Skip to main content

Table 2 Framework analysis of focus group data

From: Characteristics of outdoor falls among older people: a qualitative study

Task How performed Coding structure
First iteration of coding
Each transcript was coded for the experience of outdoor falls recounted by participants, with each outdoor fall initially broadly coded by context and impact. Each fall event was sorted by participant, the focus group they attended, and had quotations pasted into cells to justify each code. Each cell was populated with descriptions that captured the substantive content of the transcript excerpt. Context of the outdoor fall was initially split into four broad codes:
1) Characteristics of the environment (such as weather, lighting, incline, etc.)
2) Social context (such as alone, talking with another, etc.)
3) Familiarity (in familiar/unfamiliar area)
4) Attribution (perceived cause of the fall)
Impact of the fall was initially split into three broad codes:
1) Physical injury
2) Emotional reaction
3) Anxiety about falling again
Indexing. The above broad framework was systematically applied to all the transcripts.  
Consensus on number of outdoor falls recounted and refinement of inclusion criteria. Independent coding by two researchers followed by discussion. The above coding was used to reach consensus on which falls occurred outdoors and were therefore to be included in the remainder of the analysis.  
Second iteration of coding
The initial coding was then subcoded to capture the multiple variations of contexts and impacts of outdoor falls. Each cell was refined to not only capture the substantive content but the dimensions of the transcript excerpts according to the sub-codes.  
Indexing. New subcodes were generated as analysis progressed from the first to the final transcript, along with refinement of existing subcodes (e.g. splitting the subcode “season” into the four separate seasons of the year, and collapsing two similar subcodes into one subcode).  
Summaries were produced from the analysis that captured the total number of falls that occurred, and the patterns that emerged from the multiple sub-codes (known as charting). The frequency of subcodes that emerged across participants was noted.  
Third iteration of coding
While producing the summaries noted above, the sub-codes were reviewed and refined.   
Reliability checking and refinement
Independent reliability checking, with particular attention to a code that was deemed to warrant further refinement: physical injury from falls (under the broad code of impact of falls). The overall coding framework was checked by another researcher (CB). Each item coded under the broad code of impact of falls was then checked by the researcher (CB) and a falls practitioner (physiotherapist and lead for a local hospital-based falls team). In refining the codes, we arrived at four broad codes:
1) Features of the physical environment
2) Features of the social environment (including familiarity of the location of the fall)
3) Attributions
4) Impact (physical injury with emotional response including anxiety about falling)
Refinement of coding of physical injury from outdoor falls. We employed a coding framework used previously [13, 31]. In using this framework we arrived at four codes for physical injury:
1) No injury
2) Untreated injury (minor injury that did not receive medical treatment)
3) Treated injury (injury that received medical treatment, such as presenting to a family physician, hospital, or accident and emergency ward)
4) Fractures (major injury or multiple fractures requiring hospital treatment)
We made comparisons between the codes that emerged as most prevalent amongst the sample (known as mapping and interpretation). The context of outdoor falls was compared with the impact of outdoor falls in terms of (a) physical injury and (b) anxiety about falling again.