Emergency Physician Variation in Admissions for Common Chief Complaints

Abstract

The emergency department (ED) represents an increasing source of hospital admissions, placing costly decisions in the hands of emergency physicians. Emergency physicians have been shown to vary dramatically in their admission rates, suggesting some providers may admit patients unnecessarily. It is unclear whether this variation in provider admission rates exists across all types of patient presentations. Determining the degree of admission rate variation for common chief complaints may help delineate clinical presentations that increase risk of inappropriate and costly hospital admissions. This retrospective study examined board certified/eligible emergency physicians practicing at an 850-bed tertiary level 1 trauma center with 85,000 annual care visits. Physicians worked equally in all areas of the ED. Patient characteristics and encounter data were obtained from an electronic medical record and physician order entry system from August 2008 to January 2012. Urgent care visits, patients younger than age 8, and trauma activations were excluded. Physicians were excluded if they worked less than 200 clinical hours, had less than 50 observations of a given chief complaint, or worked primarily in the pediatric ED. Admission rates were modeled using Emergency Severity Index, patient age, and time of presentation with logistic regression to adjust for differences in patient populations between physicians. Adjusted admission rates were modeled for the chief complaints of abdominal pain, shortness of breath, chest pain, and headache. Odds ratios were calculated to determine whether physicians with high admission rates (upper quartile) were consistent across chief complaints. There were 193,532 eligible ED visits, 57,213 of which resulted in admission. There were 48 emergency physicians who averaged 2,494 clinical hours and saw 3,949 patients during the study period. Each physician admitted an average of 1,168 patients. The adjusted overall admission rate was 23.5% (95%CI 22.9-24.3%) of ED visits, ranging from 17.4% (95%CI 16.2-18.6%) to 33.5% (95%CI 32.3-34.8%). The upper quartile of providers ordered 87% of admissions above the group average. There was large variation in admission rates between providers for all chief complaints. Shortness of breath had the highest admission rate among chief complaints, with 36.8% (range 17.1-53.0%) of visits resulting in admission. This was followed by abdominal pain (29.2%, 16.2-41.2%), chest pain (24.9%, 11.8-38.4%), and headache (15.3%, 5.4-26.7%). The odds ratio that high admitters for chest pain were high admitters for all patients was 34.0 (95%CI 5.84-197.8). High admission rates for shortness of breath (12.8, 95%CI 2.78-58.9) and abdominal pain (8.96, 95%CI 2.03-39.6) also predicted high overall admission rates. Odds ratios show high admitters of any one chief complaint were consistently high admitters for any one other chief complaint except headache. Large variation exists in provider admission rates for common chief complaints. A small group of providers appears to account for a significant volume of excess admissions in all chief complaints. Common factors may exist in this group that guide admission decisions. Chest pain and shortness of breath demonstrated the most variation among providers and appear to be good targets for interventions aimed at reducing unnecessary admissions.