Abstract
OBJECTIVE: Overuse of diagnostic tests is of particular concern for pediatric academic medical centers. Our objective was to measure variation in testing based on proportion of hospitalization during the day versus night and the association between attending in-house coverage on the teaching service and test utilization for hospitalized pediatric patients. METHODS: Electronic health record data from 11,567 hospitalizations to a large, Northeastern, academic pediatric hospital were collected between January 2007 and December 2010. The patient-level dataset included orders for laboratory and imaging tests, information about who placed the order, and the timing of the order. Using a cross-sectional effect modification analysis, we estimated the difference in test utilization attributable to attending in-house coverage. RESULTS: We found that admission to the teaching service was independently associated with higher utilization of laboratory and imaging tests. However, the number of orders was 0.76 lower (95% confidence interval:-1.31 to -0.21, P = .006) per 10% increase in the proportion in the share of the hospitalization that occurred during daytime hours on the teaching services, which is attributable to direct attending supervision. CONCLUSIONS: Direct attending care of hospitalized pediatric patients at night was associated with slightly lower diagnostic test utilization.