Abdominal imaging ordering patterns by referring provider specialty

Andrew B. Rosenkrantz, Abimbola Ayoola, Richard Duszak

Prior work has demonstrated marked growth in the volume of abdominal imaging performed by radiologists. However, decisions to pursue imaging are largely driven by referring providers. In this study, we take the novel approach of investigating abdominal imaging utilization patterns by referring provider specialty. Data on imaging services were obtained from the 2014 DocGraph Medicare Referring Provider Utilization for Procedures (MrPUP) public use file. MrPUP contains aggregated transaction data for combinations of unique referring provider and service code. Imaging services were classified by modality and body region using the Neiman Institute Types of Service (NITOS). Each referring provider's specialty was determined using Medicare Physician Compare. Abdominal imaging ordering patterns were summarized by referring specialty. The final dataset included 5,824,754 abdominal imaging transactions. The most common ordering specialties of abdominal imaging were as follows: (1) internal medicine; (2) urology; (3) emergency medicine; (4) family practice; and (5) gastroenterology. The most common ordering specialties by abdominal imaging modality were emergency medicine for CT; gastroenterology for MRI and nuclear medicine; and internal medicine for ultrasound and radiography. While numerous specialties commonly ordered abdominal radiography and CT, urologists also commonly ordered retroperitoneal ultrasound and retrograde urography, and gastroenterologists also commonly ordered abdominal ultrasound, abdominal MRI, and esophagrams. Internal medicine, family practice, and emergency medicine providers ordered a much broader mix of imaging, including many non-abdominal imaging examinations. Referring specialty abdominal imaging ordering patterns are varied but distinct. Awareness of these patterns may facilitate focused educational and policy initiatives to improve abdominal imaging appropriateness and utilization.