High occupancy rates increase the probability of adverse events. In the U.S., the inpatient census averages an alarming 60 percent.
At Geisinger, we put a great emphasis on patient outcomes; therefore, inpatient bed planning is a critical aspect which requires the development of better strategies for data-driven decisions.
To deliver a timely and highly accurate estimate of the inpatient bed demand, our team has developed a predictive model based on the application of machine learning (ML) algorithms. The purpose of the model is to predict the hospital census for the next five days, allowing healthcare professionals to plan and act before an expected high census day. The predictive model is updated and improved consistently over the years. The most recent update improved the mean average error (MAE) from a MAE = 9.84 (old model) to MAE = 4.08 (new model). The newly developed predictive model is being used daily by operations teams at Geisinger Medical Center in central Pennsylvania and Geisinger Wyoming Valley Medical Center in northeast Pennsylvania.