The wheelchair skills test (version 2.4): measurement properties.
To evaluate the measurement properties of the Wheelchair Skills Test (WST), version 2.4. Cohort study. Rehabilitation center. A total of 298 subjects (169 wheelchair users, 129 able-bodied subjects) ranging in age from 17 to 88 years. We videotaped subjects as they attempted the 50 skills of the WST 2.4. The test-retest, intrarater, and interrater reliabilities were determined on a subset of 20 wheelchair users. We assessed construct validity by evaluating whether the WST detected expected changes and concurrent validity by seeing how well total WST scores correlated with criterion measures. The mean time +/- standard deviation taken to administer the WST was 27.0+/-9.3 minutes. There were no serious adverse incidents, and the test was well tolerated. For the test-retest, intrarater and interrater reliabilities, the intraclass correlation coefficients for the total scores were.904,.959, and.968. For individual skills, the percentage concordance ranged from 73% to 100%. Regarding construct validity, there was a slightly negative Pearson correlation between total WST score and age (-.434). Gender was identified as a significant factor on multiple regression analysis (P<.001). Wheelchair users with more than 21 days of experience scored higher than those with less experience (65.0% vs 59.6%; P=.01). Participants with stroke and related disorders had a mean score (55.0%+/-13.9%) that was significantly lower than those in other diagnostic categories (P<.05). Participants using conventional wheelchairs had lower scores than those in lightweight ones (66.4% vs 75.1%; P<.001). Regarding concurrent validity, Spearman rank correlations between total WST scores and the global assessments of the wheelchair users' therapists and admission and discharge FIM instrument scores were.394,.38, and.31. The WST 2.4 is practical and safe, and its measurement properties are very good to excellent. Further study is needed to determine its usefulness in various clinical settings.