Neoplasms of the central nervous system in Norway. I. Quality control of the registration in the Norwegian Cancer Registry.

The Norwegian Cancer registry includes reports on 8,933 cases of primary central nervous system neoplasms diagnosed in the period 1955-1984. Before submitting this data set to epidemiological analysis, errors were searched out and subsequently corrected, and a quality control was performed. First, seven categories of neoplasm records likely to be faulty were defined. In this way 109 cases (1.2% of the total) were identified for extensive study. Minor or major errors were found and corrected in 86 cases. The main source of error was misinterpretation of data by cancer registry staff (67 out of 109 cases). The second approach was to evaluate the quality of the corrected data set by a random draw of 300 cases. Errors concerning total incidence rates and rates for main groups of gliomas, meningiomas and neurilemmomas represented 0.3% of the total, and altogether 9.3% of errors of varying severity were revealed. The series is discussed with regard to prerequisites for serving as reliable data on CNS neoplasm epidemiology. The results indicate that the data from the Norwegian Cancer Registry is sufficiently valid for a thorough study of CNS neoplasms.