Autopsy: clinicopathological concordance and imaging techniques

We have assessed clinical and post-mortem diagnostic agreement, comparing two one-year periods 10 years apart (1976 and 1986), in order to evaluate the contribution of noninvasive imaging techniques to diagnostic accuracy. Excluding stillbirths, perinatal deaths and forensic cases, a total of 434 hospital autopsies were analysed retrospectively, 190 from 1976 and 244 from 1986. The overall percentage of major disagreement between clinicians and pathologists was 55% for autopsies carried out in 1976 and 34% for those in 1986. In the majority of these, the clinicians located the immediate cause of death outside the organ system identified by the pathologist. For major conditions not classified as the immediate cause of death, the concordance rates were 62% in 1976 and 75% in 1986; this significant improvement between the two periods suggests that the new modalities in widespread use in 1986 may have increased clinical diagnostic accuracy with respect to these conditions. We conclude that (i) the rate of verification at autopsy of the clinically suspected cause of death remains unacceptably low; (ii) the autopsy is indispensable as an instrument of quality control and for generating mortality and morbidity statistics; and (iii) the introduction of new diagnostic modalities seems not to have improved the accuracy of clinical diagnosis of the immediate cause of death but may significantly have improved the detection of other major diseases.