The duty to procure informed consent (IC) from patients before any significant intervention is among the pillars of medical and research ethics. The provision by the doctor of relevant information about treatment and free decision-making by the patient are essential elements of IC. The paper presents cases of IC where the free decision about treatment is not causally related to the information provided, and claims that such cases pose a difficulty parallel to that presented by the Gettier Problem in epistemology. In analogy to the original problem with the concept of knowledge, these Gettier-type cases show an indeterminacy in the concept of IC: we either need to add some explicit additional condition of causal connection between information and consent, or else we should understand the concept in a new way—specifically, since the practice of autonomy necessarily involves some consideration of the relevant information, we must understand free consent in a way that no longer refers to patient autonomy.
- Informed consent
- the Gettier-problem
Statistics from Altmetric.com
The duty to procure informed consent (IC) from patients before any significant intervention is among the pillars of medical and research ethics. The aim of this paper is to expose a hidden indeterminacy in the concept of IC by using an analogy with epistemology. A second aim relates to the association between IC and autonomy. Since the Belmont Report, the locus classicus with respect to IC, it has become somewhat of a truism that IC is justified by the respect owed to patient autonomy.1–3 This paper will contribute to the effort to rethink that association. If it does not hold, the justification of IC will need to be found elsewhere—most likely in either a libertarian-like notion of free choice or in the objective of the patient's well-being.
It is generally accepted that the meaning of IC requires the fulfilment of three necessary and sufficient conditions: that the patient be competent, that s/he is given adequate information about the proposed treatment and understands that information and that the consent expresses truly free choice. Much philosophical and legal discussion has debated the precise meanings and limits of each of these conditions, but it is generally agreed that if, under some best interpretation, all three conditions are fulfilled, then the requirement of IC has been successfully met.
Now consider the following two cases.
Case 1: Mr A sees an ophthalmologist for a persistent deterioration in the sight in his right eye. The doctor tells Mr A that he has a cataract and recommends surgery. He explains the nature of the procedure, the good chances of success and the potential risks. Mr A listens politely to the detailed explanation; he is an intelligent man and he understands perfectly well. He mumbles something about taking time to mull over it, thinking to himself that he would never let any surgeon touch his eyes, and goes home. In the evening Mr A plays poker in the local bar. Those who lose all their money can keep playing on the condition that they fulfil whatever the winner requests of them, should they lose again. Mr A feels strongly that his next hand will be a winning one and agrees to the terms. He loses again and is asked to undergo eye surgery. He returns to the doctor and declares, “I'll have that surgery!”
Case 2: Mrs B, whose heart failure has been progressing, sees a heart surgeon, Dr Fox, who suggests the prompt replacement of a leaking valve. Dr Fox explains at length to Mrs B the risks in her situation, the nature of the procedure, its prognosis, the risks of permanent anticoagulant therapy after surgery, and so on. Mrs B listens carefully and asks intelligent questions. After thinking about it, however, she tells Dr Fox that she'd rather let the disease run its natural course and is not interested in the surgery. That same night, Mrs B dreams that Dr Fox is her God-sent personal saviour. She returns to him the next day declaring, “On a second thought, Dr Fox, I do want that surgery!”
In both of these cases all three conditions for IC are fulfilled: the patients are competent, the doctors did a fine job of explaining the proposed treatments, which were well understood by the patients, who in turn gave their free consent. These then are supposed to be clear instances of IC. And yet an essential problem remains: the patients' choices, though free, are not the result of the medical information they received from their doctors. On the contrary, the information they received led them to refuse the proposed treatments; their ultimate decisions to consent were actually made despite the relevant information. What are we to make of such cases of IC?
Interestingly, these cases bear formal analogy to the Gettier Problem in epistemology. A classic theory of the nature of knowledge views knowledge as justified true belief, such that if (1) X has the belief that p, (2) p is true and (3) X is justified in believing that p, then X knows that p. Edmund Gettier was first to demonstrate cases where all three conditions hold, yet we would still be reluctant to admit as real cases of knowledge, concluding that ‘justified true belief’ is not a sufficient set of conditions for knowledge.4 An example would be the following. I am on campus; you believe that I am on campus, because you just saw my identical twin (whom you had never heard of) in the campus library. In this case (1) you have the belief that I am now on campus, (2) it happens to be true that I am indeed there and (3) you are justified in thinking I am there, since you just saw someone there who looks exactly like me. Can we be justified in asserting that you know that I am on campus? The obvious difficulty with cases of this type is that the justification for the belief is wholly accidental to that which makes it true.
The parallel problem here is that the consent is accidental to the medical information. Knowledge, whatever else it is, is true belief, and it is natural to expect that the truth of the matter (the fact of the matter in reality) be the grounds for our true belief. Similarly, it is natural to expect that the information provided by the doctor be in some significant sense the grounds for the informed consent. However, the problem in the Gettier cases is precisely that the truth is not properly related to the justification (reason) of the ‘knowledge,’ and similarly in our Gettier-type cases, the information is not operative in the justification (motivation) of the ‘IC.’ In other words, the de facto reason for the ‘knowledge’ in the former is not the relevant one for what would intuitively qualify as knowledge, and similarly the motivation for the ‘IC’ in the latter is not the relevant one for what would intuitively qualify as IC.
The Gettier Problem can in principle be resolved in one of two ways: either we accede that Gettier cases are not cases of knowledge and that a fourth condition—that presumably bridges the gap between the second and third conditions—is necessary; or we insist that such cases are knowledge after all, preserving the original definition of knowledge for the price of diluting its paradigmatic sense. These two general options are also available for the analysis of the presented cases of IC. We are faced with a question about the meaning of fulfilling the conditions for IC: does it simply require that adequate information is provided and free consent is given, or does it implicitly also require that the consent is the result of (the understanding of) the information? In other words, the question is whether the two conditions are independent or dependent (the dependency then being an extra condition). Now IC is not an a priori concept, and the conceptual question of its definition may indeed have more than one reasonable answer; which one we ultimately choose is likely to be determined by our normative considerations vis-à-vis that practice (just as we employ normative considerations in epistemology to determine the most suitable interpretation of ‘knowledge’). Whichever answer we choose, however, we face significant challenges.
It seems natural to maintain that a fourth condition is in fact implicitly assumed as part of our intuitive understanding of IC—why otherwise would we insist on ‘informed’ consent, as opposed to consent simpliciter, if the information plays no causal role? (See elaboration on causality below.) If such is our answer, then the contribution of our Gettier-type cases is in helping to make that extra condition explicit. The problem with that answer is twofold, however. First of all, it seems reasonable to argue that the cases of Mr A and Mrs B, though perhaps peculiar, do not pose any prima facie moral problem: the doctors provided the needed information and the competent patients gave their free acceptance—why worry any further? Moreover, if we were to act according to the more robust concept of IC, we would face a true puzzle as to what an authentication of the reliance of consent on the given information could possibly entail in practice. It would seem to require an obscure psychoanalysis of even the most competent patients to rule out ulterior motives for consent. But beyond the fact that the meaning of ‘ulterior’ here would by and large remain too vague, such policing of thought and motivation would seem to undermine the very spirit of the IC institution.
The second option then is to insist that the three conditions for IC are indeed sufficient, and that cases 1 and 2 are fully legitimate examples of IC. This solution is problematic in abandoning the (implicit but) intuitive requirement that the relevant information actually informs the consent, and it raises the further challenge that the standard ground for IC, respect for patient autonomy, must be revised or even discarded. The provision of accurate and reasonably comprehensive information, we are told, is mandatory for well-reasoned deliberation over treatment options, for the ability to act ‘in accordance with a self-chosen plan,’ (p121)2 for the authorship of the patient over his or her life, in short—for the practice of patient autonomy. But decisions arrived at through gambler's luck or presumed divine inspiration (as in cases 1 and 2), though free, have nothing to do with autonomous reasoning about one's medical options. (Remember that as far as autonomous deliberation over the information is concerned, the patients in our examples repudiated the possibility of giving their consent.) It might be argued that the principle of respect that grounds IC is meant just to provide people with the option of autonomous deliberation. But if this is the ground of respect, then it is not autonomy we respect in IC but a different (libertarian-like) idea of free choice, which encompasses the choice of whether to act autonomously or not. Gettier-type cases focus our attention on the specific condition of causative connection between information and decision as constitutive of autonomy in the context of IC. If we reject this condition, we ought to do so with a clear vision of what we are rejecting, for it is likely to have far-reaching ramifications in (medical) ethics in general.
Some fine points about the lack of causative connection between information and decision need clarification. Contrast first the cases above with the following cases:
Case 3: The doctor suggests to Mrs C the option of an invasive treatment that is likely to improve her medical condition but that also incurs non-marginal risks. After considering her options in detail, Mrs C concludes that the pros and cons are equal. She therefore flips a coin to determine her decision, and subsequently gives her IC.
Case 4: The doctor explains to Mr D that his morbid obesity endangers his longevity and recommends a partial gastrectomy. He provides the relevant information and suggests that Mr D gives it a good thought. When Mr D tells his girlfriend about the doctor's proposal, she opines that it will make him much sexier. Mr D yearns to be attractive in the eyes of his girlfriend and so returns to the doctor and gives his consent to surgery.
In cases 3 and 4, unlike in 1 and 2, the information is causative in reaching the decision to consent. Mrs C feels she reached a Buridan-type situation and so decides to solve it by flipping a coin. The flipping of the coin is itself the result of weighing the medical information. Similarly, while Mr D's decision is based on his desire to satisfy his girlfriend, her opinion, though expressing a different rationale from the doctor's, is indeed based on some real merits of the proposed treatment. Mr D's consent therefore indirectly refers to the medical information.
Even in cases 1 and 2, however, we still encounter the following challenge to the thesis of lack of connection between information and decision: given that the patient's decision is a competent one, once he was made aware of the information regarding treatment, his decision will inescapably be an IC, because whatever he decides, the information is perforce part of his considerations (since it now resides within the realm of his awareness). So the mere fact that a competent decision is reached under the condition of being aware of the information already provides the needed connection between information and decision, and therefore secures the decision as autonomous. It then all boils down to the relative importance the patient gives to different considerations: according to his values, he may care much more about his reputation among his poker buddies (for instance) than about the details of a proposed treatment. And of course none of this makes his decision any less autonomous or legitimate as an IC.
The Gettier-type cases purport to present a unique exception to this logic, for in them the patient's decision effectively treats the information as irrelevant. In the calculation of his preferences, the relative importance of the information is not just over-ridden by other options that rank higher for the patient, rather it is completely ignored. The medical information adds a zero to the calculation of his preferences. This is not because it really has no inherent importance (it obviously does), but rather because the patient's decision about the operation is reached by adherence to a certain element (eg, the personal—not professional—identity of the doctor) that is contingent as far as the inherent logic of the surgery goes, yet is de facto part and parcel of the concrete option to undergo the surgery in his particular case. In Gettier-type cases, consent to treatment is the result, not the conclusion, of a decision on a different matter. Owing to such a state of affairs the patient does not compute the medical information in his decision-making process. Mr A would do the eye surgery irrespective of the information about it, and Mrs B would do any minimally reasonable operation that Dr Fox would prescribe to her, whatever its merits. The information in those cases is thus not over-ridden by other considerations—rather, it is not calculated at all.
Now the question this entails is this: is total ignoring a form of consideration or of lack of consideration? But the more critical distinction is in fact a finer one, for purposeful ignoring is itself a form of consideration. Importantly, in cases 1 and 2 the ignoring of the medical information is incidental. How can a patient make a competent decision about surgery while ignoring its merits? As the examples show, by fixing on an element that is indeed glued to the surgery option but completely contingent to it—that is, has nothing to do with the essence of the procedure itself; thus the information about the surgery has zero relevance to the decision about it. This latter precisely defines a Gettier-type case. Think of the members of a jury who hear a claim by one of the attorneys—a claim the attorney was not allowed to make—and are ordered by the judge to completely overlook that claim in their deliberations. They are not autonomous when it comes to using that information when they deliberate. Now the obvious difference is that in cases 1 and 2 the patients are not ordered by any external agency to overlook the information; rather, it is the result of their own decisions. But that, we maintain, does not change the nature of the decision itself as a non-autonomous one with respect to that information. The only difference is that the patients made free decisions which, as a side effect, accord zero importance to the medical information. It follows that the idea of IC we would be endorsing (if indeed this is the option we endorsed of the two general possibilities to solve our Gettier Problem) is one that surely advances some basic idea of liberty, but not primarily the idea of autonomy. And this can only be learnt from ‘the zero cases’ through the Gettier manoeuvre.
It might be objected that patient competence is disproved by the ignoring of relevant medical information, that competence without autonomy is an oxymoron. If this were true, it would undermine the attempt to present Gettier-type cases as IC after all. To complete my argument, I will therefore explain why I believe it is not true. To be competent, a patient needs (among other things) to be able to participate in treatment decisions by means of rational thought processes. What should we say then of patients who give consent to medical treatment in ways that effectively ignore relevant medical information? The argument for competence in Gettier-type cases depends on three factors. The first is that patients do not ignore the information out of lack of capacity for understanding or logical thinking; they do exhibit rational deliberation (but, as explained, they focus their decision on a different element, which contingently but inevitably affects the medical choice). Even if not formally irrational, however, patients' decisions to ignore relevant information may seem substantively inadequate, as they have the effect of cancelling their own autonomy in the context of medical choice. (This seems especially acute in the case of Mrs B, who relegates decision-making to her doctor.) Two additional factors are therefore necessary for the preservation of competence: any retraction of autonomy must be both reasoned and well-circumscribed. I will explain.
Beyond a certain point, all patients relegate decision-making (on the specifics of treatment) to their doctor. In some trusting patient–doctor relationships this can go a very long way, without casting doubt on the patient's competence. This has limits, however, as when a patient would entrust decision-making to random strangers (the retraction of autonomy is not well-circumscribed), or would accept ‘treatments’ that a reasonable person would find obviously harmful (the retraction of autonomy is not reasonable). But surely none of this is necessarily the case. Crucially, it is imperative—and implicit in the two conditions above—that the patient retains some background awareness of the surrogate decision-making she initiated, so as to allow herself to reclaim autonomy should things pass some red line. (I therefore said that Mrs B would consent only to ‘minimally reasonable’ treatments suggested by her doctor.) Although the medical information does not function as the ground for the patient's decision, it can still be evoked to veto unreasonable treatments. This scheme combines competence with a limited surrender of autonomy. One could define competence so as to include in it the strongest notion of autonomy, and thus tailor IC to be the expression of autonomy in a way which begs the question. We need to be careful, however, not to set the bar of competence too high so as to deprive by fiat otherwise competent people of their right to choose, in well-circumscribed cases, non-autonomous patterns of decision-making. (Such choice becomes inevitable in some ironic cases where the best course of action for an agent is to give up elements of autonomy—for example, knowledge-as a necessary condition for achieving her specific goals.) It would be unduly radical to contend that any agent who chooses a limited surrender of autonomy is thereby rendered incompetent. The normative extent of such surrender can surely be debated but, even if small, it is all that is needed to allow the Gettier Problem in IC.
Gettier-type cases thus show the need for a renewed understanding of IC. It is not the purpose of this paper to develop a new theory of IC; however, alternatives to autonomy could either appeal to the patient's welfare and well-being as grounds,5 or if based on the principle of respect for persons, then that should not be explicated solely or even primarily as referring to patient autonomy, but to the more basic liberty of personal choice. That respect for people in the context of IC appeals to this idea of free choice might be explained, for instance, by a universal principle of near-total sovereignty of persons over their bodies: owing to some deep existential sentiments all humans share, others' interference with our bodies without our consent, even if for beneficial or pleasant purposes, is almost inevitably experienced as deeply humiliating; and this phenomenological truism precedes reflective deliberation (autonomy) of whatever sort. A few attempts were made in recent years to challenge the dominant traditional view of respect for autonomy as the foundation of IC6–10; Gettier-type cases offer a new argument in their support. This, however, cannot eliminate the hidden ambiguity that they expose in the very notion of IC.
The author thanks David Enoch, David Heyd, Nicole Leifer, Jacob Samet, as well as an anonymous referee of this journal for helpful comments on earlier drafts.
Competing interests None.
Provenance and peer review Not commissioned; externally peer reviewed.
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.