What health care providers know: a taxonomy of clinical disagreements.
Consider the following case: Horace Johnson is a forty-year-old, wheelchair-bound patient who has been suffering for the past ten years from type 2 diabetes mellitus. He has wet gangrene on his fifth toe. He doesn't visit the outpatient clinic for care of his diabetes and infection as he is scheduled to. The infection is so severe that his physician, Dr. Garcia, concludes that the toe cannot be saved and that if it is not amputated, Mr. Johnson could die. Mr. Johnson has been seen by a psychiatrist, who finds him eccentric but believes that he has no evidence of mental illness and must therefore be declared competent to make his own health care decisions. (1)
What is Dr. Garcia to do in this case? Given that Mr. Johnson is competent, almost everyone will agree that she cannot hospitalize him against his will. (2) Competent patients have the right to make choices about their own care, and their clinicians are, other things being equal, duty-bound to defer to their wishes concerning treatment options, even if doing so is not best for the patient. (3)
Even so, we might feel that Dr. Garcia is well positioned to tell Mr. Johnson what is best for him. She sees--as surely we all do--that choosing to die rather than lose a baby toe is bad for Mr. Johnson. In this case, we might be tempted to say that Dr. Garcia knows best. Even if she cannot make Mr. Johnson do anything, she can try to persuade him to have the toe amputated, confident in her assessment of what is best for him.
But is that really so clear? For who is Dr. Garcia to say that Mr. Johnson is wrong to prefer death to living without his baby toe? Perhaps, given his view of life and the importance of bodily integrity, dying with ten toes is more important than living with nine. If that is what Mr. Johnson thinks, are we still so sure that Dr. Garcia knows what is best for her patient?
I will not try to directly answer these questions on behalf of Dr. Garcia. Instead, I will develop a model for understanding disagreements in a clinical setting that illuminates the different types of disagreement that can take place between a health care provider and her patients (or more broadly, her patients' families). In doing so, I hope to give health care providers some guidance in understanding their role in such situations. More specifically, I aim to offer insight into what clinicians can and cannot plausibly be said to know about what is best for their patients, particularly in cases of disagreement.
I do so by developing a taxonomy of clinical disagreements. I maintain that this taxonomy helps us to see that health care providers can legitimately lay claim to knowing what is best for their patients in many cases. It reveals that, in some cases, clinicians can lay claim to having special expertise regarding treatment options, while in other cases, they can lay claim to knowing what is best for their patients, even if they cannot do so in their role as clinicians.
The second point depends on distinguishing between the idea that clinicians often do not have special expertise regarding what is best for their patients, and the idea that they do not know (or are unlikely to know) what is best. These ideas are not the same. Moreover, we cannot infer the second from the first, for reasons I discuss below. Keeping these two ideas distinct can help someone like Dr. Garcia decide how to proceed in a case like that involving Mr. Johnson.
But beyond the inherent interest and, I hope, helpfulness in providing a taxonomy of clinical disagreements, this paper has a subsidiary aim: to combat a kind of skepticism about the role of the clinician in determining patient care that can be found in the work of Robert M. Veatch. Over the past ten years, culminating in his latest book, Patient, Heal Thyself: How the New Medicine Puts the Patient in Charge, (4) Veatch has argued that the model of the physician as someone who, in a clinical setting, "knows best" is mistaken. (5) Rather, Veatch maintains, "the new medicine" will make the "patient ... the dominant or primary decision maker, leaving the physician in a much more derivative or secondary role." (6)
Veatch gives two reasons why this is so. First, other things being equal, competent patients have the right to make decisions about their own care. This is an uncontroversial and commonly accepted idea. What interests me, and what my taxonomy directly addresses, is Veatch's second reason.7 According to Veatch, every clinical decision involves making a value judgment concerning what is best for the patient. (8) But in almost all cases, claims Veatch, clinicians cannot legitimately lay claim to knowing what is best for their patient. As Veatch puts it, "The new medicine rejects the old slogan, 'Doctor knows best.' It will turn to others to make virtually all the critical choices." (9) The upshot, for Veatch, is that in almost all cases, a clinician must defer to her patient's conception of his own best interests.
My subsidiary aim, then, is to show that Veatch's conception of what clinicians can--or, more precisely, cannot--know about their patients' well-being is overly skeptical and would have us unduly limit the role of the clinician in determining treatment. I do not dispute Veatch's idea that all clinical decision-making regarding treatment options involves making value judgments at some point. Rather, my claim is that even if this is right, there are still many cases where clinicians can legitimately lay claim to knowing what is best for their patients and so, within the constraints demanded by the need to respect patients' autonomy, may guide treatment decisions.
I want to start by laying out three scenarios that will serve as the raw material for my taxonomy. We already have the case of Dr. Garcia and Mr. Johnson on board; I want to use some variations on it to present the kinds of disagreements I am interested in. Although these scenarios revolve around disputes concerning what to do about Mr. Johnson's gangrene, the lessons I draw from them can apply fairly straightforwardly to disagreements in other kinds of cases. (10) In all three scenarios, Mr. Johnson refuses to have his toe amputated. What separates them are his reasons why.
In scenario one, Mr. Johnson refuses treatment, citing his belief that the recommended treatment--amputating his toe--will not save his life. In fact, he thinks that he will get better without any medical intervention. In scenario two, the reason Mr. Johnson gives for refusing treatment is his belief that God will cure him without any medical intervention. And in scenario three, he refuses treatment because he is deeply attached to having an "intact" body, which includes having ten toes. He would rather die with ten toes than live with nine.
All of these scenarios share a competent patient who refuses the medically indicated treatment. Medically speaking, these are clear-cut cases--the treatment is likely to succeed in returning Mr. Johnson to his normal quality of life, while failure to treat is likely to result in death. So we can say with confidence that, medically speaking (I say more about what I mean by this below), Mr. Johnson is making a bad decision in each case. But since he is competent, he has the right to make a bad decision. All clear and, I hope, uncontroversial so far.
Ends versus Means
As I suggested above, these scenarios have important differences. First, consider the difference between scenarios one and two on the one hand, and scenario three on the other. In the first two scenarios, Mr. Johnson and Dr. Garcia share the same goal--a gangrene-free patient. (11) What they disagree about are the means to get there. Mr. Johnson opts for something other than the indicated treatment on the grounds that the alternative will make him gangrene-free. Dr. Garcia disagrees with that judgment while sharing Mr. Johnson's goal: she believes that the alternative treatments will not make Mr. Johnson gangrene-free. The first two scenarios, then, are cases of what I will call means-end disagreement: given a shared end (making the patient gangrene-free), a disagreement crops up concerning how best to achieve it.
The third scenario is different. Here, we might imagine Mr. Johnson agreeing entirely with Dr. Garcia about how best to rid himself of gangrene. The disagreement arises when we ask him if he wants to be gangrene-free, but without his baby toe: he says no. So what we have in scenario three is a disagreement about what goal we think Mr. Johnson should have. Mr. Johnson has the goal--and thinks he should have the goal--of dying of gangrene with his baby toe intact, while Dr. Garcia thinks Mr. Johnson should have the goal of living gangrene-free without his baby toe. The third scenario, then, is a case of end disagreement: Dr. Garcia and Mr. Johnson disagree about what Mr. Johnson should be aiming for. So when a clinician finds herself with a patient who disagrees with a proposed course of action, she can ask whether she is dealing with a dispute about ends or one about means.
While the distinction between means-end disagreements and end disagreements is quite simple, applying it in particular cases might not be. This is because end disagreements can masquerade as means-end disagreements. Consider a patient with body integrity identity disorder who wants her doctor to surgically remove her leg on the grounds that she is a legless person trapped in a legged person's body. The physician might say, "I just want you to be as healthy as you can be--physically and psychologically." The patient might respond, "I couldn't agree more, doctor. We want the same thing." And in some sense they do. What they disagree about is how to get there: losing a leg (the patient's preferred method) versus treatment to get over the desire to lose one's legs (the physician's preferred method). As such, this might seem to be a straightforward example of a means-end disagreement. But I am inclined to say that it is not a means-end disagreement at all. The disagreement is about what constitutes physical and psychological health; "health" means something very different to each of them. So what looks like a means-end disagreement is really a disagreement about ends. (12)
Medically Assessable versus Nonmedically Assessable
Contrasting the first scenario with the third develops another distinction for understanding clinical disagreements. The contrast I am interested in has to do with whether the dispute in question is medically assessable. What does this mean?
Imagine what a clinician might say to the patient in the first scenario who thinks that amputation is ineffective when it comes to treating the disease. Here, at least, is what I would say: "You claim that the medically indicated treatment is ineffective. But there is plenty of empirical evidence that this treatment is very effective. Of course, that doesn't make it a sure thing. But the evidence shows that you're simply wrong to maintain that this treatment is ineffective." We could imagine providing precise statistics to back this claim up. Moreover, a clinician could explain why this treatment is effective. She could give what we might call the medical reasoning that explains the treatment's efficacy. Medical reasoning is, simply, reasoning in light of the stock of knowledge that constitutes modern biomedicine: physiology, pharmacology, anatomy, pathology, and so on. In the case of many effective medical treatments, we have not only outcomes studies that substantiate the treatments' efficacy, but we also have a massive theoretical apparatus that explains why these treatments, and not others, are effective. This puts us in a very strong argumentative position. Of course, we might not convince Mr. Johnson. But we know that we stand on very firm medical ground when we say that amputation is the right course of action given his goal. In short, we have, in this case, a medically assessable disagreement: the issue at hand can, in principle, be resolved by appeal to medical resources.
Now consider the third scenario, in which the dispute is about ends. I suspect that, absent special circumstances, we all think Mr. Johnson's decision to die fairly soon with his baby toe rather than live much longer without it is a bad decision. But what would we say to him? Here is what we would not say: "I understand that you think the right thing to do here is to refuse treatment on the grounds that it is better, in your view, to live a short life with your baby toe than a longer life without it. But it turns out there is ample empirical evidence that this is not true--studies show that, other things being equal, it is better to live longer without your baby toe." The problem is not just that there are, in fact, no studies that answer the question, nor that there is currently no mode of medical reasoning that shows that the patient's goal is the wrong one. The point is that there could be no such studies or mode of medical reasoning: the dispute here turns on an issue that cannot be settled by appeal to medical findings. This is because the dispute concerns the value of the patient's goal. We want to say that valuing having all ten toes so much is wrong. Mr. Johnson disagrees. No empirical findings can settle this question, since empirical investigation cannot directly discover what matters in life. (13)
Of course, empirical investigation might tell us that given what matters to us, we should pursue x and avoid y, perhaps because x promotes health while y is poisonous. In this kind of case, we can say that empirical findings indirectly discover what matters in life. They can furnish an answer to the questions "What matters in life?" or "What is valuable?" only on the assumption that we have some prior specification of what matters. But without that prior specification, empirical findings alone cannot tell us what matters in life. (14) There is also no mode of medical reasoning that will resolve the question, since the resources required to answer such a question extend well beyond (indeed, perhaps do not even make contact with) the realm of modern medicine--they are squarely within the purview of moral philosophy.
In their role as medical professionals, then, clinicians have no special expertise in this matter: unlike with the dispute in the first scenario, they have no distinct ownership of the problem in virtue of being doctors, nurses, or researchers. There are no distinctly medical methods for settling this question, and there could not be: what is at issue in scenario three is beyond the purview of medicine and science. Scenario three, then, is an example of a nonmedically assessable disagreement.
Degrees of Medical Assessability
It is worth pausing for a moment to see that the space of medically assessable disputes is much larger than it might first appear. For what makes something a medically assessable disagreement is whether the dispute can in principle be settled by medical science. And this means that medically assessable disagreements will include cases where we have only partial medical knowledge.
Consider a variation on the first scenario. Mr. Johnson not only has ill-formed views about the dangers and efficacy of amputation, but he also believes that some other form of treatment, which is not medically indicated, is more effective: he wants to pursue a "natural" method of treatment for his gangrene. What might we say to him? Notice that we cannot say just what we have said to the person who simply denies the efficacy of amputation. Mr. Johnson now has an additional belief--namely, that some other, "natural" treatment will work. The problem is that a study showing the ineffectiveness of a "natural" gangrene cure is unlikely to exist because "natural" gangrene cures are not used. So we would not be able to say, "There is plenty of empirical evidence, in the form of research studies, showing that the natural mode of treatment you want to pursue does not cure gangrene."
Nonetheless, the dispute between the medical team and the patient remains a dispute about a purely empirical matter. Even if there are no studies that have taken up the question of whether a particular "natural" mode of treatment--say, a diet of uncooked vegetables--is effective for gangrene, the hypothesis would be, in principle, testable. Studies could settle the matter. And furthermore, even though the studies do not exist, we are not at an impasse! For there is still ample empirical evidence that this "treatment" will not work: given what we know about the human body and vegetables, we can show, via a process of medical reasoning, that it would not work. So our dispute, while not medically assessable by pointing to particular studies, is nonetheless assessable on medical grounds. (15) The point is not that we will be able to convince Mr. Johnson of the foolishness of the alternative treatment--we might not. The point is that we are still squarely within the realm of a medically assessable disagreement, even if we cannot muster the same argumentative resources as we can in the case of the person who simply denies the effectiveness of amputation.
Now consider scenario two, where Mr. Johnson refuses treatment on the grounds that God will cure him. In this case, we cannot appeal to medical reasoning to bolster our conviction that Mr. Johnson is wrong. Why? Because God is not a part of medical ontology. Medicine does not deny the existence of God nor the efficacy of God in treating the sick. Rather, medicine has nothing to say about it: medical ontology trades in proteins, enzymes, sugars, cells, and so on. No mode of medical reasoning can appeal to God's curative properties and remain an example of medical reasoning (it is, perhaps, medico-theological reasoning!). A physician cannot lay claim to special expertise in the case of a treatment that depends for its (supposed) efficacy on agents that are not a part of the medical ontology at all.
As such, we might think that scenario two is an example of a nonmedically assessable disagreement. But that is not right. Even if we admit that the proposed treatment in this case is outside the scope of medical reasoning, we might nonetheless think it is empirically assessable. And here everything turns on how we understand Mr. Johnson's claim. Suppose his claim is something like the following: "God generally cures people with my kind of gangrene without amputation. Therefore, there's a very good chance he'll cure me." Let's call this the "God saves!" reading of scenario two. If this is the claim, then even leaving aside the point about medical ontology, we have an empirically assessable claim: the patient is positing some kind of law-like relationship between having gangrene and being cured by God. And now we can bracket questions about how such a cure is supposed to take place and focus on the outcome--is it true that, in general, people with advanced gangrene are cured without amputation? If the answer is "no," then it follows that they are not cured by God (since they are not cured at all). When someone claims a law-like relationship between treatment y and outcome x, we have an empirically assessable dispute, since we can test whether x generally follows from y, no matter how farfetched y seems to be. While the dispute is not medically assessable via medical reasoning, it is still medically assessable via empirical testing.
The upshot is this: within the class of medically assessable disputes, there are varying kinds of medical assessability. There are disputes that are medically assessable via both medical reasoning and empirical testing, and there are disputes that are medically assessable via either medical reasoning or empirical testing, but not both. (16) The absence of either empirical evidence through testing or a line of medical reasoning concerning a course of treatment does not necessarily make disputes about that treatment nonmedically assessable.
Combining the Distinctions
With this understanding of the distinction between medically assessable disagreements and nonmedically assessable disagreements in hand, we can now combine this distinction with the distinction between means-end disagreements and end disagreements. What we have so far fills out two of the quadrants in a two-by-two matrix (Table 1).
In the upper left quadrant are medically assessable means-end disagreements, and in the lower right are nonmedically assessable disagreements about ends. The lower left quadrant--the medically assessable, end disagreement quadrant--is empty, reflecting the conclusion that there is no such thing as a (directly) medically assessable disagreement about ends. Disagreements about ends are disagreements about what matters or what is valuable, and those claims are not medically assessable. But what about the upper right quadrant--the space for clinical disputes about how best to achieve a shared end, but where the conflicting claims are not medically assessable? None of the scenarios so far discussed fall into this quadrant.
Consider scenario two again. But now suppose that Mr. Johnson is not claiming that God generally cures gangrene patients. Suppose all he is claiming is that God will cure his gangrene. Let's call this the "God saves me!" reading of scenario two. There is a sense in which this dispute is medically assessable: we can wait to see if the patient dies. If he does, we know that he was wrong. (17) So experience will, in all likelihood, tell us who was right. But this is not helpful to Mr. Johnson or Dr. Garcia. In fact, there is in principle no way to empirically resolve the dispute in a way that is helpful to anybody. Not only do we lack a mode of medical reasoning that might convince Mr. Johnson he is wrong--he does not share our medical ontology--but we also lack, and must necessarily lack, any data that can convincingly show him he is wrong. We might say, "Look: all these other people also thought God would cure them, and now they're dead. What's the difference between them and you?" (18) But inasmuch as he is convinced that God will save him, such reasoning will not resolve the matter. Our strategy of trying to show Mr. Johnson that there is no relevant difference between his case and others only suggests that he is looking at things the wrong way. If that is right, then we have an example of a means-end disagreement that is also a nonmedically assessable disagreement. The dispute is about the means to a shared end, but the dispute (God's curing this particular patient) is not medically assessable.
Health Care Providers' Limits
I think this way of looking at things is inherently interesting and, for reasons that will become clear below, possibly helpful to health care providers like Dr. Garcia. But as I mentioned at the start, the taxonomy also puts us in position to assess Veatch's views about the role of the health care provider in determining what is best for her patients. Recall that the idea we are interested in here is that the role of the physician in twenty-first century medicine is severely limited by the health care provider's lack of knowledge concerning her patients' well-being. Here's how Veatch explains the situation:
It is increasingly clear ... that there is a huge difference between being medically well-off and being well-off considering all spheres of life. If the physician's task is to focus on maximizing the patient's medical well-being, he or she must realize that rational patients usually do not want their medical well-being literally maximized (at least if that comes at the price of sacrificing goods in other spheres of life).... This poses an enormous problem for the medical professional role. (19)
Why does this pose an "enormous" problem? Veatch here skewers the physician on the horns of a dilemma. If the physician's end is only the (maximal) medical well-being of her patient, then she is not in a position to determine what is best for her patient, since "what is best" will include factors that go well beyond medical well-being. (20) But if the physician's role is to treat "total well-being"--which encompasses more than merely medical well-being--then, insofar as she makes judgments about what is best for her patients, she becomes "imperialistic." (21) This is because physicians "are really not equipped to promote our spiritual, legal, aesthetic, financial, or mental well-being." (22) Veatch's suggestion is that on a more expansive conception of well-being, physicians cannot lay claim to knowing what is best for their patients.
But according to Veatch, even if we limit ourselves to questions of medical well-being, physicians cannot lay claim to knowing what is best. Why? Because there are a myriad of medical goods that physicians are in no position to know how to weigh or trade off against each other:
By the mid-twentieth century, the typical physician had a rather simplistic view about the nature of medical good for his (or occasionally her) patients. The medical good was increasingly equated with preserving life. We had discovered antibiotics, we were aggressively pursuing polio, and we were still focused on acute illness that threatened life. The goal was to preserve life as long as possible. ... Patients, however, had a much more complex view about the medical good (as did physicians of earlier centuries). They sometimes were committed to preserving life, but also desired cure of disease, relief of suffering, and, increasingly, promotion of continued good health. (23)
As the century progressed, claims Veatch, and physicians and patients were confronted with cases like that involving Karen Ann Quinlan, "lay people realized that even within the medical sphere, there were many disparate goals that one could choose to pursue." (24) Furthermore:
There was often conflict among [these goals] and there was no definite method for balancing among these competing claims when they came into conflict. Relief of suffering might come at the expense of preserving life; preserving health might come at the expense of increasing risks for certain pain and suffering. (25)
The upshot, once again, is that:
[The physician has no] special expertise in balancing among these competing claims--even within the medical sphere. Being an expert in medicine does not imply one is an expert in the way the patient should trade off one medical good against another. (26)
Veatch's conclusion, as given in the title of one of his papers, is that in many cases, "doctor does not know best" when it comes to her patients' well-being, medical or otherwise. (27)
I agree that in many cases, and for just the reasons Veatch identifies, physicians cannot lay claim to knowing what is best for their patients. Nonetheless, I think Veatch overstates the degree to which this is true, and he overstates the problem in two ways, which map onto the two distinctions set out in the taxonomy. First, the distinction between means-end disagreements and end disagreements suggests that Veatch overstates the degree to which clinicians, in their role as clinicians, cannot determine what is best for their patients. Carefully distinguishing between decisions concerning means and those concerning ends reveals that there are numerous decisions where clinicians do know best. Second, the distinction between medically assessable disagreements and nonmedically assessable disagreements suggests that what is relevant to the question of the clinician's role in many instances of clinical decision-making is that they have no special expertise as health care providers to determine what is best but not, as Veatch would have it, that they are unlikely to know what is best. That is, by eliding the ideas that clinicians have no special expertise in many matters and that they do not know what is best in those matters, Veatch erroneously concludes that in many cases, clinicians lack knowledge of what is best, when in fact all they lack is knowledge in their role as medical experts.
Veatch maintains that in most cases doctors cannot lay claim to knowing what is best for their patients, even when dealing with medical well-being alone. This surely is often right, so long as the disputes we have in mind are end disagreements. For here, as we have seen, the dispute turns on questions about what matters in life, and, as Veatch rightly notes, physicians have no special expertise concerning this. But many disputes are means-end disagreements, not end disagreements. And with means-end disagreements, physicians can often lay claim to having special expertise about what is best for their patients. This is true whether we are working within the framework of medical well-being or total well-being: once the patient has fixed her end--whether it be the medical end of reducing suffering or the nonmedical end of playing the trumpet again--the physician can often, rather unproblematically, lay claim to knowing what is best for the patient whose medical problems (a) stand in the way of the patient achieving her end and (b) can be alleviated using a fairly well-tested method of treatment. Counter to Veatch's pessimism, then, there is still substantial room for a twenty-first century physician to determine what is best for her patient, so long as that determination takes place within the context of a shared end.
This is consistent with Veatch's position that clinical decision-making always involves making value judgments. Even means-end disagreements involve making value judgments. The point is only that in these kinds of cases, the dispute is not about those judgments. For in means-end disagreements, there is agreement on the question of ends, and it is in setting ends, medical or otherwise, that value judgments play an ineliminable part. So Veatch is right to claim that all clinical decision-making involves making value judgments, but it does not follow from that claim that doctors are in no better position than patients to make health care decisions. The decision--and the possible dispute that might arise in the face of the decision--is about the best means to a shared end. And in these cases, the physician can often lay claim to knowing what is best.
However, the distinction between medically assessable disagreements and nonmedically assessable disagreements shows that there is a limit to clinicians' expertise, even when we have a dispute about the best means to a shared end (in other words, a means-ends disagreement). The dispute might be about the patient's wish to pursue a means that not only has not received empirical scrutiny, but could not receive such scrutiny. What should we say about such cases? This brings us to the second point above and the importance of distinguishing sharply, as Veatch does not, between lacking expertise and lacking knowledge. The right thing to say about at least some cases where we have a nonmedically assessable means-end disagreement is that clinicians have no special expertise. To return to the "God saves me!" example, one's training as a clinician does not put one in a special position to pronounce on whether God will save this patient--that is a claim that is outside the space of medical and empirical assessability.
But note that it does not follow from the fact that clinicians have no special expertise on such matters as clinicians that they often do not know what the right thing to do is. Indeed, the "God saves me!" scenario nicely illustrates that the clinician may indeed have that knowledge. Confronted with a patient like Mr. Johnson who turns down amputation on the grounds that God will save him, I am sure most of us would confidently say that we know that this is a bad decision. Of course, a lot depends on what we mean by "know" here--I grant that we cannot have certainty on this matter; but we almost never mean that we are certain when we say that we know something. I do not want to get into messy issues in epistemology. All I want to block is the idea that with a nonmedically assessable means-end disagreement, the reason a physician like Dr. Garcia should be less forceful in putting forward her view on what ought to be done is because, unlike in a medically assessable disagreement (which is, if I am right, necessarily a means-end disagreement), she does not know what is best as a doctor. Therefore, any opinion she expresses on how to proceed in a nonmedically assessable means-end disagreement will not carry any medical authority with it. However, insofar as it is correct, it will carry the authority of an ordinary, sensible person who can see that it is crazy not to get a toe amputated on the grounds that God will cure your gangrene.
Seeing this distinction--between having no special expertise with regards to some claim x and not knowing whether claim x is true/false--at play in means-end disagreements allows us to understand the kinds of disputes Veatch is primarily interested in--namely, disagreements about ends--differently. Although he does not say so, Veatch's analysis clearly depends on the idea that in most cases, doctors do not know which ends are best for their patients. But as we have seen, the critical point is not whether doctors know what is best for their patients, but whether they can lay claim to knowing what is best as a doctor. We know that they often can lay claim to such knowledge when it comes to medically assessable means-end disagreements. We know that they cannot when it comes to nonmedically assessable means-end disagreements. But this is all still perfectly consistent with them knowing what is best for their patients. The key point is that whatever clinicians know about what a patient ought to do in these kinds of disagreements, they know not by virtue of their medical expertise, but simply as normal, psychologically healthy people.
But now we might ask: Why does this matter, given that on either analysis the upshot seems to be that the clinician should back off? Either she does not know what is best at all, or she does not know it as a clinician. It matters for two reasons. First, by sliding between the claims that physicians do not know what is best for their patients and that they do not know what is best for their patients as doctors, Veatch's analysis invites an overly skeptical attitude regarding health care providers' ability to make good judgments about what is good for people. On Veatch's analysis, health care providers should have no firm opinions concerning what is best for their patients. They should also be skeptical of their own assessments of what is good for their patients. On my analysis, health care providers should resist Veatch's form of skepticism and hold firm (but not hubristically or dogmatically so) to their conceptions of their patients' good, while recognizing that in many cases their confidence in what is best is not rooted in their medical expertise.
There are practical implications of seeing things this way. On Veatch's account, the physician is handcuffed because she does not know how to advise the patient to proceed. As such, the best course of action, as Veatch suggests, is to do one's best "to assist the patient in pursuing the patient's understanding of the patient's interests," no matter how mistaken that understanding might be.28 But on my way of seeing things, the physician does not suffer from a lack of knowledge (at least not as often as Veatch would have us believe). She merely lacks knowledge as a physician. And this presents a distinct challenge for the physician: she lacks authority as a health care provider to impart the knowledge. This is a problem because the default assumption for most patients, I imagine, is that the advice health care providers offer is offered in the role of medical expert. As such, the physician must be careful to not take advantage of her role as an authority on medical issues when imparting advice or ideas concerning matters beyond the purview of medicine.
But keeping in mind that she often does know what is best for her patients, various options are available to her, and not simply as someone who happens to know best, but as someone who, while perhaps not having any special expertise on the matter under dispute, nonetheless has a special responsibility to the patient as a clinician. First, she could engage the patient in a discussion about what course to pursue in a way that carefully delineates between her roles as someone who has a tremendous amount of medical knowledge and someone who is specially tasked to look out for the patient's well-being. (29) She might also reach out to others who are, perhaps, in a better position to engage the patient in a substantive conversation about what is best for her--friends, family, social worker, or priest, for example--in the hopes they will bring the patient around to a better course of action. It's not obvious that either strategy is acceptable on Veatch's analysis. (30)
My point here is not that Veatch is wrong to think that the physician's duty is no longer simply to act in her patient's best interests. The need to respect patient autonomy even when the patient is making a poor decision is enough to show that Veatch is right on this point. What I am trying to show is that even before we raise issues concerning patient autonomy, Veatch paints too skeptical a picture of the role of the clinician in disputes about treatment. For the distinctions that I have highlighted provide the clinician with a model for how to conceive of disputes they might find themselves a part of, which, in turn, can help them decide what to do beyond simply following the patient's own conception of what is best. When confronted with a patient who refuses treatment, the first thing a clinician can try to determine is whether the dispute is about the best means to a shared end or about what end should be pursued in the first place. To the extent that it is the former, the clinician can now ask whether the dispute is medically assessable. To the extent that it is medically assessable, and depending on the precise nature of dispute, the clinician can be confident that she has expertise on how best to proceed. Indeed, this is precisely part of what it is to be a clinician: to know how best to proceed in the face of various maladies, given a certain, shared end. In such cases, the clinician should not be afraid to conceive of herself as an expert who is in the best position to determine the right thing to do. (31)
In the face of nonmedically assessable disagreements, things are different. Here, the key insight is not that the health care provider should adopt an attitude of skepticism toward her own ability to know what is best for the patient. Rather, what matters is that she recognizes that the dispute cannot be answered by medical science alone and that she has no special expertise on the matter. To be truthful with her patients, the health care provider must not conflate her position as a medical authority with her position as someone who has special responsibilities for the patient's well-being and may very well know what is best for him. In no case, however, should skepticism about what is best for the patient in the face of disagreement be the default position of the twenty-first century clinician.
Thanks to Daniel Brudney, Emily Carroll, Anita Chikkatur, Jason Decker, Clara Hardy, Christopher Heurlin, Annette Nierobisz, Alan Rubenstein, and Kim Smith for very helpful comments on earlier drafts of this paper. Thanks as well to audiences at the University of Minnesota Department of Bioethics, the Mayo Clinic Program in Professionalism and Bioethics, and the St. Olaf Philosophy Department for helpful feedback on presentations based on this paper. Many of the ideas in this paper are indebted to discussions that took place during my two years as a participant in the weekly medical ethics case conference at the University of Chicago Hospitals. Finally, thanks to the editors at the Hastings Center Report for very helpful suggestions about how to frame the paper.
(1.) This case is based on a case in R.M. Veatch, A.M. Haddad, and D.C. English, Case Studies in Biomedical Ethics (New York: Oxford University Press, 2010), 125-26.
(2.) Ibid., 126. This is what Veatch, Haddad, and English imagine Dr. Garcia considering.
(3.) This is, of course, an articulation of the idea that health care providers must respect their patients' autonomy. Exactly what this means and entails is up for debate. My focus, however, is not on this point, so we need not enter into a detailed discussion of what autonomy is or why it matters.
(4.) R.M. Veatch, Patient, Heal Thyself: How the New Medicine Puts the Patient in Charge (New York: Oxford University Press, 2009). Veatch's book was reviewed in this journal by J. Lantos, "Veatch Hates Hippocrates," Hastings Center Report 40, no. 1 (2010): 46-47.
(5.) Veatch, Patient, Heal Thyself, 33. Veatch's discussion is about the role of physicians in particular. What he says applies to other clinicians (nurses and nurse practitioners in particular). As such, I will continue to move between discussing the role of the twenty-first century clinician and sometimes, following Veatch for ease of exposition, the twenty-first century physician.
(6.) Ibid., 14. Indeed, Veatch considers but ultimately rejects the idea that, given this shift, we ought to stop calling patients "patients"; Patient, Heal Thyself, 17.
(7.) Veatch also identifies a third consideration. He thinks considerations of justice also constrain clinicians' ability to always do what is in the patient's best interest. Decisions made in the clinician-patient relationship inevitably affect others, in some cases very seriously. Even if we remain agnostic concerning which normative ethical theory should govern our relations with others, the new medicine must recognize that the twenty-first century clinician is duty-bound to consider the effects on others of decisions made in a clinical setting and to admit that "there are times when the patient's interest must be sacrificed for the good of society or to fulfill duties to others"; Patient, Heal Thyself, 51. I set aside these ideas for the remainder of this paper.
(8.) Ibid., 4.
(9.) Ibid., 4-5.
(10.) Although the rich texture and fine details of real-life cases always make things more complicated. Since I am aiming to provide a model that captures a wide array of cases of disagreement, I have purposefully abstracted from the kind of details that a real-life clinician would need to pay attention to when dealing with a real-life patient or family.
(11.) Although I make no claim that they share this as a final goal. We can imagine that the gangrene patient wants to live only because she wants to outlive her middle-aged parrot. Moreover, some might be inclined to see "being healthy" as an unlikely candidate for a final end, inasmuch as "being healthy" does not give us a reason to live. As Bernard Williams would put it, the desire to be healthy is not a "categorical desire"--a desire that gives us reason to live. On the other hand, our attachment to being healthy can plausibly be seen as a basic, final end that we cannot help but care about as rational creatures (a "volitional necessity," as Harry Frankfurt calls it). Nothing in my argument depends on taking a stand on these issues. See B. Williams, "The Markropulos Case: Reflections on the Tedium of Immortality," in Problems of the Self (Cambridge, U.K.: Cambridge University of Press, 1973), and H. Frankfurt, Taking Ourselves Seriously and Getting It Right (Palo Alto, Calif.: Stanford University Press, 2006).
(12.) The general idea is this: disagreements about what constitutes achievement of an end that shares a mode of articulation (such as "being healthy") are end disagreements, while disagreements about what leads to a conceptually distinct end that the parties to the disagreement have in common are means-end disagreements. Thanks to Jason Decker for a helpful discussion on this point.
(13.) This idea is nicely expressed by Ed Pellegrino: "The choice of how we want to live our lives when we face serious illness, whether we want to reject the indicated treatment or run substantial risks of discomfort for even a small chance of benefit, are value decisions no one can make for us. The complexities of a 'good' decision are such that we cannot deduce them automatically from what may be a scientifically correct decision." E.D. Pellegrino, "Toward a Reconstruction of Medical Morality," American Journal of Bioethics 6, no. 2 (2006): 68.
(14.) It can tell us what tends to matter to people, and we might take this as evidence concerning what matters. But this is not a case of direct discovery. For a rich discussion about the relationship between nonnormative facts and normative principles, see G.A. Cohen, Rescuing Justice and Equality (Cambridge, Mass.: Harvard University Press, 2008), chapter 6.
(15.) Assuming, that is, that we have confidence in our medical knowledge. In some cases, we may well not: we might not know enough about the course of a disease to know whether a proposed form of treatment will work. Or we might be surprised to learn that a treatment we thought, via a process of medical reasoning, would not work in fact does. Such a case would demand that we reevaluate what we think we know about the human body, the disease in question, and so on.
(16.) Really, there are a series of interesting distinctions that could be made here. For, in addition to these kinds of medical assessability, we need to add the notion of degrees of medical assessability, which measures how much empirical evidence and how solid the available medical reasoning is with regards to a particular course of treatment.
(17.) Of course if he lives, it does not mean that he was right!
(18.) Notice that this is a very different kind of reasoning than that described above, where the patient sees himself as falling under a general law. Now we are trying to convince him that he should see his case as relevantly similar to other cases and not that, given that he thinks it is relevantly similar, he can expect similar results.
(19.) R.M. Veatch, "Doctor Does Not Know Best: Why in the New Century Physicians Must Stop Trying to Benefit Patients," Journal of Medicine and Philosophy 25, no. 6 (2000): 701-721.
(20.) Ibid., 704.
(21.) Ibid., 704.
(25.) Ibid., 705-6.
(26.) Indeed, Veatch goes on to suggest that physicians might "tend to make . . . medical value trade-offs atypically"; ibid., 706.
(27.) Veatch acknowledges that there are cases where physicians do know what is best (ibid., 713). On the face of it, case three seems to provide such an example: all things being equal, we can be quite confident that a person who refuses amputation to save his baby toe, even though it will cost him his life, is making a bad decision. And so we might think that in those cases where the physician does know what is best--however rare such cases might be--the physician's duty is, as it has traditionally been, to act in the patient's best interests. At this point Veatch adverts to the ideal of patient autonomy to explain why physicians should not act on their best judgment of the patient's interests and treat against the patient's will. So, according to Veatch, the full problem is this: (1) in almost all cases physicians cannot lay claim to knowing what is best for their patients; and (2) in those cases where they can, they nonetheless have a duty not to treat when a competent patient refuses treatment. I am only dealing with Veatch's argument for (1) in this paper.
(28.) Ibid., 701. This phrase is from the paper's abstract.
(29.) I thank Daniel Brudney for drawing my attention to the idea that while a clinician might have only common knowledge (that is, nonexpert knowledge) about what is best in a certain case, she might nonetheless have a special responsibility as a clinician to impart that knowledge. This also suggests that the role for clinicians in guiding patient care is greater than Veatch realizes.
(30.) Veatch might interject here and point out that my second suggestion is equally open to him: in the face of not knowing what is best for her patient, a physician might reach out to those whom she thinks have a good idea. But Veatch never suggests such a route, instead moving from the idea that the clinician has no special expertise about x to the claim that the clinician does not know x to the claim that, even apart from considerations of autonomy, it is the clinician's duty to help the patient pursue the patient's conception of what is best.
(31.) Of course, in the face of a recalcitrant patient, it might not be best, strategically speaking, to assert that one is the expert and (so) knows best. Nonetheless, the physician can proceed with confidence that this is in fact the case and turn to the question of how best to convince the patient of this.
Daniel Groll, "What Health Care Providers Know: A Taxonomy of Clinical Disagreements," Hastings Center Report 41, no. 5 (2011): 27-36.
Table 1. Medically assessable Nonmedically assessable disagreement disagreement Means-end Scenario one (rejects Scenario two (rejects disagreement amputation because it amputation because "God doesn't work), scenario saves me!") two (rejects amputation because "God saves!") End Scenario three (rejects disagreement amputation because of a desire to keep the body intact)
|Printer friendly Cite/link Email Feedback|
|Publication:||The Hastings Center Report|
|Date:||Sep 1, 2011|
|Previous Article:||No more larking around! Why we need male LARCs.|
|Next Article:||Gallows humor in medicine: medical professionals regularly joke about their patients' problems. Some of these jokes are clearly wrong, but are all...|