3 Methodology
3.7 Phase 2: Implementation and evaluation of the leadership intervention . 43
3.7.3 Data analyses
Paper II: Exploring and mapping of quality and safety challenges before implementation of the intervention
The data material in Phase 2 consisted of interviews with managers and employees from two nursing homes and two homecare services. All tape-recorded interviews were transcribed before deductive content analysis was initiated. The analysis was conducted by directed content analysis (Hsieh & Shannon 2005) according to the OQ framework (Bate et al., 2008) refined in Phase 1 (Johannessen et al., 2020). The directed content analysis (Hsieh & Shannon 2005) was used as described in Phase 1 (Chapter 3.5.3). The purpose of the analysis is to condense a large amount of text into categories that are similar in meaning (Hsieh &
Shannon 2005; Patton, 2015). The predefined categories were structure, coordination and organisational politics, culture, competence,
engagement, psychical design and technology and external demands.
The data material from each case was analysed by a within-case analysis (Yin, 2018). The transcribed interviews were read through and text was highlighted and coded after the predefined categories. In the analysis process we identified subcategories in the predefined categories, as shown in an example of the predefined category structure in homecare service 1, in Table 7.
Table 7. Directed Content Analysis of Structure in Homecare Service 1 Predefined
category Subcategory Codes Structure Struggle
with continuity
What has happened in recent years is that the groups are getting bigger. The management density is not as it was before. There have been larger units, so that each manager has greater personnel responsibilities, and is unable to be so close to the employees.
But it is clear that there will always be a lot of work anyway, new shifts and things like that! I always run out of time to adjust.
I think that the tasks take an enormous amount of time, and that is precisely why I think that this is so important in a way, to be even more in the group of the employees and where it in a way moves the most.
There is a bit of feedback from my employees as well, that they experience that I am not as much present as I have been before, and it is clear that I do not follow up in the same way I would like.
Action plan ... so it's nice that they are updated so that you have a safe place to go and check. If I'm not sure what to do, I can go there and read, but that's not really how it works today because we do not have time to update them.
Data that could not fit the initial coding category was identified and analysed later (Hsieh et al., 2005). For example, ‘contexting’ came up during data analysis. Data (meaning units) of context work that did not fit the initial coding categories were gathered in one additional category.
The authors read the transcripts and discussed theme development in several meetings. The within-case analysis was followed by a cross-case analysis to trace similarities and differences among municipalities, between nursing homes and homecare services, and between managers and employees. In the cross-case analysis (Yin, 2018) the data was sorted in a table format. The data on ‘contexting’ was used to review the data;
and we extended and adapted the OQ framework to the nursing home and homecare service. The context work that managers do in quality and safety improvement did not fit the initial coding categories, so
‘contexting’ was added to show context as an active notion rather than a frame around the organisations. The purpose of a directed approach to content analysis is to validate or extend conceptually a theoretical framework or theory (Hsieh et al., 2005). The initial coding was initially deductive but by the end it was more abductive. The abductive approach rests on the cultivation of anomalous and surprising empirical findings against a background or framework of existing theories and through systematic methodological analysis (Timmermans & Tavory, 2012).
Similar research by Stoopendaal and Bal (2013) linked ethnographic findings to theory and found that different kinds of work had to be done by human and non-human actors to displace improvements into specific organisational situations.
Paper III: Evaluation of the intervention
The data analysis in Paper III used transcripts from focus group interviews and observational notes. The data material was subjected to integrative analysis (Strøm & Fagermoen, 2012) to integrate interviews and observation notes collected throughout the 12-month intervention and analysed as a complete dataset from each organisation as described in phase I (Chapter 3.6.3). Co-researchers were not directly involved in
the analysis. However, one researcher from each intervention team was involved in the analysis to ensure sound interpretation of results and that descriptions from the organisations were recognisable and described the process accurately. Within-case analysis was conducted to capture information within each of the four cases over the intervention period (Yin, 2018). An inductive content analysis was conducted on the organisations’ implementation, changes during the intervention and mechanisms that contributed to implementation and quality and safety improvement work. All authors read the data and highlighted themes.
Meaning units were extracted from the text, sorted and categorised in Microsoft Word. Here, the meaning units were condensed and translated from Norwegian to English. These were developed by integrating data from interviews, workshop notes, and observations describing the organisations’ intervention process and changes throughout the intervention period. The analysis was sent to supervisors for review. The third step was a cross-case analysis (Yin, 2018) to compare and contrast the units’ and managers’ work practice to improve quality and safety, and to identify requirements for the intervention. The entire research team then met to agree on themes and categories. The implementation was depicted in a Microsoft Word table. A narrative of each case was drafted (Langley, 1999) and discussed with the research team. The interweaving produced a large amount of data for analysis and for validation. The analysis enabled the researchers to validate whether and how the observation and reflection corresponded with the managers’ and employees’ experiences and opinions (Strøm & Fagermoen, 2012). The integrative analysis resulted in a descriptive overview of each organisation and how the managers implemented and worked with the leadership guide.