Self-Interpretation of Imaging Studies by Ordering Providers: Frequency and Associated Provider and Practice Characteristics.

Abstract

Background. Imaging self-interpretation refers to interpretation of an imaging study by the study's ordering provider. This phenomenon warrants scrutiny given nonradiologists' limited training in imaging interpretation. Objective. To evaluate the frequency of self-interpretation of office-based diagnostic imaging by ordering providers and to assess associations of ordering provider and practice characteristics with self-interpretation, within-practice interpretation, and radiologist interpretation. Methods. This retrospective study used the CMS 5% Research Identifiable File for calendar year 2022 to identify claims for office-based nonbreast diagnostic imaging ordered by nonradiologists for a nationally representative sample of Medicare fee-for-service beneficiaries. Ordering provider and practice characteristics were extracted from the Medicare Data on Provider Practice and Specialty dataset. Self-interpretation, within-practice interpretation, and radiologist interpretation were summarized. Logistic regression models were used to assess associations with self-interpretation and within-practice interpretation for six selected combinations of ordering provider specialty and modality with high imaging volumes. Results. The study included 1,632,212 imaging claims. The unadjusted self-interpretation, within-practice, and radiologist interpretation rates overall were 43.6%, 58.5%, and 36.4%, respectively. The unadjusted self-interpretation rate was 50.4% for radiography and fluoroscopy [XR], 52.0% for ultrasound, 5.3% for CT, 6.1% for MRI, and 39.5% for nuclear medicine. The adjusted odds of self-interpretation were lower for providers with versus without a within-practice radiologist for five of six selected specialty-modality combinations (lowest odds for nonphysician practitioner-ordered XR; OR=0.44), and lower for providers in the largest-sized versus the smallest-sized practices for all selected combinations (lowest odds for orthopedic/sports-ordered MRI; OR=0.01). The adjusted odds of within-practice interpretation were higher for providers in practices with versus without a within-practice radiologist for five of six selected combinations (highest odds for orthopedic/sports-ordered MRI: OR=8.15). Across practice sizes, the unadjusted rate of radiologist interpretation for XR ranged from 15.5-36.2% and 24.9-61.1% for ordering providers in practices with versus without a within-practice radiologist, respectively. Conclusion. Self-interpretation by nonradiologist ordering providers occurred for 43.6% of office-based imaging studies and, in general, was less likely for larger practices or practices with a within-practice radiologist. Clinical Impact. Ongoing healthcare consolidation may partially address the current high rate of imaging self-interpretation by potentially allowing access to a within-practice radiologist.