Stochastic Dynamic Programming in Hospital Resource Optimization

The costs associated with the healthcare system have risen dramatically in recent years. Healthcare decision-makers, especially in areas of hospital management, are rarely fortunate enough to have all necessary information made available to them at once. In this work we propose a stochastic model for the dynamics of the number of patients in a hospital department with the objective to improve the allocation of resources. The solution is based on a stochastic dynamic programming approach where the control variable is the number of admissions in the department. We use the dataset provided by one of the biggest Italian Intensive Care Units to test the application of our model. We propose also a comparison between the optimal policy of admissions and an empirical policy which describes the effective medical practice in the department. The method allows also to reduce the variability of the length of stay.