Dunja Dreesens

104 During the pilot interviews, it became clear that not all participants might have an in-depth knowledge of the guideline PCFC. Consequently, it was decided that if the pediatrician was not familiar in detail with the guideline that the researcher would walk the pediatrician through the guideline and the SDM-recommendations at the start of the interview. In addition to the PCFC guideline, all participants were presented with two other SDM- approaches (249). First, participants were shown two patient decision aids (PDA) for children: one on anticoagulants and another on bone infection (267, 268). Second, we showed them the PCFC guideline recommendation on pain relief (Appendix 1) that we had rewritten so that it offered several options for pain relief and could stimulate discussion with the patient and/ or parents. Approval for the study protocol was granted by the Medical Ethical Committee of the University Medical Center Groningen, reference number M14.150.681/METc 2014.057 (10 February 2014). Verbal informed consent was gained from all participants, which included consent for use of anonymized quotes in publications. No follow-up interviews took place. The interviews were audio-recorded and transcribed verbatim by LV, JW and NW, with at least one author carrying out a reliability check on each transcript correcting any transcribing errors. The participants received the transcripts for comments, correction and approval (member check). Furthermore, LV made field notes before and after the interviews, and these were discussed with DD. Data analysis Data analysis was performed by using the constant comparative method (264). Three authors (LV, NW and DD) read and reread all transcripts independently. Text fragments related to the research questions were selected and coded. The research team did not identify themes or codes in advance. The codes were developed inductively, while reading the transcripts and making notes about selected fragments in the transcripts. After the first five transcripts, the researchers (LV, DD) compared their observations and developed preliminary coding categories. After analyzing individual interviews, we added and altered codes and categories in an iterative process. LV and DD discussed disagreement about the coding until consensus was reached. When all the interviews were analyzed and coded, we collected the codes in a preliminary coding scheme. Through axial analysis, we constructed (sub-) categories. LV and DD discussed and compared all the codes and categories resulting in the final coding scheme (see Appendix 3). Data analysis was supported by NVivo-software. The research team met to discuss and complete the study report and manuscript. We did not invite participants to provide feedback on the findings. Chapter 6