2.1

Recruitment process

Participants were English speaking, secondary school/sixth form college pupils, 12–18 years of age, of mixed gender and ethnicity. Purposive sampling was used to select a range of year groups and schools/colleges in different geographical and socio-economic areas of Cardiff. To ensure meaningful insights, the sampling strategy was designed to include diverse perspectives and pupils with a range of orthodontic treatment experiences. This included pupils who had discussed treatment with a dentist/orthodontist but had not proceeded, pupils who were undergoing treatment at the time of the study and pupils who had completed treatment and were in different points of retention. Data collection and sampling were conducted in parallel and the preliminary findings informed participant selection. Recruitment continued until data saturation was achieved; this was the point at which no new themes or ideas were emerging .

2.2

Conduct of focus groups

Focus groups were conducted at participants’ schools/colleges between January and April 2016. Each focus group was conducted with pupils who belonged to the same year group to facilitate naturalistic discussions. The moderator (JP) was a male, white Caucasian clinician (postgraduate trainee in orthodontics) who had completed focus groups methods training over a six-month period with experienced researchers (IJ and PG). JP kept a reflective journal to help recognise the influence of personal biases and assumptions throughout the study. The assistant moderator (IJ) was a female, white Caucasian clinician, experienced in focus group studies.

The research team designed a questioning route informed by the existing risk communication literature and guidance on focus groups ( Appendix A ). The questioning route was used consistently across groups and involved a number of approaches to explore participants’ perceptions of orthodontic treatment risks and risk information; the first involved open questions, the second involved asking participants to write down ideas on sticky notes and the third involved a sorting exercise, using picture/word cards of orthodontic treatment risks identified by orthodontists as important risks during a Delphi exercise ( Appendix B ). Additional questions were asked in a flexible manner to explore emerging issues raised by participants.

Focus groups lasted 45–60 min and were digitally audio recorded. The assistant moderator took field notes including a record of contextual details and non-verbal communication. Participants were blinded to the professional expertise of the moderators.

Audio recordings were transcribed verbatim, by one author (JP) and imported into NVivo (version 10, QSR International, Victoria, Australia) to assist with data management. Transcripts were anonymised using pupil codes instead of names and were analysed in a timely fashion to inform future data collection.

2.3

Data analysis

All data were reviewed and coded by JP and all coding was verified by IJ. Textual data and coding were discussed as a group to enrich interpretation. Themes were derived from the data using a thematic analysis approach . Initially, each transcript was read to enable familiarity with the data and coding notes were recorded detailing thoughts, observations and early interpretations. Emergent themes were then developed by clustering related initial broad-based codes and searching for connections across them and finally, data were categorised and themes defined to capture their essence. Data contradicting emerging concepts and ideas were searched for and discussed to increase the depth of the analysis.

To encourage participant validation, pupils were given the opportunity to feedback on results; verbally during group debriefs and shortly after data analysis via feedback forms. Any feedback from the respondents was recorded and used to refine theme and theory development.

2.1

Recruitment process

Participants were English speaking, secondary school/sixth form college pupils, 12–18 years of age, of mixed gender and ethnicity. Purposive sampling was used to select a range of year groups and schools/colleges in different geographical and socio-economic areas of Cardiff. To ensure meaningful insights, the sampling strategy was designed to include diverse perspectives and pupils with a range of orthodontic treatment experiences. This included pupils who had discussed treatment with a dentist/orthodontist but had not proceeded, pupils who were undergoing treatment at the time of the study and pupils who had completed treatment and were in different points of retention. Data collection and sampling were conducted in parallel and the preliminary findings informed participant selection. Recruitment continued until data saturation was achieved; this was the point at which no new themes or ideas were emerging .

2.2

Conduct of focus groups

Focus groups were conducted at participants’ schools/colleges between January and April 2016. Each focus group was conducted with pupils who belonged to the same year group to facilitate naturalistic discussions. The moderator (JP) was a male, white Caucasian clinician (postgraduate trainee in orthodontics) who had completed focus groups methods training over a six-month period with experienced researchers (IJ and PG). JP kept a reflective journal to help recognise the influence of personal biases and assumptions throughout the study. The assistant moderator (IJ) was a female, white Caucasian clinician, experienced in focus group studies.

The research team designed a questioning route informed by the existing risk communication literature and guidance on focus groups ( Appendix A ). The questioning route was used consistently across groups and involved a number of approaches to explore participants’ perceptions of orthodontic treatment risks and risk information; the first involved open questions, the second involved asking participants to write down ideas on sticky notes and the third involved a sorting exercise, using picture/word cards of orthodontic treatment risks identified by orthodontists as important risks during a Delphi exercise ( Appendix B ). Additional questions were asked in a flexible manner to explore emerging issues raised by participants.

Focus groups lasted 45–60 min and were digitally audio recorded. The assistant moderator took field notes including a record of contextual details and non-verbal communication. Participants were blinded to the professional expertise of the moderators.

Audio recordings were transcribed verbatim, by one author (JP) and imported into NVivo (version 10, QSR International, Victoria, Australia) to assist with data management. Transcripts were anonymised using pupil codes instead of names and were analysed in a timely fashion to inform future data collection.

2.3

Data analysis

All data were reviewed and coded by JP and all coding was verified by IJ. Textual data and coding were discussed as a group to enrich interpretation. Themes were derived from the data using a thematic analysis approach . Initially, each transcript was read to enable familiarity with the data and coding notes were recorded detailing thoughts, observations and early interpretations. Emergent themes were then developed by clustering related initial broad-based codes and searching for connections across them and finally, data were categorised and themes defined to capture their essence. Data contradicting emerging concepts and ideas were searched for and discussed to increase the depth of the analysis.

To encourage participant validation, pupils were given the opportunity to feedback on results; verbally during group debriefs and shortly after data analysis via feedback forms. Any feedback from the respondents was recorded and used to refine theme and theory development.