Emergency Departments (EDs) must treat growing numbers of patients quickly and efficiently. However, there are bottlenecks caused by many reasons including the lack of information to process patients timely, the lack of decision-makers and the lack of timely decision-making that is affecting the smooth flow of processes. Techniques used to address bottlenecks have yielded limited sustainability due to reliance on simplistic models as inputs which do not account for the complexities and variations in the real system. This study aimed to address bottlenecks by developing a systematic model-driven approach, for assessing ED processes for improving waiting time as measured by the 4-hour quality indicator (4HQI).

Using an exploratory framework, this study employed a mixed-method approach in examining heterogeneous data to realise its aim. Semi-structured interviews with 21 ED clinicians were conducted in a level-1 ED of an Acute Trust in the UK. Interview transcripts embedded with systems knowledge were extracted to develop role activity diagrams (RAD) to capture granularity of care processes and identify bottlenecks through process mapping. Additionally, service utilisation data were analysed using logistic regression, generalized linear model and decision tree. The impact of changes on waiting time was assessed using Discrete Event Simulation (DES).

Process mapping revealed Majors, the unit that treats complex patients to be the most problematic in the ED and also identified five bottlenecks in the unit: awaiting specialty input, test outside the ED, awaiting transportation, bed search and inpatient handover. The process maps further revealed that information available to the ED at the pre-hospital phase and before entry into Majors can be better utilised to address bottlenecks, especially those related to awaiting specialty input, test outside the ED and awaiting transportation. This led to exploring improvement suggestions that included: (1) introducing an advanced nurse practitioner at triage, (2) utilising pre-hospital information to reduce repeat testing and (3) operating a discharge lounge. Results from the qualitative and quantitative analysis were integrated into a discrete event simulation (DES) model to evaluate the improvement suggestions, leading to reductions in the length of stay (LOS) for given scenarios. Several statistical models for predicting LOS and breach of the 4HQI were also developed.

The methodology developed entailed (1) qualitative process modelling to derive the systems model, (2) quantitative analysis of audit-level patient data to understand decision-making and patient flow (3) integration of qualitative and quantitative analysis results to derive improvement suggestions and (4) simulation to analyse suggestions. RADs served as a granular process mapping technique for bottleneck identification and solution derivation in analysing complex systems. Its application helped to derive realistic models of the system This is the first study to model Majors, unit. Furthermore, a methodology for indirect mapping of RAD to DES was developed to bridge the gap between the two methods where RAD provides granular input to complement DES models. Monitoring patients’ length of stay as three-time blocks, was recommended in addition to a model-based, data-informed alert system to support decision-making and patient flow.

This study sheds light on the development of quality indicators scientifically and operationally. The Majors unit identified as the most crowded unit underscores to ED managers and policymakers as an area of focus for improvement initiatives considering limited resources. This study modelled and analysed heterogeneous data to improve process flow in the ED. Implementing the recommendations made would enhance patient flow and bottlenecks, thereby improving waiting times.