Design for patient safety: A systematic evaluation of process modelling approaches for healthcare system safety improvement
Author: Jun, Gyuchan Thomas
Supervisor: Clarkson, P. John; Ward, James
Institution: University of Cambridge, Department of Engineering
This thesis systematically evaluates various process modelling methods with the aim of providing fundamental guidance on their effective application to healthcare system safety improvement.
Outside of healthcare, other safety critical industries, such as the aerospace, nuclear power and chemical industries, have used various process models to understand how their processes operate. Likewise, a whole-process (whole-systems) understanding by healthcare workers is an essential stepping stone for safer healthcare (re)design and management. However, a very limited range of process models have been extensively used in healthcare and no evidence was found that a wide range of process models were systematically reviewed for healthcare applications. In this research, various process modelling methods originating from a systems engineering discipline were compared and categorised to understand their key characteristics and differences. Using ten different process modelling methods, which were identified through the comparison and categorisation, three different care processes were modelled to study the feasibility of the ten different process modelling methods to healthcare. The usability and utility of each process model were evaluated by 29 healthcare workers through quantitative and qualitative feedback. Flowcharts, which have been extensively used in healthcare, were perceived by the healthcare workers as the most easily understandable and useful in systems understanding and communication. However, some other types of process models were perceived to be more suitable for specific purposes or users, e.g. state transition diagrams for patient-focused descriptions, communication diagrams for interface-related issues, swim lane activity diagrams for human resource-related issues and IDEF0 for managers. Based on the insights from these findings, different sets of process modelling methods were recommended according to their contexts: users (team or individual-based), resource for model building (paper or computer-based) and modelling purposes (task issues, interface issues or role issues). This research has contributed towards applying general-purpose modelling methods to a healthcare domain, which was one of the most challenging domains considering its diversity and complexity.