Validation of self-reported colorectal cancer (CRC) screening in a study of ethnically diverse first-degree relatives of CRC cases

Abstract

Background: Evidence about the accuracy of self-reports of colorectal cancer (CRC) screening is lacking. We conducted a validation protocol in a randomized trial to increase CRC screening among high-risk individuals.

Methods: First-degree relatives (n = 1,280) of CRC cases who were due for CRC screening were included in the parent trial. All subjects who completed the follow-up interview (n = 948) were asked to participate in validation activities. Self-reports of receipt of CRC screening during the 12-month study period were verified via physicians.

Results: Although 60% (n = 567) verbally agreed, only 171 subjects (18% of original sample) returned the signed validation form with the physician name and contact information and a medical information release statement. The signed forms were mailed to physicians with a $10 incentive and the request to list the dates of recent CRC screening tests. One hundred twenty-three physicians (72% of physicians contacted, 13% of original sample) returned completed validation forms. Rates of agreement were low across all three screening types with physicians verifying self-reported screening for 29% of fecal occult blood testing, 56% of sigmoidoscopy, 55% of colonoscopy, and 57% of any screening test.

Conclusion: Validation of self-report using the type of protocol we used for subjects receiving medical care in many community settings may be unfeasible and cost inefficient. Given the overall low participation rate in validation activities and considerable challenges in collecting high quality data, conclusions about the accuracy of self-reported CRC screening are difficult to make based on the results of this study.