Using simulation to identify sources of medical diagnostic error in child physical abuse.

Little is known regarding sources of diagnostic error at the provider level in cases of possible child physical abuse. This study examines medical diagnosis as part of medical management and not as part of legal investigation. Simulation offers the opportunity to evaluate diagnostic accuracy and identify error sources. We aimed to identify sources of medical diagnostic error in cases of possible abuse by assessing diagnostic accuracy, identifying gaps in evaluation, and characterizing information used by medical providers to reach their diagnoses. Eight femur fracture simulation cases, half of which were abuse and half accident, were created. Providers from a tertiary pediatric emergency department participated in a simulation exercise involving 1 of the 8 cases. Performance was evaluated using structured scoring tools and debriefing, and qualitative analysis characterized participants' rationales for their diagnoses. Overall, 39% of the 43 participants made an incorrect diagnosis regarding abuse. An incorrect diagnosis was over 8 times more likely to occur in accident than in abuse cases (OR=8.8; 95% CI 2 to 39). Only 58% of participants correctly identified the fracture morphology, 60% correctly identified the mechanics necessary to generate the morphology, and 30% of ordered all appropriate tests for occult injury. In misdiagnoses, participants frequently falsely believed the injury did not match the proposed mechanism and the history provided by the caregiver had changed. Education programs targeting the identified error sources may result in fewer diagnostic errors and improve outcomes. The findings also support the need for referral to child abuse experts in many cases.