Validation of Self-Reported Colorectal Cancer Screening Behaviors Among Appalachian Residents

Objectives: We determined the validity of self-reported colorectal cancer (CRC) screening data provided by Appalachian Ohio residents and identified correlates of providing accurate data. Design and Sample: We conducted cross-sectional telephone interviews between September 2009 and April 2010. Our study included Appalachian Ohio residents (n = 721) ages 51 -75 years. Measures: We compared self-reported CRC screening data to medical records to determine validity. Multivariable logistic regression was used to identify correlates of providing accurate self-reported screening data. Results: About 68% of participants self-reported having any CRC screening test within recommended guidelines, whereas medical records indicated that only 49% were within guidelines (concordance = 0.76). Concordance was higher for flexible sigmoidoscopy and fecal occult blood test compared with colonoscopy, although sensitivity and positive predictive value were much higher for colonoscopy. Participants overreported CRC screening behaviors for all tests. Participants who had a regular checkup in the last 2 years (OR = 2.78, 95% CI: 1.15 -6.73), or who self-rated their health as good or better (OR = 1.88, 95% CI: 1.12 -3.16) were more likely to provide accurate screening data. Conclusions: Many participants failed to provide accurate CRC screening data, and validity varied greatly across individual CRC screening tests. Future CRC screening studies among Appalachian residents should use medical records, if possible, to determine screening histories.