ObjectiveAs part of the overall quality assurance program for the Department of Pediatrics, we determined whether there were differences in the rates of unexpected autopsy findings between pediatric intensive care unit (ICU), emergency department, and ward patients.DesignProspective, descriptive study.SettingTertiary care children's hospital.PatientsPediatric deaths (n equals 212).InterventionsNone.Measurements and Main ResultsAutopsies were obtained more frequently in emergency department patients (27/29 [93%]) compared with pediatric ICU (88/121 [73%]) and ward (42/62 [68%]) patients (p equals .03). The medical examiner's cases were more frequently from emergency department patients (22/27 [81%]) compared with pediatric ICU (39/88 [44%]) or ward (11/42 [26%]) patients (p less than .001). We found unexpected autopsy findings in 19 (12%) of 157 patients. There were no unexpected findings from the medical examiner's cases. The most common unexpected findings were unidentified infections (n equals 7 [three fungal, three viral, and one nonspecific]) and unrecognized cardiac malformations (n equals 4). Unexpected findings occurred more frequently in pediatric ICU patients (16/88 [18%]) vs. emergency department (2/27 [7%]) or ward (1/42 [2%]) patients (p equals .03). The occurrence rates of major unexpected findings (Class I and II) in pediatric ICU (7/79 [9%]), emergency department (2/27 [7%]), and ward (1/42 [2%]) patients were similar (p equals .4). There were two Goldman's Class I unexpected findings in the pediatric ICU and emergency department patients, and one Goldman's Class I unexpected finding in the ward patients.ConclusionsAutopsies were performed more frequently in emergency department patients. Class I through IV unexpected findings occurred more frequently in pediatric ICU patients compared with emergency department or ward patients. Autopsy examinations are an especially valuable diagnostic tool for pediatric ICU patients and physicians.(Crit Care Med 1996; 24:683-686)