106 Chapter 5 Semi-structured interviews were conducted from March through June 2017 with EPs of the Radboudumc. We approached all EPs, who participated in the program. Interviews were held in Dutch at the ED staff’s office by a trained interviewer (GH) and lasted between 20-30 minutes. During the interviews, the following topics were discussed: a) experiences with the program (e.g., educational approach, content, organization, learning climate), b) impact of the program on geriatric knowledge, attitude towards older adults and treatment of older adults, and c) suggestions for improvement. Interviews were digitally recorded and transcribed verbatim. The excerpts of the interviews were translated to English. Anonymity of the interviewees was guaranteed to maximize candid discussion. Interviews were held concurrently with the analysis of transcripts to determine data saturation. Data analysis Quantitative data was analyzed using SPSS Statistics version 23.0. The Shapiro-Wilk test was applied to assess the normality for the distribution of the data. Independent two-sample t-test and Mann-Whitney U test were used to compare baseline characteristics. We used basic descriptive statistics to present questionnaire and multiple choice test scores (i.e., mean and standard deviation) and medical record assessment scores (i.e., frequency and percentage). Missing data in the multiple choice test were scored as ‘wrong answer’. Paired sample t-test, chi-squared test, Fischer’s exact test, Mann-Whitney U test and Wilcoxon Signed Rank test were used as appropriate to compare before and after data. A significance level of p<0.05 was used to determine statistical differences. Interview transcripts were analyzed in line with Braun and Clarke’s outline of the thematic analysis process [21]. Detailed reading and re-reading of transcripts led to the generation of initial codes. Codes were generated in a systematic fashion across all transcripts. Subsequently, codes were combined under overarching themes that were internally coherent, consistent and unique from each other. All transcripts were initially coded by GH. Credibility checks were carried out by ÖS for quality control purposes, who read and coded all transcripts at different stages of the analysis process. We used computer-assisted qualitative data analysis software (Atlas.ti version 8) to manage and code the transcripts.
RkJQdWJsaXNoZXIy MjY0ODMw