Development of an emergency department work score to predict ambulance diversion

Objectives: The authors sought to develop and validate an emergency department (ED) work score that could be used in real time to quantify crowding and staff workload in an ED. This work score could be used by public health officials to direct ambulance traffic based on an objective measure of ED status and to track ED conditions over time. In addition, the authors sought to determine which portion of ED care was most responsible for crowding. Methods: The setting was a tertiary teaching hospital with an emergency medicine residency. A number of ED parameters were measured throughout 2003 and then matched to times that an ED was on diversion status. Odd months of the year were used to develop the standard and even months to validate the standard. A marginal logistic regression analysis was used to develop the standard. The decision to divert ambulances was used as the criterion for ED crowding. Results: The logistic regression demonstrated excellent correlation between the work score and diversion status. At the point of maximum inflection of the receiver operating characteristic curve, the work score predicted diversion status with 86% sensitivity and 80% specificity. Conclusions: An ED work score was successfully developed and internally validated. External validation should be performed before widespread use.