1917. Diagnostic Errors in Bacterial Osteomyelitis in Children

BACKGROUND: Delayed or erroneous diagnoses of bacterial infections may cause adverse outcomes in patients. Bacterial osteomyelitis has a low incidence, is only infrequently encountered by primary care pediatricians, and has obscure symptoms in children that make an early and accurate diagnosis challenging. The aim of this study was to determine the incidence and causes of diagnostic errors in pediatric patients in whom bacterial osteomyelitis was finally diagnosed. METHODS: Children who received a definitive diagnosis of acute or chronic bacterial osteomyelitis were enrolled at Tokyo Metropolitan Children’s Medical Center between April 2010 and September 2017. The initial diagnoses were retrospectively reviewed by two pediatricians to evaluate the incidence of misdiagnosis and the types of diagnostic error involved, such as system-related and cognitive errors. Each type of error was subcategorized into associated factors including patient, task, team, organizational/ management, and individual factors, work conditions, and flaws in data gathering, information processing, and verification. The cumulative results for each pediatrician were averaged. A kappa statistic was calculated to assess interobserver agreement. RESULTS: The total incidence of misdiagnosis of bacterial osteomyelitis was 36% (27/75), of which 33.3% (13/39) and 38.9% (14/36) were misdiagnoses of acute and chronic osteomyelitis, respectively. The main type of diagnostic error was cognitive (89.3%). The number of subcategorized factors was 2.5 per diagnostic error. In cases of cognitive errors, the associated factors were flaws in data processing, data gathering, and verification, at 20 (30.2%), 15 (22.6%), and 17 (26.4%) cases, respectively. Inter-reviewer agreement was substantial, with kappa = 0.993 for the primary analysis and 0.805 for the subgroup analysis. CONCLUSION: Cognitive errors were the primary cause of the misdiagnosis of bacterial osteomyelitis. Educational approaches focusing on improving data gathering, processing, and verification should improve diagnostic accuracy. DISCLOSURES: All authors: No reported disclosures.