US Geographical Variation in Rates of Shoulder and Knee Arthroscopy and Association With Orthopedist Density

Importance: Although rates of arthroscopy have substantially increased, recent data question its comparative effectiveness. Objectives: To assess time trends and geographical variations among several US states in arthroscopy rates and to assess the association of orthopedist density with arthroscopy rates. Design, Setting, and Participants: In this cross-sectional study, procedure rates were calculated for knee arthroscopy, shoulder arthroscopy, and arthroscopic rotator cuff repair using data from the State Ambulatory Surgery and Services Databases for 2006 to 2016 (as available) for the states of California, Colorado, Florida, Iowa, Kentucky, Maryland, Maine, Michigan, Minnesota, North Carolina, Nebraska, New Jersey, Nevada, New York, Oregon, Utah, Vermont, and Wisconsin. Data were analyzed from June 2017 to October 2019. Main Outcomes and Measures: Rates of knee arthroscopy, shoulder arthroscopy, and arthroscopic rotator cuff repair. Results: The combined data sets included 4 856 385 records with 2 530 840 female patients (47%); mean (SD) patient age was 49.13 (16.34) years. Rates per 100 000 persons showed large geographical variations for knee arthroscopy (from 63.31 95% CI, 5.92-198.95 to 721.72 95% CI, 633.41-806.20), shoulder arthroscopy (from 53.02 95% CI, 2.80-164.36 to 438.25 95% CI, 399.00-476.78), and arthroscopic rotator cuff repair (from 11.94 95% CI, 1.30-56.98 to 185.35 95% CI, 143.84-226.20) across US states and years. There were significant downward time trends in knee arthroscopy rates in California, Florida, Iowa, Maryland, Michigan, Nebraska, and New Jersey and upward trends for arthroscopic rotator cuff repair in Colorado, Florida, Kentucky, Maine, and North Carolina. Orthopedist density was not associated with knee arthroscopy rates (slope = 3.07; 95% CI, -9.88 to 16.03; P = .54), shoulder arthroscopy rates (slope = 2.74; 95% CI, -6.53 to 12.01; P = .47), or rates of arthroscopic rotator cuff repair (slope = 1.15; 95% CI, -2.77 to 5.05; P = .49). Conclusions and Relevance: There is large geographical variation in arthroscopy rates despite the questionable comparative effectiveness of these procedures. The reasons for increasing rates of rotator cuff surgery should be further examined.