Introduction

The role and importance of empathy in clinical practice have been widely discussed.Footnote 1 Definitions of empathy vary, ranging from the cognitive ability to understand the emotional states and perspectives of another individual [2] to a combination of cognitive and affective capacities, such as emotional resonance, and even to behavioral responses [3]. It is also an important feature of empathy in medicine that physicians be able to communicate their understanding of the patient state back to the patient. Ultimately, the goal of employing empathy in clinical settings is to produce better health outcomes for patients.

Proponents of empathy in medical practice argue that empathy plays a crucial role in the physician–patient relationship, fostering better communication, treatment, and patient satisfaction [3,4,5]. Ideally, physicians who employ empathy in communication and patient engagement will learn more about patient concerns and symptoms, given both that they are more attentive to the emotional states and perspectives of their patients and that their patients are more willing to communicate. Such engagement is thought to produce better diagnoses and treatment plans. Patients feel that they are listened to and understood and that their physician is concerned about their well-being [6]. In light of these purported benefits, some proponents of applying clinical empathy argue that empathy is one of the “essential attitudes and skills” of a medical practitioner [7].Footnote 2 For example, in its learning objectives for medical school, the Association of American Medical Colleges states that “physicians must be compassionate and empathetic in caring for patients” [8], and several prominent hospitals and medical schools have begun to adopt empathy training programs.Footnote 3

This paper focuses on the ideal of clinical empathy, as involving both cognitive and affective capacities. I argue that one widely discussed account of clinical empathy is subject to a number of objections. My main concern is that a central feature of the account—namely, emotional resonance with patients—may serve more as a liability than as a benefit in clinical settings. Additionally, relevant empirical research does not clearly support the benefits of employing emotional resonance with patients. Instead, I argue that the ideal account of empathy in medicine remains cognitive, though there is a central role for expressing empathic concern (sympathy or compassion) toward patients.

Before addressing the issue of empathy in clinical settings, I should briefly delineate the relevant class of individuals for whom empathy is therapeutically relevant. Since empathy has been proposed as an important—or, to some, necessary—feature of the physician–patient relationship, I limit the scope of my discussion on empathy in medicine to the subset of physicians who have regular, medically related personal interactions with patients as part of the patient diagnosis and treatment process. This includes both primary care physicians, such as general practitioners or family physicians, and the host of specialist practitioners (e.g., surgeons, radiologists, oncologists).

From detached concern to clinical empathy

According to Jodi Halpern [3], the emotional ideal for the physician in modern medicine has been one of emotionally detached concern [3, 11, 12]. The ideal of detached concern involves observation and projection into the inner life of the patient “without feeling … difficult emotions” [11]. This is not to deny that physicians can and will experience negative emotions, but to say that they should make efforts to avoid such experiences. For example, Mohammadreza Hojat et al. define empathy in medicine as a “cognitive (as opposed to affective or emotional) attribute that involves an understanding … of patients’ experiences, concerns, and perspectives combined with a capacity to communicate this understanding” to the patient [2, p. 1183]. Such cognitive empathy is measured by focusing on the perspective-taking abilities of the subjects.Footnote 4

There are a number of reasons that proponents of detached concern have favored this ideal. First, sharing the negative affective states of patients is believed to impede effective medical care, since such exposure can produce emotional fatigue, leading to less careful treatment.Footnote 5 Second, such intellectual engagement provides an additional source of information about a patient’s history and present state, while promoting further diagnostic communication. Finally, since detached concern is an intellectual capacity, it can be developed with educational programs and techniques in a more obvious way than can affective and emotional responses [2, p. 1183].

Clinical empathy

While detached concern has plausible advantages over less engaged modes of information gathering and communication with patients, it has been challenged in recent years. For example, Halpern [3, 11] has argued that we should replace the detached concern model with clinical empathy. Clinical empathy is presented as a response to the merely cognitive approach of detached concern and is “an experiential way of grasping another’s emotional states” [11, p. 673]. It is a skill that employs cognitive empathy but encourages and often relies on emotional resonance with patients, though sharing the emotional states of the target is not always required.Footnote 6 Halpern gives several ordinary examples of emotional resonance. For example, when “listening to an anxious friend, one becomes anxious, while talking with a coworker, one feels heavy, depressed feelings” [11, p. 671]. While emotional resonance may occur throughout an encounter with a patient, the physician must imaginatively “unify the details and nuances of the patient’s life into an integrated affective experience” [3, p. 88]. In addition, clinical empathy involves being able to communicate this understanding back to a patient as part of the therapeutic relationship.

In more recent research, Jean Decety and colleagues argue that “the underlying rationale for implementing a ‘detached concern’ approach is no longer tenable” for a number of reasons [14, p. 233]. First, they argue that patients respond better to emotionally attuned physicians and that empathetic physicians produce more efficacious treatment.Footnote 7 Second, they argue that affective engagement helps to improve empathic accuracy and understanding, since cognitive and affective processes typically work together in the experience of empathy. Finally, they argue that the view that emotional resonance (in clinical settings) necessarily produces emotional turmoil in physicians is empirically unsupported. I examine each of these claims in turn and outline several objections to each of them.

The value of emotional resonance in clinical situations

Emotional resonance and patient trust

A central claim in the literature about empathy in medicine is that by engaging in affective attunement with patients, physicians increase patient trust [15]. As a result of this increased trust, patients are presumably willing to give fuller case histories and adhere more closely to their treatment schedules. The study most widely cited to support this claim is presented by Anthony Suchman et al. [16].Footnote 8 Suchman et al. lay out a model for more effective clinical communication, based on patient interviews with primary care physicians. The study’s findings indicate that physicians who acknowledge expressions of patient emotions—a practice Suchman et al. characterize as empathic response—or probe for further information based on emotive cues achieve more complete histories from patients.

I have several concerns with the general connection between emotional resonance and patient trust and treatment adherence. In the Suchman et al. study, for example, researchers studied the extent to which physicians observed and drew attention to emotions and patient descriptions of emotional states. Yet the study provides no clear evidence that physicians who are more effective communicators in such interactions utilize emotional resonance as a means to this end. One could engage in cognitive empathy with a patient, while also acknowledging the importance of the patient’s emotional states and attending to the various verbal cues that index their emotional states. In other studies on this topic, the account of empathy in physician–patient communication involves a cognitive form of understanding.Footnote 9 In other words, the importance of giving attention to the emotional state of the patient and to patient concerns can be granted without encouraging emotional resonance in clinical situations.

Empathy and delivering bad news

Another claim is that patients cope better when their physicians deliver bad news to them empathetically [20]. This is a plausible claim, but it is important to clarify how empathy is typically used in such recommendations. One widely cited schema for breaking bad news in clinical settings—known as SPIKES [21]Footnote 10—defines empathy as an exclusively cognitive attribute. Empathy here involves observing a patient’s emotion, identifying the emotion and the reason for the emotion, giving the patient time to express the emotion, and acknowledging the emotion by connecting it with the reason for the emotion. The study cited by Halpern [11] in support of this claim does not clearly define empathy, but it does recommend that physicians give patients time to express their feelings and remind them that it is normal to feel negative emotions in such situations [20, p. 57]. Here again, while there is support for recognizing and acknowledging patient emotional states, there is not clear support for engaging in emotional sharing or resonance with patients in clinical settings.

Physician empathy improves therapeutic efficacy

Halpern cites a number of studies in support of the claim that physician empathy improves therapeutic efficacy. In one study, Debra Roter et al. [22] find that patient trust is vital to patients’ adherence to treatment regimes. However, the study shows that patient trust is dependent on whether the physician seems to be concerned about the patient’s health, not whether some measure of emotional resonance is employed. The studies that explicitly mention empathy as part of the communicative process employ cognitive accounts [19, 23]. For instance, Rainer Beck et al. [24] define empathy as a cognitive capacity, finding that additional non-verbal behaviors like mutual gaze, nodding, and leaning in predict more favorable outcomes. Kim et al. [19] define affective empathy as the ability to be responsive to and improve a patient’s emotional state. Measures include factors like whether the physician responds mechanically or shows interest in patients and their well-being. Kerse et al. [25] do not mention empathy, but instead measure how well patients feel understood and listened to during their appointments.

These studies offer good reason to think that establishing open communication with patients is therapeutically efficacious and increases patient satisfaction [26]. Further, attending to the patient’s emotional state and perspective as part of the clinical interaction is important. However, this research does not provide clear reasons to adopt clinical empathy as an emotional ideal in medical practice, or to think that sole reliance on cognitive empathy in medical contexts (alongside standard diagnostics) is incapable of producing effective treatment.

The epistemic value of emotional resonance

Emotional resonance may provide an additional source of information about the patient state. For instance, a physician who resonates with a patient’s emotional state throughout a conversation has a direct source of affective information that she might use to better understand the patient’s needs or perspective. Since emotional experiences guide and frame our ordinary ways of thinking and deliberating, more direct access to such states could provide a clearer window into patient concerns.Footnote 11 Such emotional “data” may support the current understanding of the patient or, in cases where this information conflicts with patient reports, provide grounds for investigation. For example, a patient might report feeling fine, but the physician might directly resonate with the depressed, fearful, anxious, or distressed state expressed by the patient. Such mismatch in information may prompt further questioning or examination.

I have two concerns with this line of argument in practical situations. First, in many clinical settings, such affective information is also accessible via basic observation or, perhaps, by using a process like cognitive empathy. Further, it remains unclear why it is diagnostically useful for a physician to experience—even in subtle or attenuated ways—patient states in order to properly diagnose or treat patient problems. Second, a patient’s precise emotional state in the diagnostic stage is less important than the relevant range of symptoms (or problems) he exhibits, his medical history, and the object of his emotional concern. Each of these aspects of diagnosis can be more accurately determined without additional emotional resonance.

Consider my first concern—that information about patients’ emotional states can be acquired through non-empathetic means. Many of the affective states experienced and expressed by patients in hospitals and clinics are negative. Patients are often anxious, irritable, angry, or depressed. In many cases, they are dealing with varying levels of pain, discomfort, and distress. Further, such states are often the first to be directly reported by patients when they detail their concerns (e.g., “I’m in terrible pain with these migraines” or “My kid won’t stop screaming when I touch his belly”). In this way, an additional affective source of information throughout the diagnostic process is typically unnecessary and distracting.

There are cases in which a patient may not be verbally explicit about his emotional state or concerns. Or perhaps the patient’s expressive and behavioral cues conflict with his verbal reports about his symptoms or condition. For example, a patient might share only partial symptoms with a physician while blushing or averting his eyes. In such cases, a physician can imagine what would motivate such behavior (perhaps embarrassment), prompting further gentle questioning. Perhaps the patient is averse to a particular form of examination, so he avoids sharing the symptoms related to the corresponding area of his body. Here, further questioning might be required for the optimal diagnosis. But, in such cases, basic observation, cognitive empathy, and inference appear to be sufficient for guiding the physician through the diagnostic process.

My second concern is that, epistemically, the current psychological state of patients during the physician–patient encounter is less important than the range of symptoms or problems they exhibit, the details of their medical histories, and the objects of their present concerns. Each of these elements can be properly obtained through dialogue, imagination, and detailed inference or induction from available information. This is not to deny the importance of cognitive empathy, imagining situations and problems from the patient’s perspective, or attending to the patient’s emotional state. While emotional resonance delivers some data about the present state of the individual, such data arrive without clear context or object. Since most of the explicit emotional states exhibited by patients in clinical settings are negative (e.g., fear, frustration, distress), a physician’s sharing in these states amounts to an additional cognitive burden without any clear epistemic benefit for the patient.

One response to these concerns might be that the form of emotional engagement that I have just outlined—namely, attention to emotion without sharing emotion—is just what is meant by clinical empathy. However, if this is the case, then it is difficult to see how this approach to the physician–patient relationship differs in a significant way from the detached concern model of empathy. If clinical empathy is a substantial development from the previous detached model, it seems to be because of the additional element of emotional resonance. But, as I have noted, the epistemic advantages of this additional affective element are not clearly greater than those of a purely cognitive approach utilizing basic observation and cognitive empathy.Footnote 12

Does emotional resonance necessarily produce turmoil?

It should be granted that emotional resonance does not necessarily produce emotional turmoil for physicians. Further, physicians regularly feel personal distress in ways that are unrelated to direct emotional resonance. It is typical for physicians to feel anxiety as they work to develop accurate diagnoses or prognoses [27]. One study found that the reason most cited by emergency physicians when ordering additional tests is fear of missing a correct diagnosis, closely followed by fear of litigation [28]. In other cases, physicians experience sadness at the loss of patients and feelings of failure related to their inability to help [29]. These forms of anxiety and grief, compounded by difficult and stressful institutional conditions, seem to be primary sources of physician burnout.Footnote 13

While the causal sources of personal distress in medical settings are complex, emotional resonance with patients does appear to be a contributing factor, particularly in the form of so-called compassion fatigue and secondary traumatic stress. Such secondary vicarious trauma is often pronounced in nursing staff who attend to trauma patients [30] and physicians who work in emergency, trauma, and oncology care [31, p. 249]. In a large survey of physicians, Ezequiel Gleichgerrcht and Decety [32] found a significant relationship between the experience of secondary traumatic stress and physician burnout. There is also evidence for the deleterious effects of such personal distress in multiple clinical fields, including an increased likelihood of medical errors.Footnote 14

Outside of the medical environment, an individual’s typical response to the perception of suffering in others is an aversive experience [35]. However, as part of standard medical training and experience, physicians ostensibly learn to down-regulate their negative emotional arousal, particularly when perceiving painful procedures [36, 37].Footnote 15 Further, physicians’ regulatory abilities serve an important role in decreasing personal distress. There are a number of regulatory strategies that can be employed in clinical settings. Gleichgerrcht and Decety [31] note three: exposure control, emotion suppression, and framing [20]. Examples of exposure control include scheduling shifts so that clinical staff are not required to experience constant patient distress or pain and balancing emotionally demanding medical care with less intense patient interactions. Emotion suppression involves deliberately avoiding thoughts related to the target of distress, while framing might involve reappraisal of patient distress (it is not that bad) or deliberate objectification (they are not real, they cannot feel at all).Footnote 16 Another important method for reframing involves deliberately taking the perspective of the target, a method that appears to down-regulate personal distress in subjects [39].

While emotional regulation arguably helps to protect the physician from vicarious distress, Decety and Gleichgerrcht argue that total affective detachment is not clinically desirable for two reasons. First, “excessive regulation has been shown to lead to personal distress and increased anxiety with both physiological and sociopsychological consequences that included increased blood pressure, disrupted communication, and reduced rapport” [31, p. 250]. Second, Gleichgerrcht and Decety argue that some measure of personal distress is necessary to “attune to and empathically understand the patient’s emotions” [31, p. 252],Footnote 17 as well as to benefit fully from the positive aspects of patient care, culminating in what is sometimes called compassion satisfaction [33, p. 268].Footnote 18

It may be the case that total affective detachment in clinical settings is undesirable, but the evidence cited against such detachment does not clearly support this. In the primary study cited in support of this claim, Emily Butler et al. argue that “in some contexts … suppressing emotion disrupts communication, hinders the development of social bonds, and is physiologically taxing for both the suppressor and her social partner” [40, p. 62]. Emotional suppression involves not showing or communicating emotional states that are being experienced. It is also worth noting that while this form of suppression has costs, the authors nonetheless observe that in other contexts the benefits of suppression outweigh the associated costs. There are good reasons for thinking that this is likely the case in many clinical interactions. Many of the negative emotions experienced in clinical settings would be unsettling to patients if they saw them expressed by their physicians. In some cases, such emotions are irrelevant to the patient interaction, and in others, they are likely to be distressing to the patient. Consider a physician who shares a bad prognosis with a patient and expresses anxiety at the results. Such a response would likely exacerbate the already difficult situation for the patient. Or consider how unpleasant it might be for a patient who suffers from an embarrassing ailment to see the empathic embarrassment in the faces of his care providers.

Perhaps some measure of emotional resonance is an inevitable feature of working with patients (particularly those in distress). Further, in cases where emotional resonance is clinically useful, communicating shared emotional understanding back to the patient seems merited. In other cases, however, the distressing emotions received via resonance have no additional diagnostic or therapeutic benefit. Additionally, for some pervasive and difficult emotions, the ideal outlet for the physician’s expression falls outside the physician–patient relationship.Footnote 19 For example, this need could be better met if physicians readily shared their negative emotional states with other medical professionals, support groups, or support staff like therapists.

Gleichgerrcht and Decety’s second claim also needs to be qualified, since I do not think that clinical empathy involving emotional resonance is either motivationally or epistemically required for careful and effective medical treatment. What seems beneficial with respect to compassion satisfaction is that physicians experience empathic concern for their patients. Psychologists often define empathic concern as an “other-oriented emotion elicited by and congruent with the perceived welfare of someone in need” [42, p. 11]. This experience has a different emotional shape than affective resonance or personal distress. For example, mere emotional resonance with someone’s anxious state is not sufficient for me to feel sympathy for her. However, if I feel bad that she is terrified of her upcoming procedure, I can feel sympathy for her without resonating with or sharing her anxious state. My emotional state is other-oriented in that my feeling is directed toward her, in her situation. In this case, my sympathy is congruent with her perceived welfare—which is reduced, given her fear. Sympathy is distinct from personal distress, with the latter involving an emotional state that is self-oriented.Footnote 20 I might be said to experience personal distress if another person’s anxious state causes me to feel distressed and my focus then shifts to my own affective experience—that is, how distressed I am feeling.Footnote 21

While emotional resonance might serve an epistemic role in motivating a concerned response, people often feel sympathy without sharing any of the actual affective states of the target or focusing on our own discomfort. Epistemically, what seems necessary for sympathy is that one perceives another being in a state of need that, in a broad sense, bears negatively on the welfare of that being. In addition, people often feel an other-oriented emotional state congruent with this perceived state. While affective engagement can help to track these cues, this information is available through direct observation and cognitive empathy.Footnote 22 For the reasons noted above, it is not clear that physicians need to directly share in their patients’ affective states or experience self-directed personal distress in response to their patients in order to engage in empathic concern or sympathy for them.Footnote 23

Refining clinical empathy

The above-mentioned literature does give reason to think that many of the purported benefits of empathy in clinical communication and information gathering are the result of cognitive engagement with patients, which sometimes involves employing cognitive empathy. More recent research also supports this claim. Jessica Ogle et al. [43] argue that third-party institutional reviews of physician–patient interactions show a close relationship between a physician’s ability to place herself in the position of her patients and to see the world from their perspective—which I understand as a form of cognitive empathy—and her general competence as a clinical communicator.

Previous research similarly suggests that a physician’s empathetic abilities, understood as her abilities to take the perspective of the patient, correlate with more favorable health outcomes for her patients [26], as well as increased patient satisfaction [44]. Moreover, Martin Lamothe et al. [45] have found evidence that as the perspective-taking abilities of physicians increase, the rate of physician burnout decreases. I think that there are a number of plausible reasons for these relationships between perspective-taking and physician communication, patient outcomes, and lower rates of physician burnout.

More deliberate perspective-taking functions to reorient the attention of an individual to the perspective of the other. In a clinical interaction, when the physician more carefully attends to the perspective of the patient, the focus of the conversation remains on the patient’s concerns. Further, a skilled physician utilizes empathy as part of a feedback process, so the questions that the physician uses to probe for further information typically facilitate the flow of conversation. Since many health issues have a deeply subjective nature, it seems plausible that a more engaged manner of approaching patients in clinical interactions would leave patients feeling as though they have been listened to and their problems have been understood. In any human interaction, such efforts to achieve interpersonal understanding are the plausible basis for an increasing level of trust, a generalization that also seems to extend to clinical settings.

Another interesting feature of perspective-taking often mentioned in the literature is its relationship with the regulation of personal distress. One explanation for this relationship involves the other-oriented nature of perspective-taking. When the clinician is flooded with distress on a regular basis, such distress becomes emotionally burdensome and, along with other factors, can contribute to emotional burnout. Perspective-taking reorients the focus of the clinician onto the source of distress (you are feeling this way), and the deliberate nature of this reorientation ostensibly serves to inhibit the clinician’s experience of vicarious distress.

In addition, perspective-taking and engaging in the inner lives of others are effective means of encouraging empathic concern (sympathy or compassion). While I have argued against the clear value of shared emotional resonance in clinical settings, my argument does not entail that emotional engagement with patients is unimportant or should be avoided. I propose that the ideal account of empathy in clinical settings be revised to include both cognitive forms of empathetic understanding and the emotional response of empathic concern. Some measures focus on both the cognitive and affective features of empathy in medicine. For instance, the Consultation and Relational Empathy (CARE) measure captures patient perceptions of physicians’ active listening, attention to narrative, and expression of concern. Recent research has found a significant relationship between these measures and compliance with treatment [46], improved health outcomes [47], and greater patient enablement [48].

Working to encourage medical practitioners to actively care also appears to be beneficial for healthcare providers themselves. For example, both Gleichgerrcht and Decety [32] and Martin Lamothe et al. [45] have found that along with higher levels of perspective-taking, higher levels of empathic concern are associated with greater compassion satisfaction and lower rates of burnout among physicians. Many—though probably not all—of those who practice medicine are motivated by concern for the health of others. To be (and feel) engaged in the process of providing care is clearly very satisfying for many individuals. Doctors are, after all, human beings, and human beings are often motivated by the desire to do some good in this world.

There remains the issue of whether such empathetic capacities can be taught or developed, particularly in the stressful environment of modern medicine. Research in this area is relatively recent, but there is reason to believe that cognitive empathy can be improved with training and that one’s capacity for empathic concern can be cultivated. For example, Riess et al. [10] developed a training program in empathy that teaches physicians how to more accurately interpret emotional cues in facial and bodily behaviors, how to regulate their own emotional states in clinical situations using perspective-taking, and how to deliver bad news empathetically. Patients rate physicians who receive this empathy training as significantly more empathetic than those who do not receive such training. Some research also suggests that deliberately cultivating compassion toward a target in a state of suffering reduces the negative neurological response of the subject while producing positively valenced emotions, even as the subject continues to perceive the target in distress [49]. Such results could have important implications for the practice of employing empathic concern in clinical settings, particularly in terms of strategies for mitigating negative arousal in the presence of suffering. However, more research in clinical settings is required to support this claim.

Conclusion

I have argued that a widely cited account of clinical empathy—involving both cognitive understanding and emotional resonance with patients—is subject to a number of objections. Such an account is not clearly supported by the relevant empirical literature, and there is reason to believe that it may serve more as a liability than as a benefit. However, it does not follow that empathy is unimportant in clinical settings. Instead, I have argued that the ideal account of empathy in medicine remains cognitive, though empathic concern should continue to be cultivated in healthcare providers and employed in the context of physician–patient communication.