Autopsy rates have dramatically fallen worldwide over the past few decades. The reasons for this decrease are many, including the suppression in 1970 of a minimum autopsy rate needed for accreditation of U.S. hospitals, the costs and the risks of the procedure for busy pathology departments, potential legal repercussions and fear of lawsuits should misdiagnoses be discovered, doubt on its real value in the era of modern diagnostic techniques, and the legal obligation in many countries to seek family permission before the examination. However, major discrepancies between premortem and autopsy diagnoses, an important element of quality-of-care evaluation, are still reported in up to 30% of the cases in recent intensive care unit series. Autopsies also remain useful for the evaluation of novel diagnostic and therapeutic procedures, which are frequently used in the intensive care unit setting, for the understanding of the pathophysiology and epidemiology of complex diseases and for the surveillance of emerging diseases. Additionally, using autopsy results in teaching situations is still of utmost educational value. For autopsies to remain a valid monitor of diagnostic performance and an efficient epidemiologic tool, efforts should be made to increase rates of the procedure by improving the consent process and providing physicians and pathologists with more information regarding the value of the examination. Finally, defining a representative group of hospitals funded to perform a large number of unselected autopsies may be sufficient to derive accurate data on major diagnostic errors in the era of cost-containment policy.