Outcome prediction after mild and complicated mild traumatic brain injury: External validation of existing models and identification of new predictors using the TRACK-TBI pilot study

Hester F. Lingsma, John K. Yue, Andrew I.R. Maas, Ewout W. Steyerberg, Geoffrey T. Manley, Shelly R. Cooper, Kristen Dams-O'Connor, Wayne A. Gordon, David K. Menon, Pratik Mukherjee, David O. Okonkwo, Ava M. Puccio, David M. Schnyer, Alex B. Valadka, Mary J. Vassar, Esther L. Yuh

Research output: Contribution to journalArticlepeer-review

142 Scopus citations

Abstract

Although the majority of patients with mild traumatic brain injury (mTBI) recover completely, some still suffer from disabling ailments at 3 or 6 months. We validated existing prognostic models for mTBI and explored predictors of poor outcome after mTBI. We selected patients with mTBI from TRACK-TBI Pilot, an unselected observational cohort of TBI patients from three centers in the United States. We validated two prognostic models for the Glasgow Outcome Scale Extended (GOS-E) at 6 months after injury. One model was based on the CRASH study data and another from Nijmegen, The Netherlands. Possible predictors of 3-and 6-month GOS-E were analyzed with univariate and multi-variable proportional odds regression models. Of the 386 of 485 patients included in the study (median age, 44 years; interquartile range, 27-58), 75% (n=290) presented with a Glasgow Coma Score (GCS) of 15. In this mTBI population, both previously developed models had a poor performance (area under the receiver operating characteristic curve, 0.49-0.56). In multivariable analyses, the strongest predictors of lower 3-and 6-month GOS-E were older age, pre-existing psychiatric conditions, and lower education. Injury caused by assault, extracranial injuries, and lower GCS were also predictive of lower GOS-E. Existing models for mTBI performed unsatisfactorily. Our study shows that, for mTBI, different predictors are relevant as for moderate and severe TBI. These include age, pre-existing psychiatric conditions, and lower education. Development of a valid prediction model for mTBI patients requires further research efforts.

Original languageEnglish (US)
Pages (from-to)83-94
Number of pages12
JournalJournal of neurotrauma
Volume32
Issue number2
DOIs
StatePublished - Jan 15 2015
Externally publishedYes

Keywords

  • GOS-E
  • TBI
  • prognostic models
  • validation

ASJC Scopus subject areas

  • Clinical Neurology

Fingerprint

Dive into the research topics of 'Outcome prediction after mild and complicated mild traumatic brain injury: External validation of existing models and identification of new predictors using the TRACK-TBI pilot study'. Together they form a unique fingerprint.

Cite this