Table 4.
Process of data analysis.
| Data analysis | Process | Outcome |
|---|---|---|
| Stage 1: focus groups | All focus group discussions were recorded and transcribed. A preliminary coding list was developed inductively with further iterations of the master code list developed by the research team. Master code list applied to transcripts (ASD). Four transcripts were independently analysed by TLS and MC. Codes were grouped into categories around the study objectives and compared. Preliminary categories were reviewed by the project steering group and research team with further analysis reaching data saturation. |
Six transcripts. Preliminary coding list. Master code list of 22 codes. Five categories: experience, views, patient characteristics, barriers/facilitators and concerns. |
| Stage 2: telephone interviews | All recorded telephone interviews were transcribed. A research fellow (AS) commenced an inductive content analysis of interview transcripts under the guidance of two senior researchers (TLS and ASD). Initial coding across the entire data set was completed. A second researcher (ASD) coded three randomly chosen interview transcripts. Both coders worked systematically through the data to achieve data saturation. Data management and coding were facilitated by a QDA software NVivo v10. |
21 transcripts indexed and Initial coding list generated. Final master code list generated. Three categories developed: Working towards a decision with five subcategories: non-staff action, staff cognitive work and interaction. Enacting transfer, with three subcategories: leadership, decision-making and knowledge acquisition. Post-transfer, with two subcategories ICU/community interface and reflections. |
| Stage 3: integration | The interview coding list was compared to the focus group coding list, and a combined master code list was then agreed and applied deductively to all interview transcripts. | Combined master code list. Three global themes were generated: Should we do it? Can we do it? How do we do it? |
| Activities to enhance rigour | Development of an audit trail of all meetings held with associated aims and outcomes. Review of individual researchers’ coding notes and notes of Iterative discussion about the coding decisions, discrepancies and agreed definitions of codes. Discussion of generated codes, categories and final themes with clinical partners and research advisory group. | |
QDA: qualitative data analysis; ICU: intensive care unit.