P129: Variability in ordering of computed tomography (CT) scans among emergency physicians

Abstract

Introduction: The increased availability and increased utilization of Computed Tomography (CT) imaging as a diagnostic tool has in the past several years led to concerns regarding the unknown and potentially harmful effects of ionizing radiation exposure to patients, as well as the increased cost to the health care system. Multiple education campaigns (e.g. Choosing Wisely) and institution-wide interventions have been implemented in order to limit the use of potentially unnecessary CT imaging. Two specific modalities CT head and CT angiography to rule out pulmonary embolism (CT PE) have been identified as potential targets of these interventions due to their likely overutilization in the clinical ED setting. The objective of this study was to determine the interphysician variability in the ordering rates of CT head and CT PE, and to determine if any correlation existed between CT head and CT PE ordering rates among physicians. Methods: Data was collected on all diagnostic imaging ordered by ED physicians at two very high volume community hospitals during the 4-year period between 2013 and 2016. Analysis was limited to those physicians who worked at least 3 of the 4 years at either site and saw at least 1000 patients per year. The ordering rates for each physician were calculated by dividing the number of the imaging modality ordered over the total number of patients seen. Correlation coefficients (r values) were calculated to determine if a linear correlation existed between increased CT head and increased CT PE ordering rates. Results: The DI ordering data for a total of 44 ED physicians were analyzed. Results show average 4-year ordering rates for CT heads among ED physicians ranging from 4.0% to 13.9%, and CT PE ordering rates ranging from 0.1% - 1.7%. The correlation coefficient between CT head and CT PE ordering rates was positive for all 4 years, with a statistically significant (p<0.05) correlation coefficient of 0.53. Conclusion: There is a wide degree of variability in DI ordering patterns among physicians working within the same clinical environment. Further exploration of this interphysician variability will be helpful in designing strategies to mitigate overutilization of diagnostic imaging.