Hospital personnel are commonly tasked with implementing innovative and evidence-based practices. However, successes are often limited and short-lived. One likely explanation is that implementation processes vary between sites, leading to differences in implementation outcomes. In this dissertation, I built on the organizational and implementation science literatures to improve our understanding of implementation processes and how they unfolded in small, rural hospitals in Iowa.
I adopted two theoretical perspectives – process and variance models. Process models explain change as a series of steps or phases that organizations go through when implementing changes, while variance models explain variation in change outcomes as a relationship between variables. More specifically, I examined Kotter’s process model of change and tested the proposition that performance on earlier steps influences performance on subsequent steps. I then built on the literature on the Promoting Action on Research Implementation in Health Services (PARIHS) framework to examine the implementation processes from the variance model perspective. I first developed a typology of internal facilitation activities that hospital change agents engaged in throughout the implementation process, and then built on organizational implementation models to assess the influence of management support, time availability and team viability on sustainment and non-sustainment of facilitation activities.
This study was based on a qualitative longitudinal evaluation of TeamSTEPPS implementation in critical access hospitals in Iowa. Our research team recruited 17 hospitals attending TeamSTEPPS Master Training in 2011, 2012 and 2013 and followed them for a period of two years, interviewing key informants quarterly to inquire about their goals, strategies and activities, barriers and facilitators, and the progress they were making. My analytic samples for the dissertation included 8-10 hospitals and varied depending on the research questions. For each construct, a group of student coders read and coded the interview transcripts (two coders per transcript) using both inductive and deductive coding approaches. The coded content was reviewed and disagreements discussed in a group meeting until differences have been resolved. To examine the Kotter model, hospitals were scored on their performance on the three phases, which allowed for assessing whether their performance was consistent across the three phases as the model proposes. To develop the typology of facilitation activities, I compared and contrasted the different types of activities to identify the characteristics that distinguish them. To test sustainment of facilitation activities, I used the fuzzy-set Qualitative Comparative Analysis method to calibrate and test the relationships using set-theoretic methods. As needed, I also qualitatively re-examined the cases to identify exemplar cases or identify additional factors that helped develop our understanding of the implementation processes.
I found that the Kotter model helped explain the implementation processes in half the hospitals, while the other hospitals followed different trajectories, depending on implementation scope. Next, four types of facilitation activities were identified – Leadership, Buy-in, Customization and Accountability. They are distinguished by who engaged in the activities, what or whom they targeted, and the timing patterns of the activities. I also found that facilitation activities were sustained in hospitals with both senior and middle manager support and whose facilitator team remained viable throughout the implementation process.
These findings contributed to our understanding of implementation processes. Individual findings and their implications were discussed. Overall, both process and variance model perspectives provide useful but different insights into implementation processes. I concluded that both perspectives are needed to inform practice and future research.