A decision support system for estimating short-term hospital inpatient demands

The prediction of hospital inpatient demands and bed capacity planning has been one of the major challenges for healthcare decision makers. For hospitals where the majority of inpatient admissions come from the Emergency Department (ED), it is essential to estimate real-time inpatient demands to allocate hospital resources efficiently. The objective of this research is therefore to develop a decision support system for healthcare managers to predict short-term inpatient demands. In this research, typical inpatient admission and discharge processes are studied to consider the length of stay at the ED for inpatients. Different process parameters, such as forecasting time, triage level, and ED and nursing unit census, are identified and statistically analyzed to predict real-time inpatient demands. Based on these various input parameters, the number of patients waiting and the number of current patients at different nursing units are estimated. The decision support system has been developed and validated with the support from a major U.S. hospital in the state of New Jersey.