Accuracy and epidemiological implications of the death certificate diagnosis of pulmonary embolism

The death certificate (DC) diagnosis of pulmonary embolism (PE) has been compared to autopsy findings in 2398 patients dying between 1966 and 1976. Accuracy of the death certificate and the effect of diagnostic misclassification errors on the association between PE and potential risk factors were analyzed in this study. Sensitivity of the DC diagnosis of PE was less than 40% and the predictive value of a positive diagnosis was less than 50%. The epidemiological associations derived using the DC diagnosis of PE were very unreliable for myocardial infarction and malignant neoplasm because errors of diagnostic misclassification were not randomly distributed. Associations with risk factors derived using massive embolism were conservative estimates of the associations derived by using any embolism found at autopsy. We conclude that the death certificate diagnosis of PE is extremely inaccurate and should be used with great caution to study the epidemiology of PE.