Dependence and caring in clinical communication: the relevance of attachment and other theories.

OBJECTIVE: Clinical relationships are usually asymmetric, being defined by patients' dependence and practitioners' care. Our aims are to: (i) identify literature that can contribute to theory for researching and teaching clinical communication from this perspective; (ii) highlight where theoretical development is needed; and (iii) test the utility of the emerging theory by identifying whether it leads to implications for educational practice.
METHODS: Selective and critical review of research concerned with dependence and caring in clinical and non-clinical relationships.
RESULTS: Attachment theory helps to understand patients' need to seek safety in relationships with expert and authoritative practitioners but is of limited help in understanding practitioners' caring. Different theories that formulate practitioners' care as altruistic, rewarded by personal connection or as a contract indicate the potential importance of practitioners' emotions, values and sense of role in understanding their clinical communication.
CONCLUSION: Extending the theoretical grounding of clinical communication can accommodate patients' dependence and practitioners' caring without return to medical paternalism. PRACTICE IMPLICATIONS: A broader theoretical base will help educators to address the inherent subjectivity of clinical relationships, and researchers to distinguish scientific questions about how patients and clinicians are from normative questions about how they should be.

Introduction

As a moral enterprise, clinical communication teaching and research has placed patients’ rights at the centre of clinical practice against a historical background of medical paternalism. As a science it has reminded biomedicine of the need for compassion in clinical care. However, conflation of morality and science has allowed assumptions to develop in the field that constrain the theoretical framework within which communication researchers and educators work [1]. Our aim is to identify areas of theory and research that can inform clinical communication and to highlight areas where further conceptual development is needed. Then we test the potential practical utility of this exercise by examining its implications for educators.

The patient: autonomous or vulnerable

The imperative to involve patients as partners in care shapes health policy and dominates teaching and research in clinical communication [2]. As a moral statement, promoting patients as equal partners enhances respect for their autonomy and human rights. As a scientific concept, typically operationalised as the need for information and choice [3], it proves problematic. Appreciation of choice (such as between treatment options) is rare when patients describe their care [4-8]. Although some value information and choices, many do not [9-18], are too impaired by illness to exercise informed choice or feel overwhelmed by it [19-21]. People who are seriously ill want less involvement in care decisions than people who are healthy think they would want if they became ill [16,17,22,23]. Many patients want practitioners to exercise the authority associated with a traditional biomedical mode l [24], or have better outcomes with such care [25,26]. For practitioners, allowing patients to make potentially damaging decisions can feel incompatible with caring for them [27]. By prioritising autonomy above other considerations, ideas of patient partnership and empowerment therefore risk diverting care from the ethic and expectation of beneficence [28].

Of course, many patients do want information and choice, but not necessarily as ends in themselves [15,29,30]. Faced with serious illness, information can build hope or trust [11,30-33] or help patients accept what doctors know is in their best interests [4,34]. These findings are unsurprising if we acknowledge that patients mostly seek health care because they have – or fear – illness and want practitioners to help them. Reflecting this inherent asymmetry in the patient–practitioner relationship, clinical communication theory could, as a first approximation, model the patient as vulnerable and dependent and the practitioner as expert and caring [1,35].

Towards a theory of dependence

Attachment theory

Although we shall see that it provides a very limited theory of caring, attachment theory is an obvious starting point for understanding dependence because it concerns people's propensity to seek relationships with individuals who can help them feel safe when they are vulnerable [36]. ‘Attachments’, or ‘attachment relationships’, are emotional bonds that lead an individual to seek proximity to a safe or powerful person (the ‘attachment figure’) when threatened. Attachments are thought to be based on ‘mental models’ (i.e. enduring belief systems) of oneself and others. These models are the product of how individuals have learned to regard themselves and care-givers during childhood, and become generalised to other dependent relationships [37]. Each mental model varies from positive (trust of others; confidence in oneself) to negative (distrust of others; anxiety about oneself). These mental models shape individuals’ expectations about relationships. They underlie ‘attachment styles’, which are consistent and enduring patterns in how an individual relates to people in dependent relationships. A ‘secure’ attachment style indicates that an individual has a positive mental model of self and others, while an individual with an ‘insecure’ style has negative mental models (Fig. 1).

Fig. 1

Different theories of attachment describe different specific attachment styles and label them differently, but all distinguish ‘secure’ from various types of ‘insecure’ attachment. Here, following Ciechanowski et al. [57,98], secure attachment is distinguished from three styles of insecure attachment, each style being defined by the intersection of mental models of self and other. Each style has different implications for behaviour in dependent relationships, and examples for clinical relationships are shown.

Attachment theory originated in Bowlby's [38] studies of young children's relationships with their parents, particularly mothers. He argued that adults also rely on attachment at times of vulnerability and noted that, beyond childhood, the focus of attachment shifts from parents to romantic partners or close friends [39]. Empirically, there is evidence that attachment styles persist from childhood into adulthood [40,41], although this is controversial [42,43]. Continuity can be explained not just by the enduring and generalized nature of the mental models but because these lead to behaviours in interactions with others which perpetuate them. For example, the guarded or over-dependent behaviour of someone who has experienced poor parental care as a child will elicit behaviour in other people that perpetuates that individual's negative mental models [44].

Applying attachment ideas to adults exposes important differences between adult and child relationships [45,46]. Classically, the currency of attachment is seeking and giving comfort and proximity; material care is secondary. In other words, it is a theory about feeling safe rather than being safe [47]. Whereas, from a child's perspective, being comforted will be the main indicator of safety, adults will seek more evidence of the protective power of a potential attachment relationship and, in particular, of the other party's ability to ameliorate the threat. With age also comes increased ability to use symbolic representations of attachment figures to substitute for physical proximity; for example having a meeting arranged with an attachment figure could provide a sense of safety [48]. Adults also seek security in appropriate distance and separation as well as intimacy [46].

Attachment theory is not inimical to respect for individuals’ autonomy. Instead, attachment can promote a person's autonomous self-regulation of distress [48] by strengthening self-concept and enabling the experience of affiliation and caring from others [49]. Understanding the dialectic between dependence and autonomy will, however, require a broader scientific view of autonomy than the current emphasis on information and choice in health policy and related literature [50-52].

Attachment in clinical relationships

Even in the absence of serious illness, symptoms provoke fears of mortality or threaten assumptions about life [53]. Recent theorists have emphasised that thoughts and images can activate attachment needs [48] so we should expect that illness is an arena in which intense attachment needs arise. Bowlby envisaged these being directed to family, longstanding romantic partners and close friends [39]. However, when illness is the threat, these are rarely the people who can offer safety because they do not have the expertise to manage illness. Several writers have begun to apply attachment theory to clinical relationships because, when threatened by illness, it is practitioners whom patients regard as having the expertise and power to provide safety [9,31,53-60].

Attachment style is related to health-care seeking, particularly in the absence of serious disease [57,61], and to symptom report and readiness to attribute symptoms to psychological factors [57,61-63]. Attachment also helps to understand patients’ behaviour when they consult. Patients with attachment styles characterised by negative mental models of self or others are more likely than others to be seen as ‘challenging’, or as non-adhering to advice [54,64], perhaps because they find it hard to trust that practitioners have their interests at heart. In addition, the framework of attachment styles can be used to categorise patients’ different presentations in clinical settings and associated communication problems (see Fig. 1) [56,64,65]. Attachment theory points to patients’ previous experiences as potential influences on clinical relationships. For example childhood abuse, which disrupts the formation of positive mental models of oneself and others [66], impairs patients’ abilities to form supportive clinical relationships during primary treatment for breast cancer [67].

Safety and hope in clinical relationships

Attachment theory places patients’ seeking of safety at the centre of clinical relationships, and emphasises their need to trust practitioners’ expertise to provide that safety. However, pairing ‘communication’ with ‘safety’, ‘trust’, ‘expertise’ or feeling ‘cared for’ as terms for searching clinical research literature is rewarded by many fewer hits than pairing with ‘information-giving’ or ‘choice’. Current literature does, of course, recognise the emotional quality of clinical relationships, and practitioners are already encouraged to meet patients’ emotional as well as clinical needs [68]. Indeed, distinguishing emotional from biomedical agendas is firmly embedded in communication research and teaching [69], and practitioners are criticised for prioritising biomedical ones. Attachment theory challenges this division because it suggests that patients can be helped emotionally – to feel secure and comforted – by practitioners’ technical or biomedical care. It thereby helps to explain why patients value doctors’ concern with emotional factors and ‘small-talk’ less than communication experts might expect [70,71], why psychosocial talk with the wrong patient or at the wrong time can be experienced as intrusive [72-74], and why being offered choice about treatment can damage trust in the practitioner [9]. Indeed, a doctor whom a patient thinks is prioritising psychosocial concerns over biomedical ones would seem dangerous rather than reassuring.

The pursuit of safety, alone, is insufficient to understand patients’ commitment to clinical relationships because practitioners cannot always guarantee the safety that patients seek. Further theoretical development will be needed in this area, but pointers are available. In some situations safety is probably redefined, for example as the prospect of painless death rather than avoiding death [75]. Perhaps alliance or companionship with practitioners offers a kind of safety in its own right. It is well known that patients and practitioners often collaborate in building hope [76], and it has been suggested that this is the way that doctors and patients avoid the inherent tension between patients’ need for safety and doctors’ frequent inability to provide this [77]. If hope is the basis for many clinical relationships, we will need to understand more about practitioners’ part in building or supporting it. Portmann [77] suggested that hope relies heavily on medical technology, which contrasts with the argument that technology is inimical to feeling cared for [78].

Subjectivity in clinical relationships

Classically, attachment relationships are thought to be formed through repeated interactions with care-givers, and to be called upon in response to each new threat by seeking proximity to the care-giver. However, clinical care-giving usually means brief encounters with unfamiliar practitioners who cannot offer ‘on-demand’ proximity. Theoretical development will be needed to understand how some patients can feel protected in a caring relationship with someone that they barely know, while others cannot. Here, again, there are clues. One suggestion is that patients subjectively construct their image of the practitioner to meet their own safety and dependency needs and in line with their own mental models of self and other [9,31].

This view predicts that, in general, the more threatening the illness, the more positive will be the image of the practitioner that a patient constructs. This view would also explain why patients seem often to think of practitioners in terms of global characteristics, such as their personality, rather than specific behaviours [20]. Some practitioner behaviours, such as appearing inexpert, could prevent patients trusting in them and might preclude patients feeling cared for [9,79-81]. Beyond this minimum, patients will substantially construct the practitioner they need, albeit having to work harder with some practitioners than others. Therefore the link between practitioners’ behaviour and patients’ sense of relationship might be weak [82]. Indeed, patients can describe positive experiences that are not always evident in the interaction [83]. Similarly ‘improvements’ in practitioners’ communication, such as greater attention to psychosocial problems, do not necessarily translate into a better patient experience [84,85].

Individuality and authenticity in clinical relationships

In attachment theory, an attachment figure (i.e. the care-giver whose proximity is sought at times of vulnerability) is unique and irreplaceable. Attachment theory therefore points to patients’ need to feel that they are in relationships with practitioners who genuinely care for them. This can explain why a sense of authentic caring can matter to patients more than polished performance of generic communication skills and why patients often value idiosyncratic aspects of practitioners’ behaviour [9,86-88]. Attachment theory usually also envisages that attachment figures see those seeking their comfort as unique and irreplaceable. However, while there is little evidence to show whether practitioners view patients in this way, it seems unlikely given that they typically see more patients than they know by name.

Therefore, given the importance of practitioners’ authenticity for patients [89], conceptual and empirical work will be needed to understand what authenticity can mean in a clinical context. A bond between two people who each see the other as uniquely important will usually be unattainable – and undesirable – in clinical care, but patients might recognise other kinds of authentic commitment. For example, they might understand that a practitioner's job means seeing many patients in a clinic, but experience authenticity in seeing the practitioner's character as promising commitment to the best possible care for every patient.

Towards a theory of care

Attachment and care-giving

Bowlby described a ‘care-giving’ behavioural system centred on parental roles and complementary to the attachment system [38], so attachment offers a theory of caring as well as dependence. Accordingly, a history of good parenting and socialisation, and a secure attachment style (i.e. positive mental models of oneself and others) enhance adults’ disposition for caring and their sensitivity to others’ needs [90-96].

Attachment in practitioners’ caring

Researchers have begun to apply these ideas to health practitioners. Specialty choices reflect attachment style, securely attached medical students (i.e. with positive mental models of themselves and others) being particularly likely to choose primary care, perhaps because of its emphasis on longer-term relationships [97,98]. Recent studies of doctors’ interactions with patients presenting medically unexplained symptoms suggest that attachment style affects doctors’ behaviour once they are working in primary care too. Doctors with a more positive mental model of themselves proved to be most ready to contradict these patients’ ideas about their illness or treatment [99]. Similarly, in mental health care, more secure case managers were more likely to respond to their patients’ underlying needs, whereas others were more easily diverted by patients’ own attachment needs, e.g. providing reassurance to insecure patients who need, instead, to learn to be independent and not to rely on reassurance [100].

Finding out the relationship between practitioners’ behaviour and attachment style has already helped to challenge assumptions about why they sometimes communicate in ways that communication specialists deprecate. GPs frequently offer somatic investigations and treatments when patients with unexplained symptoms disclose psychological problems [101]. Viewing this as psychologically defensive led to the prediction that the GPs most likely to do this would have negative mental models of themselves and others. Finding that they had more negative models of themselves but more positive models of others suggested a different view: that GPs provide somatic intervention because they value patients (positive model of others) but devalue what they can offer psychologically (negative model of self) [102].

It would be over-simplistic to propose that secure attachment in practitioners always favours effective care. Indeed, research in mental health care suggested that more effective relationships result where practitioners have styles complementary to patients’, and so can resist and change those styles [103,104]. For example, a patient who is hesitant to trust a practitioner would be ill-served by one who tends to withdraw from signs of others’ dependence, whereas the same practitioner would be well-equipped to avoid being drawn into unproductive relationships with a demanding patient who needs to learn independence.

Although attachment theory assumes some continuity of attachment style from childhood to adulthood, attachment security can be manipulated. For example, being prompted experimentally to recall one's own experiences of care can enhance compassion or care-giving [94,95,105,106]. Attachment theory therefore directly confronts researchers and educators with the importance of individual differences amongst practitioners and patients, and empirical and ethical analyses are needed to address educational implications that arise from acknowledging these. For example, can or should educators seek to change practitioners’ attachment style, or should they seek to make practitioners aware of their own styles and to learn to use them, or to compensate for them?

While practitioners’ attachment styles can explain some individual differences in caring, they do not explain caring [49,94]. That is, attachment theory does not identify the motivation for caring. Indeed, its neglect of motivation for caring reflects its grounding in the maternal relationship and associated assumptions of innate caring, and it has therefore been criticised for trivialising the conflicts and choices that care-givers experience [107,108]. One solution to the motivational lacuna in attachment theory has been to propose an autonomous emotion and motivation of ‘caring’ [107]. However, this is circular because the only evidence of the motivation is the phenomenon that the motivation is imputed to explain [108]. Similarly, Staub [96] proposed that caring can be a way to satisfy a basic human need for transcendence of self but, again, this circularity offers little help to educators, who need to understand motivation to care in a way that explains differences between practitioners and identifies processes that can be targets of intervention to enhance caring. We have reached the limits of what attachment theory can offer in understanding why practitioners care, and why some might care more than others, and we need to consider other theories that have been proposed to explain why practitioners care.

Understanding practitioners’ motivation to care: looking beyond attachment theory

The reward of relationships

In clinical literature, an influential view has been that caring behaviour is rewarded by the sense of personal fulfilment in ‘connecting’ with patients [109,110]. The view has been extended even to claim that the therapeutic property of the relationship arises from its meeting the needs of both patient and practitioner for connection and meaning in their lives [109].

As a scientific account of a possible source of some clinicians’ motivation, this view gains plausibility from evidence that empathy and sympathy in non-clinical contexts promote caring and other pro-social behaviour [111]. Moreover, relationships do, indeed, matter to practitioners, who can feel hurt when patients end them peremptorily, for example by switching from one practitioner to another [112]. However, as a normative account of how practitioners should feel, to link clinical caring to positive feelings imposes a demanding requirement to feel positive about all patients at all times and implies not being able to care for a disliked or emotionally distant patient. Moreover, norms that govern caring in the personal domain diverge from ethical principles, such as equity (treating all people as having the same value) or accountability (the need to justify one's care to others such as professional bodies or employers), that must apply professionally [78,113]. Also, if practitioners prioritise relationship above other considerations, this can prevent them from addressing patients’ needs, for example where it reinforces inappropriate dependence [114]. Portmann [77] criticised the ‘pretence’ that the clinical relationship is an affectionate one, suggesting that both patients and practitioners work to construct this illusion as a defence against the uncertainties of treatment and the threats of morbidity and mortality that pervade the relationship. A motivation for practitioners to continue seeking close relationships with the large number of patients that they have to see is also hard to reconcile with broader evidence about people's motivation for connection in relationships, which is normally satisfied by a few close relationships [115].

Altruism

In striking contrast with the view that caring is based on affection or relationships from which practitioners directly gain emotionally is the influential view that practitioners are altruistic [113,116]. Altruism is a tricky concept to define, and behaviour that appears altruistic can often be explained according to selfish interests, such as the desire to enhance reputation [117]. Moreover, the face validity of the current culturally prevalent view of practitioners as altruistic may be weakening as generational changes shift practitioners’ concern to their own well-being [118]. Bishop and Rees [119] suggest that claims to altruism are generally best understood as social construction to mask the reality of financial reward for care. Moreover, Glannon and Ross [120] argued that clinical practice is no more altruistic than other types of work that involve obligations to others; altruism arises only where a practitioner goes beyond the role prescribed by their fiduciary relationship with patients. Altruism is therefore not a plausible theory to understand motivation for care. However, it does focus attention on the extent to which practitioners value patients and their well-being, which is likely to be important in a complete account of professional caring.

Caring as contract

Glannon and Ross [120] and Portmann [77] offered a more straightforward formulation of motivation for clinical care, and Meagher provided a similar analysis in relation to social care [121]. That is, practitioners are mostly paid to care as one side of a contract with their clients. On this reasoning, although practitioners may draw on the norms of the personal sphere, personal affection for their patients is not part of their contract. Indeed, as we noted above, practitioners necessarily need to draw on contrasting societal norms including fairness, equity and freedom from reciprocal obligations associated with receiving care.

This formulation leads to other problems, however. Whereas the concept of contract implies agreed expectations, entered into by equal parties [122], patient–practitioner relationships are inherently asymmetric. Therefore, what is defined by the clinical contract is likely often to be ambiguous or contested, reflecting the intersection of professional norms, contract and employment law and, where public provision is concerned, relationships between the individual and the state. The nature of the contract will depend on situational factors, and practitioners’ understanding of their professional role, too. For example, whether practitioners’ ‘contracts’ include responsibility for patients’ social and emotional well-being in surgery is unclear, although this would be generally assumed in hospices [78,123].

The authenticity of practitioners’ care

Extending beyond attachment theory to considering these other diverse attempts to explain practitioners’ caring still does not provide a complete account of caring. The main value of these different theoretical approaches is to identify the issues that communication researchers and educators will need to consider and that go well beyond the prevailing concern with practitioners’ ‘communication skills’ or ‘relationship skills’. Attachment theory points to the role of practitioners’ emotional security in understanding their motivation to care; and ideas of caring as ‘emotional connection’ point to the potential importance of practitioners’ own emotional experiences with patients. Discussions of altruism point to the need to consider practitioners’ attitudes, or values, and accounts of professional caring as contract point to the need to understand practitioners’ views of their own contracted role.

These different perspectives suggest an approach to practitioners’ authenticity to mirror the analysis that we proposed of patients’ need for authenticity. Different sources of practitioner motivation would each imply a different type of authenticity in caring—from feeling affection for patients, through valuing them as fellow humans, to determination to fulfil the professional role as expertly as possible. Investigating how these compare with what patients seek might help to turn the conundrum of authenticity in clinical communication into an area of research and theorising with practical benefits [124].

Practice implications for enhancing and assessing communication

We previously argued that a broader theoretical base is necessary to understand clinical relationships from the perspective of patients’ dependence on practitioners’ expertise [1]. Here, we have shown that relevant theory is available, although theoretical lacunae and contradictions point to priorities for clinical communication researchers. These will only be worth pursuing if practical implications are likely, so we conclude by demonstrating that significant implications already arise.

A functional approach to teaching and assessing communication

Whereas current consensus emphasises types of communication, such as breaking bad news, providing information or attending to psychological cues, we suggest focussing on its functions [125], such as hope or feeling cared for [31,126]. This will complicate teaching because a single type of communication, such as providing information, can serve different, conflicting functions. Educators can, though, draw on a developing literature about how communication addresses the needs that matter to patients and it will be important to be guided by the clues that this contains rather than by what has mattered to researchers and educators. Recognising patients’ need to trust practitioners’ authority and expertise as a starting point may bring ‘new’ communication strategies into teaching, such as persuasion or argumentation [127]. Recognising that practitioners sometimes need to counter inappropriate dependence, being able to neglect rather than facilitate psychological cues might sometimes be the important skill [128,129]. Because tension is inherent in many practitioner–patient relationships where patients seek safety that practitioners cannot provide, educators need to help practitioners go beyond merely diffusing or avoiding conflict, to conduct it effectively to meet patients’ needs when these diverge from their requests [130].

There will need to be corresponding changes in assessment, too. In many situations, successful communication of safety, trust, caring and hope, or identifying and managing conflict, will be clinically valid subjects for assessment. However, this will be practically more challenging than assessment of current priorities such as providing information or responding to psychological cues.

Teaching and assessing subjectivity

The areas of theory reviewed here, particularly attachment theory, are incompatible with the currently prevalent view of relationships as objective things that, once ‘built’, will be experienced similarly by each party. Both patients and practitioners can construct their own sense of relationship, reflecting their own needs and only partly visible in communication that passes between them. Instead of recycling the metaphor that communication skills build relationships, educators need to help practitioners to recognise and work with the subjective sense of relationship that patients might have, reflecting their vulnerability and dependency needs.

Acknowledging the subjectivity of relationships opens new directions for assessment, too. The need for reliability and fairness puts a premium on ‘objective’ behaviours, but assessing these can only be clinically valid for the few behaviours or situations that leave little room for subjectivity in how they are experienced. In situations where patients’ subjective experience of communication is likely to depend on their or their practitioners’ individual characteristics, assessment will need to make more use of ‘global’ – i.e. subjective – ratings [131]. A second implication is that quality of communication does not follow a linear scale. Certain communication behaviours will preclude patients from constructing any sense of relationship [9] but, in the absence of these, the quality of relationship that patients sense might be only loosely related to practitioners’ communication. This is recognised, but not yet explicitly, in current assessment approaches in many schools which emphasise, not the reward of excellence, but the detection of unacceptable communication.

Teaching and assessing individuality and authenticity

Reflecting patients’ concern with practitioners’ authenticity and individuality, educators need to find ways to nurture this in the context of generic training programmes, and to acknowledge that different practitioners might have different motivations for caring and that different patients might look for different sources of practitioners’ authenticity. As a start, attachment theory provides a framework that can give individual differences a central place so that communication teaching is tailored according to clinicians’ character as well as individual patients’ needs [132].

Implications for assessment are profound because the importance of authenticity challenges the validity of simulation in assessment [133]. Perhaps, at best, simulated-patient encounters might only detect students who lack the ability, or are too overwhelmed by anxiety to appear to be concerned for a patient [134]; at worst, by rewarding pretence, they might inadvertently teach inauthenticity [89]. If educators are not to give up on measuring the quality of practitioners’ clinical relationships, they may need to focus on real patients’ experiences of their students, with all the complexities that will entail.

Expanding the conceptual framework: beyond practitioners’ skills

Communication teaching is dominated by the concept of communication as skill. While this has helped to establish communication in clinical curricula alongside traditional skills, over-emphasis on skill detracts from patient benefit [1,125,131]. To apply the ideas outlined here, we need to expand the conceptual framework for understanding practitioners’ communication. Attachment theory introduces emotional mechanisms but is a cognitive theory, too, in which mental models are central. Just as the strength of practitioners’ motivation for technical tasks drives clinical practice and outcome [135], educators need to be concerned with their motivation to care. Although motivation remains a confusing field, it is clear that it will mean addressing practitioners’ values, attitudes and sense of professional role.

In practice, many communication curricula already include teaching at the levels of cognition, emotion and attitudes, with topics such as ‘understanding the patient's perspective’, ‘dealing with emotion’ or ‘cultural competence’ [136-138]. Whereas allegiance to the concept of communication skill obscures these levels, future educators could be explicit about them so that educational interventions are tailored to each, and sequenced in a way that is theoretically grounded in understanding how they interact. Concern with practitioners’ emotions and values, in particular, promises ethical and educational challenges in choosing the goals and methods of teaching [137,139]. Epstein suggested that compassion is promoted by ‘mindful practice’; that is, critical, but non-judgmental attention to one's own mental processes when with patients [131,140]. Others [131,141,142] have suggested that empathy is achieved by ‘deep acting’, whereby cognitive processes (e.g. imagining what the patient is experiencing) change emotional reactions so that behaviour changes naturally, or that curricula could explicitly teach and assess moral reasoning [143]. Others go further in advocating immersion in patients’ narratives [89] or even in patient roles [144]. Educators and researchers need to examine whether such approaches achieve the authenticity that patients seek or whether addressing values directly is necessary for any kind of empathic work [145].

Our analysis again complicates assessment. Conceptualised as behavioural skill, communication quality can be objectively measured by sampling micro-skills in simulated-patient encounters. On the present analysis, educators will need to find ways to assess, not only behavioural demonstration of skills, but also whether the student knows when to use them (cognitive level), is sufficiently at ease with patients to use them (emotional level), and cares enough about patients to use them in everyday practice (values).

Conclusion

By extending the theoretical grounding of clinical communication, researchers and educators can accommodate patients’ vulnerability and practitioners’ caring without returning to the moral and scientific constraints of medical paternalism. The theory and research that we have reviewed presents new dilemmas. Some will be resolved by further research and theoretical development about how patients and clinicians are, but others may lead to normative questions about how they should be. There are already potential implications for how teaching and assessment can become more theoretically grounded and better represent patients’ interests.

Even though the patient–practitioner relationship shares the asymmetry in expertise and vulnerability that defines classic attachment relationships in childhood, attachment theory is incomplete as an account of clinical relationships. Therefore, it is important not simply to replace over-simplistic but morally appealing metaphors, like ‘patients as partners’ or ‘building relationships’, by new ones, like ‘attachment’ or ‘authenticity’, which will also be over-simplifications or will prove to have restricted applicability [49]. The inherent contextual specificity and time-varying nature of clinical relationships [146] will demand a diversity of theories [20]. Arguably, the field of clinical communication has reached a stage of maturity in which theoretical pluralism offers more than does adherence to theoretical hegemonies. Nevertheless, it will be important to ensure that we do not lose sight of the vulnerability and dependence that are intrinsic to being ill and needing health care.

Conflict of interest

The authors are aware of no conflict of interest arising from this work.