Methods
In this study, we used constructivist grounded theory9to explore how physicians interacted and reacted to a variety of data elements presented in ED audit and feedback. This project is the second in a multiphase needs assessment aimed at developing a new approach to A&F. In prior work, we conducted a regional needs assessment for A&F for emergency physicians, soliciting which practice data elements were most desired regarding practice measurement.10 The results of this regional needs assessment were stratified into the subgroups of respondents. These were divided by gender, duration of practice, and type of practice (community vs academic). By asking physicians to provide their own reactions as well as their perceptions on differences we observed between subgroups, we looked to gain insight into how they may react to their own practice data.
Reflexivity. Before engaging in any analyses, our investigatory team met to discuss our inherent stances and assumptions in order to assist us in achieving reflexivity. Each member of the analysis team sought to declare their perceptions about the role and value of A&F and the predicted view of A&F by the participating subgroups. We also discussed our prior training (if we had any), and the sensitizing concepts/frameworks brought to the analysis process. The research team then made a plan for the theme extraction of interviews.
Population/Recruitment . A total of 15 participants were recruited for interview participation. Intentional sampling was performed to ensure that participants represented a variety of practice characteristics that were identified in the needs assessment (gender, community vs. academic practice, and duration of independent practice). Participants were emergency physicians practicing in the Hamilton region amongst 7 practice sites which were included in our needs analysis. All interviews were conducted between July 15th and September 2nd, 2019.
Interview Procedure. The interview guide was constructed by a single author (SM) based on the needs assessment survey data. This was refined and finalized by a second author with significant qualitative research experience (TC). Interviews were conducted by our first author (RK) after undergoing training with the more experienced members (TC, SM) of the team and receiving feedback. A training interview was conducted with this author before the inclusion of the research subjects. Impressions and insights from participants were collected during a single phase of interviews with all participants. Interview scripts are available in Appendix A.
Ethical review. T his project received an ethical exemption for the Hamilton Integrated Research Ethics Board (HiREB).
Analysis. I nterviews were audio-recorded and then transcribed without identifiers. The author who interviewed participants was the only author who was privy to the identity of the respondents. We analyzed the data using a constant comparative method to identify these, initially starting with an open coding process for all of the transcripts. Once this was complete, we proceeded with axial coding to group relevant themes and subthemes.
For the first transcript, the entire research team met to normalize the process of open coding for all. Thereafter, two authors (SM, RK) met and coded all the subsequent transcripts. With each analysis meeting, codes were discussed, added, or collapsed our previous coding. Our team opted to resolve disagreements through a discussion and consensus-building process. We eventually came to a consensus on all themes presented in our final analysis. Two authors subsequently met (SM, TC) for the axial coding of the data into larger themes.
Methods to Increase Rigor. As a form of member check, we sent the penultimate version of our newly developed model to our participants for their review and comment. We asked them to review the figure and associated explanations and provide insights about whether this model suited their worldview. Amendments and questions were used to revise the model to generate our final version.