Andrea Clark And Medical Futility

by hilzoy

In a comment on an earlier thread, DaveC asked me to comment on the case of Andrea Clark, a Texas woman whose hospital had planned to discontinue life-support treatment. Yesterday, the hospital decided to continue treating her. But that doesn't make the underlying issues go away, and since this is neither the first nor (in all likelihood) the last such case, I'll discuss them below the fold. But it's important to recognize one crucial point about cases like Andrea Clark's: They are not about whether or not a hospital can decide to kill people. In this country, no hospital can do that legally. They are about whether a hospital should be required to go on providing care that it believes is futile.

Under current US law, competent adults are allowed to decline any medical treatment. The reason for this is straightforward: most medical treatments, if performed against a patient's will, constitute a form of assault. There are cases in which medical treatments can be performed on normally competent adults without their consent -- e.g., in emergencies in which the patient is unconscious -- but treating patients over their objections can be done only when the patient is incompetent. This is why it's OK to vaccinate toddlers with their parents' consent, even if the toddler screams that she doesn't want to have a shot: by law (and in fact), toddlers are not competent to make that decision. It's also why I get to make an advance directive specifying that under certain circumstances, I do not want further medical treatment: I am competent to decide, and I get to refuse to allow people to do things to my body if I so choose, even if they think it's "for my own good."

While I get to reject any medical treatment that someone proposes to give me, however, that does not mean that I get to require that someone give me medical treatment, whether they want to provide it or not. That raises entirely different questions. (In this respect, medical treatment is like a lot of other things. Absent some special circumstance like a contractual obligation, the laws against kidnapping mean that I cannot be transported somewhere against my will; but the fact that I can refuse to be taken to Paris against my will obviously does not imply that I can demand to be taken to Paris, and that other people are then obligated to take me there.)

This is why doctors sometimes turn away patients who can't pay, even though they want treatment. (Hospitals are required to provide care in life-threatening emergencies, and to women who are giving birth.) And it's also why an advance directive that says that a person wants a certain medical procedure does not obligate anyone to actually provide that procedure to her.

Once someone is actually in a hospital, of course, things change. If someone is already on a surgical ward, receiving treatment, the hospital will generally try to provide the treatment that the patient wants, within reason. The question involved in cases like Andrea Clark's is: should that "within reason" clause be there at all? If so, what should it mean?

I think that there should clearly be some limits to what a hospital can be required to do when a patient wants them to do it. And I'm going to start with a deliberately extreme example. This is not because I want to imply that Andrea Clark's case is like this; it's just to explain why I think that there are some cases in which hospitals can rightly refuse to offer the treatment someone wants.

So: suppose that I am hospitalized with pneumonia, and I become convinced that my pneumonia will not be cured unless someone amputates my leg. Amputating my leg will not, of course, help my pneumonia at all. It would just add a whole new range of medical problems to the one I already have. Moreover, it would eat up an enormous amount of resources -- not just money, but things like the time of surgeons and nurses, the operating theater, the resources I will need while recovering from my operation, and so on. These resources are scarce: hospitals do not have an infinite number of ICU nurses, beds, and so on, and there are other patients who need them. For these reasons, no hospital would agree to cut off my leg, however much I wanted them to, and however strongly I believed that it would cure my pneumonia.

Moreover, they would probably refuse even if I could pay for the procedure, on two separate grounds. First, whatever I believe, cutting off my leg would not only not cure my pneumonia; it would actually harm me, and most hospitals would not want to participate in it for that reason. Second, even if I can pay for my amputation, money is not the only scarce resource involved here. As I said above, the mere fact that I want some procedure and am willing to pay for it does not obligate anyone to actually perform it on me, especially when this would require using scarce resources for no good reason.

***

If you're with me so far, here's where we are: if I'm a competent adult, I get to decline any treatment. However, a hospital does not have to perform any treatment I want, even if I can pay for it. There are many cases in which, if i want something done and can pay for it, a hospital should accede to my request. But there are some cases -- like my hypothetical amputation -- in which it would clearly be wrong for a hospital to do so. This means that we have to answer the question: Where should we draw the line?

The generally accepted answer is: when the treatment offers no medical benefit. This explains my hypothetical amputation case. But it leaves open the question: what, exactly, counts as a "medical benefit"? This question is most problematic in end-of-life cases in which the patient is already hospitalized, the patient or her guardians or proxy requests a treatment that the hospital does not think is appropriate, and the treatment in question would prolong her life. If she isn't already hospitalized, she or her relatives can seek out a facility where she can receive the kind of treatment she wants, and no one hospital's policies determine what care she receives. And if the treatment isn't needed to prolong her life, then the results of her not getting it are much less serious, and the issue is more likely to be the straightforward question: is this treatment indicated for a patient like her, or not?

Now, you might ask: if a treatment would prolong a patient's life, why on earth would a hospital think it wasn't appropriate? And why, in particular, wouldn't they think it provides a "medical benefit"? Isn't prolonging life the most obvious "medical benefit" imaginable?

Hospitals will not fail to prolong someone's life if that person is competent to decide what treatment she wants, and if she requests that treatment. Nor, in general, will they fail to treat if there is any real prospect that the patient will regain competence. The question only arises when the patient is, as far as they can tell, permanently incompetent; in practice, it only arises when someone is not just marginally incompetent, but severely mentally impaired.

Moreover, the default assumption is, of course, that prolonging someone's life is a benefit to her. The question whether to continue treatment only arises when there is some specific reason to question that assumption. And the most obvious such reason is serious pain: either pain caused by the treatment itself (e.g., an operation that would prolong a patient's life for a few days, but would cause her to suffer a lot), or pain that the patient is already experiencing, and that cannot be relieved by palliative care.

Consider, for instance, an infant with a disease that is inevitably fatal by age three, that involves excruciating pain that cannot be medicated away, and that would kill her now without intensive medical care. In any case involving an infant's life, it's easy to understand why a family might want everything possible to be done to prolong it. But in a case like this, you can also see why a doctor or a hospital might not agree. The treatment will prolong the child's life, but not beyond the age of three. In the meantime, she will suffer enormously. Being an infant, she has never asked for her life to be prolonged even if she has to suffer excruciating pain, nor has she had a chance to develop views on this matter that others could report. When treatment is given in cases like this, medical staff often say things like: I'm participating in torture.

(This is similar to the case of Sun Hudson, another person whose life support was terminated under the Texas law. He was born with an incurable genetic disorder that's generally fatal within the first few hours after birth, causes pain and severe retardation in those who live long enough, and whose sufferers can never live off life support. His mother said that "her son just needed time to grow and to be weaned off of the ventilator he was on since birth." That's not true: it is, as I said, an incurable genetic disorder; the medical literature contains reports of "rare long-term survivors", but in this context, "long-term survivors" refers to "a male aged 4.75 years and a female aged 3.7 years at last follow-up". The hospital's view was that "it would be unethical to continue with care that is futile and prolongs Sun's suffering.")

The fact that the treatment might also use up scarce resources is a consideration that the medical profession is profoundly conflicted about; doctors are unwilling to say that this should be a factor, but looking at the profession from the outside, I'm not sure that I see any principled reason why it shouldn't be. Kevin Keith has written a thoughtful post about the resource issues:

"Providing futile treatments is a problem. Often, they are treatments for the most badly-debilitated patients, and such patients often require extensive clinical staff involvement - nursing, physical or other therapy, doctors’ visits, and other staff assistance on a regular basis - that takes staff away from aiding patients who can benefit more. Also, such treatments may involve the use of scarce resources such as exotic equipment or scarce drugs, thus preventing other patients from obtaining them. And, finally, there is of course the question of money: insurance companies may not pay for futile treatment, and patients’ families may not be able to pay, leaving hospitals with sometimes very large costs for intensive treatment that is justified by nothing more than wishful thinking."

***

In any case, here's how the AMA addresses the problem:

"When further intervention to prolong the life of a patient becomes futile, physicians have an obligation to shift the intent of care toward comfort and closure. However, there are necessary value judgments involved in coming to the assessment of futility. These judgments must give consideration to patient or proxy assessments of worthwhile outcome. They should also take into account the physician or other provider's perception of intent in treatment, which should not be to prolong the dying process without benefit to the patient or to others with legitimate interests. They may also take into account community and institutional standards, which in turn may have used physiological or functional outcome measures."

The AMA then lays out a series of procedural requirements for such cases: physicians must consult with the patient and/or her caregiver, guardian, or proxy; if they cannot agree, the hospital's ethics committee should get involved; if disagreement persists, they should try to get the patient transferred to another hospital, etc. However, if all these steps fail to yield agreement on what should be done, the AMA concludes: "the intervention need not be offered."

The Texas law (pdf, see p. 12) under which Andrea Clark's case unfolded says the following: if a physician decides not to provide care that a patient or her family, proxy, or guardian has asked for, a lengthy process of consultation ensues. The patient's family, proxy, or guardian must be informed of this process, and can attend the meetings involved. Life-sustaining treatment must be provided throughout.

If, at the end of the process, the hospital still thinks that life-sustaining care is "inappropriate", then the hospital must keep the patient on life support for ten days, while it looks for another facility that will provide the care it thinks is inappropriate. But if neither it nor the family, proxy or guardian can find such a facility, after ten days it is allowed to take the patient off life support. (Generally, hospitals will extend this period to try to work with the families, and it can also be extended by a court.)

Personally, I think that this is a very good law. (I'm not saying it's perfect; I can imagine the odd tweak here and there. But it's very good.) Someone might think: ah, that's just because you're a liberal bioethicist. If so, it's worth noting that this law was drafted with substantial input from Texas Right To Life, and signed by George W. Bush. In any case: it recognizes the right of hospitals to decline to pursue treatment they think is futile; it puts on place a serious consultative process to ensure that all voices are heard; and it requires efforts to place any patient whose life support will be terminated over the objections of her family, guardian, or proxy. These are all, I think, good things.

In particular, the requirement of consultation is crucial. Disagreements about the care someone should receive are horrible for all concerned; and I cannot imagine what it must be like for the family of someone whose care is discontinued to believe that that person has died for reasons they do not understand. Extensive consultation is the best way to prevent this, or at least to make sure that each side has a chance to hear and understand the concerns of the others.

Likewise, the ten day grace period is also crucial. If some other hospital can be found to take the patient, that's good. If not, at least all sides will know that the reason treatment was discontinued wasn't just that that one particular hospital had a bizarre ethics committee. No decision like this should be vulnerable to the mistakes that one group of people might make, and by allowing everyone to seek out another hospital, the law minimizes the chances that it will be.

***

I'm not sure what to make of Andrea Clark's case in particular, since there are aspects of her case that I'm not really clear on. For one thing, what is her condition, and what is her prognosis? Here's a newspaper article:

"Clark, 54, a "blue baby" upon whom Drs. Michael E. DeBakey and Denton Cooley performed pioneering open-heart surgery on the then new heart-lung machine in the late 1950s, has been at St. Luke's since November.

Her condition deteriorated after open-heart surgery in January and bleeding in the brain this month, Dixon said.

The conflict erupted April 19, when Dr. Robert Carpenter Jr., chairman of St. Luke's ethics committee, wrote Ward and Dixon to say that the committee had unanimously agreed with Clark's attending physician that "the life-sustaining treatment currently being provided to your sister is inappropriate and should be discontinued."

Although St. Luke's officials said federal confidentiality laws prevent them from talking about the case, the letter stated that Clark was "experiencing substantial pain and suffering.""

And here's her sister:

"Andrea, until a few days ago, when the physicians decided to increase her pain medication and anesthetize her into unconsciousness, was fully able to make her own medical decisions and had decided that she wanted life saving treatment until she dies naturally. We have learned that this is part of the process, when hospitals decided to declare the "medical futility" of continueing treatment for a patient. But, this is not a Terry Schiavo case; not anything like it. Andrea, when she is not medicated into unconsciousness (and even when she is, and the medication has worn off to some degree) is aware and cognizant. She has suffered no brain damage to the parts of her brain responsible for thought and reason, or speech. She has only suffered loss of some motor control. The reason that the physician gave to medicate her so much is that she is suffering from intractable pain in the sacral region (in other words, she has a bedsore that causes her pain). This is not reason enough, in our books, and we are trying, as we speak, to get Andrea's medication lowered so that she can speak to us.

There is also some disagreement as to whether Andrea is really in that much pain, as well. When she is not medicated to this degree, and she sees her son, Charles, she smiles. She also mouths words (Andrea is very vocal, normally, even with a trach, and asks for food, etc., when she is not medicated to the gills). Once again, this is not like the Shiavo case, where there was brain death. Andrea has voiced her wishes, over and over again, and if she were not on so much pain medication, she would voice them again."

Elsewhere, we learn this: "I just spoke to Andrea's sister Lanore and she verified that not only does Andrea have bedsores, but one is "as big as a dinner plate."" That's a very big bedsore, and probably accounts for a lot of the pain. (It is my understanding that, contrary to comments on some of the blogs, while some bedsores result from bad treatment, not all do.)

I think that a lot would depend on the answers to questions that are not answered in any reports I've seen. What is her prognosis? How much pain is she in? What is her mental state, exactly, and how much brain damage has occurred? If she will die within a few months, whatever happens, and she is both permanently incompetent to decide on her own treatment and in severe pain, then I would be inclined to agree with the hospital. If, on the other hand, some or all of these things are not true, then I would not. None of the news sources really gives enough information for me to decide.

It is worth noting, though, that the hospital has decided to continue her treatment. This is one moral of the story: in general, hospitals are understandably reluctant to cut off life support to their patients, especially when those patients' families do not agree. Even a hospital staffed by horrible malevolent people without a decent bone in their bodies would be reluctant to do this because of the publicity; hospitals staffed by decent people who are trying to help their patients would be all the more reluctant.

On the other hand, there are cases -- and, as I understand it, Sun Hudson's was one -- in which it's very hard to make a case that prolonging a patient's life is anything more than high-tech torture. This isn't true when the patient is competent: if I can think, then I can give my pain some meaning, and thus, however excruciating it is, it will not be true that it is nothing but torment. And it isn't true when the patient might recover competence, for the same reason. But when prolonging a patient's life means postponing, for a relatively short period, a death that will come quickly in any case, and when that postponement brings her nothing but suffering, then I think that a hospital can legitimately decline to provide treatment.

But it's crucial, in this context, that we're talking about declining to provide treatment, not killing the patient outright. No one has the right to decide that a patient must die. The only right I claim that doctors and hospitals should have is the right to decide that they do not wish to provide treatment when they believe that it would be futile, given a law that provides extensive consultation, and that allows them and the patient's family or proxy to try to find someone else who will provide it. Because, as I said at the outset, while competent adults have the right to refuse any medical treatment they do not wish to undergo, it does not follow that they have the right to require that others provide any medical treatment they want to receive.

***

One last thing: any case in which questions of futility arise are excruciating for all concerned. My thoughts are with Andrea and her family, and also with those who are caring for her.

Comments

[rilkefan]

Is refusing to rescue someone drowning in a puddle not murder? If the drowner is already dying of cancer?

Just a useless link for hilzoy.

[lily]

I am struck by the family's phrase that Andrea should die naturally. Isn't that what's being prevented by medical treatment?
I don't mean to be critical of her family. It really isn't clear any moe when a death is natural and when it isn't. My grandmother died naturally I suppose after my dad told the hospital to stop treating her for anything but pain. She was unconscious and had pnuemonia. She was also over ninety.

[Francis]

Is refusing to rescue someone drowning in a puddle not murder? yes, it is not murder. There is NO duty to come to the aid of another except in limited circumstances.

[rilkefan]

Should have specified "not in legal terms".

Well, but what if they tripped on your foot? On your off-leash dog?

[rilkefan]

Anyway, I think if there were a universal ethic it would treat shooting someone and not rescuing them from a puddle as nearly equivalent.

[Jeff Eaton]

It really isn't clear any moe when a death is natural and when it isn't.

I once said to a friend, as we discussed this very question, that the only 'natural' way to die is being taken down by wolves when you can't keep up with the rest of your tribe. It's harsh, but we have to remember that medicine is all about preventing things that would normally kill us from doing so. As our capabilities grow, this inevitably raises questions. Hilzoy laid them out quite eloquently, but the 'Let someone die naturally' has never struck me as terribly compelling.

[hilzoy]

rilkefan: I don't think it's murder in non-legal terms. Or at least: it's not murder to come to the aid of someone whose life you can save. If it were, then every dime we spend on anything not strictly necessary for the preservation of our own lives would be murdering one of the people whose lives we could save by sending it (with the other dimes) to Oxfam. I have pretty serious views about obligations to give to charity, but they don't go that far.

Besides, it's disanalogous in all sorts of ways. Competence, for one.

Here's a different case: suppose you get some horrible disease that renders you (a) incompetent to choose (suppose sever brain damage), (b) in serious pain, and (c) terminal -- you have, at best, a few months. Suppose further that you have left no advance directive, and your wishes are unknown. Finally, suppose that you get some illness that would kill you more quickly if not treated -- say, pneumonia. Is it murder not to treat you?

[hilzoy]

'severe', not 'sever', brain damage.

[Slartibartfast]

'sever' brain damage would be 'severe' by definition, no? Fatal, even.

Unless it's on Star Trek, and Dr. McCoy hasn't forgotten the so-simple-a-child-could-do-it brain-reconnection technique.

[Dantheman]

"'sever' brain damage would be 'severe' by definition, no? Fatal, even."

For humans, maybe, but not all species. See Ringworld for a counterexample.

[bleh]

IANABioethicist, but FWIW I'll point out that pneumonia long had the nickname "the old man's friend," and that, in practice, doctors and families quite often DO make the conscious decision to take action to end lives sooner than they might otherwise end by increasing dosages of IV morphine provided to treat pain to the point of fatal respiratory depression.

The implications for me are that the boundary between action and inaction, when it comes to medical treatment that, as noted upthread, is all about NOT allowing people to "die naturally," is thin indeed, and to my mind matters far less than consideration of possible outcomes (including both benefits and costs) of possible courses of action. Along those lines, not pulling someone out of a puddle, which would save a life at the cost of a little bit of muscular effort, is morally not far away from outright murder, while not providing expensive or rationed (i.e., nonzero opportunity cost) treatment to a fatally ill patient whose quality and quantity of life thereby would not be extended much if at all, thus incurring a certain high cost for an uncertain, limited, and possibly nonexistent benefit, would be much less morally clear.

[rilkefan]

"If it were, then every dime we spend on anything not strictly necessary for the preservation of our own lives would be murdering one of the people whose lives we could save by sending it (with the other dimes) to Oxfam."

If Oxfam uses the money efficiently, then I think you're wrong, based on my guess at the supposed ultimate ethics. What e.g. would you think from behind the veil of ignorance? I at least don't see an obvious stopping point before radical equality. (Aside: as I understand it the opposite is rather like trickle-down theory.)

In the brain damage scenario, the "severe brain damage" bit suggests to me that "murder" isn't applicable. If Oxfam used my dime to support hydrocephalics, then probably I'd want my dime back.

[emmanuel goldstein]

Hilzoy,

....Finally, suppose that you get some illness that would kill you more quickly if not treated -- say, pneumonia. Is it murder not to treat you?

I want to tentatively answer yes. Suppose that someone has a terminal disease and will die in five years. But then they develop TB now. I’d say it’s impermissible to refuse them treatment for TB now, merely because they’ll likely die of some other disease. For, surely the (likely) medical effectiveness of a procedure is determined by how good it is at curing/alleviating its target ailment in the patient, not how likely it is to make the patient a whole, healthy human being. If so, and if a particular treatment for the TB would be medically effective, it’s impermissible to refuse to carry it out b/c the patient will die anyway.

Now, if in that situation it’d be OK to treat the patient, then why wouldn’t it be OK to treat the patient who’ll die in (5 yrs. – 1day), and so on, to your preferred sorites cut-off point. (So, expected time of death is probably irrelevant.)

Whatever one replaces TB with, either it’ll be curable (at the time) or not. If it’s not curable, then no dilemma. If it is curable, I’m not seeing why the fact that the patient is near death from some other disease makes it impermissible to carry out potentially medically useful procedures etc. So, if one denies treatment for curable ailment A because the patient is going to die anyway from ailment B, one is culpably negligent. But if one’s culpable negligence leads to a reasonably foreseeable death, then that’s some form of culpable killing.

[hilzoy]

emmanuel g: I was thinking that two crucial facts were the unremitting suffering, combined with the loss of so much mental functioning that that suffering had to be unredeemed. (I think that if I suffer, now, I can alter my suffering by what I make of it. I can turn it into a blessing of some sort, or I can make it more of a curse than it ever had to be. In either case, I give it some meaning above and beyond itself, by what I make of it. But that takes some amount of thinking -- not lots and lots, not the sort you have to be an Einstein to engage in, thinking that's within the grasp of the vast majority of people -- but not thinking that's within the grasp of someone with severe brain damage. So her suffering will be suffering, pure and simple.)

The role of time does matter, at least insofar as the longer someone will live, the less grounds you have for certainty about the outcome. That a child who is born now with Tay-Sachs will die a horrible death after a horrible life is pretty certain. That someone with an incurable illness will die ten years from now, and that no relief will be forthcoming during that period, is a lot less certain, though there are things that could make it more so (e.g,, certain forms of massive damage to the body.)

I am not, of course, denying the Sorites. Still, in most real-life situations, I think that the thing to do with Sorites paradoxes is to recognize that you need to draw a line. This is what we do with drivers' licenses: of course there's nothing magical about a kid's sixteenth birthday (or whatever driving age is today), and of course she's not markedly less likely to be a good driver if we let her start driving one day earlier, or one day before that... etc. But since letting her start driving at birth would be a bad idea, we draw a line.

[Anarch]

I am not, of course, denying the Sorites.

I do, but only because I regard the underlying binary logic as inapplicable. [Removing an object from a pile decreases its "pileness" by a fraction and increases its "not-pileness" by a similar fraction; it's just that we tend to regard extremal cases as definitive while casually disregarding the blurry area in the middle.] When we have some external need, e.g. the drawing of a legal line to distinguish a driving age as per your example, we can draw the line as you said, somewhere in the middle where it seems reasonable to do so. As you said, the line's necessarily arbitrary but the key point IMO is to acknowledge 1) precisely that it is arbitrary and 2) that we're drawing this lines for extrinsic reasons, often barely related to the issue at hand before 3) moving on to more important matters.

I'm not entirely sure what this has to do with the topic, but since I've got nothin' on that count anyway...

[Harald Korneliussen]

If this woman genuinely wants treatment, and this is the USA, right? She even pays personally for her treatment (through insurance). No, the hospital is not obliged to give it to her, nor is the insurance company I suppose, that depends on the contract, but if either refuse, prudent people should shun them.

And also, by the same reasoning refusing to save a life is not murder, prolonging a painful life is not torture.

[john miller]

Harold, FWIW, in most of these cases the patients are not able to pay for it, even if they have insurance. Insurance companies have their own criteria and will frequently stop paying for treatment that is futile or with no prospect for improvement of the patient's condition. Exceptions are made for treatment such as hospice, or purely palliative care (pain medications, etc.)

And this brings us back to one of hilzoy's points about appropriate use of resources, including financial, that a treating facility has available.

I think in this case, one of the factors in the facility deciding to keep the patient is that at least one doctor felt that there was reason to believe that the patient may recover, after a thorough examination of the patient.

I, quite simply, am very glad I am not placed in a situation where I would have to make the determinations necessary.

[DaveC]

Thanks hilzoy, good explanation of the issues. Wonder if there are not-exactly-hospice, not-exactly-hospital facilities that could handle this sort of case.

[Fred Sutherland]

The sister's complaint indicated overmedication for pain by the doctor. This is supposedly keeping her sister from being competent. Why doesn't she refuse this treatment? Then her sister would, according to her, become competent to speak for herself. Since "competent" patients will get whatever treatment will prolong their life, why doesn't the sister take action to ensure that this competence can show forth?