Burn center verification requires the use of autopsy as one method of quality assurance in a burn center. Because of the decreasing rates of autopsies worldwide and improved diagnostic accuracy in our critical care units, we tested the hypothesis that autopsy diagnosis would not alter our clinical diagnosis. A chart review of all deaths (N = 94) that occurred during a 6-year period (1989-1994) was performed. The clinical diagnoses from the hospital charts and autopsy reports for the patients were reviewed, and diagnostic discrepancies were classified as class I or class II errors. Class I diagnostic errors might have altered the clinical outcome. Class II errors were attributable to the burn injuries but were believed to have had little impact on the clinical outcome. The overall autopsy rate was 93.6% (n = 88). Clinical diagnostic errors were found in 16 (18%) of 88 patients. Five class I errors were found in 4 patients (4.5%), and 15 class II errors were found in 13 patients (14.7%). Although the rate of potentially serious errors was low (only 4.5% of the patients in this study) postmortem examinations revealed clinical diagnostic errors. The results of this study support the continued use of autopsies as a means of quality assurance, despite our ability to closely monitor our critically ill patients with burns.