The accuracy of primary care patients' self-reports of prostate-specific antigen testing

Background: The accuracy of reports made by male primary care patients of previous screening for prostate cancer with the prostate-specific antigen (PSA) assay is important for clinical practice, surveillance, and research. Methods: As part of a clinical trial that evaluated a decision aid for prostate cancer screening, 133 male primary care patients, aged 45 to 70 years, were contacted 1 year after enrollment and asked whether they had received PSA testing. To corroborate these self-reports, each patient's clinic and hospital medical records were reviewed for evidence of PSA test results. Results: The raw percentage agreement between self-reported PSA testing and the medical records was 74.4% (κ=0.48). When compared to the findings of the medical record reviews, apparent overreporting by some patients was explained by problems with recalling the test within the 1-year time frame (4.5%) and by having been tested outside of the medical center (5.3%). Uncertainty about having been tested was reported by six (4.5%) patients with PSA test results in their medical records. Misreported testing could not be explained for 11.3% of patients. Patients who had not completed high school were 2.7 times more likely to misreport PSA testing (95% confidence interval, 1.61-3.87). Conclusions: The accuracy of primary care patients' self-reports of previous PSA testing is questionable. Errors in reporting are due largely to problems with recalling a test within a specific time frame or to testing outside of the primary care setting. Under-reporting is a problem for patients uncertain of previous testing.