Clinical diagnostic accuracy audited by autopsy in a university hospital in two eras.

A retrospective analysis of 3042 autopsies was carried out to determine and compare the discrepancy rates between clinical and autopsy diagnoses as well as the sensitivity and specificity of clinical diagnostics in 10 diseases in 1977/78 and 1987/88. The autopsy rate decreased from 80% to 39%, which might explain the increased discrepancy rate from 22% to 27% regarding the diagnoses of major, principal diseases. The accuracy of clinical diagnostics seemed to improve during the period for some of the diseases (pulmonary embolism, peptic ulcer, infectious carditis, peritonitis), while it worsened for others (acute myocardial infarction, thrombosis of the mesenteric artery, ruptured aortic aneurysm, tuberculosis) or remained unchanged (cirrhosis of the liver, malignant tumours regarded as a whole group). The findings underline the importance of autopsies and their cumulative studies in assessing the accuracy and providing data for the determination of necessary fallibility of clinical diagnostics.