiting our ability to be fully present to patients, family, and colleagues. It has the potential to violate and damage us indeed.

  • Peter Wicks says:

    I think Risa raises a crucial issue here, although I don’t think it quite makes the case that providing such treatment damages the providers’ *integrity*. At least not their professional integrity, as the term is usually understood. But that it damages them *emotionally*, first as people, but also with implications for their subsequent performs as health providers (who need, after all, care for their patients on an emotional level as well if they are to provide the best overall care), seems to be a valid argument that is distinct from the ones that Dominic has analysed.

  • Dominic Wilkinson says:

    Thanks Peter and Risa for your comments

    (Peter, apologies for tarring you with the non-cognitivist brush – your view appears to be a form of moral subjectivism – with which I also disagree!)

    Risa, I am glad that you picked me up on the question of professional integrity. I have responded to your concerns in another post ( )

  • David Li says:

    My name is David Li. I am Mann Kee’s brother.

    As an update, Mann Kee sadly passed away on Mon 25/10 at 4.30pm. I am struck by how many blogs have touched upon my brother’s case…I suppose increased debate on end-of-life issues is a silver lining on the cloud our family is currently going through. This blog caught my eye as it’s from the UK (like Mann Kee originally) rather than North America.

    I wanted to make a couple of clarifications as this case highlighted to me not just ethical issues, but, at the most fundamental level, plain misunderstanding of law even for major hospitals engaging major lawyers…we came up against a major hospital who displayed startling ignorance of their obligations under Ontario law. Had they been more aware, much of the aggravation, & legal expense on both sides, could have been avoided.

    The main facts before the court were:
    1) the hospital offered full code / CPR in their medical plan
    2) Mann Kee made a clear, informed choice to accept it
    3) under Ontario law, changing an accepted medical plan requires consent of the patient (or substitute decision maker), or failing this, a judgement from the “Consent & Capacity Board” (CCB)

    Based on points 1 & 2, the hospital was on the hook under Ontario law, to provide full code / CPR. Under Ontario law, the only way they can get out of this is to either persuade Mann Kee to change his mind, or otherwise go to the CCB for resolution.

    As Mann Kee’s substitute decision makers, we were obliged to advocate for his views on full code / CPR, & precluded from imposing our views upon him, our personal preferences were irrelevant…i.e. although I personally might prefer a quiet, peaceful death, Ontario law prevents me from making a decision based on my personal belief, I could only advocate based on Mann Kee’s clearly expressed views, or failing that, my best understanding of his thought process / value system.

    The big beef before the Superior Court was that Sunnybrook unlawfully changed the agreed plan, ignored clear statements from us that they were breaking Ontario law by reneging on an accepted medical plan without Mann Kee’s agreement, & not following any of the processes mandated under Ontario law, such as going to the CCB.

    The Superior Court recognised this by staying the DNR order, & ordering for a CCB…which we had told Sunnybrook to do from the beginning.

    While the issues presented to CCB panels are inevitably complex & thorny, the role of the CCB itself to resolve disputes is clearly understood.

    It is a pity, & waste of both our family’s money, as well as tax-payer $$, that we had to take Sunnybrook to court before they followed the law on the CCB.

    As a post-thought, given I was there in person, I have also been struck by how much is “lost in translation” in the various blogs …while most have delved into ethical / value judgements, the case itself was actually driven primarily by the hospital’s ignorance & disregard for Ontario law.

    Most blogs suffered from being “lost in translation” which handicapped discussions…there are many points I could clarify on the various blogs (including this one), as a trivial sample I thought it at least worth clarifying that my brother was 46, not 41 (!)…but most blogs would have benefited from discussing the prevailing law, if any.

  • Peter Wicks says:

    I guess the problem with discussing prevailing law, if the purpose (as on this blog) is to use such cases as a launchpad for discussion of wider ethical issues, is that there are so many different possibilities that this could actually obscure the bigger picture. I also think that it is possible and legitimate to discuss those wider issues *without* knowledge of the law in the specific case. Laws can be changed, after all, and among the practical outcomes of forging consensus on the ethical issues would be to promote appropriate changes in the law. So the question here would be not so much whether the hospital broke the law (that was a matter for the Superior Court to decide, not this blog…) as whether the law is just, and also what criteria a body such as the CCB should use in providing its judgements.

    But this has in any case been a touching reminder that the cases we use as inspirations for our discussions on this blog are not abstract thought experiments, but real issues involving real people…

  • Dominic Wilkinson says:


    thank-you for your comment. It is kind of you to provide an insight into the actual issues involved in Mann Lee’s care from his family’s perspective – at a time that must be extremely difficult for you and for your family. As Peter notes, it is important for ethicists and philosophers to remember when discussing real cases, that there are subtleties and nuances that cannot be gleaned from media reports, and that our commentary may sometimes add to the distress of those directly involved. I am sorry for any inaccuracies in my description in the blog above.
    As Peter also notes, although there are genuine questions of law at stake, our focus on this blog is usually on the underlying and wider ethical questions. That is partly because we (speaking for myself at least) are not legal experts, partly because there are prior issues about what the law should be, and partly because it is useful to discuss questions that are relevant in different parts of the world (where laws may be different).
    thanks again for your contribution to our blog, and my deepest sympathies to you and your family.


Subscribe Via Email