One hundred sixty-two consecutive adult autopsies (87 of subjects over age 65 years and 75 of subjects aged 23 to 64 years) performed at a university hospital were studied retrospectively by six internists to determine (1) if diagnostic errors were quantitatively or qualitatively different between the two age groups; (2) if the underlying causes of error (divided into nine categories) were different or age related in any way between the two groups; and (3) any aspects of care that related age to clinical outcome. We found the frequency of major clinical/autopsy discrepancies to be similar to those in previous studies (35%), but in only 7% of cases were these likely to have affected therapy/outcome. Ther was no difference in the frequency of major discrepancies between age groups. There were significantly more "unexpected" minor discrepancies in the older patients, probably related to the multiplicity and complexity of their problems, but these would have affected therapy/outcome in only 1 (3%) of 37 cases. The most common causes of 136 clinical "errors" in 151 autopsies were, in order of frequency: a diagnostic "blind spot," a conscious decision not to pursue a clinical finding (not a real "error"), failure to account for a symptom or sign, atypical presentations, and inadequate follow-up of abnormal laboratory findings. There were no differences between the geriatric and adult groups in terms of frequency or cause of the errors. We conclude that (1) there is no difference in the diagnostic accuracy regarding cause of death between geriatric and nongeriatric patients in the acute hospital environment, and (2) closer attention to basic knowledge and clinical skills and a special focus on judgment and reasoning skills, utilizing autopsy findings among other things, will lead to even further improvement in clinical care at all ages.