Data analysis

Audio recordings of English interviews were translated using Descript⁽²²⁾ and checked for accuracy by the first author. Interviews conducted in other languages (Arabic, Nepali and Dari) were translated and transcribed simultaneously by the community researcher who conducted the interview.

Analysis was undertaken by the first author (TG) and community researchers, of which four are authors (RA, NE, FN and EP). Transcripts were analysed using the ‘best-practice framework for collaborative data analysis⁽²³⁾.’ This framework involves people with lived experiences in the coding framework co-production and includes three phases: preparation; collaborative data analysis (co-production of coding framework) and application (integration, refinement, publication and reflective learning)⁽²³⁾. The preparation phase was completed earlier in the project, where heterogeneous community researchers were hired from preliminary co-design workshops. Prior to analysis, de-identified transcripts were circulated among the community researchers for review. Community researchers met to complete the collaborative data analysis phase, and the following steps were completed for each transcript.

First, the interview setting was described by the community researcher who conducted the interview, and then the group was asked to independently read the transcripts while considering the objectives of the project and their own experiences. Next, a group discussion was held where the community researchers were encouraged to highlight interesting quotes and interpret their meaning, relate this to their own experiences and comment on the significance to communities. They also explored themes across the transcript and, for subsequent transcripts, connections to quotes and discussion points from transcripts already analysed. The first author recorded discussion points on a white board displayed in the room, noting where themes converged. During the application phase, the first author elaborated on discussion points from the previous phase by drawing on quotes from the transcripts and a vignette was created for each interviewee, all of which was then distributed to the research team for further feedback.