Abstract
We describe data analysis undertaken as part of a patient admissions prediction project underway through Emergency Departments (EDs) at two hospitals in the Southern Area Health Service within Queensland Health. The project involves the modelling and software development of forecasting tools that predict ED admission for time and day of the year. These tools may assist with the allocation of inpatient beds to aide in alleviating two major problems of most EDs: overcrowding and access block.1 The modelling data consists of fi ve years of ED presentations and admissions (1/7/02 - 30/6/07) from the Gold Coast and Toowoomba hospitals which were chosen for their different demographic characteristics. Toowoomba refl ects an entire population (~90,000) served by one ED with a fairly stable population, unlike the Gold Coast, which has one of the busiest EDs in the state, a large itinerant population (~500,000) and numerous other EDs serving the area. Many useful characteristics which can help shape health management practices have been identifi ed from the data. For example, the date and time when admitted patients leave the ED, indicating the times of highest demand on hospital beds; patient arrival time in the ED, which represents a staffi ng impact with workload; and the days of the week which represent higher ED workloads and hospital bed demand. The data also enables the analysis of 'frequent-fl yers' - those patients who presented multiple times during the analysis period. From the analysis of this data, we have been generating forecast estimates and associated confi dence intervals based on several forecasting approaches and validating the forecasts against actual data. The project also includes packaging the most accurate technique into a standalone software application.