[Anatomical and clinical correlates statistical analysis].
A retrospective study of the clinico-pathological correlation was undertaken on 1000 consecutive autopsies required by a university department of medicine. For each case the following data were collected on a standardized card: information on the patient, the hospital stay, the therapy; the main clinical and pathological diagnoses, classified according to their importance; judgements on the clinico-pathological correlation related to each single diagnosis as well as to each case as a whole. The data were analyzed by computer. Two samples of the results are presented her. 1. 18 cases of active tuberculosis were observed. In 7, among which 5 were autopsy discoveries, the tuberculosis was rated as the cause of death. 2. General estimate of clinico-pathological correlation. In the cases where the clinical picture was obscure the autopsy was able to solve the problem completely in 2 cases out of 3. Even when the clinival situation was thought to be obvious, the post mortem examination brought to light important errors or omissions in just over 30% of the cases. On ther other hand, in one case out of 10, the autopsy could not solve in part or in whole the clinical problems involved. There results show that routine autopsies are generally still useful today, but fail fairly ofter to meet the clinician's expectation. These failures can probably be in part corrected.