Cause-of-death query in validation of death certification by expert panel; effects on mortality statistics in...

The correctness of selection, coding and registration of underlying cause-of-death is important for the quality of mortality statistics. One measure to improve quality is the query to the certifier for verification of the underlying cause-of-death. In Finland, 3478 death certificates, 7.1% of total 49074 certifications in 1995, were considered questionable by statisticians. The expert panel at Statistics Finland was able to resolve 2813 (80.9%) of them. However, 665 (19.1%) certificates needed to be further queried from the certifier. Of these, 318 (47.8%) were re-assigned to another ICD-9 category or to the applicable three-digit category within the main category of heart and vascular diseases, resulting in changes from a 17.00-fold increase in rheumatic heart diseases (ICD-9 codes 390-398) to a decrease of about one-half (0.45-fold change) in unspecified neoplasms (codes 235-239). However, a statistically significant impact on national mortality statistics was not observed in any of applied ICD categories. Among all questionable death certificates, most prone to query of the certifier, and with a statistical significance of P<0.05, were those with no cause-of-death specified, those stating underlying cause-of-death as non-specified neoplasms (with a observed/expected ratio, O/E, of 1.69), and heart and vascular diseases (1.45), with its subcategories of ischaemic heart diseases (1.33) and other heart diseases (2.92). Death certificate validation, by expert panel consultations and query to the certifiers, and the importance of estimation of the validity of cause-of-death information on death certificates are strongly pointed out in a continuous strive for correct and reliable mortality statistics.