Digital Twins for Decision Making: Combining Ambulance and ED Facility Resources

Clinical Snapshot

Key Findings

• P Value: Not applicable — simulation study; no formal statistical hypothesis testing reported

• Effect Size: Not formally quantified; results reported as qualitative patterns of depletion and recovery. Reduction from 43 to 34 TS (a 21% capacity reduction) was associated with failure to recover TS availability within the 24-hour simulation window

• Primary Outcome: Patterns of treatment space (TS) depletion and recovery over a 24-hour period under three capacity scenarios (43, 38, and 34 TS) with five ambulance bays, including identification of scenarios where TS recovery failed to occur within 24 hours

• Nnt Or Sensitivity: Not applicable in traditional sense; the key operational metric is patient off-stretcher time (POST) and ambulance bay occupancy as proxies for system performance. Specific POST values not reported in abstract

• Confidence Interval: Not reported

Clinical Application

Implementation of a digital twin requires sustained data infrastructure investment, including real-time or near-real-time feeds from ambulance dispatch systems, ED information systems, and bed management platforms. This is feasible within well-resourced metropolitan health networks but may be challenging in regional or rural settings. The DES algorithm itself is described as established, suggesting the computational component is not a primary barrier. Highly relevant to the Australian context. Ambulance ramping and ED overcrowding are recognised national patient safety priorities, with the Australian Commission on Safety and Quality in Health Care and state health departments actively seeking capacity planning solutions. The study was conducted using Queensland hospital data by CSIRO, directly informing Australian health system planning. There are no PBS or TGA regulatory considerations for this type of modelling tool. RACGP and ACEM (Australasian College for Emergency Medicine) guidelines on ED access and flow are relevant benchmarks. The National Emergency Access Target (NEAT) — requiring 90% of ED patients to be admitted, discharged, or transferred within four hours — provides a policy framework within which this tool's outputs should be interpreted. State-level ambulance ramping data (particularly prominent in Queensland, Western Australia, and Victoria) underscores the operational urgency of this research. Emergency departments with mixed ambulance and walk-in presentations, particularly those operating near or at treatment space capacity. Most directly applicable to mid-sized Australian public hospitals (200–300 beds) with high ambulance demand.

Abstract

Emergency department (ED) walk-ins and ambulance arrivals are often modelled independently, overlooking their interdependence and the impact of fixed treatment space (TS) capacity. An established discrete event simulation algorithm has been tailored as the basis of a digital twin (DT). Driven by regular streams of incoming ED presentations, patient off-stretcher time (POST), ambulance bay and TS occupation across a 24-hour period contingent on patient flow and ED TS capacity was simulated. The simulation was calibrated using 19,570 ambulance records and data from a 250-bed hospital (2021-2022). Scenarios of five ambulance bays with 43, 38, and 34 TS capacities were investigated, revealing patterns of TS depletion and recovery, including cases where recovery failed within 24 hours, underscoring the sensitivity of patient flow to resource constraints and the value of DT-based simulation for capacity planning.

References

1. 1.Tan, D., Cespedes, M., Hassanzadeh, H., Riahi, V., & Khanna, S. (2026). Digital Twins for Decision Making: Combining Ambulance and ED Facility Resources. Studies in Health Technology and Informatics. https://doi.org/10.3233/SHTI260965