The U.S. Radiologist Workforce: An Analysis of Temporal and Geographic Variation by Using Large National Datasets

Abstract

Purpose: To determine recent trends related to temporal as well as national and statewide geographic variation in the U.S. radiologist and radiology resident workforce.

Materials and methods: This retrospective HIPAA-compliant study was exempted from the internal review board. Federal Area Health Resources Files and Medicare 5% research identifiable files were used to compute parameters related to the radiologist workforce. Geographic variation and annual temporal trends were analyzed. Pearson and Spearman correlations were assessed.

Results: Nationally, the number of radiology trainees increased 84.2% from a nadir in 1997 (3080 trainees) to 2011 (5674 trainees) and showed high state-to-state variation (range, 0-678 trainees in 2011). However, total radiologists nationally increased 39.2% from 1995 (27 906 radiologists) to 2011 (38 875 radiologists), and radiologists per 100 000 population nationally increased by 7.5% from 1995 (10.62%) to 2011 (11.42%), while showing high state-to-state variation (highest-to-lowest state ratio of 4.3). Radiologists' share of the overall physician workforce declined nationally by 8.8% from 1995 (4.0%) to 2011 (3.7%), with moderate state-to-state variation (highest-to-lowest state ratio of 1.7). Radiology trainee numbers exhibited weak-to-moderate positive state-by-state correlation with radiologists per 100 000 population (r = 0.292-0.532), but moderate-to-strong inverse correlation with the percentage of radiologists in rural practice (r = -0.464 to -0.635).

Conclusion: Although the number of radiology trainees dramatically increased, radiologists per 100 000 population increased only slightly, and radiologists' share of the overall physician workforce declined. State-to-state variations in radiologist and radiology resident workforces are high, which suggests a potential role for geographic redistribution rather than changes in the overall workforce size.