Design and methods
Participatory action research
We used participatory action research as our methodological approach (Reason & Bradbury 2001, Stringer & Genat 2004). Action research dates back to the 1940s when Kurt Lewin introduced action research as a dynamic, collaborative process for helping people address social issues influencing their lives. Through cycles of planning, acting and reflecting, people were able to bring about change and social improvement. This approach was chosen because it compliments the project’s goal of engaging nurses–nurse leaders in shared decision making. Participatory action research relies on building relationships between researchers and the ‘experts’, or those individuals who are most aware of their environment. Researchers become participant observers, and they act primarily as facilitators to enhance relationships that enable others to gradually take control. This approach allowed us to connect theory to practice.
Methods
Similar to qualitative methods described by Kramer et al. (2008), we used participant observation (i.e. two facilitators taking field notes at each meeting), leader interviews and project team and staff focus groups to collect our data. Ethics approval was obtained from the principal investigator’s university ethics board and the ethics boards of the health authorities representing our project sites. After recruiting project teams, consents were obtained from them to take field notes during team meetings, and to digitally record and utilize de-identified comments from team discussions. Consents were also obtained from team members, nurse leaders and nursing staff participating in end-of-project interviews and focus groups. Project team members and staff were invited to participate in focus groups of 3–10 individuals, and operations leaders were individually interviewed. We used semi-structured, open-ended questions such as: ‘Have you noticed any results in your unit/facility as a result of the workload project?’ The interviews and focus groups ranged in length from 20 to 40 minutes. We digitally recorded these sessions and took notes. Digital recordings were used primarily as an audit trail: we did not transcribe and categorically code meeting or interview/focus group transcripts as part of our cross-case analysis.
Analytic techniques
We used methods described by Miles and Huberman (1994) to construct cross-case SPO matrices for display and analysis of qualitative data from four project sites. There are inductive or deductive approaches for building matrices and exploring causal relationships. We followed a deductive approach by selecting initial SPO components based on theoretical/empirical literature. We organized or case-ordered data from our four sites according to successful outcomes. Successful outcomes were based upon project deliverables and end-of-project interview and focus group data.
We used participant observer field notes and team discussions to validate the SPO components and linkages in our original framework and to identify any missing processes/steps important to nurse–nurse leader shared decision making. Similar methods were used by Kramer and Schmalenberg (2005) to test and refine the SPO components and linkages associated with magnet hospitals. To organize our field notes data, we independently created SPO summaries or narratives that we compared among ourselves to consensually identify the most important components and linkages corresponding to each project team meeting. We used examples and quotations from team discussions and team-operations leader discussions to support our participant observations. At the time of project completion, separate focus groups with the teams and the nursing staff and interviews with operations leaders were used to validate our SPO matrices and summaries. Credibility was established through triangulation of field notes findings with interview/focus group findings. We maintained audit trails by referring back to original field notes and digital recordings of focus groups.