The association between emergency department crowding and time to antibiotic administration

Study objectives: There is concern that crowding in our nationʹs emergency departments (EDs) leads to lower quality of care, but there is little evidence that this is the case. Time to antibiotic administration (TTAA) for patients with community acquired pneumonia (CAP) is a national hospital quality measure used by the Centers for Medicare and Medicaid and the Joint Commission on Accreditation of Healthcare Organizations where ED performance is critical. Our objective was to determine the relationship between ED crowding and TTAA for CAP patients. Methods: We conducted a retrospective observational cohort study using the ED encounter database from a medium-sized community hospital in Connecticut to identify patients admitted with CAP from December 2001 through December 2002. “Crowding” was defined in 2 ways: (1) ED census when the CAP patient arrived exceeded ED bed capacity; and (2) mean ED length of stay for patients in the ED when the CAP patient arrived exceeded national mean (4 hours). Our outcome measure was based on national performance measures of TTAA for CAP (240 minutes). We reviewed medical records to determine TTAA for CAP patients. CAP cases were included in TTAA determination if age was older than 18 years, ED admitting diagnosis was CAP, hospital discharge diagnosis was CAP, and admission was from home or a nursing home. CAP cases were excluded from TTAA determination if antibiotic therapy was given before presentation, the patient was in the hospital during the previous 10 days, or if immunodeficiency, Pneumocystis carinii pneumonia, or tuberculosis was suspected. We sought a sample size (N=310) to detect a 1-way difference of 12 minutes in TTAA with α=0.05 and β=0.80. Bivariate analyses were completed using simple linear regression and t test. Secondary analyses were also completed stratifying for pneumonia severity index (PSI) and the shift the CAP patient registered. Results: We identified 438 CAP patients, of whom 302 met criteria for inclusion in the national quality measure. Patientsʹ mean age was 71.8±16.8 years; 133 (44%) were women; 88 (29%) had a PSI of 90 or less (stable for outpatient treatment); and 264 (87%) were treated between 7 AM and 11 PM. The correlation coefficient for mean ED census versus TTAA was R=0.05 (P=.37) and for mean ED length of stay versus TTAA was R=0.02 (P=.69). Using “census greater than or equal to capacity” definition of crowded, the difference between TTAA during crowded and uncrowded periods was 2.4 minutes (P=.91). Using “ED length of stay greater than or equal to national mean” definition of crowded, this difference was 8.8 minutes (P=.75). Data stratified by PSI of 90 or less versus greater than 90 and by arrival between 7 AM and 11 PM versus 11 PM to 7 AM revealed no statistically significant associations. Post hoc analyses, although statistically not significant, show that TTAA was shorter at lower census (<70%). Conclusion: ED crowding, by ED census or ED length of stay criteria, does not have a statistically significant association with TTAA for patients admitted with CAP. These results demonstrate that ED crowding is a complex problem, and more research is needed to understand its effect on patient care.