Visits to the emergency department as transactional data

Patients arrive at the hospital emergency department for treatment on a random basis. The amount of time required for treatment is a function of the triage level, the patient diagnosis, and the congestion that exists in the emergency department (ED) at the time of patient arrival. The implementation of electronic medical records in the ED permits the accurate tracking and examination of time to allow for improved scheduling of personnel and for the development of protocols for diagnoses that occur on a daily basis in the ED. The SAS Institute in Cary, North Carolina, has developed a method called High Performance Forecasting System that allows for the prediction of time series with random time points. The target variable is the amount of time needed to treat individual patients from the time they enter the system through triage to the time they are discharged. Variability in treatment time by ED personnel can also be examined. Results indicate that there is a "cut point" during the day when the time needed to treat each patient increases. In addition, some specific diagnoses can be anticipated on a regular basis, and the amount of time needed on a daily basis can be predicted by using the forecasting system. Personnel can be scheduled in accordance with the peak treatment time.