Rasch Properties of the Cognitive and Linguistic Scale and Optimization for Outcome Trajectory Modeling in Pediatric Acquired Brain Injury

Abstract

Objectives: To (1) determine if items on the Cognitive and Linguistic Scale (CALS) follow a Rasch distribution and (2) explore the relationship between Rasch-derived Cognitive Ability Estimates and outcome trajectory parameters using a nonlinear mixed-effects modeling approach.

Design: Retrospective study.

Setting: Pediatric inpatient rehabilitation hospital.

Participants: A total of 252 children (N=252) aged of 2-21 years (median, 11.8 [IQR, 6.4-15.9] years) consecutively admitted to an inpatient rehabilitation brain injury unit (2008-2014) for a first inpatient admission after acquired brain injury.

Interventions: Not applicable.

Main outcome measures: Rasch-derived Cognitive Ability Estimates from the CALS and associated outcome trajectory parameters.

Results: The CALS demonstrates adequate interval-scale properties with removal of scores from the arousal and responsivity items. Rasch-derived Cognitive Ability Estimates were associated with age (β=0.025, P<.001) such that older age was associated with a faster rate of recovery and more complete ultimate recovery. Slower recovery initiation was associated with a less complete overall cognitive recovery (Spearman ρ=-0.31; P<.001).

Conclusions: The Cognitive Ability Estimates derived from the CALS and associated outcome parameters (eg, rate of recovery) may serve as an ideal outcome measure for clinical trials evaluating interventions for acquired brain injury in a pediatric rehabilitation setting.

Keywords: Brain injuries; Cognition; Linguistics; Pediatrics; Psychometrics; Rehabilitation.