Abstract
Child physical abuse is a leading cause of morbidity and mortality in young children. Identification of abused children is challenging, and can affect risk-adjusted benchmarking of trauma center performance. The purpose of this project was to understand diagnosis coding capture rates for child abuse and develop a standardized approach to clinician documentation to improve trauma registry capture. A retrospective cohort was obtained including all admitted trauma patients with injuries from known or suspected abusive mechanism in 2017. Patients who received forensic workup for child physical abuse were classified as "no abuse," "suspected abuse," and "confirmed abuse" using narratives from social work notes. Our trauma registry was used to abstract International Classification of Diseases, Tenth Revision (ICD-10) diagnostic and external cause codes for each patient. Abuse classifications defined by chart review were then compared with coding in the registry using crosstabs. A total of 115 patients were identified as having a forensic workup for child physical abuse. Patients who underwent forensic workup were classified as: 40% no abuse, 37% suspected abuse, and 23% confirmed abuse at the time of discharge. Three patients (6%) with a negative forensic workup were overcoded as suspected abuse in our trauma registry. Among patients with clinically confirmed abuse, our trauma registry identified only 63% by diagnostic codes and only 33% by external cause codes. Child physical abuse is frequently undercoded, and clear clinical documentation of the level of suspicion of abuse at discharge is needed to accurately identify abused patients.