Completeness of reporting and accuracy of the diagnosis of ovarian cancer from one health region in Norway to the Cancer Registry were examined. Data kept by the Cancer Registry were evaluated against discharge diagnosis data from all 8 hospitals in the health region during the period of 1987-1996. The assessment of the accuracy of the diagnosis recorded in the Cancer Registry was based on review of all medical records in the hospital setting and on slide review of all histologic diagnoses. The overall completeness of reporting ovarian cancer to the Cancer Registry was 99.6%. The organ specific completeness of registration of histologic verified ovarian cancer within the Cancer Registry was 95.3%; 0.9% was erroneously coded and 3.5% had their diagnosis changed to ovarian cancer at re-evaluation. Of all ovarian cancer cases registered at the Cancer Registry, 91% had a primary histologic diagnosis. Among 591 cases identified with a histologic diagnosis in the Cancer Registry, the accuracy of the diagnosis was estimated at 92%. Coding errors were found in 2% of these cases, while in 6% of the cases it was not possible to reproduce the original diagnosis of ovarian cancer at re-evaluation. In order to provide data of high quality for cancer surveillance a cancer registry needs several data providers, such as histopathologic laboratory reports and clinical reports. In addition, assessment of reported data through stringent quality assurance procedures within the registry are necessary for reaching a nearly 100% completeness of registration as found for ovarian cancer in the Cancer Registry of Norway.
Copyright 2002 Wiley-Liss, Inc.