Geographic variation among medicare beneficiaries in the use of colorectal carcinoma screening procedures

OBJECTIVES: To study geographic variation in the use of recommended screening procedures for colorectal carcinoma. METHODS: All Medicare claims for fecal occult blood testing (FOBT), flexible sigmoidoscopy, colonoscopy, barium enema and two reference procedures, upper endoscopy, and upper gastrointestinal series in patients greater than or equal to65 yr were obtained from the 1997-1999 physician-supplier and outpatient files. State-level procedure rates and the change in rates from 1997 to 1999 were calculated. RESULTS: The median annual rates were as follows: FOBT, 14.54%; flexible sigmoidoscopy, 3.03%; colonoscopy, 6.22%; barium enema, 2.21%; upper gastrointestinal endoscopy, 4.88%; and upper gastrointestinal series, 2.78%. Whereas there was at least a 50% difference between the 25th and 75th percentiles for state rates of FOBT and sigmoidoscopy, the variation in other procedures was more modest. State-level rates of colonoscopy and upper gastrointestinal endoscopy were highly correlated (r = 0.79; p < 0.0001), as were the rates of barium enema and upper gastrointestinal series (r = 0.68; p < 0.0001). Universal increases in colonoscopy use (median +25.0%) and decreases in barium enema use (median -20.9%) from 1997 to 1999 were observed among individual states. CONCLUSIONS: There is variation at the state level in the use of some, but not all colorectal procedures, as well as fairly consistent temporal trends. The etiology of the differences is not clear, but the correlation in the rates of selected procedures suggests the presence of local practice patterns.