This Small Business Innovation Research (SBIR) Phase I project will address the problem of short-term hospital census prediction. Fluctuating occupancy can impact many operational metrics: direct staffing costs, emergency center ambulance diversions, medical errors triggered in understaffed patient areas, quality of care, and job satisfaction for the dwindling nursing pool. This project will result in a computer based decision support system for predicting short-term patient occupancy at a nursing unit level over 72-hour time horizons. The short-term forecasts will be tested in a 226-bed community hospital. The first objective is to quantify sources of error in an alpha-version of the model and to develop an improved model. The objective is to prove that the improved model generates accurate predictions of patient census. Next, the research will quantify the potential impact of short-term forecasts on nurse staffing and scheduling. Surveys with hospital personnel will gauge the interest in and utility of the forecast information. The national shortage of nurses demands better use of this limited resource. Improvement Path Systems' objective is to build commercial forecasting software to be licensed to hospitals. The daily forecasts of hospital occupancy will allow hospitals to make better staffing decisions and operate more efficiently. The proposed decision support system can be generalized to other dynamic systems involving people flows.