Hourly emergency department census: A simple measure of crowding

Study objectives: Although emergency department (ED) crowding is widely believed to be a pervasive problem, the lack of a commonly accepted measure of crowding hinders research on the subject. We propose hourly ED census (census) as a measure of crowding. We measure ED crowding by calculating the census using the patientʹs ED arrival and departure times and demonstrate that census correlates with ambulance diversion and with patients leaving without being seen (LWBS), 2 adverse outcomes of crowding. Methods: This was a retrospective study of all patients treated in the main ED of an urban teaching hospital in 2002. We developed a computer program to calculate the census for each hour in 2002 using 2 variables from administrative databases—each patientʹs time of ED arrival and departure. Using logistic regression, we calculated the probability of a patient LWBS as a function of the census when the patient registered. We also used logistic regression to calculate the probability of ambulance diversion as a function of census. Results: Median census varied more than twofold over the course of the day, and the 90th percentile census at a given hour was approximately twice the 10th percentile census. LWBS rates were associated with census. The logit odds ratio was 1.05 (95% confidence interval [CI] 1.04 to 1.06). The LWBS rate ranged from zero at a census of 1% to 7.8% (95% CI 6.5% to 9.2%) at a census of 42 patients (175% of available treatment spaces). Diversion rates were near zero at a census of up to 30 patients (125% of available treatment spaces) and were higher above that. Ambulance diversion was strongly related to census, with an odds ratio of 1.10 (95% CI 1.07 to 1.12). Conclusion: ED crowding varies widely from day to day and from hour to hour. It can be measured by hourly ED census, which can be calculated using administrative data. The hourly census is a predictor of the LWBS rate and of ambulance diversion. Hourly census provides an objective, quantitative measure of crowding that may facilitate further research on the topic.