Operational Foresight in Hospital Flow: A Dual-Approach Study Using Data-Driven Heuristics and System Dynamics

Project Details

Student(s): Elias El Ghossein, Jhonny Harb
Advisor(s): Dr. Marc Haddad
Department: Industrial & Mechanical
Academic Year(s): 2025-2026

Abstract

This project is focused on the creation of a proof-of-concept model using the system dynamics methodology to optimize patient flow in a French hospital. The work consists first of developing data analysis heuristics to clean and validate raw event logs so as to retain only clinically coherent stays from patient admission to transfer and discharge. A System Dynamics model for two hospital units (Neurosurgery/Orthopedics and Intensive Care) was then created to capture occupancy-driven slowdown through feedback loops and calibrated discharge suppression. A real-time, non-simulative heuristic - the Symmetric Lens - was finally applied to compute a Response/Pressure ratio from inflow (admissions + transfers) and outflow (transfers) windows. A Domino Risk Alert signals imminent cross-unit congestion, enabling proactive coordination across clinical pathways. Together, these methods allow deep structural insight and rapid operational decision-making for optimizing patient flow.