How Could the Radiologist Shortage Have Eased?

Abstract

In 2000, a severe shortage of diagnostic radiologists existed in the United States. We seek to explain how the shortage eased greatly by 2003, despite the fact that the total imaging workload usually grows much faster than the number of radiologists in practice, which would be expected to intensify the shortage. We measured the contribution of eight possible explanations, predominantly using simple quantitative analyses. We analyzed published data, data on the volume of imaging from Medicare and from the Medical Expenditure Panel Survey, data on residents and fellows from the American College of Radiology's (ACR) membership department, data on residents from the American Board of Radiology, data from the ACR's 1995 and 2003 Surveys of Radiologists, and data from interviews about nighthawk services. From these data sources, we determined the following. Total imaging and imaging by radiologists continued to grow rapidly--by > 20% from 2000 to 2003 (measured in relative value units), which was somewhat faster than in the years preceding 2000 when the shortage was building. Foreign imagers took on a negligible portion of the workload. No reductions in retirement occurred among radiologists during 2000-2003, a 10-20% decrease in the annual number of residency graduates occurred, and no increase in residents going directly into the workforce rather than taking a fellowship was noted. Radiologists' average annual work hours were relatively constant, increasing by perhaps 2%. Work done per hour--that is, productivity--increased sharply (by approximately 15%) during this period. Increased productivity is the predominant explanation of how the radiologist shortage eased. The contribution of other factors was, in comparison, small or even in the opposite direction.