Table 1.

Steps in thematic analysis using ‘Framework’ [30]

Steps in Analysis	Procedures
Familiarisation and transcription	Preliminary analysis of recordings, field and reflective notes allowed for the iterative adjustment of the interview schedule with the inclusion of further probes. This step enabled an overall view of the data and a list of 72 preliminary codes were initially devised
Identifying a framework	The framework categories were based around the interview schedule which was underpinned by the literature, theoretical model and informed by the survey data. The thematic matrix initially represented the general content of the interviews and provided the ‘thick description’ that is needed to interpret the data relating to other contexts [33].
Indexing	The matrix provided a structure enabling discussion and collaboration among a diverse research team. A challenge was the overlapping nature of the data. While data relating to feelings and emotions can cause problems relating to categorisation, [34] the theoretical perspective of the SEM had sufficient breadth to ensure the data were not restricted, initially the same data were coded in several categories
Charting	This step involved the summary and charting of the data within the categories allowing between-case and across-case analyses allowing for refinement of categories and themes
Mapping and interpretation	As a test of authenticity, the researchers read across the data in each case to check that they got a sense of the interview looking between themes for patterns. [31] First attempts at association and explanation were drafted by the researcher and discussed with the team. Following discussion and redrafting of the analysis, a consensus in interpretation was reached