RESUMO
Mild traumatic brain injury (mTBI) is highly prevalent in children. Recent literature suggests that children with mTBI are at considerable risk of persisting neurocognitive deficits, threatening post-injury child development. Nevertheless, clinical tools for early identification of children at risk are currently not available. This systematic review aims to describe the available literature on neurocognitive outcome prediction models in children with mTBI. Findings are highly relevant for early identification of children at risk of persistent neurocognitive deficits, allowing targeted treatment of these children to optimize recovery. The electronic literature search was conducted in PubMed, EMBASE, CINAHL, Cochrane, PsychINFO and Web of Science on February 9, 2022. We included all studies with multi-variate models for neurocognitive outcome based on original data from only children (age <18 years) with mTBI. Following Preferred Reporting Items for Systematic Reviews and Meta-analyses (PRISMA) guidelines, two authors independently performed data extraction and risk of bias analysis using the Prediction model Risk of Bias Assessment Tool (PROBAST). This systematic review identified eight original studies (nine articles) reporting prediction models for neurocognitive outcome, representing a total of 1033 children diagnosed with mTBI (mean age at injury = 10.5 years, 37.6% girls). Neurocognitive outcome assessment took place between 1 month and 7 years post-injury. Models were identified with significant predictive value for the following outcomes: memory, working memory, inhibition, processing speed, and general neurocognitive functioning. Prediction performance of these models varied greatly between weak and substantial (R2 = 10.0%-54.7%). The best performing model was based on demographic and pre-morbid risk factors in conjunction with a subacute neurocognitive screening to predict the presence of a deficit in general neurocognitive functioning at 12 months post-injury. This systematic review reflects the absence of robust prediction models for neurocognitive outcome of children with mTBI. The findings indicate that demographic factors, pre-morbid factors as well as acute and subacute clinical factors have relevance for neurocognitive outcome. Based on the available evidence, evaluation of demographic and pre-morbid risk factors in conjunction with a subacute neurocognitive screening may have the best potential to predict neurocognitive outcome in children with mTBI. The findings underline the importance of future research contributing to early identification of children at risk of persisting neurocognitive deficits.
