Using Prediction to Improve Patient Flow in a Health Care Delivery Chain

Abstract

Often, in a health care delivery chain, lack of coordination has been detrimental to timely, high quality care. This paper focuses on the two steps of the hospital health care delivery chain, an emergency department and a hospital’s inpatient units. Past research into this chain has suggested that early prediction of patient need for admission can be used to better align flow between the two departments. This chain and the nature of prediction in health care delivery are discussed as well as a how prediction may be useful in this context. Finally tools for making admission predictions are tested and their possible implications are explored. The results of this exploration show that both expert opinion and a Naïve Bayesian statistical approach have predictive value in this context.