A virtual reality environment for evaluation of a daily living skill in brain injury rehabilitation: reliability and validity

Abstract

Zhang L, Abreu BC, Seale GS, Masel B, Christiansen CH, Ottenbacher KJ. A virtual reality environment for evaluation of a daily living skill in brain injury rehabilitation: reliability and validity. Arch Phys Med Rehabil 2003;84:1118–24. Objective: To establish the stability and validity of information collected in a virtual reality environment from persons with traumatic brain injury (TBI). Design: Prospective correlation design to examine 3-week test-retest results for equivalence reliability between computer-simulated and natural environments. Setting: A residential rehabilitation center for brain injury. Participants: Fifty-four consecutive patients with TBI who received comprehensive rehabilitation services and who were at different stages of recovery. Intervention: An immersive virtual kitchen was developed in which a meal preparation task involving multiple steps was performed. The subjects completed meal preparation both in a virtual reality kitchen and an actual kitchen twice over a 3-week period. Main Outcome Measures: Time and errors on task completion using virtual reality assessment, actual kitchen performance (analogous to the virtual reality environment), occupational therapy (OT) evaluation, and neuropsychologic tests. Results: The stability of performance using the simulated virtual environment was estimated with intraclass correlation coefficients (ICCs). The ICC value for total performance, based on all steps involved in the meal preparation task, was .76 ( P<.01). The construct validity of the simulated environment was examined by correlating performance in the virtual environment with that in the actual kitchen ( r=.63, P<.01), the OT evaluation ( r=.30, P=.05 for meal preparation; r=.40, P=.01 for cognitive subskills), and neuropsychologic tests ( r=.56, P<.01 for the full-scale intelligence quotient [IQ]; r=.40, P<.01 for the verbal IQ; r=.56, P<.01 for the performance IQ). Finally, a multiple regression analysis revealed that the virtual reality environment test was a good predictor for the actual assessment kitchen (β=.35, P=.01). Conclusion: The virtual reality system showed adequate reliability and validity as a method of assessment in persons with brain injury.