Demand analysis and capacity management for hospital emergencies using advanced forecasting models and stochastic simulation

Abstract

Demand forecasting and capacity management are complicated tasks for emergency healthcare services due to the uncertainty, complex relationships, and high public exposure involved. Published research does not show integrated solutions to these tasks. Thus, the objective of this paper is to present results from three hospitals that show the feasibility of routinely applying integrated forecasting and capacity management with advanced operations research tools. After testing several forecasting methods, neural networks and support vector regression provided the best results in terms of variance and accuracy. Based on this forecasting, a logic for managing hospital capacity was designed and implemented. This logic includes the comparison between the forecasted demand and the available medical resources and a stochastic simulation model to assess the performance of different configurations of facilities and resources. The logic also provides hospital managers with a decision tool for determining the number and distribution of medical resources on emergency services based on a cost/benefit analysis of resources and service improvement. Such results support the task of assigning doctors to different kinds of boxes, defining their work schedules, and considering additional doctors. The contribution of this paper consists of an integrated solution designed to implement the abovementioned logic. This solution combines forecasting, simulation for capacity management, process design, and IT support, facilitating the practical routine use of complex models. The integration explicitly considers a solution that also has adaptation capabilities to facilitate use under changing conditions. The solution is also general and admits adaptation and extension to other services. Thus, we have already performed similar work for ambulatory and surgical services.