Although the clinician sample was largely based on convenience, an attempt was made to include health professionals with different backgrounds; thus 2 nurses, 2 fellows, one physician’s assistant (PA), and one medical student were interviewed in addition to the family practice residents (n=9). The goal of the patient sample was to recruit and interview a patient mix representative of the ethnicities and sex of the patients in each clinic. Most patient interviews were done in the clinic before or after appointments, while clinician interviews were generally done in an office. Seventeen clinician and 22 patient interviews were taped and transcribed for review.
Participant Observation
The research team also observed 8 clinical precepting sessions involving family physician attendings and residents, students, or PAs who were seeing patients with diabetes. During these sessions the observers took field notes or tape-recorded the sessions; the encounters were written up or transcribed in their entirety and input into Nud.ist (Qualitative Solutions and Research, Melbourne, Australia), a qualitative software package, for subsequent analysis. Following the precepting sessions, the observers talked informally with the attendings about the encounters and other aspects of diabetes care.
Taped Interactions Between Clinicians and Patients
Three encounters between clinicians and patients were audiotaped or videotaped with the informed consent of both parties and transcribed for analysis.
Data Analysis
As transcripts became available they were analyzed for content by members of the study group, who coded the data independently and met to discuss their interpretations in monthly data-analysis sessions. Comments were coded using P (precepting session), I (interview), M (medical consultation), L (lecture), and C (informal conversation). The diversity of professional training in the group (a family physician, a psychologist/family therapist, 3 cultural anthropologists-2 of whom were also family therapists-and a PhD-level nurse brought multiple perspectives to a single data source, a qualitative technique known as researcher (or theoretical) triangulation.21 Most of the data collected through these 3 techniques was also input into Nud.ist and indexed according to format (interview, precepting session, and so forth), speaker characteristics (ethnicity, sex, professional status), clinical site, and other attributes.* The results presented here are grouped by themes derived from common content analysis techniques, such as counting, and clustering.22
Results
The most significant themes revolved around control and frustration. Within these broad analytic categories, important subthemes that distinguish clinician narratives from those provided by patients include: the complexity of the disease; the importance of internal processes, such as nephropathy; long-term outcomes or end points; and profound pessimism. Also, while individual autonomy or control appears to be emphasized in the medical consultation, the impediments to control (social, cultural, and biological) and the importance of medication are noted in private discourses.
Control
A dominant, recurring theme, and a major source of confusion, was the use of the term “control.” This word was frequently used by clinicians and patients, with a wide range of meanings.
Metabolic Control. Both clinicians and patients talked about diabetes (the disease) being “in control” or “out of control,” meaning that the blood sugar or glycohemoglobin was in or out of the appropriate range: “His hemoglobin A1c is 10.8, his urine was 500 mg per deciliter of glucose…diabetes is not very well controlled” (resident, P).
Compliance. Clinicians also talked about the patients themselves being “out of control,” referring to patients’ inability or unwillingness to control their behaviors or diet. A resident uses the word in this sense when she points out that her patient has “no control…[she] had a donut and coffee for breakfast” (P).
Strategic or Situational Control. Finally, patients at times spoke of being in control in ways that clearly demonstrated their need to be in charge of their daily activities.
I was just talking to a doctor, and I said when I check my blood sugar before I eat it’s like 80 or maybe below…. He told me not to do that, and I, I want to do that and keep it low, because when I feel bad I want to go get something to eat and make that bad feeling go away…. For some reason the low blood sugar feeling is a lot easier to deal with than the high blood sugar; I think because I have control. (I)
Frustration
Physicians repeatedly expressed frustration in conjunction with both diabetes and caring for patients with diabetes. An attending physician said, “I think especially NIDDM is frustrating to take care of” (I), while a nurse said, “They’re [people with diabetes], I think, the hardest people to work with.” (I) A PA who works extensively with patients who have human immunodeficiency virus (HIV) remarked, “Diabetes is probably one of the most frustrating things I have to deal with…HIV patients are much easier to deal with.” (I)