New research published online Feb. 23 in the Journal of the American College of Radiology suggests that Medicare Physician Fee Schedule (PFS) spending on non-invasive diagnostic imaging services has been declining over the past decade.    

David Levin, MD, Laurence Parker, PhD and Vijay Rao, MD, from the Center for Research on Utilization of Imaging Services at Thomas Jefferson University Hospital in Philadelphia, conducted the study.  

The researchers use Medicare Pat B Physician Supplier Procedure Summary Master Files from 2003 to 2015 for data and total approved payments to physicians for all medical services were calculated for each year, according to study methods. Procedure codes for noninvasive diagnostic imaging were also selected and then determined aggregate approved payments to physicians for the codes.  

"Medicare’s provider specialty codes were used to define payments to four provider categories: radiologists, cardiologists, all other physicians, and independent diagnostic testing facilities together with multispecialty groups," the researchers wrote.  

According to study results, researchers came to the following conclusions:  

• Total Medicare-approved payments for all physician services under the PFS increased progressively from $92.7 billion in 2003 to $132.9 billion in 2015. 
• In 2003, the share of those payments attributable to NDI was 9.5 percent, increasing to a peak of 10.8 percent in 2006, then progressively declining to six percent in 2015. 
• All four provider categories saw the same trend pattern—a peak in 2006 but then decline after. 
• By 2015, the shares of total PFS payments to physicians that were attributable to NDI were radiologists (3.2 percent), cardiologists (1.2 percent), all other physicians (1.2 percent) and independent diagnostic testing facilities or multispecialty groups (0.4 percent).  

Overall, researchers found that the proportion of Medicare PFS spending on physician services regarding non-invasive diagnostic imaging has been on a steady decline over the past decade.