[Survey of clinical diagnoses and autopsy findings. Autopsy findings and sensitivity, specificity and predictive values for clinical diagnosis during the periods 1 July 1980-30 June 1081 and 1 July 1990-30 June 1993].

The sensitivity, specificity and clinical accuracy of clinical diagnoses were determined and compared for two periods of time: 1.7.1980-30.6.1981 and 1.7.1990-30.6.1993 based on the analysis of 286 and 138 autopsies respectively. The autopsy rate decreased from 82.7% in the first period to 11.2% in the second. The first period shows a generally higher sensitivity and accuracy for positive diagnosis. Both periods reveal the lowest sensitivity for pulmonary embolism and the lowest accuracy for positive clinical diagnosis of pneumonia/bronchopneumonia. For malignancies and arteriosclerotic heart diseases significant discrepancy between the periods was demonstrated using the chi 2-test. The results are influenced by low autopsy rates causing fewer true-positive diagnoses and a declining sensitivity. This type of study is a useful tool for demonstrating changes in the diagnostic procedure. The present investigation demonstrates a need for further analysis of malignancies to explain the simultaneous decrease in sensitivity, specificity and accuracy in spite of an increasing number of malignancies in autopsy findings.