Increase in patient mortality at 10 days associated with emergency department overcrowding.

OBJECTIVE To quantify any relationship between emergency department (ED) overcrowding and 10-day patient mortality. DESIGN AND SETTING Retrospective stratified cohort analysis of three 48-week periods in a tertiary mixed ED in 2002-2004. Mean "occupancy" (a measure of overcrowding based on number of patients receiving treatment) was calculated for 8-hour shifts and for 12-week periods. The shifts of each type in the highest quartile of occupancy were classified as overcrowded. PARTICIPANTS All presentations of patients (except those arriving by interstate ambulance) during "overcrowded" (OC) shifts and during an equivalent number of "not overcrowded" (NOC) shifts (same shift, weekday and period). MAIN OUTCOME MEASURE In-hospital death of a patient recorded within 10 days of the most recent ED presentation. RESULTS There were 34 377 OC and 32 231 NOC presentations (736 shifts each); the presenting patients were well matched for age and sex. Mean occupancy was 21.6 on OC shifts and 16.4 on NOC shifts. There were 144 deaths in the OC cohort and 101 in the NOC cohort (0.42% and 0.31%, respectively; P=0.025). The relative risk of death at 10 days was 1.34 (95% CI, 1.04-1.72). Subgroup analysis showed that, in the OC cohort, there were more presentations in more urgent triage categories, decreased treatment performance by standard measures, and a higher mortality rate by triage category. CONCLUSIONS In this hospital, presentation during high ED occupancy was associated with increased in-hospital mortality at 10 days, after controlling for seasonal, shift, and day of the week effects. The magnitude of the effect is about 13 deaths per year. Further studies are warranted.