To determine the extent of agreement on underlying cause of death between death certificates and autopsy reports, we analyzed 272 randomly selected autopsy reports and corresponding death certificates from among all such data on autopsies performed in Connecticut in 1980. In 29 per cent of the deaths, a major disagreement on the underlying cause of death led to reclassification of the death in a different International Classification of Diseases major disease category. In an additional 26 per cent, the death certificate and autopsy report agreed on the major disease category but attributed the death to a different specific disease. Deaths due to neoplasms were most accurately diagnosed, with a sensitivity of 87 per cent and a positive predictive value of 85 per cent. Deaths resulting from diseases of the respiratory or digestive system were associated with the highest rates of disagreement. Diseases most commonly overdiagnosed were circulatory disorders, ill-defined conditions, and respiratory diseases. Diseases most commonly underdiagnosed as the cause of death on the death certificate were specific traumatic conditions and gastrointestinal disorders. The autopsy remains an important method for ensuring the quality of mortality statistics.

In reviews of malignancy and occupation, cancer of the nose and paranasal sinuses (sinonasal cancer (SNC) is frequently mentioned. In a case-control study of SNC among subjects who died in Connecticut in the period 1935--1975, occupations from death certificates and city directories were compared to job titles taken from previous literature. The results do not support an association of the cancer with nickel (odds ratio = 0.7, 95% confidence limits (CL): 0.4, 1.5) but do support an association with cutting oils (odds ratio = 2.8, 95% CL: 1.4, 5.7) and wood dust (odds ratio = 4.0, 95% CL: 1.5, 10.8). Actual exposure to these agents was not documented, but the results were consistent with other features of the study and with previous literature. In a search for other occupational correlates, SNC was also found to be positively associated with cutters and with construction workers.

No abstract