48 Chapter 3 Data collection Only patients with known email addresses were eligible for inclusion. After the consultation, patients received an email invitation to an online questionnaire. Questionnaires were handled anonymously through a secured survey platform. No reminder emails were sent. No demographic patient data were recorded. Patient questionnaires were collected over 20 weeks, from April to July 2019. Residents received their personal patient feedback at three time points: after week 4, 10, and 20. At 10 weeks, there was a joint meeting in which questionnaire results were briefly discussed. At the end of the study period, all residents were interviewed individually by one experienced interviewer (LS). Interviews were transcribed verbatim. Analysis Response rates were calculated for all questionnaires sent out. Due to anonymity restrictions, it was impossible to perform a non-response analysis. We reported both mean scores and percentages of top scores, as CollaboRATE top scores are hypothesized to discriminate better between individual doctors’ performances than mean scores. [18] A CollaboRATE top score is reached when a patient scores 4 – ‘every effort was made’ on all three CollaboRATE items. Differences between individual scores were tested for significance by performing the Kruskal Wallis non-parametric test. Group scores were reported for both the first and the second 10week study period. Differences between group scores were tested for significance by paired t-testing (level of significance = 0.05). For the two open-ended questions, percentages are reported, reflecting how many respondents provided feedback. All statistical tests were performed in SPSS statistical software version 25. [21] Interview transcripts were analyzed with the software program Atlas.ti. [22]A thematic content analysis approach was used to analyze data. [23] Answers were open-coded by two researchers (LS and JA), focusing on residents’ views on the new tool, SDM, and receiving patient feedback. Codes from both researchers were compared and commented on. Finally, a code list was built. As limited previous research showed that patient feedback could lead to reflection on behavior and thus skills improvement, [12, 13] we used the reflexivity framework to further thematize the code list. The five elements within the framework are explained in more detail below.
RkJQdWJsaXNoZXIy MTk4NDMw