• Mehdi Moazzez Lesko, Sarah J O'Brien, Charmaine Childs, Omar Bouamra, Timothy Rainey, and Fiona Lecky.
• The Trauma Audit and Research Network (TARN), Salford Royal Hospital , Salford , UK .
• Brain Inj. 2014 Jan 1; 28 (7): 987-94.

Primary ObjectiveTo identify which tool (a model, a biomarker or a combination of these) has better prognostic strength in traumatic brain injury (TBI).Design And MethodsData of 100 patients were analysed. TBI prognostic model B, constructed in Trauma Audit and Research Network (TARN), was run on the dataset and then S100B was added to this model. Another model was developed containing only S100B and, subsequently, other important predictors were added to assess the enhancement of the predictive power. The outcome measures were survival and favourable outcome.ResultsNo difference between performance of the prognostic model or S100B in isolation is observed. Addition of S100B to the prognostic model improves the performance (e.g. AUC, R(2) Nagelkerke and classification accuracy of TARN model B to predict survival increase respectively from 0.66, 0.11 and 70% to 0.77, 0.25 and 75%). Similarly, the predictive power of S100B increases by adding other predictors (e.g. AUC (0.69 vs. 0.79), R(2) Nagelkerke (0.15 vs. 0.30) and classification accuracy (73% vs. 77%) for survival prediction).ConclusionA better prognostic tool than those currently available may be a combination of clinical predictors with a biomarker.

25 keywords...

hide…