Evaluation of claims, medical records, and self-report for measuring fecal occult blood testing among medicare enrollees in fee for service.

BACKGROUND: There is no agreement on the best data source for measuring colorectal cancer (CRC) screening. Medicare claims have been used to measure CRC testing but the validity of using claims to measure fecal occult blood tests (FOBT) has not been established.
METHODS: We compared ascertainment of FOBT among three data sources: self-reports, Medicare claims, and medical records. Data were collected on FOBT use during the study window (1/1/1998 - 12/31/2002). Our study was conducted with North Carolina Medicare enrollees (N = 561) who had previously responded to a telephone survey on CRC tests. FOBT information was abstracted from respondents' physician office medical records and compared with self-reported FOBT use and Medicare claims for FOBT. Data sources were assessed for accuracy and completeness of FOBT reporting using sensitivity, specificity, positive predictive value, negative predictive value, and agreement.
RESULTS: Reporting of FOBT use in the prior year in medical records and Medicare claims agreed 82% of the time [95% confidence interval (95% CI), 79-85%]. FOBT 1-year use rates from self-report agreed with test use found in medical records 70% of the time (95% CI, 66-74%). The lowest agreement was between self-reported 1-year FOBT use and Medicare claims, which agreed 67% of the time (95% CI, 63-71%).
CONCLUSIONS: No data source could be established as providing complete and valid information about FOBT use among Medicare enrollees, showing the difficulty of ascertaining test use rates for noninvasive, low-cost procedures conducted in multiple settings. Caution should be used when attempting to measure FOBT use with self-report, Medicare claims, or medical records.