Multidisciplinary Approach for Managing Complex Pain and Addiction in Primary Care

A Qualitative Study

Randi G. Sokol, MD, MPH, MMedEd; Rachyl Pines, PhD; Aaronson Chew, PhD


Ann Fam Med. 2021;19(3):224-231. 

In This Article


We obtained our institutional review board's approval for the study. We used the Consolidated Criteria for Reporting Qualitative Research (COREQ) checklist[19] to ensure rigor in our methodology. All 3 authors (R.G.S., a female family medicine and addiction physician; R.P., a female visiting researcher; and A.C., a male clinical psychologist) were involved in data collection and analysis.

Data Collection

We used 2 sources of data for this study. First, a research assistant performed a medical record review and compiled a list of the consultation questions submitted to PASS by PCPs over the past 4 years. Second, we interviewed referring physicians who had consulted the PASS team. A member of the research team (R.G.S.), who knows the referring PCPs, invited them to voluntarily participate via e-mail. Another author (R.P.), who was not a member of the PASS team or previously known by the PCPs, then conducted interviews to maintain unbiased data collection.

Fourteen PCPs agreed to be interviewed: 11 attending physicians, 2 family medicine residents, and 1 physician assistant. Interviews were conducted by telephone or in person at the clinician's primary care clinic and ranged in length from 30 to 45 minutes. All interviews were audio recorded and then transcribed by a professional transcription service. Identifying names were deleted during transcription. The transcripts were not returned to the participants for their feedback before analysis. The interviewer (R.P.) also made field notes during the interview process to guide subsequent interview content and to conclude the interviews when data saturation was reached. Each interview was conducted in a semistructured format (using a semistructured interview guide available on request) and began with asking the PCPs about their role at their organization and about their PASS referral generally. This question was followed by questions about their experience with the referral such as, "What did you find most valuable about the recommendations?" and "How could the PASS referral process be improved?" and "How have you used this referral since (in providing care for this or other patients)?"

Data Analysis

We coded all data using Dedoose version 8.3.17 software (SocioCultural Research Consultants) and used qualitative thematic analysis[20] to analyze both the referral questions and interviews.

To assess PCPs' expressed needs, we analyzed 66 referral questions. Two of the 3 researchers (R.G.S. and R.P.) reviewed the first 14 referral questions to generate a list of agreed-on codes to serve as an initial coding framework for the remaining referral questions. One researcher (R.P.) then completed analysis of all referral questions to generate a total of 14 codes. The other 2 researchers (R.G.S. and A.C.) then completed analysis of all referral questions, and the research team met to reconcile differences. During this process, the 3 researchers reconciled discrepancies through consensus and added in 2 more codes, resulting in 16 distinct, clearly defined expressed needs from the PCPs' referral questions.

To assess actual needs, we performed thematic analysis of the 14 semistructured interviews. Each researcher (R.G.S., R.P., and A.C.) independently coded the same 3 initial interviews, deriving their own themes for each research question, resulting in a total of 76 data-driven codes for actual needs. On review of the codes from each researcher, many overlapped in ways that suggested moving up a level of abstraction, resulting in 11 themes. Each of the 3 researchers then used the agreed-on set of themes to code the remaining 11 interviews. After coding all interviews, the researchers met to rectify discrepancies through consensus (a complete coding tree is available on request).