Estimating patient discharge from hospitals using both historical and physician-supplied estimates combined in a...

The effectiveness of admissions scheduling systems for optimal resource allocation in hospitals is directly dependent on the ability to estimate patient discharges accurately. Efforts to estimate discharges to date include two general methods: estimating from historical data, and physician-supplied estimates. This paper takes the position that there are essentially two criteria in evaluating methods of discharge estimation: accuracy of information and cost of information. A cost/benefit concept is used to argue that an optimal mixture exists between historical-based estimates and physician-supplied estimates, and a model is developed which combines the best aspects of each method. Data from a 300-bed hospital in Michigan are analyzed to provide the parameters for the model. The discharge estimating model was designed to be an integral part of an admission scheduling system recently implemented in the study hospital. Results of the discharge estimating model and its participation in admissions scheduling are discussed.