Dead men tell no tales: Analysis of the use of autopsy reports in trauma system performance improvement activities

摘要

PURPOSE: To analyze the influence and use of autopsy report review on preventability judgments as part of trauma system performance improvement activities. Methods: All cases trauma fatalities occurring across one state within 1 year were reviewed. Preventability judgments were first analyzed by multidisciplinary panel consensus without benefit of autopsy report. Deaths were then reanalyzed after the panel was provided with autopsy findings. Changes in panel determinations of preventability and cause of death were noted. Results: A total of 434 cases were reviewed, autopsies were performed in 240 (55%) patients. Autopsy rate was 83% for prehospital deaths (PHDs) and 37% for hospital deaths (HDs). A complete examination (CA) was performed in 166 (69%) cases, and 74 (31%) cases were limited internal or external examinations only (NCA). Of autopsies performed on HD, 60% were CA versus 75% in PHD. Autopsy review changed preventability determination in four cases (1%). All changes were from nonpreventable to possibly preventable. For all patients with autopsy, the panel felt that the autopsy should have been of sufficient quality to analyze the cause of death in 83%. The autopsy was felt to actually establish a specific cause of death in 70% of all patients with autopsy, 71% in patients with NCA, and 74% in patients with CA. The autopsy changed the panel's preautopsy review-determination cause of death in 31% of all patients with autopsy (37% in the CA group; 13% in the NCA group). For PHD, autopsy changed the panel-determination cause of death in 44% and in 13% for HD. Conclusion: Review of autopsy reports adds little to the trauma performance improvement process. It does not significantly change death review panel determinations. It may, perhaps, be most useful in PHD. Ardent initiatives to expend resources on autopsy performance and acquisition of autopsy reports in all patients with trauma is unwarranted.