Markov Decision Process Model for Patient Admission Decision at an Emergency Department in Disasters

Abstract

Efficient use of medical resources is essential to satisfying the surge demand experienced in the aftermath of a disaster. In hospitals, inefficiency exists in the use of bed resources when less urgent patients pre-occupy the beds so that there is no bed for more urgent patients arriving later. We develop a decision model to minimize such inefficiency in the event of a disaster through admission control of patients entering a hospital. We use Markov Decision Process (MDP) to make admission decisions for a finite horizon. It models the time-dependent arrival of disaster victims and their time-dependent survival probabilities. We numerically solve the MDP model using a discretization technique. The results of experiments conducted using virtual patient arrival data, in which the efficiency of our MDP solution was compared with that of other operating schemes, indicate that our proposed MDP model can improve the efficiency of current operations.