Completeness of hospital cancer case reporting from the SEER program of the national cancer institute

BACKGROUND: To ascertain the quality of data entering a population-based reporting system, an essential requirement is to study levels of completeness of case-ascertainment and reporting. This study represents an effort to quantify completeness of case reporting in the SEER (Surveillance, Epidemiology, and End Results) Program of the National Cancer Institute. METHODS: Hospitals in each of the participating SEER areas were stratified according to their annual hospital cancer caseload for the year 1987. Within each caseload stratum, a random sample of hospitals was selected for inclusion in this study. Files in the medical record, pathology, and radiation oncology departments in each hospital were reviewed for SEER reportable cases. These cases were then matched against SEER case listings to identify unreported cases. RESULTS: The crude estimated completeness of reporting for 1987 in the six participating SEER areas was 97.7% and the registry-caseload standardized rate was 96.8%. Variation was noted by SEER registry, hospital cancer caseload, and casefinding source (hospital department). Three-quarters of unreported cases were of invasive disease and one-fourth were in situ, primarily of the cervix uteri. CONCLUSIONS: There is variation in completeness of casefinding among SEER registries, hospital size, and hospital department source. Additional factors that appear to be related to case ascertainment are cancer site or type and who performs the casefinding function (hospital registry or central registry staff).