Classification of Traumatic Brain Injury Severity: A Neuropsychological Approach

Abstract

Classification of traumatic brain injury (TBI) severity is of great interest because it may assist in guiding treatment as well as predicting course of recovery and outcome. However, classification of TBI severity has presented a number of unique challenges owing largely to the heterogeneity in neuropathology that can result from the injury and associated heterogeneity in clinical, cognitive, and behavioral disturbances. Clinicians and researchers have proposed classification systems based on lesion location and causative processes, physical mechanisms that cause TBI, distinctions between primary vs. secondary injuries, and clinical signs such as coma duration or post-traumatic amnesia, among others. Neuropsychological approaches to classifying brain injury severity have also been proposed, some that are based on a general index of impairment derived across multiple neuropsychological test scores. Cluster analysis of neuropsychology test scores has been accomplished and identified subgroups of individuals with TBI who are to some extent distinguished by severity of neuropsychological impairment. This chapter reviews issues relevant to TBI severity classification and cluster analysis of neuropsychological test scores and then goes on to present results of a study of children with TBI wherein the Trail Making Test (TMT) is subjected to cluster analysis in order to derive brain injury severity groups. These groups are compared on a number of relevant outcome variables, and the TMT classifications are compared to injury severity classifications based on the Glasgow Coma Scale (GCS). Results indicate that three TMT clusters best fit the data, with clusters corresponding to mild, moderate, and severe impairment. Differences between the TMT clusters were present on a number of relevant outcome variables. There was little correspondence between CGS and TMT severity classifications. These results are discussed in the context of current classification methods, limitations of the current findings are discussed, and directions for future research are suggested.