Understanding Healing Relationships in Primary Care

healing relationships


PURPOSE Clinicians often have an intuitive understanding of how their relationships with patients foster healing. Yet we know little empirically about the experience of healing and how it occurs between clinicians and patients. Our purpose was to create a model that identifies how healing relationships are developed and maintained.

METHODS Primary care clinicians were purposefully selected as exemplar healers. Patients were selected by these clinicians as having experienced healing relationships. In-depth interviews, designed to elicit stories of healing relationships, were conducted with patients and clinicians separately. A multidisciplinary team analyzed the interviews using an iterative process, leading to the development of case studies for each clinician-patient dyad. A comparative analysis across dyads was conducted to identify common components of healing relationships

RESULTS Three key processes emerged as fostering healing relationships: (1) valuing/creating a nonjudgmental emotional bond; (2) appreciating power/consciously managing clinician power in ways that would most benefit the patient; and (3) abiding/displaying a commitment to caring for patients over time. Three relational outcomes result from these processes: trust, hope, and a sense of being known. Clinician competencies that facilitate these processes are self-confidence, emotional self-management, mindfulness, and knowledge.

CONCLUSIONS Healing relationships have an underlying structure and lead to important patient-centered outcomes. This conceptual model of clinician-patient healing relationships may be generalizable to other kinds of healing relationships.

Keywords: Physician-patient relations, communication, primary health care, healing


Wild azaleas bloom in my garden every spring, reminding me of the botanist who gave them to me and our journey through his suffering and eventual death from prostate cancer. During this relationship and others like it I (J.G.S.) came to understand the powerful healing connections forged between doctor and patient. I realized the quality of the relationships I created with patients was as important as the pills I dispensed, and that relationships with patients sustained me through the difficult and sometimes frustrating tasks of practicing family medicine. Although many physicians have an intuitive understanding of the importance of healing relationships, there are few systematic studies in the medical literature that empirically examine what healing relationships might look like and how they are built by clinician and patient. 1

Research in other disciplines shows the importance of healing relationships. Anthropologists have explored healing as a cross-cultural phenomenon and distinguished categories related to healing. 2 In psychotherapy, research finds that the nature of the therapist-client relationship accounts for approximately 45% of the effectiveness of therapy. 3 Nurses have carried out research on healing for many years. Although there has been considerable theoretical development in this literature, most empirical work has focused on interviews with nurses. 4 Patient interview studies have focused on particular aspects of the nurse-patient relationship, especially caring. 5

Most of the existing theoretical models of healing relationships are based on interviews with health care professionals or patients, but not both. 4 – 6 Using a grounded theory approach, my fellow authors and I interviewed clinicians and their patients to depict how healing relationships are created, structured, and maintained.


This study was designed to explore healing in the context of ongoing clinician-patient relationships in which healing was recognized by the clinician. We realize that there may be many other situations in which healing occurs that are not connected to clinician-patient relationships, 7 and that healing may occur in the context of clinician-patient relationships without the clinician’s knowledge. We focused on healing in clinician-patient relationships because of its potential to change and improve clinician behavior and facilitate the development of the “continuous healing relationships” recommended by the Institute of Medicine’s report on quality of care. 8

Sampling Strategy

To enhance the probability of observing the phenomenon under study, it was necessary to choose physicians who were most likely to create healing relationships with patients. Physicians believed to be exemplars in developing and maintaining healing relationships were purposefully selected based on an assessment of publications, reputation, and awards. In addition, we recognized that even exemplar clinicians would not have healing relationships with all their patients, and the phenomenon we wanted to explore required that clinicians be aware healing had emerged in the context of their relationship with patients. For these reasons, each clinician was asked to choose adult patients who they perceived had experienced healing. Healing was purposely left undefined to allow the definition to emerge from the participants’ experiences. Sampling proceeded iteratively, with analysis of each interview informing and refining the interview guide and the interview process for subsequent interviews. Interviews continued until

the analysis team determined that saturation had been reached.


To minimize analysis complexities associated with differing world views and experiences of multiple qualitative interviewers, the analysis team decided that the first author should conduct all interviews. The potential bias introduced by having a single physician interviewer was managed as follows. Before conducting the study interviews, the first author interviewed 5 of his former patients to increase his own and the analysis team’s awareness of his experience as a clinician and healer; to allow him to discuss, analyze, and gain a greater awareness of his personal beliefs about healing; and then to manage and control these preconceptions, as best as possible, during interviews. 9 The physician’s ideas about healing and the nature of healing relationships were recorded in a journal format, and self-reflective field notes were dictated after each interview. The analysis team reviewed all of this material. In addition, the analysis team critiqued each interview, particularly during the early interviews, pointing out to the interviewer questions and approaches that revealed preconceptions, as well as physician-centric biases.

www.annfammed.org/cgi/content/full/6/4/315/DC1 ) that contained several grand tour questions 10 designed to elicit healing stories from physicians and patients. Additional questions examined physicians’ role as healers in the context of the ongoing relationship with patients and the effect of relationships on healing processes. Interviews lasted 1 to 2 hours.

Interviews were recorded and transcribed. Physician D5 invited the spouses of 3 of his patients to be present during the interviews. Although asking a spouse to be present was not part of the original study design, the interviewer chose to view it as an opportunity to observe how such an arrangement might influence accounts of healing experiences. The analysis team, however, found no substantive differences in the content of interviews when spouses were present compared with interviews with patients alone. Transcripts were checked for accuracy. Digital voice files and transcripts were imported into qualitative analysis software, Atlas ti. 11 The Robert Wood Johnson Medical School Institutional Review Board approved the study.


The analysis team consisted of a family physician with 21 years’ experience in private practice (J.G.S.), a medical anthropologist with years of experience in primary care research (B.F.C.), a nurse who had extensive experience in home and hospice care (B.D.B.), and a specialist in communication science with expertise in qualitative methods (D.C.).

Interviewing and analysis proceeded iteratively. 12 As transcripts became available, the analysis team listened to and discussed interviews as a group. After the group had listened to a number of interviews, common issues or themes began to emerge. The group discussed these themes, making our understanding of them richer and deciding how insights would guide subsequent data collection. Data collection and preliminary analysis continued in this fashion until saturation was reached. Saturation occurred after interviewing 5 physicians and 23 patients. 12

The first author used an open coding process 13 to tag data excerpts the group identified as interesting. The analysis team read and reread these excerpts in the context of the larger interview to construct case studies describing the nature of the relationship of the clinician-patient dyad. Insights were discussed, refined, and developed into a coherent case study of each physician and all of his/her patients. Case studies were analyzed across physicians to identify common themes and to develop a preliminary model of healing relationships.

Although patients’ comments about the study clinicians were uniformly positive, patients made numerous negative comments about other clinicians. These negative comments served as a contrast to highlight what participants took to be components of healing. Because of space limitations we report only the components of healing, but we examined these contrastive comments in depth in our analysis and in our construction of the conceptual model.

Two authors (W.L.M. and K.C.S.) who were unfamiliar with the details of the team’s preliminary analysis helped refine the model during 2, 2-day retreats. Finally, the analysis team performed member checking 9 by soliciting feedback on the model from 2 physician interviewees.


Sample Demographics

Demographic characteristics of participating physicians and their practices (Table 1 ) and their patients (Table 2

) are shown below. Although it was not a requirement for participation in this study, most physicians selected patients who had either current or chronic illnesses, including human immunodeficiency virus (HIV) infection, ischemic heart disease, chronic pain syndrome, recurrent pulmonary emboli, diabetes, valvular heart disease, history of sexual abuse, history of drug abuse, and breast cancer.

Similar articles: