It is here that the physician must show herself capable of empathy. According to Carl Rogers ’ definition, this means

to perceive the internal frame of reference of another with accuracy and with the emotional components and meanings which pertain thereto as if one were the person, but without ever losing the “as if” condition. Thus, it means to sense the hurt or the pleasure of another as he senses it and to perceive the causes thereof as he perceives them, but without ever losing the recognition that it is as if I were hurt or pleased and so forth. (Rogers 1959)

The idea of empathy is at the root of person-centered therapy, developed by C. Rogers (see Prochaska 1994, 133–162) and of the general concept of counseling, which is used particularly to improve adherence in illnesses such as AIDS .

And when the patient perceives in the physician the desire to show empathy towards her, perhaps she will feel the symmetrical desire to attempt to share beliefs . This description of the doctor -patient encounter finds an echo in the description of the therapeutic relationship by Michael Balint :

It is on this basis of mutual satisfaction and frustration that a unique relationship is established between a general practitioner and those who stay with him. It is very difficult to describe this relationship in psychological terms. It is not love or mutual respect, or mutual identification or friendship, although all these elements are present in the relationship. We have called, for lack of a better word, a ‘company of mutual investment’. What we mean by this is that the general practitioner progressively acquires precious capital invested in his patient and, respectively, the patient acquires a precious capital which he deposits in his doctor . (Balint 1988, 265)

The sharing of beliefs implies a dialogue aimed first at discovering the other’s beliefs concerning her representation of illness and health , her expectations and fears concerning the treatment and the physician. This conversation is the beginning of a negotiation, which will lead to a necessary therapeutic alliance between the physician and the patient .

Here we can quote an enlightening passage from Jean-François Bloch-Lainé ’s text on therapeutic alliance :

The conditions required to achieve a therapeutic alliance when treating drug addiction are, at first glance, the same as the conditions generally required in any medical situation: the patient is identified as ill, the therapist is identified as the one offering treatment. The therapeutic alliance between the patient and the physician can exist only as long as there is no contempt, no complicity, no indifference or complacency, no hate or love , but only genuine mutual respect. The respect felt by the patient towards the physician cannot be dissociated from the required trust that the patient has in the physician, everyone agrees with this banality. The respect of the patient felt by the physician is more difficult to formulate. To respect the patient supposes that one respects the person treated and the reality of her illness (Bloch-Lainé , Internet).

The first part of this text refers to empathy, according to the definition given earlier. One wonders if it is not easier to formulate the physician’s respect for her patient if the former recognizes the legitimacy of the patient’s perspective on the illness in addition to the medical perspective. Is it not a display of empathy to show the patient that her perspective on the illness is legitimate? This may lead the patient to accept that the physician’s perspective is legitimate as well. Trust , in the context of the exchange of representations, becomes the basis for mutual respect.

Here it is useful to clarify more precisely the concept of belief . Following Pascal Engel , there are several types of beliefs , defined according to the degree of objective guarantee attached to the representation and according to the degree of subjective confidence the agent has as to the truth of this representation. Thus we can speak of hypotheses or conjectures when the beliefs are likely to be true or to have an objective basis, or when they are being verified; of convictions, when there is a strong subjective feeling but its basis is not guaranteed; of prejudices and superstitions when the objective guarantee of the opinion is very weak or nonexistent, although the agent may feel a very strong conviction to the contrary; and finally of confidence (in someone) or faith (in something) when, despite a very weak objective guarantee, the degree of subjective certitude is very strong and goes further than what the facts and guarantees allow one to affirm (Engel 1995, 10–11).

Using this classification, we can distinguish those that concern more or less verifiable facts and those that concern a priori unverifiable representations, because they concern the future , which is essentially contingent.

First, there are matters of fact of which patients may nonetheless feel uncertain. Patients’ knowledge often includes some uncertainty and must be seen as beliefs of one of the two first categories, hypotheses, conjectures or convictions, as described above. This does not mean that their propositional content , i.e., what they believe, is not true. But we can also suppose that the degree of certitude plays a role in the belief ’s motivational force . We have studied the effect of the degree of certitude concerning diabetic patients’ knowledge on the adjustment of insulin doses (Reach et al. 2005).

Recall that when a belief is part of an action ’s reason/cause, it is typically a belief about the outcome of that action. For instance, the active belief in the intention to drink tea in order to alleviate urinary burning is a belief about the action effected: Namely, it’s the belief that the action will have the desired effect. Since these beliefs concern the future , and since the future is uncertain, such beliefs are also of the second type. They belong more to the last two categories of beliefs, trust or faith, but also prejudice or superstition.

But we can note that the same may be true of numerous medical ideas, which, not having been proven and the physician not being able to affirm that they are certain, are deep down merely beliefs on which they base their therapeutic strategies. This is what justifies the emergence of evidence-based medicine , which aims to eliminate uncertainty from the development of therapeutic strategies, or, more precisely, to quantify the uncertainty. It uses large scale controlled clinical studies and meta-analysis of all the available information produced by these studies: The principle of statistical demonstration by a controlled study is to limit the risk of error by depending primarily on probability claims. Ever since the DCCT study, I don’t simply believe that control of diabetes lowers the risk of complications, I think that I know it; thus I feel more authorized to propose this therapeutic strategy to you. The advent of evidence-based medicine could come down to applying Ramsey ’s principle, stipulating that the degree of truth of a belief represents the probability that it will lead to a successful action .

The preceding nonetheless implies that the sometimes antagonistic relationship between physician and patient is not a conflict between knowledge on one side and beliefs on the other, but rather a confrontation between two webs of beliefs, i.e., between two holisms. The patient’s web of beliefs is greatly shaped by her past and her personal experiences. That of the physician is vaster and better documented, but only in what concerns the prediction for the future of an actual case. The physician believes that she can make an analogy between what she knows about what happened in the past with other patients, whether that experience be personal or from books, and from the evidence-based medicine, although this evidence is essentially statistical.

According to Hippocrates :

Physicians come to a case in full health of body and mind. They compare the present symptoms of the patient with similar cases they have seen in the past, so that they can say how cures were affected then. But consider the view of the patients. They do not know what they are suffering from, nor why they are suffering from it, nor what will succeed their present symptoms. Nor have they experience of the course of similar cases. Their present pains are increased by fears for the future . They are full of disease and starved of nourishment; they prefer an immediate alleviation of pain to a remedy that will return them to health. Although they have no wish to die, they have not the courage to be patient. Such is their condition when they receive the physician’s orders. Which then is more likely? That they will carry out the doctor ’s orders or do something else? (Hippocrates , Prognosis)

The physician must take into account her patient ’s past, present, and future . The Persona^® test which places people in the categories of “analyzer, facilitator, promoter, controller” suggests that the ‘facilitating’ individuals are those who face at the same time the past, the present and the future. It is in this category that we find most doctors…

It is because the physician also has this vision of the future that she can act on her patient ’s behalf according to her own principle of foresight , and we will precisely see in the next chapter that failing to abide by this principle may represent a cause of doctor ’s clinical inertia . And her most difficult task is to help her patient acquire, little by little, the principle that will lead her on the road to adherence. In so doing, the physician fosters acceptance of propositional attitudes that form the holistic richness of her mind, the facts that will protect her future rather than those, often more seductive, that would allow her to enjoy the present moment.

It is precisely here that there is a difficulty due to a difference between the functioning of the physician’s mind and the patient ’s: If we return to the concept of time discounting described earlier, we can calculate that the discount rate and its curve must sometimes be different for the physician and the patient.

Steven Feldman clearly showed the reasons for these differences (Feldman et al. 2002). For instance, because the physician has chosen the medical profession and practices it, she likely has a weaker discount rate for the future than the patient , or attributes more importance to health . In short, physicians may be more able to delay gratification and overplay the importance of health. In particular, patients from cultures where the present moment is important might have a very different future discount rate. Certainly, we have insisted on the fact that the operational mode of thinking is universal: The role of beliefs , desires and emotions in the genesis of actions is certainly the same for all humans. But this is fundamentally true only from the qualitative point of view, and quantitative differences—say, for the future discount rate—may have important consequences.

We can mention two conditions where the discount rate of the patient and the physician are different: Patient’s social deprivation and serious depression. Here, the feeling of the future has more or less disappeared (the future discount rate approaches infinity) and a physician who proposes a treatment of indefinite duration is sure to fail. This obviously does not mean that all therapeutic attempts are impossible; it merely illustrates the importance of taking into account considerations of this type. In this article, the authors suggest that the discount rate might not play an important role in the short term. This may explain the higher frequency of adherence in cases of acute diseases.

Two aspects of the therapeutic relationship emerge at the conclusion of this analysis. First, we saw that one cannot decide, all of sudden, to believe something, just as ‘one decides to go away for the weekend’ as Pascal Engel puts it; on the other hand, one may believe what one is told—the patient believes what the physician tells her: As we saw, the sharing of beliefs (and desires ) between the physician and patient is possible. Caring is sharing: Saint Martin doesn’t give his coat to the beggar, he shares it.

This sharing does not concern only beliefs and desires . It can also include emotions . Here we can again quote Pierre Livet :

the sharing of emotions plays a double role here. On the one hand, it is an important emotional consequence of the preference being tested. To discover that others do not share the emotions connected to our preference provokes a strong negative emotion. Conversely, the fact that others place a certain value on it, which we have not considered, is an important source of emotion and revision of our preferences. On the other hand, it is an empirical ersatz of universalisation. If we can suppose in advance that our emotion is not shareable, we cannot give value to our preference, even if for us it is a deeply rooted preference (Livet 2002, 185).

What this remark suggests is that the patient and the physician, during their interaction, are led to confront arguments of an emotional nature. For example, when the patient invokes her pleasure of smoking , the physician fears the appearance of long-term complications if the patient does not give it up. And according to the definition of empathy, one must be capable of feeling what the other one feels. In the next chapter, we will revisit the very concept of empathy and show how, paradoxically, its use can lead to clinical inertia . To try to understand someone else’s emotions means checking whether she is ready to revise her preferences. For Livet , the debate about preferences is actually a debate about values. If the patient is resistant to the idea of a revision of preferences, that means that she places more value in smoking than in the idea of health . The first role of the physician is then to demonstrate to her the value of health, for which the physician is a sort of representative. That is why the physician certainly has the right to smoke , but smoking in front of patients is highly reprehensible. Health might be seen as a universal value, or even the archetype of what is valued: According to Canguilhem ,

for man health is a feeling of assurance in life to which no limits is fixed. Valere, from which value derives, means to be in good health in Latin. Health is a way of tackling existence as one feels that one is not only possessor or bearer but also; if necessary; creator of value; establisher of vital norms (Canguilhem 1989, 201).

But often the discussion falls short and the patient does not admit the necessity of revising her preferences. Then we come to a conflict, and nonadherence is likely to be its expression. I will show that in fact it is not a conflict between the physician and the patient. It is a deeper issue of a conflict between two current trends in medicine : One that consists in desiring to benefit from the progress of modern science and ‘to improve adherence’, in the name of the patient’s well-being, and one that wants to promote patient autonomy .

Up to now we have assumed that the physician’s desire to improve adherence is self evident. In a recent report by the World Health Organization (2003) devoted to adherence in cases of chronic disease we read that finding a solution to nonadherence would be more beneficial than any other medical advance. However, at a time when medicine is becoming more and more effective and sometimes even manages to prove its effectiveness according to the criteria of evidence-based medicine, society is becoming more sensitive to patient rights and autonomy . There looms a possibility of conflict created by medicine itself: What attitude to adopt when the physician’s and the patient’s points of view diverge? How far can the physician go in her desire to convince the patient?

Etymologically, a person is autonomous if she chooses herself (autos) which rules (nomos) she is going to follow and applies them, in the same way that an autonomous government writes its own laws and has the freedom to enforce them. This supposes that the person, like the government, intends to have control over her actions and does not give the right to control her action to anyone else without express permission. Analyzing the notion of autonomy then comes down to analyzing the control that the person has over her actions. And inasmuch as an action can be characterized by its intentional nature, analyzing the notion of autonomy comes down to analyzing the control a person has over her reasons and her capacity to act according to these reasons. More precisely, the autonomous person has the capacity to perform an action or to not perform it, or to perform one action rather than another. Thus she is endowed with the capacity to choose. However, we must still clarify whether we are speaking of an autonomous person or an autonomous action. Indeed, there are persons whom we would not consider autonomous who nonetheless perform autonomous actions and, inversely, autonomous persons who from time to time perform non-autonomous actions.

When we go from the generic concept of personal autonomy to the specific concept of personal autonomy in the context of a medical decision —making a choice affecting one’s health —the concept is somewhat modified. Call the modified concept therapeutic autonomy. We are now no longer considering only the person. Into the discussion is introduced the fact that the person is engaged in a therapeutic, or binary, situation. In this situation, the person became ill and comes to consult a doctor ; she has become a patient who will be dealing with a medical team: The question of autonomy now needs to be considered not only from the point of view of a person, but of a person in a therapeutic relationship with another. The person will have to be autonomous not only in relation to herself, i.e. to have or not have control over her actions, but also in relation to the physician, to another, who may want to control her actions (in fact, the relationship may be more complicated than simply binary between the patient and the physician, as it may involve, for instance, the patient’s family and friends, who might also want to intervene in the decision making process).

This has a major implication: Recognizing that the patient , as a person, is an autonomous being implies for the physician a particular behavior in regard to her autonomy . And, in an attempt to benefit the patient, this behavior can override the patient’s decision (this is often called ‘paternalism’); or, it can seek to respect the patient, which for now can be termed ‘autonomist behavior of the physician’.

In a psychological , or philosophical , conception of autonomy , the autonomous individual is capable of choosing according to her values and preferences which are an expression of her individuality: For instance, a patient could choose to quit smoking because she places more value in the idea of health than in the pleasure of smoking. Hence it is the notions of value and preference that it is now time to analyze: What does it mean to assign a value to something, which will give it a certain weight during decision making?