Emergency medicine 4.0: an integrated data-driven approach to improve emergency department performances (n. P20222XM58)
Type
PRIN 2022 PNRR, D.D. n. 1409 del 14 settembre 2022
CUP
C53D23008000001
Principal Investigator
Alessandro Stefanini (Università degli Studi di Pisa)
Other research units
UO1 - Alessandro STEFANINI - Università degli Studi di PISA
UO2 - Mattia CATTANEO - Università degli Studi di BERGAMO
UO3 - Massimiliano DE LEONI - Università degli Studi di PADOVA
Duration
30/11/2023 - 29/11/2025
Description
Emergency department (ED) overcrowding is a worldwide issue that entails excessive waiting time, increased length of stay, clinical risk/stress of the ED staff, and overall patient dissatisfaction and poorer care. Given its economic and social relevance, governments
and health authorities, with the help of the scientific world, are paying special attention to it. For example, the efforts of the Italian Government to propose new forms of territorial medicine (i.e., Community Healthcare Centres) and to prioritise the problem within the PNRR framework should be seen in this direction. Thanks to the rising availability of process and patient-related data stored in Healthcare Information Systems, researchers have started focusing on new advanced data-driven techniques for analysing and improving healthcare processes. However, studies tackling the emerging data-driven methods for supporting real-time or quasi-real-time decision-making in ED systems are still scant and needed.
Trying to address this gap, the main goal of this project is to reduce ED overcrowding and improve ED performances by enabling real-time process monitoring and dynamic (real-time) management of patient-flows and ED resources through the most recent data-driven analysis techniques (e.g., Machine Learning, Process Mining, Statistical Learning, and Process Simulation). The final aim is to propose an integrated data-driven approach supporting the management of emergency medicine at a territorial or regional
level.
To this aim, the project intends:
(1) To develop and empirically test new Machine Learning (ML) models and techniques for time and workload prediction in EDs – e.g., patient arrivals, waiting times, and service times – to monitor ED processes in real-time.
(2) To estimate the potential changes in ED patient arrivals related to: i) the re-routing of ambulance patients in the EDs of the same geographical area based on the predicted ED workloads and time performances; ii) the introduction of Community Healthcare Centres (CHCs), which can take charge of minor patients and redirect the most serious ones to the less crowded EDs by exploiting the aforementioned predictions.
(3) To develop a simulation system customized for a single ED that allows to evaluate in real-time different ED configurations, in terms of resources and processes (e.g., fast tracks, See&Treat), based on the expected ED status (e.g., waiting times, service times, and
workloads). The project will be carried out by exploiting real datasets and field-testing thanks to the involvement of hospitals and local health organizations (see the attached letters of endorsement).
Related publications
1) F. Vinci, G. Park, W. M. P. van der Aalst, M. de Leoni: "Reliable and Configurable Process Simulations via Probabilistic White-Box Models". Proceedings of the 23rd International Confeence on Service-Oriented Computing (ICSOC 2025)
2) F. Vinci, G. Park, W. M. P. van der Aalst, M. de Leoni: "Online Discovery of Simulation Models for Evolving Business Processes". Proceedings of the 23rd International Conference on Business Process Management (BPM 2025)
3) F. Vinci, M. de Leoni: "Balancing Fitness and Precision in Process Model Repair: Framework Formalization, Assessment, and Benefits for Simulation". Process Science, Volume 2, Article Number 3, 2025, Springer
4) F. Vinci, M. de Leoni: "Repairing Process Models through Simulation and Explainable AI"
Proceedings of the 22nd International Conference on Business Process Management (BPM 2024)
