Ignoring the diagnosis


Introducing the Example

In this essay, I will consider an ethical problem regarding the doctor-patient relationship that has far reaching practical implications. Let’s take the following abbreviated example as a guide to our discussion:

Q: A 49-year old individual (S) has voluntarily chosen to see a neurologist at the recommendation of her internist, but refuses to learn of the subsequent diagnosis. In fact, the diagnosis — which her internist knows — is positive for Pick’s disease[1]. Should the internist inform S of the diagnosis?

We might not have strong or pressing intuitions one way or the other about Q — but this makes it perhaps more important that we should construct some criterion for how a doctor who faces this situation should react. I will not explicitly address (i) what responsibilities immediately fall on the family members of S, and (ii) how confidentiality of the information should bear on the doctor’s decision making. There are two immediate questions I want to pursue here:

1. What are the moral obligations of the internist?

2. What is within the scope of moral permissibility for the internist, and what epistemic problems arise in real-world, nuanced situations?

In order to establish a proper review of these questions, I will address this from the standpoint of paternalism vs. patient autonomy. I will first outline the general arguments before focusing on how they inform the answers to (1) and (2).

Paternalism and Patient Autonomy

As this example is not representative of the typical debates in this field, we will turn to the usual arguments first in order to get a good grip of the complexities in Q. In the ethics of health care and specific diagnoses, the conflict is usually framed as a tension between paternalism and patient autonomy.

Paternalism is the idea that the doctor’s chief basis for action comes from an overriding responsibility to medically benefit the patient, even against the patient’s active consent. A doctor acts paternalistically by doing X when X takes away a liberty the patient regularly has, but the goal of X is to promote the patient’s well being. Though paternalism might sound like a harsh philosophy when put so starkly, we take it for granted in other domains of life. For example, enforced road safety measures and parents’ treatments of children are ordinary and well-accepted examples of the principle. It is also interesting (but perhaps not ethically informative) that the Hippocratic Oath along with orthodox cultural views of medicine do not contain any reference to honoring patient consent (or even telling the truth to patients)[2]. Rather, the Hippocratic slogan “do no harm” might quite correctly be interpreted to be in line with a paternalism-guided view of medical practice. A compelling reason in favor of paternalism is that doctors attend medical school for years and have much richer understanding and experience, whereas it might be impossible to impart that medical knowledge (or even basic understanding of probability, false positives, testing procedures, etc.) to a patient. Proponents of paternalism, whether guided by an underlying notion of utility or otherwise, will be comfortable in some situations with not being honest with the patient.

The other opinion (though of course still intending the benefit of the patient) in the debate is that doctors should always respect the autonomy of the patient, first formalized as a burden of the physician in the work of Beauchamp and Childress. Autonomy is closely tied with ‘truth-telling’. In the vast majority of cases, doctors preserve patient autonomy by not making independent judgments about treatment decisions. The foundation of the principle of autonomy can be traced back to Mill’s conception of ‘liberty’: in On Liberty, Mill argues that

‘the only purpose for which power can be rightfully exercised over any member of a civilized community, against his will, is to prevent harm to others’ (621).

This requires that doctors respect the ‘personal right’ — in most cases where the patient does not pose a threat to others — of the individual to come to her own decisions without any influences that constrain that freedom. In the health care debate, this is usually a matter of giving the patient relevant information regarding the diagnosis in an understandable way.

Generally applied, both paternalism and autonomy face numerous objections. Paternalism ostensibly has an air of dominance and arrogance, but more importantly can conflate medical expertise with moral expertise or falsely anticipate the knowledge of complex preferences. Gillon argues that doctors who act paternalistically and deceive (or do not consult) patients in the name of beneficence will have an extremely hard time judging the overall extent to which the information might interplay with the patient’s life. Sound medical advice should not be confused with universal advice without an intimate understanding of the relevant factors that only the patient can have. On the other hand, others have argued that paternalism can be justified on the grounds that it is more humane: patients often react poorly to grave diagnoses. Collins takes an even stronger stance, arguing that the overall happiness and productivity of patients’ lives may depend on lying unless we can determine diagnoses and options without imprecision.

It is important here to note that Q is a different scenario than in the framework above. Whereas in most cases honoring autonomy implies truth-telling and paternalism implies proceeding without full consent, honoring the patient’s autonomy in Q will require the internist to withhold the diagnosis. The above principles should nevertheless inform our assessment of Q. In this case, they suggest two important questions:

A. If patient autonomy promotes the idea that S is better equipped to make treatment and general ‘life’ decisions with all the relevant information, how can we reconcile this with S rejecting this information in the ‘first-order decision’[3] of learning the diagnosis? In other words, what qualifies a desire as autonomous?

B. Paternalism (and the overwhelming responsibility to the good of S) will dictate that the internist tell S the truth for her own long-term good. Is this a morally justifiable route?

I will treat (A) and (B) separately in the following sections; (A) will lead to definite progress on the question of moral obligations, whereas (B) will address (2) above.

(A) The First Order Decision

The decision I want to consider in this section is S’s refusal to learn the diagnosis (hence “first order”). In the previous section, we established some arguments for why the doctor might be principally obliged to respect this wish as a consequence of the right of personal autonomy. However, since respecting this decision will lead to limiting life-sustaining treatment in Q, we should consider Savalescu’s arguments that characterize situations when such a right is genuinely protected (and perhaps when it ought not be). Savalescu argues that we should not be under obligations to honor a patient’s desires unless they are rational. This rationality requirement is close to what other authors have termed ‘decision making capacity’, and presents the following constraints[4]:

C1. S has the facts that are required to make the decision.

C2. S has the logical capacity to evaluate the facts.

C3. S can ‘vividly’ imagine the possible outcomes of her decision.

I want to apply these constraints not to a treatment decision about Pick’s disease, but rather her decision about the diagnosis (in particular, the facts in C1 need not concern Pick’s at all). Each of these constraints seems to be well-justified by taking examples to the contrary. Moreover, I will demonstrate that violating any of these three constraints is self-defeating to the decision D whose autonomy we want to uphold:

C1. Suppose S is misinformed about the facts of the matter, and believes that knowing the diagnosis will cause her family to abandon her or that the diagnosis will legally bind her to treatment (false). In this case, S is not making the relevant decision, but rather a different decision D’ concerning separate outcomes and implications.

C2. Suppose S is somehow logically inconsistent internally: she believes that the diagnosis will give her a better sense of her future’s timeline, and expects no negative consequences. Yet, from these premises, she still concludes some subjective harm from learning the diagnosis. It is clear that the decision D, if arrived at without proper reasoning, is not a decision at all but a simple belief (and it would certainly be a larger burden on the proponent of autonomy to honor all beliefs rather than autonomous decisions).

C3. If S cannot imagine the outcome of knowing the diagnosis, then we are left at a point where the guiding notion of autonomy is violated. Proponents of autonomy depend on an element of planning for self over time. We lose this if proper imagination of the outcomes is lacking, because without being able to imagine herself knowing or not knowing the diagnosis, S cannot make a life plan.

Savalescu’s elements of capacity can be refuted on the basis that they are (i) practically and epistemically too demanding for the internist and (ii) are too stringent, even in principle.

While I think (i) is certainly true and perhaps an unfortunate consequence of these difficult situations, it should not be construed to absolve the doctor of the responsibility. Rather, we should take it as a reason to train doctors not only in their relevant medical expertise, but also with an ability to communicate facts and develop a doctor-patient relationship over time that leads to an understanding of the patient’s intellectual abilities and values. To avoid trivializing the endeavor (which I believe lies in the domain of psychology)[5], I will only repeat Applebaum’s suggestion that we ask patients to restate the information and their role in the decision process, work with them to spell out their logic, and take help from individuals close to them to understand their imaginative state.

I think (ii) presents a stronger case against holding these evaluative criteria for decision making capacity. In particular, requiring rationality as in C2 might disqualify most decisions we normally make from being autonomous. Humans by nature have finite reasoning powers. Even if we believe some set of propositions, it’s quite likely that we will not immediately believe a sufficiently long tautology or logical entailment of our beliefs. As behavioral research by Kahneman, Tversky, and others has repeatedly shown, it is not that our judgments are simply mistakes around a rational basis — even in explicit questions, we display systematic biases and irrationalities. Though this conclusion might sound catastrophic for Savalescu, I think abandoning the criteria would be overreaching. Not being able to explicitly spell out a proper sense of ‘rationality’ should only motivate the thought that a uniform psychological test for rational beliefs will escape us; as Freedman does, we can still establish very basic criteria and then rely on a case-by-case assessment.

Perhaps the most important thing to keep in mind for any assessment of decision making capacity is that disagreement from medical counsel alone (“S, it is my professional opinion that you should learn the diagnosis”) should not imply irrationality.

I want to briefly discuss how this first-order-decision is related to the second-order decision. Clearly, if S autonomously decides to reject the diagnosis, we can assume that this inherently implies a decision about (lack of) treatment, which is then not autonomous because it does not satisfy C1-C3. We might claim that the second decision about the treatment of (and S’s reaction to) her diagnosis of Pick’s disease is more important than the first order decision, and so honoring the first decision is morally reprehensible. Without some moral principle to justify why either decision is more pressing, this argument fails to do justice to the above criteria for the first-order decision. If we are to accept that the first decision is autonomous, the patient can genuinely imagine what it means to not know the diagnosis. There might be values, or non-descriptive beliefs that cause S to reject learning of the diagnosis. The second order decision should then inherit the status of autonomy from the first.

The attempt to elevate the second decision’s priority is nevertheless forceful; this requires us to revisit our notions of paternalism in a more nuanced way to address (B).

(B) Paternalism, revisited

Though I have presented paternalism in the second section as a unified concept, the stance is not so simple. For Mill and Kant, autonomy is not simply a right but a moral imperative or duty — as a consequence, an individual is not ‘free’ to give up her right to autonomy. Therefore, it seems that Mill would condone paternalism under the pretense that the long-term autonomy of S is impaired by the honoring S’s short-term autonomy for the diagnosis. Mill explains that allowing such decisions would be self defeating:

“[she] therefore defeats, in [her] own case, the very purpose which is the justification of allowing [her] to dispose of [herself]… the principle of freedom cannot require that [s]he should be free to not be free”

Framed this way, (weak) paternalism might be the only way that the internist can truly support S’s autonomy by informing her of the diagnosis.

Conclusion and Recommendations

In light of the above discussion, I propose that the internist:

a) Has the moral obligation to assess the capacity of the first order decision. Decision making capacity, especially in the case of a possible instance of neurodegeneration, will play a central role in determining the autonomy of the decision and establishing whether the desire is rational.

b) Is morally justified in being paternalistic if she can properly demonstrate that the first order decision is not autonomous or on the normative basis of weak paternalism as put forth by Mill. In either case, the moral justification should stem from a stance that intends to uphold the autonomy of the patient.

These questions often do not lend straightforward answers, and my approach here has been inevitably incomplete. I have not looked at the case from a deontological vs. consequentialist stance, and I certainly have not factored in any empirical extrapolations of whether it is reasonable for S to reject the diagnosis. No philosophical moral principle or obligation can address the inherent intricacies of any real-world situation, and it is the burden to act most responsibly under difficult circumstances that the doctor must bear.

[1] A neurodegenerative disorder similar to Alzheimer’s. I will assume that this is terminal, but has some course of therapy or management that is ‘life-sustaining’ in a broad sense.

[2] Marzanski (2000) points out that ‘purity and holiness’ (or more generally virtue) as stated in the Hippocratic oath may imply veracity. Nevertheless, the point remains that the oath is at least conflicted in its guidance.

[3] From S’s viewpoint, we should differentiate the first order decision to be whether S should learn the diagnosis, and the second order decision regarding treatment of Pick’s disease.

[4] Other authors have included even more criteria, such as the existence of specific values and an ability to express choice. See Buchanan & Brock 1989.

[5] It should be noted that while the underlying theory may be controversial, there are empirical tests for decision making capacity. For a short sample, see Etchells, 1999; Markson 1994; Applebaum 1988.


Appelbaum, Paul S., and Thomas Grisso. “Assessing Patients’ Capacities to Consent to Treatment.” The New England Journal of Medicine. U.S. National Library of Medicine, n.d. Web. 05 Dec. 2016.

Charland, Louis C. “Decision-Making Capacity.” Stanford Encyclopedia of Philosophy. Stanford University, 15 Jan. 2008. Web. 05 Dec. 2016.

Collins, Joseph, and Helga Kuhse. “Should Doctors Tell the Truth?” Bioethics. Ed. Peter Singer. 2nd ed. N.p.: Blackwell, n.d. 605–10. Print.

Drickamer, and Lachs. “Should Patients with Alzheimer’s Disease Be Told Their Diagnosis? — NEJM.” New England Journal of Medicine. N.p., n.d. Web. 05 Dec. 2016.

Dworkin, Gerald. “Paternalism.” Stanford Encyclopedia of Philosophy. Stanford University, 06 Nov. 2002. Web. 05 Dec. 2016.

Elliott, Carl. “Amputees by Choice.” Bioethics. Ed. Peter Singer and Helga Kuhse. 2nd ed. N.p.: Blackwell, n.d. 625–33. Print.

Entwistle, Vikki A., Stacy M. Carter, Alan Cribb, and Kirsten Mccaffery. “Supporting Patient Autonomy: The Importance of Clinician-patient Relationships.” Journal of General Internal Medicine 25.7 (2010): 741–45. Web.

Etchells, Edward, Peteris Darzins, Michel Silberfeld, Peter A. Singer, Julia Mckenny, Gary Naglie, Mark Katz, Gordon H. Guyatt, D. William Molloy, and David Strang. “Assessment of Patient Capacity to Consent to Treatment.” Journal of General Internal Medicine 14.1 (1999): 27–34. Web.

Gillon, R. “Paternalism and Medical Ethics.” Bmj 290.6486 (1985): 1971–972. Web.

Hermann, Helena, Manuel Trachsel, Bernice S. Elger, and Nikola Biller-Andorno. “Emotion and Value in the Evaluation of Medical Decision-Making Capacity: A Narrative Review of Arguments.” Frontiers in Psychology 7 (2016): n. pag. Web.

Higgs, Roger. “On Telling Patients the Truth.” Bioethics. Ed. Peter Singer and Helga Kuhse. 2nd ed. N.p.: Blackwell, n.d. 611–15. Print.

Jackson, J. “Telling the Truth.” Journal of Medical Ethics 17.1 (1991): 5–9. Web.

Kant, Immanuel. “On a Supposed Right to Lie from Altruistic Motives.” Ed. Helga Kuhse. Bioethics. Ed. Peter Singer. 2nd ed. N.p.: Blackwell, n.d. 603–04. Print.

Marzanski, M. “Would You like to Know What Is Wrong with You? On Telling the Truth to Patients with Dementia.” Journal of Medical Ethics 26.2 (2000): 108–13. Web.

Mill, John Stuart. “Chapter 5: Applications.” On Liberty. New York: Norton, 1975. N. pag. Print.

Murgic, Lucija, Philip C. Hébert, Slavica Sovic, and Gordana Pavlekovic. “Paternalism and Autonomy: Views of Patients and Providers in a Transitional (post-communist) Country.” BMC Medical Ethics 16.1 (2015): n. pag. Web.

Savulescu, Julian. “Rational Desires and the Limitation of Life Sustaining Treatment.” Bioethics. Ed. Peter Singer and Helga Kuhse. 2nd ed. N.p.: Blackwell, n.d. 646–63. Print.

Veatch, Robert. “Abandoning Informed Consent.” Bioethics. Ed. Peter Singer and Helga Kuhse. 2nd ed. N.p.: Blackwell, n.d. 636–45. Print.