Background. Using the medical records of a large, prepaid health plan, our purpose was to assess the accuracy of self-reported data on cancer detection practices and to evaluate any differences in accuracy between Hispanics and non-Hispanic whites. Methods. Data were collected on six procedures: mammography, clinical breast examination, Pap smear, sigmoidoscopy, fecal occult blood tests, and digital rectal examination. We audited the medical records of 815 randomly selected Hispanic and 483 non-Hispanic white members of the Kaiser Permanente Medical Care Program in Northern California. All responded to a telephone survey. Using a standardized form, we recorded any of the six target examinations that had been performed within 5 years of the telephone interview. We then compared the times of the most recent screening tests as reported in the telephone survey with those in the medical record. Results. For the 2 years before the telephone interview, self-reports for screening tests overestimated the actual rates at which these tests were performed as documented in the medical records. Lack of agreement between self-report and medical record audits was substantial for all screening procedures except sigmoidoscopy. Hispanic members had fewer procedures done for most tests, but the differences were not statistically significant. Overestimates in self-reported screening rates did not differ substantially between Hispanics and non-Hispanic whites. Conclusions. Self-reported data on early cancer detection procedures can substantially overestimate their occurrence among both non-Hispanic white and Hispanic populations. These findings confirm earlier findings for Pap smear and affect decisions on cancer prevention procedures that rely on the accuracy of self-reported data. (C) 1995 Academic Press, Inc.
