Medicinal relationships: caring conversation


  • Lisa Mikesell

    1. Center for Health Services and Society, Department of Psychiatry and Biobehavioural Sciences, University of California Los Angeles, Los Angeles, California, USA
    Search for more papers by this author

Dr Lisa Mikesell, Center for Health Services and Society, UCLA Psychiatry and Biobehavioural Sciences, Los Angeles, California 90024, USA. Tel: 00 1 717 357 2390; E-mail:


Context  Good social relationships are crucial to well-being and to health in particular. The perception of having supportive social relationships has effects on reducing morbidity and mortality comparable with those of a good diet, regular exercise and cessation of moderate smoking. This suggests that supportive, trusting relationships with doctors could have a substantial direct biomedical effect on patients’ health.

Methods  A critical review of the patient–doctor relationship (PDR) literature is presented, along with a review of relevant interactional studies that examine doctor–patient interactions from the perspective of conversation analysis (CA). This literature shows how patients respond to doctors’ verbal and non-verbal behaviours in systematic ways that affect how they disclose and how they relate to doctors.

Results  Findings from the CA literature suggest that clinicians might consider several important interactional features to improve the PDR and perhaps also patient health outcomes: (i) the use of open-ended questions (e.g. ‘What brought you in today?’) and positive polarity items (e.g. ‘Is there something else you wanted to talk about today?’) elicits patient concerns and addresses unmet concerns more effectively than the use of closed questions and negative polarity items, respectively; (ii) eye gaze suggests availability and an attending recipient, and patients indicate that doctor attentiveness at crucial parts of their problem presentation is important, and (iii) verbal dysfluencies are one practice speakers employ to gain the attention of a non-attending recipient. Doctors may want to pay attention to patients’ dysfluencies to better understand when their attention is valued.

Conclusions  Constructing supportive relationships with patients often does not require a great investment of time, but it does require commitment to ‘being there for patients’. This review suggests that when doctors attune to language and social practices during medical consultations, the relationships they develop with patients may substantially improve patients’ health and be intrinsically rewarding for both doctors and patients.


If a doctor could significantly improve a patient’s diet, increase the amount the patient exercised, or help the patient to stop smoking, any doctor would readily devote a few minutes to doing so. Although it is often difficult to have much impact on these health factors, there is another factor that has an effect on health that is comparable with those of diet, exercise and smoking cessation; this is a factor that doctors often could substantially address but don’t always attend to. This factor concerns loneliness versus social support. There is solid evidence that the health of individuals is substantially affected by their perceptions of whether they are socially isolated and whether they have at least one relationship with a person who ‘is there for them’. If doctors could ‘be there’ in supportive relationships with patients, would such social support from the doctor have direct and substantial biomedical benefits for the patient – and the doctor, too? What kinds of relationship would be important and how would they be formed? What sorts of conversational practices would promote the formation of relationships that would be effective, supportive and therapeutic?

Whether the goal is to foster meaningful social relationships for their direct therapeutic benefits or to foster good relationships with patients simply because they are instrumental in effective biomedical intervention, it is essential to understand the nature of social relationships and the interactions that build and maintain them. My collaborator, Alan P. Fiske, UCLA Department of Anthropology, and I see two approaches to forming constructive patient–doctor relationships (PDRs): from the bottom up, and from the top down. The top-down approach, only briefly introduced here, uses the framework of relational models theory (RMT), which posits (with extensive confirmatory evidence) that social interaction is built out of just four elementary systems of coordination, called ‘relational models’ (RMs). Meaningful, trusting and mutually satisfying relationships are only possible when the people interacting are both organising their actions with reference to the same RM, implemented in the same manner. Explicit knowledge of these implicit systems can help doctors create and sustain supportive relationships.

This article focuses on the bottom-up approach. It draws primarily from the field of conversation analysis (CA)1 to examine the moment-to-moment construction of relationships in conversation, and to look at how interactional and discourse practices can be conducive to good relationships, particularly good PDRs, and how other kinds of practices and response patterns may undermine social relationships. I aim to show that there is more to conversation than simply conveying information and that conveying information is only one aspect of jointly constructing meaningful, trusting and mutually satisfying relationships that are likely to contribute to good health.

Patient–doctor relationships

Medical care providers are exhorted to treat patients as human individuals, attending to ‘the whole person’. However, there is more to humans than their individual, isolated personhood. For over a half-century, the PDR has been explicitly understood to be a critical element of modern medicine and holistic care, and there is a considerable literature dedicated to defining and improving it. Goodyear-Smith and Buetow argue that a productive PDR fosters autonomy and accountability and is characterised as an ‘adult–adult’ rather than an ‘adult–child’ relationship.2 Consideration of the patient perspective has become a priority,3 in part because it is repeatedly linked to patient outcomes. Researchers have underscored the importance of gaining patient trust and eliciting the patient’s story and psychosocial features that are relevant to the patient’s health.4,5 The effects of culture and language on the PDR have been explored in attempts to improve doctor–patient communication and trust6 and there is increasing discussion about how best to incorporate relationship and empathy training into medical school curricula.7–9 Patient–provider communication and the PDR are increasingly valued, so much so that Turner et al.10 suggest that what patients and providers bring to the clinical visit may be more important than the actual treatment. Given this current climate, Roter has proposed that the ‘molecular and chemistry oriented sciences’ that defined the medical paradigm of the 20th century be replaced by a relationship-centred medical paradigm.11

Despite the growing interest in and awareness of the potential benefits of the PDR, much of the current research merely states that it is important to establish a PDR characterised by empathy and good communication, without clearly stating why it matters or how it can be achieved.12 Some studies highlight the importance of understanding patient preferences and expectations because of their impact on patients’ relations with their doctors,13 but refrain from explicitly stating why the relationship is important. Some seem to take an ‘it can’t hurt’ approach, assuming that the benefits are apparent, whereas others disregard the effects as ‘non-specific’ or ‘placebo’ effects.12,14 Studies that overtly address the benefits of the PDR tend to focus on the effects the PDR has on patient and doctor satisfaction15,16 or on positive patient or doctor behaviours. Studies highlighting the effects of the PDR on patient behaviours emphasise that the relationship is important because it influences the patient to follow the prescribed biomedical treatment plan (e.g. medication adherence) or because it affects the likelihood of follow-up or termination, and the likelihood of litigation.17–21 Similarly, studies highlighting effects on doctors’ behaviours focus on how a positive PDR enables doctors to follow the biomedical model, for instance, by helping them to better ‘understand and interpret reported symptoms, examination findings, and diagnostic test results’.22 In sum, this health services research tends to highlight the indirect effects of the PDR on health and recovery. These are important outcomes of a productive PDR and this focus on the PDR is a positive development.

Good social relationships support health

However, current discussions largely disregard the potential direct effects of the PDR on patients’ health and well-being. Although the PDR has been considered a form of medicinal therapy and an integral part of the treatment plan since the time of Hippocrates,12 this insight seems to have faded from current visions of doctoring, both in medical education and in practice. This is despite the fact that for 35 years there has been strong evidence that social support – the perception of having good relationships – has powerful direct effects not only on psychological well-being, but also on many kinds of morbidity and mortality.23–29 The largest meta-analysis of the effects of stronger social relationships examined 148 studies of 308 849 participants over a mean of 7.5 years and found a weighted mortality odds ratio of 1.50.30 The results are robust across a wide variety of types of measures of social relationships. The data from the 61 studies that used the most valid multifaceted measures of relationship quality showed an odds ratio of 1.91, and even these results represent minimum values, limited by the proportion of the variance in relationship quality that was actually measured.30 This means that the feeling that one has at least one person who ‘is there for one’ reduces mortality by at least 33% and probably cuts mortality rates in half. The person who is there for the patient could be the doctor.

Helping patients establish or repair their social relationships with others is rarely part of the treatment plan. Perhaps with the exception of psychotherapists, few would know how to do so if they wished to try. More immediately, doctors infrequently consider the possibility that their own relationships with patients might have a direct therapeutic effect. The therapeutic encounter itself has great potential for somatic healing.31,32 Furthermore, many patients have psychological as well as somatic illnesses, which a good relationship can often alleviate. In psychotherapy, the provider’s relationship with the client, the therapeutic alliance, or the cohesion of the therapy group have substantial effects on recovery, often exceeding the benefits of any specific therapy technique.33–36 Yet ‘relatively little is known about how to create and sustain the relationship and about why the relationship works’37 or how to go about strengthening it.

Medicinal relationships: perspectives of the PDR

The patient’s perspective

The patient movement, the adoption of a consumerist model of medicine, and patients’ increased access to medical information and willingness to challenge medical authority38 have stimulated growing interest in patient preferences and patient satisfaction. This research concludes that patient satisfaction is highly influenced by attributes of the PDR, more so than by characteristics of other domains such as the medical care facility.39,40 Yet, it is not clear just what satisfies patients. Information exchange and decision making have been frequently explored and, in general, experts argue for increased patient autonomy and control and more shared decision making.41 However, it is not evident that all patients desire this. Some researchers conclude that older patients prefer doctors to be in control of health care and treatment decisions.42–44 Others claim that women prefer to be given more information by doctors than men,45,46 and that patients of higher socio-economic status and education seek more information and involvement in the decision-making process.47 Patient preferences for involvement in decision making have also been found to vary according to the patient’s health status and coping strategies; patients with more severe conditions and less active coping strategies have been found to prefer to take a more passive role.48 Accordingly, several experts warn that doctors should not make assumptions regarding patient preferences, but should ask patients what information they want and how they want to make decisions.13,44

Although patient preferences about attributes of the PDR may vary, what seems to be clear is that patients want to feel cared for and listened to. Vick and Scott44 found that patients of all socio-economic levels wanted more than anything else to be listened to during clinic consultations. Importantly, this preference was not necessarily about the patient’s role in the decision-making process, but about whether the patient felt that he or she had been heard. Additionally, patients’ willingness to return to a doctor is influenced by ‘feelings of being cared about by the physician, the degree to which they took time with them, explained and listened to them, and were accessible when needed’.49 Although not all patients may desire the same kind of relationship with their doctors, the relationship itself is important to them and may be even more important than relationships with spiritual advisors and co-workers.50 Patients value their interactions with doctors, often preferring to make in-person visits rather than telephone or e-mail consultations,51 and patients seem to highly value doctors’ respectful attention to what they have to say.

The doctor’s perspective

Although there are fewer studies investigating clinicians’ perceptions of the PDR, some research indicates that the PDR is as important for clinicians as it is for patients. In fact, Kearley et al.52 found that, in the context of non-minor illnesses, a higher proportion of general practitioners (GPs) than patients in their sample reported valuing personal relationships between themselves and their patients. Clinicians indicate that knowing the patient is important for medical treatment and diagnosis. For instance, GPs noted that knowing their patients had significant therapeutic benefits, especially for complex and psychological problems.52 Even beyond diagnosis and treatment, there is evidence indicating that clinicians value humanistic PDRs in which they feel connected to patients.53–55 Clinicians value patients’ perspectives: they report being highly motivated to implement shared decision making56 and to foster cooperative relationships with patients.57,58 Medical students also value the PDR: they report desires for practical training and mentorship on how to foster good PDRs, and feelings of discouragement when they face difficulties in maintaining these values in real-world settings.59,60

There are many system constraints that may create obstacles to the development of longitudinal and personally deep relationships with patients.61 In any circumstance, improving the PDR can start with clinicians paying closer attention to their interactions with patients in the clinic office. However, accurately perceiving what actually happens during the clinic visit can be challenging.62,63 Waitzkin,45 for instance, found that doctors underestimated patients’ desire for information and the amount of time they dedicated to information giving. In 336 encounters, doctors dedicated an average of 1 minute 20 seconds to providing information or explaining problems to patients.45 Although this amounted to merely 9% of the entire clinic visit, doctors believed that they were dedicating much more time to giving information and estimated that they spent nearly 9 minutes doing so.45 Thus, these doctors perceived that they were devoting 575% more time to providing information than they actually were. I am not suggesting that clinicians should spend more time providing information in all cases, but improving clinicians’ awareness of what they are accomplishing interactionally within the visit is likely to benefit both clinicians and patients.

Acknowledgements:  none.

Funding:  none.

Conflicts of interest:  none.

Ethical approval:  not applicable.

A bottom-up approach: an interactional view of the PDR

The medical visit consists of more than just what the doctor and patient each separately perceive and desire. Considering patients and doctors independently ignores the importance of the interactional negotiation that emerges within the medical visit. Features of the PDR and communication seem to be equally, if not more, important for patients’ overall perception of their clinic experience and more highly correlated with patient satisfaction than doctor characteristics such as gender or years of experience.64–66 Kaptchuk also notes that patient characteristics and personality traits do not consistently correlate with the effects of placebos. He suggests that how the doctor and patient come together during the clinical interaction is more important than the attitude or characteristics of either party, stating that ‘reciprocal expectations need to be negotiated and joined in the patient–physician duet’.14 In short, a social interaction consists of patterns of coordination based on a relational framework that is not reducible to the attributes of the individual participants67,68,70 (A P Fiske, L Mikesell; ‘Medicinal relationships: an application of relation models theory’; unpublished manuscript, 2013).

It is the interactional moments that cumulatively build, maintain, define and transform the type of relationship that emerges and evolves between the doctor and patient. Relationships are neither static nor predefined. Research on doctor–patient interactions illuminates what makes the interaction productive and satisfying. In a sample of 35 clinic visits, Barry et al. found four types of interaction styles or communication patterns, defined by how patients and doctors oriented to two factors.69 They refer to these using Mishler’s70 distinction69 between the voice of the lifeworld, which represents ‘the patient’s contextually grounded experiences of events and problems in her life’, and the voice of medicine, which ‘reflects a technical interest and expresses a “scientific attitude”’.69 The four types of interaction, respectively, reflect the following contexts: when doctors and patients both talk exclusively in the voice of medicine; when the patient’s voice of the lifeworld is ‘blocked’ by the doctor; when the voice of the lifeworld is ‘ignored’ by the doctor, and when both the doctor and the patient talk primarily in the voice of the lifeworld.69 When doctors and patients shared the same orientation, clinical visits had better outcomes as measured by 10 factors including the other-rated presence or absence of major misunderstandings, self-reported adherence, and both patient and doctor satisfaction as expressed in follow-up interviews.69 Although several studies have argued that patients are more satisfied with ‘patient-centred’ interactions, what seems to matter most is that doctor and patient agree upon and utilise a similar communication or relationship style.71 However, the communication style may shift. Parties may implicitly but collaboratively agree to move in and out of ‘voices’ throughout the visit. More than any static qualities of the individual parties involved, what seems to be key to successful relationship building is for the doctor and patient to collaboratively and dynamically define the framework of their interactions.

Although more research prioritising an interactional approach is needed, current studies using CA suggest that discovering the patient’s orientation to the PDR does not need to be arduous. For instance, doctors can provide a space for patients to provide their orientation to the PDR early in the visit, which may require very little time and effort. It may even consist of a single communicative action. In a study of 302 primary care visits, Heritage and Robinson72,73 found that when doctors opened the visit with an open-ended question (e.g. ‘What can I do for you today?’), patients responded with more symptoms than they did when asked a confirmatory question ‘request[ing] confirmation of a generalised gloss of patients’ medical problems’ (e.g. ‘I see you’re here today for X’). Patients who were asked open-ended questions took only 16 additional seconds to present their symptoms (27.1 seconds versus 11.3 seconds) than patients who were asked closed questions. The authors argue that open-ended questions provide patients with the authority to define their health problems, whereas confirmatory questions assume prior knowledge and encourage passivity from patients.72

A follow-up study73 found that, after controlling for age, sex, race, education, problem type, visit length and health status, patients were also significantly more satisfied with their visits when doctors began visits with open-ended rather than closed questions. Taking more time for problem presentation, because it occurs early in the clinic visit, may allow patients to demonstrate how they are orienting to the PDR and which ‘voice’ they intend to adopt. Of course, it is not enough to simply provide this interactional space. Clinicians must listen to not only the reported symptoms, but the ways in which patients approach the medical encounter. That is, the clinician must orient him or herself early in the visit to the type of relationship to which the patient is orienting and be attentive to shifts in relational framing (A P Fiske, L Mikesell; ‘Medicinal relationships: an application of relation models theory’; unpublished manuscript, 2013).

Equally important is how the doctor follows up on additional patient concerns immediately after the patient’s first expressed concern. Patients frequently have multiple concerns when coming to clinic, but clinicians often address only the problem(s) introduced at the beginning of the visit.74,75 In 204 clinic visits, Heritage et al.76 found that patients were significantly more likely to express additional concerns other than the chief concern when they were asked about additional concerns using the positive polarity item ‘some’ (‘Is there something else you want to address today?’) than when asked a question using the negative polarity item ‘any’ (Is there anything else you want to address today?). More specifically, the authors concluded that the ‘some’ condition eliminated 78% of patients’ unmet concerns.76 Moreover, the percentage of concerns left unmet in the ‘any’ condition was indistinguishable from that in the non-intervention condition (37%). In the non-intervention sample, fewer than 5% of patients were asked about additional concerns, so that doctors failed to learn about unmet concerns that included chest pain, neck and back pain, angina, weight loss and medication questions. Neither the ‘any’ nor the ‘some’ condition had a significant effect on the length of the visit, which averaged 11.4 minutes.76 In short, asking negative polarity questions such as ‘Do you have any...’ may be no better than not asking a follow-up question at all.

Although conversational practices are vital for establishing and maintaining relationships, gestural and bodily orientations are also crucial components in demonstrating one’s orientation to others. Eye gaze, for example, has been shown to be an important way for recipients to show that they are attentive to and listening to the speaker.77–79 Goodwin showed that when speakers lacked an attentive hearer, they systematically produced restarts, pauses and hesitation markers in their speech.77 These verbal hiccups were produced until the speaker secured an attending recipient via gaze. Once a recipient gazed toward the speaker, indicating attentiveness, the speaker continued his or her utterance without verbal dysfluencies.77 Although such dysfluencies are often thought to result from problems in cognitive processing, Goodwin has shown that they are systematically used as interactional resources that can be used by speakers to secure an attentive recipient.77,78

Heath found that participants use verbal dysfluencies in medical consultations to secure a recipient’s gaze just as they do in ordinary interactions.80 Often, however, clinicians need to attend to medical records and other documents80 (which Robinson refers to as ‘the patient inscribed’79), as well as to the actual patient (‘the patient embodied’79). Ruusuvuori81 suggests that doctors’ focus on the patient inscribed promotes a doctor-centred orientation, whereas doctors’ focus on the patient embodied facilitates a patient-centred orientation. She also argues that the timing of doctors’ shifts in eye gaze or focus during patients’ narratives is relevant to indicating the doctor’s availability.81

In her sample of consultations, Ruusuvuori81 found that patients were especially dysfluent when doctors showed disengagement during a patient’s presentation of his or her problem by turning the gaze, head or body away from the patient, evidently interrupting the patient’s narrative. These dysfluencies elicited an attentive response from the doctor only 59% of the time; 41% of the time, the doctor continued to focus on the medical records, possibly disappointing or offending the patient.

Disengagement was treated by patients as interruptive in particular environments (Ruusuvuori discusses four such environments81) that highlight the importance of attending to spatial and temporal features of patients’ narratives. I focus here on the two environments highlighting temporal features. Doctor disengagement was problematic when: (i) the patient’s problem narrative reached a crucial moment in the telling, which Ruusuvuori describes as ‘the core of the problem’, and (ii) the patient was coming up to the completion of her problem narrative or ‘the very final point of her description’. Patients treated doctors’ disengagement within turns and during the narrative climax as interruptive and employed interactional resources, especially verbal dysfluencies, to regain the doctor’s attention and to get the doctor to look at them.

Although clinicians must divide their attention between patients and medical charts, these studies show that it is important to orient to patients while they are presenting their concerns. Orienting to medical charts at the beginning and ends of problem presentations or at narrative boundaries respects ordinary interactional expectations and therefore appears less problematic for patients. It may also help for the doctor to say something at this point that shows respect for these ordinary interactional expectations and respects the patient’s patience in holding his or her thought while the doctor looks at the chart, such as: ‘Okay, that’s important, let me check the chart for a moment before we discuss this more.’

In sum, there tend to be two main aims in most PDR literature. The first is to uncover what attributes of the PDR patients (or patients of a certain demographic) desire. The second is to argue that doctors should treat each patient as a unique individual with unique needs. Both aims are intended to be patient-centred and reflect conceptualisations of the patient as a selective consumer embedded within an individualist culture. However, the first body of literature aims to aggregate individuals to find similarities in the distinct types of PDR patients want, whereas the second argues that attention to individual personhood is essential to good medicine. These two bodies of literature send clinicians (and medical educators) confusing, often conflicting messages. Although sensitivity to patterns of gender, race and culture within a patient population is important, perhaps the most effective approach is to adapt to the unfolding interaction, to assess patient expectations and relationship orientations early in the clinic visit, to be sensitive to shifts and boundaries, and to make sure there are no unmet concerns at the end.

Acting on relationships

Clinicians may respond to this discussion with frustration. They are under a lot of stress, and endure challenging workloads and pressure to turn over higher numbers of patients. They may also feel they are losing clinical autonomy as a result of the corporatisation of medicine82 and may feel powerless, inadequate or frustrated as they treat more chronically ill patients.83 Thus, doctors may experience great challenges in meeting their moral responsibilities. Additionally, although the period of ‘doctor bashing’ seems to have reached its height several decades ago,82,84 clinicians, doctors especially, may still feel they are experiencing increasingly harsh criticism. It is not despite these constraints, but because of them that medical training and practice should aim to foster supportive PDRs.

Many of the suggestions for improving PDRs might seem to require investments of time that doctors simply do not have in a typical 15-minute consultation.85 In fact, implementing most of these practices requires little additional time. Genuine expressions of caring may not necessarily demand more talk, but may require different forms of expression. A prosodic contour, eye gaze, facial expression, gesture or posture may be enough to convey the clinician’s genuine concern and intent to hear the patient. Initially, this effort may feel stressful or even awkward simply because the doctor already has very challenging responsibilities to meet. Clinicians may similarly feel that by following these suggestions, they are merely acting as if they cared. However, research and anecdotes indicate that most clinicians do care and want to listen, albeit that their verbal and non-verbal cues may sometimes fail to convey this caring to patients. Furthermore, with practice, being mindful of the relationship with patients and using sensitive conversational practices may become second nature. As this happens, the PDR will become more rewarding and more meaningful to the doctor, as well as to the patient.

Some clinicians may also find these suggestions unrealistic in view of changing patient expectations; patients may now come to clinic armed with information assembled from various Internet sources86 and direct-to-consumer advertising.87 Some of these patients may even view the clinic visit as a service transaction, treating the clinician as a pill dispenser, and the consult as simply a means to the procurement of medicines. Doctors often find these explicit requests for treatment ‘inappropriate’ and ‘damaging’,87 and perhaps unsatisfying and offensive as well. Taking a more relationship-oriented approach may help the doctor engage in a manner that is more meaningful to both doctor and patient in these encounters. For instance, imagine a patient who has diagnosed her problem and identified a treatment that may be effective. The doctor can still engage this patient by acknowledging and respecting her initial relationship orientation, but offering more genuine social support. Good social relationships typically are multifaceted: although the framework for obtaining the prescription may be market-like, this does not preclude simultaneous demonstrations of support that enrich the PDR.

In sum, findings from the CA literature provide concrete interactional communication practices that require little additional time but may help clinicians provide a space within the clinic visit in which patients can make known their orientation to the PDR and the type of relationship they are seeking in this context68,70 (A P Fiske, L Mikesell; ‘Medicinal relationships: an application of relation models theory’; unpublished manuscript, 2013). This orientation can be made public if the clinician allows space for it to be voiced and listens to and engages with what the patient says. Open questions and body language (e.g. eye gaze) have been shown to be crucial, especially in the first moments of the clinic interaction. These practices demonstrate that the clinician is paying attention to the patient and is not attempting to multi-task or rush into a new phase of the clinic visit (e.g. the diagnostic phase) before the patient has been able to demonstrate his or her orientation to the PDR and present all of his or her concerns. Because not all patients will have the same relationship expectations and because expectations may change over the course of an interaction, clinicians must be receptive to patients’ varying ways of approaching the PDR, just as they might be with other types of relationship. The outcomes are likely to be helpful for fostering a productive and effective PDR that has the potential to provide direct and indirect health benefits, as well as to improve clinician and patient satisfaction.