Physician variation in ordering of transthoracic echocardiography in outpatient pediatric cardiac clinics

Background: The pediatric Appropriate Use Criteria (AUC) for outpatient transthoracic echocardiography (TTE) aim to reduce practice variation. Little is known on variation in TTE use between physicians. Understanding this variation will help identify areas for improvement in standardization of TTE use. Methods and Results: This is a retrospective review of initial outpatient visits at 6 pediatric cardiology centers in the United States prior to AUC release. Variation in TTE use was examined using multilevel generalized mixed effects models. Forward selection identified combinations of variables that explained the most variance in TTE use between physicians. Due to collinearity, physician compensation model and center were analyzed separately. Of 2883 encounters, the most common indication was murmur (36%), followed by chest pain (15.2%). Overall TTE use was 41.9%, and varied widely between centers (22.9%-52.6%), and between physicians within centers. Center alone explained 29% of this physician variance. Adding physician characteristics increased the variance explained to 57%, which only minimally improved by adding patient characteristics. The variance explained was driven by subspecialty. The center-based multivariable model explained more variance over compensation model. Conclusions: Center was the single largest determinant of physician variance in TTE use, followed by physician subspecialty. Efforts to reduce practice variation, such as the AUC, should be employed across centers and all pediatric cardiac providers. Center appears to have a stronger impact on variance than compensation model, though in this dataset the effect of center and compensation are hard to separate from each other and deserve further evaluation.