2.4. Data analysis

We followed the principles of reflexive thematic analysis throughout this study, involving a process of theme generation which involves immersion in the data through reading, reflecting, imagining, wondering, questioning, writing, retreating and returning to interpretations of the data (Braun & Clarke, 2021). The first author was immersed in the data by reading and rereading all the transcripts, reflecting on the interviews, and discussing the development of themes with the wider research team. A critical realist approach was adopted to understand the perspectives and experiences of participants throughout. All coding was inductive, derived from the data and not determined by pre-existing theories. The first five transcripts were provisionally coded and uploaded along with the remaining transcripts to NVivo 12 for Windows. Subsequent transcripts were coded into this framework, which was revised and expanded throughout the coding process. Similar codes were combined to explore commonalities between lower-level codes. The final stage of analysis involved conceptual linking of themes which was achieved through discussion with the wider research team, and through consideration of feedback on drafts of the evolving analysis.