Findings
In this study, we explored the views of 20 internists as to what constitutes a ‘productive interaction’ (the core concept of the CCM) and integrated their descriptions into a conceptual model of a productive interaction. Although we had anticipated that participants with a general focus would have different views from the specialists, the findings indicated that the views of both groups were similar. Physicians in both groups tended to define a productive interaction in terms of the same four goal orientations, and ‘clicking with the patient’ was generally viewed as a main condition. The participants mostly differed in how they talked about reaching goals for patients with medically clear versus medically unexplainable complaints. Before turning to a detailed discussion of our results, we introduce cultural model theory as a lens through which we will discuss their views. Cultural models or schemas or understandings refer to relatively stable cognitive structures ([
21], p. 54). Within a social group, people will have collective understandings based on shared experiences and a shared identity. By applying this perspective we were able to recognize
collective understandings that refer to the identity and meaning system of the participants as part of their medical profession [
21].
First, we identified the shared principle of intentionality regarding achieving outcomes and satisfaction, as well as implementing the medical process and successfully collaborating with the patient. These intentions pertain to medical conduct and identity in general, as well as to
a collective, functionalist understanding of the medical task based on professional standards [
26,
27]. In order to cure and heal, physicians need to solve medical problems effectively, have answers to a patient’s complaints, and support the patient in coping with illnesses and relieving suffering. As such, they need to follow procedures for resolving problems efficiently, give clear treatment advice, interact humanely, and satisfy the patient [
28]. In addition, external influences on the participants’ intentions were also recognizable. For example, in their goal orientations, the participants reflected their natural biomedical discourse as well as the current discourse on patient engagement, patients caring for themselves, and quality of life.
Second, we discerned, as a
collective understanding, that the participants shared how they related a goal to the medical context. They generally divided the context into medically clear and explainable (specific) complaints, and medically vague and often-unexplainable (nonspecific) complaints. Patients with specific complaints were often considered seriously ill and needing the help of the physician. Solving the problems of these patients seemed a natural task, although not always easy, due to issues such as lack of time or the patient’s inability to understand the advice. Solving the problems of patients with nonspecific complaints often appeared to be an unrealistic and exhausting ambition. This division of patients into these two groups has been recognized elsewhere [
29‐
31], with the latter chronically ill often labeled as patients with Medically Unexplained (Physical) Symptoms (MUS) [
32,
33].
The third, and the central,
collective understanding we found among the participants was that ‘clicking’ with the patient influences all the goals, as well as the initial intention to collaborate. By ‘clicking,’ the participants seemed to refer to something on the intuitive level, to transference and/or to a deeper existential dimension that establishes trust and makes it easier to empathize. In addition, several participants stressed the importance of a genuine contact as a fundamental basis for collaboration with a patient. It has been argued that people use particular statements, such as ‘we really clicked’ and ‘we had chemistry,’ to describe interpersonal interactions that go exceedingly well and to indicate good rapport [
34,
35]. Rapport reflects the quality of an interaction between people, and is evaluated in terms of emotional positivity and as a perceived unity in the interaction [
36], or as a whole that is more than the sum of the parts. As such, rapport is a nonverbal phenomenon, as well as a social construct that comprises the following components:
positive affect,
mutual focus of attention, and
interpersonal coordination. Positive affect concerns the feeling of liking each other, which is associated with trust, mutual understanding, caring, and giving support. The other two components pertain to harmony, agreement, and accord, or to synergy in a genuine interaction [
36,
37]. Finally, rapport is seen as an essential part of the therapeutic relationship and bond, and one that has a positive influence on the clinical outcome [
38], although situations can differ [
35] and changes are possible during the relationship [
39]. Similarly, rapport is probably an important aspect of the ‘the whole and shared mind’ concept that Epstein recently introduced to enhance clinical decision-making [
40]. Rapport, however, does not only exist in therapeutic relationships but is also part of non-therapeutic working alliances, friendships, and romantic relationships [
34]. We also found this: participants used the term ‘clicking’ not merely in the medical context but also to give expression to a broader
collective understanding of authentic relationships between people.
Our final remarks concern our model of a productive interaction. The four main elements of the CCM, and self-management support programs in particular, are broadly implemented in primary care in Western countries [
41]. In the CCM, self-management support is seen as a condition needed to involve and activate patients in productive interactions, with collaborative management of the illness, and ultimately better health outcomes, as results [
2,
42]. Wagner and others have argued that involving patients in their own care shows more clearly positive results on health outcome, as does improving attitudes and/or communication skills of physicians, which is often a focus of patient-centered approaches [
42‐
45].
In our model, conversely, the participants viewed collaboration as conditional for reaching medical process goals, and ultimately, health outcome and satisfaction; and 'çlicking with the patient' was viewed as a spontaneously emerging phenomenon that catalyzes collaboration. Their view of collaboration also echoes the therapeutic relationship as earlier described in the work of Roger and Balint, and is consistent with the more recently developed relationship-centered models [
46,
47]. In this view doctors do not stay neutral but create a bond needed for empathetic understanding of a patient’s problems as a person [
48].
A central issue in our model is that the participants explained clicking as on the level of affect and as a fact of life, something that either happens or does not, and that one has to accept and work with in this reality. This view contradicts current models on patient-physician interaction, which imply that establishing rapport can be learned [
16,
27,
49]. Moreover, the experienced difficulty of building a relationship with patients with medically unexplained symptoms seemed to justify lower levels of support and follow-up in such cases. However, the literature [
50,
51], including guidelines [
52], indicates that these patients would also benefit from a therapeutic relationship with physicians because, like persons with more easily treated conditions, they too have diverse and complex problems that impact quality of life [
50‐
52].
Our findings contribute to a better understanding of productive interactions from the viewpoint of physicians. Creating awareness among physicians that rapport can be learned [
53] and that the medical context; i.e., the nature of the complaint influences the process of building rapport [
54,
55] may deserve greater attention within intervention and implementation programs of the CCM. Educators could use these findings in continuous medical education in primary care as well as in medical specialty care.
Strengths and limitations
A strength of our research approach and conducting in-depth interviews in a natural way is that it enables one to derive shared intentions and collective understandings, and to construct a model [
23,
24]. These understandings are likely to reflect broader groups of internists, and perhaps also of primary care physicians, because they basically share the same profession and largely the same patient groups [
23]. The findings, however, should be validated for different contexts.
A basic issue is that what the participants understand as a professional productive interaction is not necessarily the same as what they experience or achieve in practice. Further, it is likely that gender, age, and experience, as well as individual beliefs or motives, will affect strategies and behaviors. Although we have not rigorously analyzed such differences, we did gain the impression that gender, and perhaps other aspects, do have an influence on participants’ interaction strategies or styles.
We should also caution that the ‘goal counts’ shown in Table
2 are not intended to imply the significance of the goals per se, but are offered as a closer look at the content of the text fragments. Nonetheless, we do believe that the relative frequencies with which the various goals were mentioned give some indication of their relative importance. Hence, the relative frequencies may function as a guide for further investigation.
To avoid recognition of colleagues in the interview data no one from the internal medicine department was involved in data analysis. However, inside knowledge of the internal medicine practice appeared to be essential for the interpretation and during the discussion of the findings. A risk is that an interviewer may not have been able to create a suitable interview climate or to establish sufficient rapport to get full and honest answers from the interviewees. However, several participants in our study commented positively on the interview process, so we do not see this as a major concern. Some participants mentioned that answering questions on productive interactions was a useful opportunity to talk about and reflect on their interactions with patients. Nevertheless, it is always possible that the interview setting influenced some answers, and that some participants may have found it difficult to provide full answers to unexpected questions. However, our approach is the only way to elicit the views and understandings of participants beyond a framework of predefined questions [
23], and we believe that if conversations are conducted in a respectful, open, and natural way, with enough space to elaborate on thoughts that arise, this goal can be achieved.