42 Chapter 2 one introductory meeting to meet their facilitators, be informed about MCD and learn about morality, dialogues, and moral dilemmas. The subsequent MCD sessions were scheduled for 120 to 180 minutes each. At the start of each session confidentiality was agreed upon. Each team had a fixed set of MCD facilitators; a total of 18 were involved. The researchers were part of local steering committees at the participating locations. These Committees consisted of a member and/or managers from the teams that participated, the management team and a coordinator from the Educational Institute. The aim was to monitor and evaluate the research and implementation process of the MCD sessions and make adjustments when required. Data collection To evaluate the content of the MCD sessions, two types of evaluation forms were used: one for MCD participants and one for MCD facilitators. For this paper, we analyzed responses to the question: ‘Please describe the discussed case and write down the moral dilemma’. The MCD participants’ form only asked the case-presenter to answer the question, and to also ‘briefly describe the situation and the decision (A or B) you had to make.’ The following inclusion criteria of MCD sessions were applied: a) one or more moral dilemmas discussed in-depth, and b) answers provided sufficient information to code a main theme. The exclusion criteria of MCD sessions were based on the availability of sufficient and reliable information. Sessions with missing evaluation forms or noninformative data in the forms were excluded. After the first steps of our analyses, we excluded MCD sessions based on content criteria (see Figure 1), e.g., sessions based solely on the exchange of emotions or opinions, without in-depth and systematic deliberation of at least one moral dilemma. Data analyses Taking a qualitative approach with a thematic content analysis, we created ‘a map’ of the content across the dataset ‘summarizing the variation and regularities in the data’ (Green & Thorogood, 2013). In our ‘conventional content analysis’ we attempted to stay close to the respondents’ words and avoided ‘preconceived categories’ (Hsieh & Shannon, 2005). In the first step of analyzing all cases of moral dilemmas we distinguished three levels: a) the main discussed moral dilemma or question in the case, b) the main moral theme in the case, and c) the most important subthemes derived from the context of the case. Here we used an inductive process; when possible, the moral dilemma or question was copied directly from the evaluation forms (Ryan & Bernard, 2003). In the main themes we tried to grasp the essence of the case in one word or concept, and subthemes were