Prediction of Mental Health Complications Following Mild Traumatic Brain Injury.

Abstract

Prognostic models can support prevention of mental health complications after mild traumatic brain injury (mTBI). The present study aimed to identify risk factors and develop prognostic model(s) for mental health complications following mTBI. This secondary analysis of data from a randomized controlled trial included 513 adults presenting to emergency departments/urgent care centers. Candidate predictors were demographic, injury-related and health history information collected during medical chart review and eligibility screening, and scores on questionnaires completed at 2 weeks postinjury. The primary outcome was presence/absence of new or worsened major depressive disorder, anxiety disorders, and post-traumatic stress disorder (PTSD), determined with a structured psychodiagnostic interview (Mini International Neuropsychiatric Interview) at 3 and 6 months after mTBI. Logistic regression assessed the prognostic value of 22 pre-, peri-, and early postinjury factors. Least absolute shrinkage and selection operator (LASSO) was used to select predictors in prognostic model development. Younger age, identifying as a person of color, prior mTBI(s), maladaptive illness perceptions, and greater PTSD, and depression and anxiety symptom severity measured at 2 weeks postinjury were significant predictors of new/worsened mental health complications 3-6 months following mTBI. A comprehensive model (with 9 LASSO-selected predictors) showed strong discriminability for predicting mental health complications (optimism-corrected area under the receiver operating characteristic curve [AUC] = 0.80), outperforming a basic model that included only variables commonly collected as part of usual clinical care (optimism-corrected AUC = 0.71). Certain pre-injury and demographic characteristics are associated with increased risk of mental health complications after mTBI. Assessing for early postinjury illness beliefs and psychological symptoms can further improve prognostic accuracy.