Breaching trust secretly

The following imaginary situation illustrates that sometimes, informed consent does not protect trust.

The patient is asleep. The room is empty. It is late at night. The IV is connected, and the nurse could slip some unwanted, potent, and—in all likelihood—medically beneficial drug into the bag. There would not be adverse drug interactions or physical traces and chances are that the drug would improve the patient's life overall, and not just medically. No one except the medical team would know that a potent drug was administered despite the patient's refusal. In such circumstances, imposing treatment is more likely than not to be medically and otherwise beneficial. It would also clearly violate informed consent norms. While the requirement of informed consent might not obtain across the board, whether it obtains in a given context intuitively seems rather independent of whether the intrusion took place late at night. However, given that no one would know about an intrusion that takes place late at night, such an intrusion would be unlikely to diminish public trust. It is possible, then, to violate standard informed consent requirements without undermining public trust in the medical community.

Some may explain that the only reason why public laws and regulations surrounding informed consent cannot permit imposing care “late at night only” is that the public, noting these public laws, would then lose its trust in the medical system; and to hide those laws would violate publicity requirements. This answer fails to substantiate the putative moral duty of informed consent as commonsense morality construes it; only the moral duty to create and protect laws and regulations that demand informed consent. Commonsense informed consent requirements are usually understood to be about how doctors and nurses should behave, and not only about how to tell them to behave.1

Another response to my imaginary situation challenge is that when it seems clear to doctors that no one will discover the imposed intervention, still someone will occasionally discover it. Then, precisely because earlier it was unlikely that anyone would discover it, people might suspect that far from the public eye, doctors impose care left and right. Public trust in doctors could sharply decline. Thus, the reply goes, doctors must never impose care—not even in seemingly ‘safe’ conditions.

This ingenious response (suggested to me by Dan Wikler, evoking ideas from R F Harrod) cannot be quite true—or the following clearly-false claim would be true: A perfectly rational egoist who stands to gain a lot from cheating people, but who stands to gain much more from maintaining their trust, will never cheat them, even when she has strong subjective reason to expect to get away with it; if ever she were caught cheating on occasions when she could expect not to, she would lose too much public trust in her. Surely this claim is hyperbolic. Sometimes the subjective probability of being caught and the expected loss from being caught are low enough to keep the net prospect from cheating positive, and cheating, advantageous for an egoist. Similarly, sometimes the subjective probability for doctors of being caught imposing beneficial care is low enough to keep the net social prospect positive.

Too much trust

Society and the medical system could not function without a minimum of trust. But it may remain the case that we can afford to lose some trust. In fact, champions of informed consent accept that there can be such a thing as too much trust in doctors. Many depict the era of medical paternalism as one of irrational, excessive trust in doctors; and many lament the therapeutic fallacy, which involves research participants’ excessive trust that medical investigators aim primarily to benefit them. Carolyn McLeod concludes, [whether] to hammer home to the public the message that physicians and other medical professionals actually help people and care about them… is morally appropriate would… depend on whether the resulting trust would be justified. In general, cultivating trust is only wise if trusting would be wise in the circumstances.23

This conclusion is not quite right. Cultivating irrational trust can also improve social cooperation (directly, or by encouraging third parties’ rational trust), a potentially strong reason to cultivate irrational trust. But McLeod is surely right that, whether we should cultivate social trust in medical institutions depends on whether we are above, or below, the (local) optimum of such trust. This holds whether we define that optimum in terms of the rationality of placing such trust, the social utility of its existence, or any other standard.

Some writers for whom informed consent matters because trust matters seem convinced that we happen to be approaching a dangerous level of distrust. We are, so to speak, far below optimum. However, little evidence supports such a bold claim. What needs to be shown is that rising distrust is pushing medical systems close to the brink of collapse, or something in that ballpark. Only that, it seems, might offset the combined badness of countless compromises of clinical outcomes that non-paternalistic practitioners make every day.

O'Neill, who discusses a crisis of trust in the medical system, admits other times that in the general population, many of those who supposedly lack trust in fact use that system regularly. Nor does she seem able to name any other truly major harm from increasing distrust. Bok does point out some damages, especially ‘difficulties in gathering census data and other types of information,’ as well as some public suspicion of medical investigations.10 But are these damages really greater than the damage of countless wasted opportunities to improve clinical outcomes by foisting beneficial care on unwilling patients, a damage wrought by informed consent? If our intuitive disapproval of imposed care is justified, another ground must make it justified.

The reality of our location in relation to the trust optimum is probably that some of us have insufficient trust in doctors while others have excessive, naïve trust in them. One possibility therefore would have been to institute fully informed consent for some of us and not for others. But different informed consent policies (given the statistical correlation with distrust, perhaps more rigorous informed consent standards for African-American research participants than for Caucasians?)27 would have obviously been problematic in their own right, partly for seeming unfair and further undermining trust.

Damage to trust

In some ways, standard informed consent requirements diminish interpersonal trust in clinicians and investigators. Their overall impact on trust in the medical system may turn out to be mixed, or negative.

Since the culture of informed consent often presents the caretaker as someone to be suspected and monitored closely, some worry that it ‘crowds out’ interpersonal trust in the doctor.2 ,15 ,19 ,39 Consider a doctor who tells her patient, ‘Don't trust me: you decide!’ Her words could carry two effects. As an invitation of audit and answerability, they make her more trustworthy and potentially more trusted.40 However, as a statement that she and her colleagues should not be trusted, they may decrease trust in them. Empirical researchers, not arm-chair bioethicists, ought to determine which effect is stronger.

Furthermore, if some (negative) factors might deter patients and research subjects or directly engender distrust, full disclosure must include them—with clear risk to trust. For this reason, it is unclear that practitioners and investigators who lay out every last detail of their conflicts of interest, for example, increase patients’ and candidate subjects’ trust in them. Imagine, ‘I will probably launch my career and earn a fortune if you undergo this experimental operation, but in all honesty I would not recommend it unless it were promising for you.’ The overall effect is complex, not purely assuring. It is true that if the public knew that disclosure was full, this would give it full assurance that nothing worse was hidden from it. But the public can never be sure that the disclosure was full. It may suspect that worse details remain hidden.

Indeed, frightening risks, juicy tales of doctors’ and investigators’ conflicts of interests, and other disclosed information might remain in the public memory and diminish many people's trust, long after the mundane fact that they were openly disclosed disappears from memory. By analogy, suppose that the worst sides of good humanitarian organisations were regularly disclosed to potential donors and the general public. A constant trickle of gossip on the inevitable petty fights among humanitarian workers, their little failures and occasional corruption was constantly on the news, open for demagogical opponents to exploit. It is far from clear that this full disclosure would boost public trust in humanitarian organisations. Short-term and longer-term trust in them has waned when they released such stories.41 Perhaps the public should have realised that transparency warrants trust. But the public doesn't.

The same part of standard informed consent requirements may also make clinicians and investigators less trustworthy. Full disclosure sometimes renders investigators overconfident in their righteousness—and consequently less trustworthy.42 ,43 This might turn out to apply to full disclosure of the risks from research participation and to other aspects of informed consent.

Finally, when trust starts out very low, the standard consent process may diminish it further. In interviews with African-Americans about medical research, participants believed the purpose of the consent document was to protect hospitals and doctors from any legal responsibility. As one participant stated, ‘If you give consent, then you don't have any legal rights. When you sign that paper, you sign all of your rights away because they have disclaimers all neatly typed up, reviewed by their lawyers to protect themselves from being sued.’27 ,29

If indeed standard informed consent practices in some ways damage trust, then their net effect on trust could turn out to be negative, a matter for empirical investigation. Interestingly, ‘the decline in public (confidence in, and respect for, doctors) after 1950 coincided with the call for accountability of the medical profession.’40 One possible explanation for that decline would be that informed consent and other accountability measures had a negative net effect on trust in doctors.

Overemphasising trust

Intuitively, it seems justified to adhere to certain practices even at the price of damage to trust. Take the following, motley assortment of examples. ‘Participatory democracy, better public education, the attention of the media and mistrust of authority and experts in general—all have weakened the trust relationship’, especially between doctors and patients19—but clearly it remains permissible to maintain participatory democracy, better public education and so forth: we should be willing to live with some loss of trust in doctors. Likewise, although atheists may be America's most distrusted minority,44 it remains permissible to hire openly atheistic physicians for American communities, even if that turned out to undermine communal trust in doctors. Finally, ‘the sense of trust is diminished as the doctor-patient relationship becomes more specialised, impersonal, and short-lived,’11 ,20 ,24 ,40 and yet surely it remains permissible to offer patients specialist care. If the need to maintain trust is not strong enough to transform our obligations to cultivate participatory democracy, hire atheists and offer specialist advice, why assume that it is strong enough to generate informed consent requirements which commonsense morality construes as powerful enough to cost patients their lives?

Underemphasising trust

One response to some of the challenges noted so far would have been to say that standard informed consent policies promote trustworthiness, not trust. Naïve patients trust practitioners, but they only have reason to trust trustworthy ones. Bigots distrust atheist doctors, but they have reason to trust many, who are trustworthy. Indeed, for Onora O'Neill, ‘Informed consent requirements are one aspect of trustworthy institutions’ (my italics).2

This response, however, would raise a different challenge. How is improving trustworthiness addressing the bad outcomes of low trust? As O'Neill admits, trustworthiness does not always increase trust: ‘We may need more than improvements in trustworthiness if we are to have any restoration of trust.’2 For O'Neill, justified informed consent works only with reasonable people. Obviously, however, they need not constitute the majority. If the problem is real-world distrust crises, why increase trustworthiness, which may turn out to increase trust only in an ideal world, populated by reasonable people? Shouldn't we resolve the actual crisis?

Conceivably, O'Neill would deem it (utility-maximising, but) unduly intrusive to manipulate people into trusting doctors who are worthy of distrust. A Kantian, she may prefer a policy of giving people sound reason to trust doctors—say, by never deceiving them but instead letting them decide on their own how much to trust doctors. This, nevertheless, would already assume that doctors have strong independent (Kantian?) reason against manipulating patients into trusting them when they are untrustworthy. This strong independent reason may then ground informed consent, for example the prohibition on manipulating patients, sufficiently, leaving no role for the case from trust-promotion.

Self-distrust

There is no informed consent if coming to see the doctor cedes sovereignty over one's body, and the doctor can then proceed with intervention despite one's concurrent refusal of that intervention. Even if the decision to come see her was informed and voluntary, commonsense informed consent standards are violated. Nonetheless, some patients would find such violations of standard informed consent requirements an attractive means of self-commitment. So their trust and medical adherence would only increase. Let me illustrate.

Imagine a devout Catholic who expects to see a decline in her currently strong faith in the course of prolonged treatment. She may expect the nuns who treat her to remain closer to her present world view throughout, including when it is time to either continue or stop the treatment in the light of religious considerations. It would make sense for her to write a binding advance directive that allows the nuns to force treatment on her future self, autonomous as she may be in the future. And it would make sense to skip form filling by choosing to get care at a convent that imposes Catholic care on its patients.

Or picture a capacitated and self-conscious, but also obstinate patient who knows that in the event of disagreement with the medical team, she would insist even without warrant. Knowing that about herself before visiting the clinic, she may find imposed care in violation of standard informed consent requirements attractive, as self-binding that would help her fulfil her true (in this case) future desires.

A worry we noted earlier, that without informed consent patients would stay away from clinics because coming there would risk unwanted care, may therefore conflate two decision nodes of these patients. One decision takes place early, and it concerns whether to come see the doctor. Another occurs late, and concerns whom to follow, the doctor or oneself, when disagreement with the doctor arises. Informed consent protects the freedom of the second decision, but that need not affect the first decision.

While usually people trust their own future selves more than they do doctors, this is not a conceptual truth. There may be pockets where such extreme alienation from one's expected future instructions are commonplace. There, trust-promotion would count against standard informed consent requirements.

Trust and small lies

As mentioned, Bok objects to deceptive psychological research among other things because it impedes trust in scientists and psychologists: to the extent that [research participants] learned, after the fact, of deceptive studies such as the one to which they had unwittingly been subjected, they would have greater reason than before to be skeptical of the hospital in question… and possibly of health professionals more generally. Such doubts could add to their reluctance to seek medical help in the future and to speak candidly enough with health professionals to receive the best possible advice.10 ,11 ,45

I doubt that psychological research really undermines subjects’ trust in these other respects. As Bok mentions, most subjects of such research are psychology students. This population seems to use medical and psychological services no less than the general public.

Bok cites only rather limited negative outcomes from the often trivial deceit of psychology students—namely, that ‘college students, who supply the majority of the subjects for social science research, have developed their own defensive strategies to confound investigators targeting them for deceptive research.’ She uses these outcomes as grounds for far-reaching conclusions: Thus, for professionals, as for public servants, it is more important than ever to win back the active trust that they no longer can count on receiving automatically. Moreover, winning this active trust requires them to conduct themselves in a way that leaves no doubt about their acceptance of ordinary standards of honesty and fair treatment.10 ,11

The sense of urgency informing the latter quote is unwarranted by its rationalisation in the quote that preceded it. Small lies rarely translate into full distrust of everything someone says or does.

Take another example of a lie small enough to leave trust in place. Jackson, who thinks doctors should be told never to lie to patients lest trust in doctors diminishes, concedes, ‘You may brush aside a patient's apology for calling you out in the night with ‘No problem!’ however unwelcome the call.’ Jackson's response to the challenge is that these are hardly lies: ‘No one relies on such conventional expressions as indicative of a person's real feelings…’4 But the challenge persists, because even if some patient were known to take such statements literally, it would remain morally permissible, or praiseworthy, to answer ‘No problem!’

Bok's and Jackson's reasoning here may misconstrue the nature of trust: ‘trust is generally a three part relation: A trusts B to do x (or with respect to x).’46 Psychology students who were deceived during research might stop trusting psychology researchers to tell study participants the truth during trials; they could continue to trust them to write truthful articles and to be caring and truthful practitioners.47 Patients who see that the physician who answers ‘No problem’ is tired might lose trust in her polite answers, but these patients’ trust in her dedication and professionalism may grow.

Bok also cites examples of specific deceitful studies which had deep adverse effects on subjects’ and on the public's trust in investigators. Nancy Kass and colleagues are of the same mind: ‘Examples from the past highlight that abuses of human subjects have a lasting and devastating effect on individual's trust of biomedical research, and also on entire communities.’5 Both are right that the grossly abusive breach of trust in Tuskegee, or in studies that stigmatised non-cognisant participants, may have had a crushing effect on communities’ trust in researchers. But that does not show that very minor breaches of trust like the ones involved in most psychological research and corrected shortly thereafter have a similar effect. Tuskegee-scale violations can undermine trust in the basic decency of trust violators and hence in virtually anything they and their professional communities do or say. But deceit in psychological research rarely involves such violations. It violates standard informed consent requirements without jeopardising public trust in the system.

Trust and pertinent information

Fully informed consent to risky interventions requires, among other things, disclosure and/or comprehension of the relevant risk, and of some other pertinent facts. To justify the standard requirement of informed consent, the need for trust would have had to justify full disclosure or comprehension of all such facts.

Nevertheless, disclosure or comprehension of such pertinent facts might be unnecessary for keeping doctors and investigators either trusted or trustworthy. What seems more necessary is easy access to a full account of the facts, should the patient or study participant decide to ‘audit’ her caretaker or investigator. Why? Because the mere possibility of an audit may deter caretakers and investigators from abuse—for the same reasons that audit possibilities deter abusers in general: they make abuse risky. It follows that if rational clinicians and investigators know that patients and study participants have full ability to kick off an audit, and no inhibition about doing so, leading to potentially severe penalties, then those patients and study participants are protected from abuse. They can afford to trust medical and research institutions not to abuse them and thus enlist in trials and so forth. This trustworthiness is accomplished without required disclosure or comprehension of all pertinent risks.

Trust and competence

What commonsensical informed consent requires is highly sensitive to the patient's (or the study participant's) level of competence for rational decision-making. Fully competent patients enjoy a right to fully informed consent. Patients with compromised decisional capacity are offered something very different: a proxy decision-maker, presumed-consent decisions made by others, or a best-interests test.

The trust-promotion argument is less sensitive to considerations of patient competence. Consider patients with paranoid schizophrenia. Many lack competency for rational decision-making and therefore lack some informed consent rights. On the other hand, because these patients are distrustful, considerations of trust-promotion may apply to them in full force, perhaps more than to other patients. If the need to protect trust provided a sufficient justification for informed consent (as the trust-promotion argument says it does), then, contrary to standard formulations of the informed consent requirement, these paranoid schizophrenic patients would probably have to have full informed consent rights. But according to commonsense they lack some consent rights. So the need to maintain trust cannot be a sufficient ground for the standard requirement of informed consent. Something like autonomy—a capacity that many such patients lack—seems more relevant here.

Indeed, children lack autonomy and some rights to consent. But Sissela Bok, who emphasises the role of truthfulness and truly informed consent for protecting trust, demands truthfulness even for children—as a matter of ‘trust in the family.’11

It might be replied (as one referee for the journal has) that most of us feel more secure if we know that, should we come to have paranoid schizophrenia, we would be treated with drugs against our wish. On this understanding of the argument, it is the trust of people in general that matters, not the trust of the people who have already developed their paranoia, perhaps because, as the referee put it, ‘they are few, we are many.’ However, commonsense morality does not take the rules governing decision-making for incapacitated patients to depend primarily on the proportion of incapacitated patients in the population, for example, on whether there are relatively many, or only a few, paranoid schizophrenic patients.

Sacrificing for trust

According to commonsense morality, when a named person is in need of being rescued, we should rescue her. Normally we should not abandon her. Moreover, we should rescue her even when her abandonment would have reserved scarce resources for rescuing a few more people in the future. One worry about grounding the standard informed consent requirement in trust promotion is that the requirement then regularly recommends what those relying only on trust to ground informed consent must take to be the abandonment of a patient in need. Envisage a patient who needs care now, but has misguided wishes about her care. Furthermore, her doctors happen to support the trust-promotion argument and to see it as the sole valid ground for informed consent. They accept her wishes simply to preserve social trust in the medical system. What the doctors do would seem to abandon a named individual in order to preserve social trust in the medical system. Contrary to commonsense norms about rescue, they abandon her for the sake of other social goals, including unidentified patients’ care. They omit to provide what they take to be valuable care for a sick patient now, only as a means to preserving trust in the system, for other, unidentified patients’ sakes. Commonsense morality likes to hold that a clear and imminent danger for a concrete patient normally takes first priority—even if diminished trust in the system would cost more future patients direly.

Of course, outside this stylised example, even doctors who see trust promotion as a sufficient ground may accept other grounds for informed consent. Still, the stylised example helps by questioning whether the fact that there are benefits to future patients from the preservation of trust would warrant informed consent policy even absent other valid grounds. The trust-promotion argument says that it would.

Trust and prior commitment to informed consent

The very need to preserve patients’ trust in the medical system may presuppose independent reasons to condition medical intervention on consent. The trust of patients and research subjects in that system would have been far less important if doctors forced citizens to undergo checkups and potential treatment or experiments—in violation of standard informed consent requirements. Our moral repugnance at this prospect and our consequent reliance on patients and research subjects to trust physicians and approach them voluntarily presuppose at least some parts of the standard requirement of informed consent, and cannot fully ground it.

An implication is that, strictly speaking it is false that only if the trust that is placed in investigators ‘is deserved can the research enterprise survive.’5 Outright coercion or naïve trust could have sustained the research enterprise even without deserved trust. We need deserved trust in researchers and practitioners only given some prior commitment to informed consent. But if we have that commitment, trust-building might not be our most fundamental ground for informed consent.