Comparison of the predictive power of socio-economic variables, severity of injury and age on long-term outcome of traumatic brain injury: sample-specific variables versus factors as predictors.

The primary objective of this study was to measure the predictive power of pre-injury socio-economic status (SES), severity of injury and age variables on the very long-term outcomes of traumatic brain injury (TBI). By applying a within-subjects retroactive follow-up design and a factor analysis, the study also compared the relative power of sample-specific predictors to that of more commonly used variables and conceptually based factors. Seventy-six participants with severe TBI were evaluated at an average of 14 years post-injury with an extensive neuropsychological battery. The results show that pre-injury SES variables predict long-term cognitive, psychiatric, vocational, and social/familial functioning. Measures of severity of injury predict daily functioning, while age at injury fails to predict any of these variables. Sample-specific predictors were more powerful than more commonly used predictors. Implications regarding long-term clinically based and conceptually based prediction, and those regarding comparisons of predictors across samples are further discussed.