Article Text

Download PDFPDF
Modelling early recovery patterns after paediatric traumatic brain injury
  1. Rob J Forsyth1,
  2. Cynthia F Salorio2,
  3. James R Christensen2
  1. 1Institute of Neuroscience, Sir James Spence Institute, Royal Victoria Infirmary, Newcastle University, Newcastle upon Tyne NE1 4LP, UK
  2. 2Kennedy Krieger Institute, Johns Hopkins School of Medicine, Baltimore, Maryland, USA
  1. Correspondence to Dr Rob J Forsyth, Sir James Spence Institute, Royal Victoria Infirmary, Newcastle University, Newcastle upon Tyne NE1 4LP, UK; r.j.forsyth{at}


Objective To describe the range of early recovery patterns seen in children admitted for inpatient rehabilitation after traumatic brain injury and to build simple predictive models of expected recovery.

Patients 103 consecutive paediatric admissions to a neurological rehabilitation facility after closed head injury.

Methods Children's recoveries were defined by repeated scores on the WeeFIM (a validated paediatric measure of functional independence) assembled into recovery trajectories. Non-linear mixed effects modelling was used to define ‘typical’ recoveries and to identify useful simple predictor variables.

Results WeeFIM recovery curves showed a characteristic sigmoidal form with an initial slow phase followed by a mid-phase of fastest improvement and a late plateau. Final WeeFIM scores ranged from 18 to 125 (median 105, IQR 87–117). The time taken to reach 50% final WeeFIM score ranged from 5 to 145 days (median 27, IQR 17–46). Both final WeeFIM and time to reach 50% final WeeFIM correlated with time to follow commands (TFC), defined as the post-injury day on which a child was first observed to follow two simple commands in a 24 h period.

Conclusions Simple models predicting outcome trajectory can be built incorporating early rate-ofrecovery indices (such as TFC) as proxies of injury severity. Such models allow informed discussion with families of likely rates of progress and the CI on these estimates. Models of this nature also potentially allow identification of children making better- or worse-thanexpected recoveries.

Statistics from

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.


  • Provenance and peer reviewed Not commissioned; externally peer reviewed.

  • Competing interests None.

  • Ethics approval The Johns Hopkins School of Medicine Institutional Review Board approved the study.