Data analysis followed an inductive thematic approach from
open to selective coding (Liamputtong and Ezzy, 2005). Initial
line-by-line coding identified initial theoretical strands, which
were organized into descriptive themes from which analytical
themes were developed (Thomas and Harden, 2008). The interview
schedule was updated iteratively throughout the research
process. All transcripts were initially coded by JS. JMS independently
coded the first four interviews, with differences compared
and discussed to engender coding reliability. These and additional
methodological details are documented in the online Supplementary
data.