Insights from combining Action Research and System Dynamics when facilitating change and improvement processes in healthcare

Abstract

BACKGROUND Healthcare is a complex system with multi-professional staff and multiple patient care pathways. Time pressure and minimal margins for error makes it challenging to implement new policies or procedures, no matter how desirable. Changes in healthcare also requires the participation of the multi-professional staff. Action Research (AR) can provide staff engagement and local adaptation of practices. System Dynamics (SD) simulations can lead to shared systems understanding and allows for the development and testing of new scenarios in silico before actualising solutions.

AIMS OF THESIS To deepen the understanding of the interplay between SD and AR when applied to problems in healthcare.

To identify possible working patterns and principles when integrating SD into AR to support multi-professional groups in healthcare to find solutions to work-related challenges.

An additional objective was to explore potential benefits from using SD for policy planning of disease prevention in Swedish healthcare.

METHODS In Papers I-II the five studied improvement cases were facilitated using an approach where SD was integrated into AR. The cases were re-analysed in depth by a multidisciplinary work group (SD, AR, medical sciences) using iterative abductive qualitative analyses. Frameworks for studying consultative projects were used to identify steps in the workflows of the cases. Socio-analytical questions were used to bridge between the AR and SD perspectives. In Papers III-V, SD methodology was used as part of two healthcare research projects. The two research groups (medical researchers/practitioners, SD experts) used SD to study disease characteristics and preventive effects by different interventions. The SD models were based on epidemiological data from disease-specific quality registers, from scientific publications and from hospital systems.

RESULTS The five improvement cases had an overall pattern of two major divergent and convergent phases that began with an extensive AR-inspired inventory of problems/objectives and ended with a SD-facilitated experimental phase where mutually agreed solutions were tested in silico. Each group moved towards a shared developed point of reference for the problem/objective and solution, a case-specific multi-professional knowledge repository. The facilitator had a neutral and catalytic role, acting as an "interpreter" between the group and the model, ensuring that the model elucidated the issues of the groups.

AR contributed to high levels of engagement among the participants and to the building of confidence in and ownership of the results. AR also ensured that the SD model was adequate, relevant, and rooted in reality. SD provided a coherent and consistent systems overview of the complex and complicated structure of each improvement case, offered causal rigor, and provided ample opportunities for reality checks. During the process, the two methods were deeply integrated and present all the time in experiential learning processes.

The two research projects were epidemiological in nature and the simulations allowed studying phenomena which were difficult to isolate and examine in reality. The projects resulted in models depicting disease trajectories which were used to test different scenarios and suggest relevant epidemiological interventions.

CONCLUSIONS Combining SD and AR to problems in healthcare has the potential to achieve outcomes that are more useful, comprehensive, and robust than applying either approach in isolation. The identified results are, by design, calibrated to local needs and circumstances and when compared to traditional top-down implementations of change processes, improves the likelihood of sustained actualisation. The key work pattern of this approach was to build multiprofessional knowledge repositories through divergent and convergent phases of problem identification and solution exploration. This was supported by work principles that ensured that the voice of each participant was heard, inspired engagement, interaction, and exploratory mutual learning activities.

SD is well-suited for policy planning of disease prevention in Swedish healthcare. The methodology is cost effective and allows for simulations to be carried out in silico for testing without risk to patients or organisational efficiency. It also increases the understanding of systemic interdependencies between various patient-related and intervention-related factors for different diseases. Policymakers can for instance be assisted in choosing the intervention with greatest preventive impact by being presented with likely effects from expected or plausible scenarios.