• Ethics

In our paper, we pointed to several problems, both practical and theoretical, with the Substituted Judgment Standard (SJS)—at least when the SJS is understood as literally requiring that surrogates always make the decision that the incompetent patient would have made, if competent.1 These problems show that the SJS, so implemented, does not respect patient autonomy. Others have considered these same problems and found them decisive, concluding that the SJS ought to be abandoned. In contrast, we argued that the SJS is best understood not in terms of replicating the decision that the patient would make if competent, but in terms of the standard's underlying purpose: respecting the patient's values and allowing them to continue, as nearly as possible, the sort of life they found worth living for themselves. We called this the Endorsed Life Approach. Understood in this way, the SJS neither consists in nor always requires asking what the patient would decide if competent, and so is not subject to the same challenges; it also offers a way to respect the autonomy of patients even after they have become incapacitated. We believe that this way of thinking about the standard is consistent with how the SJS is often understood and applied in practice.

Dresser and Chan question whether our approach represents a significant departure from the standard interpretation of the SJS.2 ,3 Dresser notes that, given the messy realities of clinical practice, most surrogates and clinicians will respond to roughly the same concerns “no matter which interpretation of substituted judgment holds sway”. She concludes that “properly interpreted, the original formulation of substituted judgment is a justifiable and workable approach”. Chan claims that our Endorsed Life Approach does not “significantly differ from the standard interpretation” since, in implementing the standard interpretation, the surrogate must take into account the patient's values.

We agree with a good deal of this and do not take ourselves to be proposing a radical revision to the way the SJS is implemented in practice. Indeed, many clinicians and surrogates likely already implement the SJS in a manner consistent with our approach. Our goal is not to critique the SJS itself, nor to completely reject use of the ‘what would the patient decide’ question that some take to constitute it. Rather, we aim to provide the sort of ‘proper interpretation’ that Dresser calls for, making sense of why this particular question so often is the appropriate one to ask, and providing a theoretical foundation that is able to withstand challenges intended to show that the SJS should be dispensed with altogether.

The Endorsed Life Approach yields different recommendations than the ‘standard’ interpretation, taken literally, in a number of cases, such as cases in which the patient, if competent, would make treatment decisions inconsistent with her own values. As a result, we recommend different practices in these cases to surrogates and clinicians who may have been misled by standard interpretations of the SJS into giving insufficient consideration to the sort of life the patient valued. However, this does not mean that our view is in the first instance revisionary; our goal is to vindicate the sort of thoughtful decision-making that Dresser and Chan see in typical implementation of the SJS.

Dresser additionally challenges our view that the life a patient valued, rather than the life the patient lived, should be the basis for surrogate decision-making. She notes that “one could argue that the way a person lived is the most genuine indication of the life that patient valued”. Of course, using the life the patient lived as evidence of what she valued is entirely compatible with our proposal. Moreover, it may often be that we have no good reason to think that the life the patient valued differs from the life she lived, at least in any way material to the decisions confronting her surrogate. But we may sometimes have such reason, derived for instance from things the patient told us. The fact that the patient's actual life is generally more evidentially accessible than her valued life is no reason to think that the former should take ethical priority in cases where we recognise that the two come apart.

Brudney, Chan and Dresser all raise questions about the importance of values other than autonomy in surrogate contexts—in particular, the importance of the patient's best interests and the interests of the family.2–4 While our discussion of the SJS focuses on the value of autonomy, we agree that it cannot always be the only relevant value. If, for instance, a competent patient wishes to receive care that will do her no good at all, and will be extremely expensive, the importance of respecting her autonomy does not necessarily trump the costs to others.

The question of how exactly to balance the importance of autonomy against these other values is an immensely complex one, and one that is not peculiar to the context of surrogate decision-making. We do not purport to have a complete answer, let alone the space here to present one. Our primary concern, in the piece under discussion, was to consider how, within a society that gives priority to respecting the autonomy of competent patients, the SJS can be seen as a way of extending that same respect to those who are no longer competent. We argued that respect for autonomy does not become completely irrelevant at the point at which the patient becomes incompetent. We fully agree with Brudney, Chan and Dresser that this claim leaves for future research the very important question of how respect for autonomy should be balanced against other values, and whether the proper balance in this regard differs depending on whether the patient in question is competent or not.

Unlike Dresser and Chan who argue that our approach is very similar to the standard SJS, Brudney suggests that our view is similar to his proposed alternative to the SJS, at least when his approach is properly and charitably understood. To the extent he is right, we welcome this conclusion. As we suggested in our paper, we think that he gets a great deal right, and we are moved by many of the same considerations he discusses. Nonetheless, there are some significant points on which we seem to disagree. Most importantly, Brudney's account focuses on the life that the patient in fact lived. In contrast, we draw a distinction between this and the life that the patient endorsed for herself, and argue that, for the purposes of respecting patient autonomy, the latter takes priority. With this distinction made explicit, Brudney suggests in his commentary that he may agree, providing reason to think that, with further elucidation on both sides, our views may be very similar. (We appreciate and accept his clarification that the authenticity with which he is concerned is of a synchronic rather than diachronic variety—ie, that he is concerned with the fit between a patient's life and his or her present, rather than past, beliefs and values. However, we ourselves would not endorse the claim that the beliefs and values properly taken into account in treating an incompetent patient must be those that can be currently attributed to him or her.)

Sulmasy and Sulmasy go further than Brudney, arguing that our view is essentially the same as their proposed alternative to the SJS, and inviting us to “acknowledge, formally, that [we] were mistaken in calling [their] model a best interests model, and simply to endorse the Substituted Interests Model”.5 In our paper, we nowhere refer to Sulmasy and Sulmasy's approach as a best interests model. The closest we come is to note that in a certain light their view “may sound like another way of characterizing the best interests standard” before going on to describe various ways in which it differs from paradigmatic best interests views.1 Our paper does describe Sulmasy and Sulmasy as being concerned with (a particular understanding of) patients’ best interests. But they are patently concerned inter alia with patients’ ‘real interests’. We interpreted them as, in part, arguing that paradigmatic best interests views are mistaken about what is really in patients’ best interests. If they find the language of best interests to be too theory-laden, we are happy to defer to their terminological choice, but do not believe that in attributing to them a concern with patients’ best interests we were misrepresenting their published views.

As Sulmasy and Sulmasy wrote in their original (2010) paper (and repeat in their commentary), on their proposal “the surrogate is asked to apply the patient's authentic values and real interests, including the patient's known preferences”.5 We were, and remain, uncertain about the exact nature of the intended relationship between these three factors: authentic values, real interests and known preferences. Are they independent considerations that must be weighed against each other to ascertain what would “best promote the good of this patient”? If so, the emphasis on the independent importance of known preferences and real interests represents a significant difference from our Endorsed Life interpretation of the SJS. Or, are the patient's known preferences to be understood as a subset of her authentic values or a subset of her real interests? What role if any do the patient's authentic values play in determining her real interests?

We find no clear answers to these questions in Sulmasy and Sulmasy (2010); still, there are ways of answering them that would produce a view in substantial agreement with the Endorsed Life Approach. This is precisely why in our paper we wrote that “the extent to which the two approaches will endorse the same treatment choices depends on the extent to which our interests are determined by our values”.1 We had imagined that, given their decision to list these as separate factors, Sulmasy and Sulmasy endorse the view that one's interests are not necessarily and wholly determined by one's values; for instance, a patient might conceivably have a pro tanto interest in extending the length of her life, whether or not she values doing so. Sulmasy and Sulmasy's response in their commentary to our Jehovah's Witness example suggests that this is not what they had in mind, and that they are committed to the ancillary claim that a patient's real interests are determined by her values to such an extent that it must always be in her interests to live the sort of life she valued while competent—it is in the Jehovah's Witness's interests to live an extra 50 years only insofar as she herself values that extended life.

If Sulmasy and Sulmasy intended to pair their Substituted Interests model with such a radically subjective understanding of patients’ interests, it likely will yield the same verdict in most cases as our own view (though we would still want to know which of a patient's values count as ‘authentic’). However, Sulmasy and Sulmasy's model could just as easily be paired with a less subjective (and, we think, independently more plausible) account of an individual's interests. For example, most commentators regard cases like the Jehovah's Witness as difficult precisely because the patient's values (appear to) conflict with her interests. To the extent one admits of more objective factors as relevant to determining what is in a patient's interests, the Endorsed Life Approach and the approach endorsed by Sulmasy and Sulmasy will be importantly different.

There is one point on which Sulmasy and Sulmasy make it clear that they do disagree with us. We argue that, although the SJS is not simply identical to the practice of making surrogate treatment decisions based on what the patient would decide if competent, nevertheless asking surrogates what the patient would decide if competent may in many cases be a valuable heuristic, a tool that can be used by surrogates and caregivers to reach appropriate decisions. Sulmasy and Sulmasy seem confused by this; they point out that we ourselves present cases in which asking what a patient would decide if competent may lead to inappropriate decisions. But this, of course, is just what one would expect of a heuristic. The existence of such cases is precisely why we think that this question is merely a heuristic, rather than a sufficient standard for surrogate decision-making in its own right. Asking the question is sometimes useful, sometimes not; like most clinical tools, it is valuable if it can be applied in such a way as to be sufficiently helpful in a sufficiently high percentage of cases. By articulating the underlying goals that the heuristic, when appropriately employed, is able to serve, we hope to provide a basis for recognising cases in which other ways of pursuing those same goals are likely to prove more successful. Of course, the question of which clinical methods are most successful for pursuing an agreed-upon set of goals is, ultimately, an empirical one; neither our belief that asking surrogates what a patient would have decided is often useful for working out what sort of life the patient valued, nor Sulmasy and Sulmasy's apparent conviction that asking this question is never helpful in determining a patient's ‘real interests’, should be accepted on faith. If empirical results showed that employing the ‘what would the patient decide’ heuristic is less effective than some viable alternative for arriving at surrogate decisions consistent with the life the patient valued, in most or all cases, then we would support further reducing or ceasing the heuristic's use.

In summary, the varying responses by the commentators to our paper provide, we think, both an illustration of the present predicament of research into surrogate decision-making and grounds for optimism about a way forward. While a great deal has been written on the SJS, there is little consensus about just what the essential elements of the standard are; thus the disagreement between the commentators who suggest that we are not departing in any interesting way from the standard interpretation and those who suggest that we are departing so radically that they call on us to join them in denouncing the SJS. What we have tried to do is to capture the important normative insights behind both criticisms of the SJS and the practice of those who are already implementing the SJS in thoughtful, nuanced ways. We believe that the fundamental agreement between the critics and the practitioners of the SJS is often deeper than both may realise. We hope that our framing of the SJS may aid in clearing up these misunderstandings and showing the way towards a common ground from which to help surrogates and clinicians with the important and often wrenching decisions they face.