Table 5 “InDepth” data collection process

Identifying interviewees

Sampling of candidates for interviews follows a ‘criterion strategy’[46] in order to reach individuals at multiple levels of the institutional hierarchy and in key departments such as hospital administration, infection control program, and the intensive care unit (ICU). Additional interviewees are being selected following a ‘snowball sampling strategy’ based on recommendations of key informants by interview partners. Typical interviewees include individuals at top-management, ICU, and infection control level. See Table3 for typical interviewee profiles. We planned for 10 to 15 interviews per site visit.

Semi-structured interviews and interview guide

Interviews are approximately one hour long and follow a semi-structured format, according to a ‘general interview guide approach’[46]. Two researchers participate in each interview, with one person leading the conversation and the other backing up for recording, note-taking and occasional jump-in questions.

Interview guides are prepared before each phase of the inquiry, listing questions and issues to be explored (Table3). This approach ensures that the same lines of inquiry are explored, while still leaving freedom for each interviewee to elaborate on topics of particular interest. Interview sections and questions for the first series of site visits (before the intervention launch) were selected to understand background, position and network of the interviewee; the interviewee’s perception of healthcare-associated infection as well as those of other members of the organization; the experience of the institution with past implementation experience to estimate ‘implementation fitness’ and, more concretely, pre-existing achievements in catheter-related bloodstream infection prevention; rounding up with a broader view of safety and organizational culture.

Interview guides for the second series of site visits (one year after the launch of the initiative) will be designed according to the cross-case analysis of the first series and in line with potential suggestions from the literature at that point in time.

Ethnographic observations

Ethnographic observations are useful as a tool for methodological triangulation. These direct, personal contacts, help us to better understand the context within which people interact, to obtain frank accounts from spontaneous, brief interviews, and to enrich data with a non-subjective view in agent behavior and sensemaking[49]. Two researchers perform observations at a time, taking notes when possible about the setting, activities taking place, who participates, and meaning of what was observed. Immediately after the observation session of 2 to 3 hours, observers individually produce narratives of the observation that serve as the primary material for the analysis alongside the interview transcripts. We perform 2 to 3 observation sessions per site visit.

Photographic documentation

Photography allows for documentation of rich, vivid accounts of the reality[46]. We take pictures of locations and important items, such as insertion sets and hand hygiene dispensers. Shots of the town and countryside further help to recall the context during the analysis. Additionally, we produce drawings of the unit layouts.

Artifacts collection

During site visits, we collect artifacts such as guidelines, written protocols, data collection and audit forms. Often, these have to be translated for the analysis.