Preferences for colorectal cancer screening among racially/ethnically diverse primary care patients.

BACKGROUND Incorporating patients' preferences into colorectal cancer (CRC) screening recommendations has been identified as a potential mechanism for increasing adherence. This study used conjoint analysis to describe variation in CRC screening preferences among racially/ethnically diverse primary care patients. METHODS We recruited patients ages 50-80 of a large practice-based research network stratified by white, African American, or Hispanic race/ethnicity to complete a preference assessment instrument. Participants were asked to rate 8 hypothetical CRC screening test scenarios comprised of different combinations of 5 attributes and 6 scenarios designed to depict guideline-recommended CRC screening tests (eg, fecal occult blood test, flexible sigmoidoscopy, colonoscopy, and double contrast barium enema) including new technology (eg, virtual colonoscopy, fecal immunochemical test). Responses were used to calculate the overall importance of test attributes, the relative importance of attribute levels, and to identify factors associated with preferences. RESULTS Two hundred twelve primary care patients were recruited to the study (74 white, 60 African American, 78 Hispanic). Of the guideline-recommended tests, 37% preferred COL, 31% FOBT, 15% BE, and 9% SIG. Ratings of new technology tests were significantly (P < 0.05) higher than ratings of guideline-recommended tests. The order of the importance of attributes was: what the test involved (37%), accuracy (19%), frequency (17%), discomfort (15%), and preparation (13%). Part-worth utilities for 1 attribute showed that collecting a stool sample was most preferable and endoscopy without sedation least preferable. Multivariate regression found that race/ethnicity and specific test attributes were independently associated (P < 0.05) with test preferences. CONCLUSIONS Primary care patients have distinct preferences for CRC screening tests that can be linked to test attributes. Racial/ethnic variations in test preferences persist when controlling for attributes. Tailoring screening recommendations to patients' preferences may increase screening adherence.