A Model for Inpatient Ward Simulation and Forecasting

Abstract

Making decisions on capacity requirements in hospital environments has long-term consequences. Sound estimates of expected mid and long-term demand for hospital beds are required to avoid oversupply or deficits. Though in many cases, forecasts of admissions and Length of Stay evolutions for specific pathologies are available, current models rarely model hospital environments on the pathology level, rendering them useless. Additionally, existing models are most often unit type or facility-dependent. This work describes a model to estimate care demand and capacity needs per ward of a generic hospital, in which pathology-specific parameters can be used, and that supports patient stays structurally segmented across multiple wards. The methodology enables an increase in model complexity while explicitly guarding the accuracy of parameter estimates.