LAUSR

Abstract Existing inventory models for medicines in the healthcare domain presume demand as a random variable independent of any environmental factors. Conversely, various randomly varying factors, such as changing patient condition, uncertain reaction of the patient to treatment, uncertain length of stay and transition from one type of hospital care unit to another at different stages of treatment, may have a significant impact on the demand of medicines. In dealing with such a problem, a Markov Decision Process model is developed for determining the optimal inventory control policy for medicines under stochastic and non-stationary demand scenario. The required data are collected from a multispecialty hospital situated in urban India. Solving the problem by a stochastic dynamic programming approach, the results as obtained demonstrate that the proposed inventory model using the knowledge of patient condition-based medication demand characteristics has a significantly lower total inventory-related cost than those based on historical daily demand without consideration of patient condition characteristics. This modelling may help optimize and support the functioning of any hospital system more effectively.