Kant, curves and medical learning practice: a reply to Le Morvan and Stock

Abstract

In a recent paper published in the Journal of Medical Ethics, Le Morvan and Stock claim that the kantian ideal of treating people always as ends in themselves and never merely as a means is in direct and insurmountable conflict with the current medical practice of allowing practitioners at the bottom of their “learning curve” to “practise their skills” on patients. In this response, I take up the challenge they issue is and try to reconcile this conflict. The kantian ideal offered in the paper is an incomplete characterisation of Kant's moral philosophy, and the formula of humanity is considered in isolation without taking into account other salient kantian principles. I also suggest that their argument based on “necessary for the patient” assumes too narrow a reading of “necessary”. This reply is intended as an extension to, rather than a criticism of, their work.

This is a reply to a paper by Le Morvan and Stock,¹ in which they try to establish a conflict between medical learning practice and the kantian ideal. Arguably, the conflict they try to establish between the kantian ideal of treating people always as ends and never merely as a means and the current medical practice of allowing practitioners at the bottom of their “learning curve” to “practise their skills” on patients is based on an incomplete characterisation of that kantian ideal. This reply is an extension to, rather than a criticism of, their work.

In Medical learning curves and the kantian ideal, Le Morvan and Stock¹ argue that the kantian ideal, characterised in terms of Kant's formula of humanity (4:429),² is in direct and insurmountable conflict with the modern medical practice of allowing novice clinicians to “practise” their skills on patients. They summarise their argument, in brief, as follows:

“…[D]eveloping and acquiring medical expertise requires practising on patients, with one's performance improving along a learning curve. Thus, patients treated at the beginning of the learning curve are exposed to higher risk, with later patients benefiting from this exposure. Using patients in this way is, at least prima facie, a violation of Kant's Principle [of treating Humanity as an end in itself]” (Le Morvan and Stock,¹ p 513)

The basic argument comprises two claims. Firstly, the claim that any medical practitioner who practises his or her skills on patients necessarily uses patients merely as a means to develop skill and therefore does not treat patients as an end in themselves (similarly, the supervisor who uses a patient to allow a novice to learn a procedure). Secondly, they claim that to treat a patient in this way is in direct conflict with the kantian ideal that instructs us never to treat other people solely as a means to some end, but always also as ends in themselves.

Before I begin, however, I should make clear exactly what I am not objecting to. I am not objecting to the claim that current medical learning practice places patients at more or less risk, depending on the position of the clinician or student along his or her learning curve. It is hardly disputable that patients undergoing any surgery are statistically more at risk of complications when less experienced surgeons have their hands on the scalpel. Neither am I suggesting that to subject a patient to the inexpert physical examination of an incompetent or insufficiently qualified clinician for teaching purposes, where the patients have expressly withheld their consent to be a “teaching case”, is not in conflict with the kantian ideal. Furthermore, I accept absolutely that cases such as medical students practising a gynaecological exam on an anaesthetised woman, without her prior consent, is a clear case of treating the patient solely as a means, as there is no therapeutic benefit for her and tacit consent cannot reasonably be assumed.

What I will argue is that allowing a suitably qualified and competent clinician or student to learn a procedure under the direction of a suitably qualified and competent clinical supervisor, where the procedure is of direct therapeutic benefit to the patient, and the patient has not explicitly withheld consent, is not in conflict with the kantian ideal, properly characterised. Where these conditions are not met, it is reasonable to say that the kantian ideal is being violated, but this violation is contingent on the actions of the individual clinician, and is not a necessary feature of the practice itself. As such, this paper is not aimed at justifying current practice, but at showing that current practice is not in conflict with the kantian ideal. As an anonymous reviewer of this paper has pointed out, whether this is sufficient to morally justify that practice is another question entirely.

Le Morvan and Stock have, as one would expect, considered and responded to several rationales that may be offered to resolve the supposed conflict. I will now examine a few of these responses.

Not merely as a means

As Le Morvan and Stock¹ acknowledge, the kantian ideal does not forbid us from ever using another person as a means, but only from using them merely as a means. To use their own example, the clinician is used by the patient as a means to get well, but he is not used merely as a means because “as rational agents, medical practitioners have agreed to participate in the health care system” (p 515). What is not clear is why they fail to extend that same rationale to the patients. It seems plausible, even before their main arguments get under way, to think that a patient also, as a rational agent, agrees to participate in the healthcare system. Both clinician and patient choose to participate in a system in which each try to achieve their own ends and both are used as a means by which the other can attain their ends.

One notion that may help us establish the legitimacy (on kantian grounds) of this kind of mutual use is that of “reciprocity”, which can be found in Kant's views on sex. Kant considered sex to consist of each partner seeking to objectify and possess the other. This would normally be objectionable on kantian grounds because by yielding to the sexual impulse we are allowing ourselves to be objectified and possessed by another and this is inconsistent with having respect for our own humanity. However, it is deemed permissible, as Korsgaard³ notes (p 195), due to the reciprocal nature of the act.

If I allow myself to become the object of another, and that other allows herself to become my object in return, then, according to Kant, I “win myself back…the two persons become a unity of will”.³

A reviewer of this paper helpfully pointed out that the analogy between the sexual act and the doctor–patient relationship is not perfect, but even so, and notwithstanding the fact that Kant's views on sex are neurotic and objectionable, it is enough to show that woven into the fabric of the kantian ideal is the notion of “reciprocity”. If the doctor–patient relationship can be considered to be reciprocal, in that both use and are used, this may be enough to solve the conflict (even if both parties do use one another solely as a means, which I think is doubtful, but difficult to establish either way).

Korsgaard³ (p 195) has noted that reciprocity requires “equality”, and so if we are to appeal to reciprocity to reconcile the conflict, the relationship between clinician and patient must be sufficiently equal. We may think that they are not equal, and that the clinician is the more powerful party by virtue of his expert status and because the patient is seeking his help. However, the power relationship between the two is not so imbalanced. Although the clinician is powerful by virtue of being an authority figure and expert, the patient is powerful in another way. The patient can sue; the patient can complain. Without patients and their illnesses, the doctor is nothing. The patient can choose to accept or reject any treatment offered. If the doctor acts unprofessionally he can be struck off, and the mere accusation of wrongdoing is enough to do a great deal of damage to a clinician's career. It may once have been true that doctors were powerful authority figures, but in this age of patient‐centred medicine and intense public scrutiny, it is naive to think that the patient does not also have power—and if the patient is sufficiently powerful to use the clinician, he is sufficiently powerful to enter into a reciprocal relationship.

UNNECESSARY EXPOSURE OF RISK

Le Morvan and Stock¹ suggest, in response to a dual motive argument (footnote i), that even if the learner is not treating the patient merely as a means, the clinical supervisor is. The argument offered claims that by allowing a novice to carry out a procedure on a patient, even if the procedure is carried out perfectly, the patient is not treated as an end because he or she is exposed to a risk that is not necessary (where “necessary” refers to necessary for the patient and not societal interests). It is claimed the risk is not necessary because more experienced clinicians, who would otherwise supervise the novice, can easily perform the procedure themselves, at less risk to the patient. This, it is argued, requires not treating the patient as an end because

... by not taking the most effective route to providing … care—namely doing the procedure herself, or assigning it to somebody with experience … the supervisor … exposes the patient to risk of harm, not for the benefit of the patient, but for the benefit of the…[novice's]…future patients” (Le Morvan and Stock,¹ p 515).

There are, I believe, two claims in this argument, only one of which is relevant to the conclusion.

1. The suggestion that the patient is treated as a means by the supervising practitioner because the most efficient route to providing care is not taken is irrelevant to the argument. This is simply because the supervisor has no duty in the first place, particularly on kantian grounds, to provide the best care possible for that patient. The duty to promote the interests of others is, for Kant,⁴ an imperfect duty (6:390–391), which means we have a degree of choice as to (a) whose interests we help promote, (b) which interests should be promoted, (c) when we promote those interests and (d) how much time we spend doing it. It is, therefore, consistent with the kantian ideal, although perhaps not with other ideals thought to govern medical duties that supervisors may have an imperfect duty both to help patients recover and to help the novices learn, both for their own sake and for the sake of their future patients, and that the supervisors choose to discharge those imperfect duties across all three parties in the best way they see fit. Thus, the claim that the supervisors have failed in their duty to the patients by not providing the most efficient care possible is no objection on kantian grounds.

Furthermore, given the kantian claim that we should not promote the interests of others if those interests go against their moral well‐being (6:394),⁴ we may think that we are under no obligation, on kantian grounds, to refrain from allowing a suitably qualified novice to practise on a patient (under adequate supervision), because any interest of the patient to the contrary is based on an end that is contrary to his or her (kantian) moral interests.

1. The claim that is relevant to the argument is that the risk the patient is subjected to is unnecessary for the patient. The suggestion is that because patients in no way benefit from the risk they are exposed to, they are being used merely as a means to benefit the novice (and the novice's future patients). This is true, they claim, both when the novice is supervised and when left to his or her own devices.

The objection I have to this lies in the very narrow definition offered of “necessary for the patient”. For Le Morvan and Stock, “necessary for the patient” means “necessary for this procedure, at this moment in time, given my immediate interest in receiving the best possible care I can get”, but this fails to take into account the broader long‐term interests of patients, such as having qualified doctors able to repeat the procedure in 20 years time should they need it, or having experienced doctors available in the future to care for their loved ones. Clearly, should the patient explicitly refuse to have a novice perform a procedure, and should the supervising practitioner duplicitously allow the novice to practise on the anaesthetised patient, this would constitute a breach of the kantian ideal. What is not clear is why, in the absence of explicit instruction, we should assume that we interpret “necessary for the patient” in the narrow sense offered by Le Morvan and Stock and not in the broader sense. In fact, it would seem to be more consistent with the kantian ideal, taken as a whole, to assume the broader reading, because it would involve a contradiction of conception⁵ to will that only experienced clinicians at the top of their learning curve perform a procedure on me (because if that maxim became a universal law it would not be possible to be treated by such a clinician, as there would be none). It would not, therefore, be considered rational, on kantian grounds, to will it on any occasion.⁶ Given that patients, in this situation, may not be well placed to act rationally, in so far as their immediate situation may prevent them from accurately recognising that their ends are not in accordance with rational moral law, and given that Kant's own moral philosophy was largely influenced by, and to some extent aimed at remedying, the anthropological “fact” that people are overly influenced by their inclinations and desires and are only imperfectly rational,⁷,⁸ it seems likely that by assuming the broader reading we are treating patients as rational agents who are an end in themselves, as opposed to mollycoddling them on the working assumption that they does not have rational ends. Of course, if the patient makes his wishes clear there is no need to assume either reading. To disregard the patient's instructions would normally be a breach of the kantian ideal. I say normally, because unless it is possible to accommodate the patient's wishes, we cannot be morally obliged to do so—according to the kantian doctrine of “ought implies can” (142).⁹

The only question here, then, is whether or not we are obliged by the kantian ideal to bring it to patients' attention that they may be treated by a novice under instruction. If it is a form of deception to allow a novice to treat a patient without explicitly making the patient aware of that fact, then this does seem to be a violation of the kantian ideal. However, to be an act of deception, the default expectation of patients would have to be that they are treated by expert clinicians at the top of their learning curve. There would be no deception unless the clinician agreed to treat a patient in the knowledge that the patient thought the clinician was at the very top of the professional learning curve and was only consenting to be treated on that basis. Furthermore, the clinician would have to propagate this false belief and not take steps to correct it.

I would suggest that the default position is that patients expect to be treated by a doctor who is competent and sufficiently qualified to treat them. Here, “sufficiently qualified” is a fuzzy notion, but in practice it tends to be clear. We would not normally consider a young clinician fresh out of medical school suitably qualified to perform a quadruple heart bypass; but we would consider a senior registrar who has performed countless surgeries before, and who has a sufficient grasp of the theory of the procedure, to be suitably qualified to learn that procedure under the supervision of a more experienced clinician. Similarly, we would not consider a student, 1 month into his course, competent or suitably qualified to take blood from a patient. However, by his or her third year, after having studied the necessary theory, and practised on models, students are as sufficiently qualified as anyone could be to take blood for the first time, under supervision. In the case of medical students, however, patients should be made aware. The medical student is not a qualified doctor and therefore does not fall under the scope of the patient's normal expectation.

An anonymous reviewer of this paper has suggested that when we look at General Medical Council documents, “there is a real sense that society expects that they get the highest quality care”. I agree that patients would like to be treated by clinicians at the top of their learning curve, but it seems naive to think that they actually expect it. I do not expect to be treated by a doctor who is at the top of the learning curve, although I do expect to be treated by a doctor who is competent and sufficiently qualified to treat me. Although patients may well expect high‐quality care, I remain convinced that their conception of high‐quality care is limited to care from clinicians who are competent to do their job. There is clearly scope for some empirical work here, which would make a very valuable contribution to this debate.

The fact that patients do not expect to be treated by clinicians at the top of their learning curve for every procedure means that it is not an act of deception to not inform them that they are being treated by a suitably qualified novice. O'Neill¹¹ has noted that, on Kant's account, to use another person merely as a means is not to do something which the person may not want to consent to, but to do it in such a way that the person cannot consent to it. Clearly, given that patients do not expect to be treated by a clinician at the very top of the learning curve, but only expect to be treated by a qualified and competent doctor, they are in a position to consent to treatment because they know what to expect. They expect to be treated by a practitioner who is judged competent to treat them by the relevant professional bodies—and this expectation is met, even when a practitioner is learning a new procedure, provided the learner is sufficiently qualified.

An exception to this rule may be that if patients were undergoing high‐risk surgery, they would normally expect to have a highly qualified and experienced clinician perform it. But this is perfectly consistent with the notion that the higher the risk of the procedure, the more qualified the patient would expect the clinician to be.

Arguably, even so, the kantian ideal dictates that clinicians ought to take the trouble to explain to patients that they may be treated by a novice and give them the option to decide for themselves whether they want to be treated by a novice or a clinician at the top of their learning curve. However, for the clinician to act on such a maxim would lead to a contradiction of conception. If the intention was to give all patients the option of being treated by a clinician at the top of the learning curve, and this option was taken (which it arguably would be), there would soon be so few top clinicians around that this option could no longer be given. Thus, that maxim would be self‐defeating—a violation of the formula of universal law (4:402).²

Considering that (a) there is no kantian duty on the part of the clinician to provide the patient with the most efficient care possible; (b) to not inform a patient of the learning curve position of the clinician is not an act of deception, and therefore is prima facie not a violation of the formula of humanity; and (c) it may sometimes be a violation of the formula of universal law for the clinician to give the patient the explicit choice, we seem to have good reason for supposing that there is no necessary conflict between the kantian ideal and current medical learning practice.

Conclusion

In this reply to Le Morvan and Stock,¹ I have taken up their challenge and attempted to show how the supposed conflict between the kantian ideal and current medical learning practice can be resolved. I have argued that the conflict they identify stems from an incomplete characterisation of the kantian ideal. By considering only the formula of humanity in isolation, the kantian ideal thus expressed was divorced from other salient kantian principles, such as the distinction between perfect and imperfect duty, “ought implies can” and the notion of “reciprocity”. I have also suggested that their argument was based on a very narrow reading of “necessary for the patient”, which was unreasonably, and unnecessarily, assumed.

Given that clinicians are not obliged to give the very best and most efficient care possible to every patient and given that patients have no expectation that they will be treated by a clinician at the top of the learning curve, and in the absence of any explicit direction from the patient to the contrary, we have no reason to think that the practice of allowing novices to learn new skills on patients is necessarily in conflict with the kantian ideal of always treating people as ends in themselves, when this kantian ideal encompasses more than an isolated reading of the formula of humanity.