Stark Law loophole allows ordering physicians to self-interpret hundreds of thousands of images each year

Officed-based physicians are interpreting hundreds of thousands of medical images they ordered themselves each year, despite lacking the rigorous training of radiologists.  

This raises potential concerns about quality of care, experts charged in a new Neiman Health Policy Institute study published Wednesday. Nonradiologists working in physician offices ordered over 1.6 million imaging exams for beneficiaries covered by traditional Medicare in 2022. Of those, they interpreted about 43.6% themselves (or nearly 712,000 exams), rather than sending them to radiologists, researchers detailed in the American Journal of Roentgenology. 

Study co-author Vijay M. Rao, MD, believes their findings signal the need for potential policy reform in Washington. Specialties such as OB-GYN and emergency medicine receive only targeted training on certain imaging and not the full gamut of modalities. 

“Our results raise potential implications for quality of patient care. Nonradiologist specialties, aside from cardiology, lack the rigorous and comprehensive training in imaging interpretation that occurs during the four years of a radiology residency program,” Rao, senior VP of enterprise radiology at Jefferson Health in Philadelphia, said in a statement April 2. “The large differences between radiologists and nonradiologists in interpretation training could lead to differences in diagnostic accuracy,” she added later.  

For the study, researchers used a 5%, nationally representative sample of Medicare claims, extrapolating the findings for a nationwide figure. They identified claims for all office-based diagnostic imaging ordered by nonradiologists—excluding breast exams—coming up with a total of 1,632,212. Of those, about 58.5% were interpreted by another clinician within the same practice as the referrer, and 36.4% by radiologists. By modality, ordering physicians self-interpreted about 50.4% in radiography and fluoroscopy, 52% for ultrasound, 39.5% of nuclear exams, 6.1% of MRIs, and 5.3% of CT. (These figures were not adjusted to account for various mitigating factors, the authors noted.)

With these adjustments, the odds of a physician self-interpreting images were typically lower if they had a radiologist working within their practice. They also were lower for those in larger practices employing 500-plus physicians (24.2%) compared to others in ones with fewer than 10 (48.9%). Of the self-interpreted exams, orthopedic/sports medicine specialists accounted for the largest share at 75.7%, followed by cardiologists (73%), physicians in various other specialties (38.4%), nonphysician practitioners such as NPs (30.5%), and primary care providers (19.9%). 

There are currently minimal restrictions around imaging interpretations handled by nonradiologists, the authors noted. Any licensed, Medicare-enrolled physician can receive payment for this work, regardless of specialty, if their claim is accompanied by a written report. Concerns about potential conflicts of interest prompted the enactment of the Stark Law decades ago. However, its in-office ancillary exception allows providers to refer patients for certain services such as imaging within their own practice, if the examination is performed on equipment in the same building. 

“To address ongoing concerns regarding self-referral, some health plans have instituted privileging programs that limit the imaging that may be performed by certain specialties, for example limiting orthopedic surgeons to performing skeletal [X-rays] or OB-GYN physicians to performing gynecologic [ultrasounds],” the authors wrote. “Nonetheless, given reported growing rates of imaging interpretation by nonradiologists, potentially associated with self-interpretation, broader federal action may be warranted, including possible modification of the in-office ancillary exception to the Stark Law to more strongly limit nonradiologist interpretation.” 

The study did not ascertain why physicians are interpreting images by themselves. However, it’s possible radiologist workforce shortages are compelling some to take on the task. When providers have access to an in-practice rad, the odds of self-interpretation decreased for all but one of the combinations of specialty and modality, the authors noted. 

“The associations related to the presence of a within-practice radiologist are also relevant to the current trend of increasing healthcare consolidation,” the authors advised. “If such consolidation occurs in a multispecialty fashion that includes radiology, then such centralization may encourage referral of imaging services to the within-practice radiologist in lieu of potential self-interpretation.”