Clinical Decision Support for Managing Children with Mild Head Injuries and Intracranial Injury
BACKGROUND: Pediatric traumatic brain injury TBI (TBI) leads to approximately 600,000 emergency departments each year in the United States, with 90% of new diagnoses considered mild (mTBI). Among children with mTBI, the acute evaluation is primarily focused on identifying children with intracranial injury who may be at increased risk of neurological decline. GAP: While substantial attention has been devoted to developing clinical decision support to guide the need for CT imaging in children with mTBI, far less effort has been dedicated to stratifying the risk of neurological decline among children with intracranial injury on CT. HYPOTHESIS: An evidence-based risk prediction model can reliably predict the risk of neurological among children with mTBI and intracranial injury. Through integration into the electronic health record, this predictive tool can improve the safety and efficiency of managing these patients. METHODS: A multicenter dataset will be used to update and externally validate the a risk prediction model to predict the risk of neurological decline among children with mTBI and intracranial injury. Employing both sociotechnical analysis and usability evaluation techniques, this study will evaluate the implementation context for and develop a prototype of electronic clinical decision support based on this predictive tool. RESULTS: To date, we have updated and externally validated a risk prediction model for risk-stratifying children with mTBI and intracranial injury. This model, the KIIDS-TBI model, showed excellent sensitivity and moderate specificity for predicting neurosurgery, prolonged intubation for TBI, or death from TBI. In addition, we have completed a sociotechnical analysis identifying key influences on the implementation of the KIIDS-TBI model. Through multidisciplinary focus group interviews, we identified the following five primary themes: 1) clinical impact; 2) involvement of stakeholders and users; 3) CDS content; 4) clinical practice integration; and 5) post-implementation evaluation measures. We are currently studying the usability and acceptability of electronic CDS based on the KIIDS-TBI model. IMPACT: The external validation of the KIIDS-TBI model served as a critical step in establishing its clinical utility, and the development and evaluation of a prototype electronic clinical decision support tool will facilitate the clinical integration of this risk model. Together, these efforts will enable a future multicenter randomized trial to definitively demonstrate the improved safety and resource-efficiency of utilizing this evidence-based decision guide.