Assessing the accuracy of death records and pre-mortem clinical diagnoses in children diagnosed with brain tumors: A retrospective chart review of children in British Columbia, Canada

The advantages of autopsy have been demonstrated in pediatric oncology; however, it is unknown to what extent the utility of autopsy is in deceased children diagnosed with a pediatric brain tumor (PBT). The purpose of this study was to describe the frequency of autopsy, prevalence of clinical discrepancies, and accuracy of cancer registry death records for deceased children diagnosed with a PBT in British Columbia, Canada. A retrospective chart review was performed of medical records and autopsy reports of pediatric patients diagnosed with a PBT that died between 1982 and 2012 in British Columbia. Clinical discrepancies between pre- and post-mortem findings were classified based on a modified classification system of the Goldman Criteria. The overall autopsy rate was 15.5% (32 of 206) during 1982-2012, with a significant (P = 0.001) decrease of 22.4% observed between decade 1 (32.8%) and decade 2 (10.4%) and a further slight decrease (4.5%) between decade 2 (10.4%) and decade 3 (5.9%) (P = 0.379). A third of patients had discrepancies between pre-mortem and post-mortem clinical diagnoses, with slightly over 10% of these cases revealing information that would have altered the probability of survival had it been known prior to death. More than half (59.3%) of cases had discordant cause of death as recorded in the cancer registry when compared to autopsy findings. Autopsy for children diagnosed with a PBT can provide health care professionals with important information about the accuracy of their diagnoses and evaluate the efficacy of therapy.