Concurrent validity of single and groups of walking assessments following acute spinal cord injury.

Abstract

Retrospective analysis of prospectively collected longitudinal data. Variables of interest are timed and untimed walking assessments (10MWT, 6MWT, TUG, WISCI, SCIM3a, SCIM3b) and lower extremities motor scores (LEMS) from both sides' lower limb motor segments, measured five times within the first year after acute spinal cord injury (SCI). Assessing concurrent validity of single and groups of walking assessments in comparison with LEMS in SCI patients. European Multicenter study about Spinal Cord Injury, a collaboration of 22 centers. Canonical correlation analysis (CCA) was applied to single and groups of assessments at each time point, separately for patients able to perform timed walking assessments (less impaired; patient subgroup I) and for all patients (no selection; patient subgroup II). In patient subgroup I, SCIM3b, WISCI, 10MWT and 6MWT all had high and similar concurrent validity one year after injury. Among all groups of three walking assessments, SCIM3a, WISCI and 10MWT had highest concurrent validity, similar to all six walking assessments together. Timed walking assessments generally had higher concurrent validity than untimed ones. In patient subgroup II, WISCI distinctly had highest concurrent validity one year after injury, similar to all three untimed walking assessments together. CCA can assess concurrent validity of single and groups of assessments. Minimal sets of walking assessments with comparable concurrent validity as all assessments together were proposed. As these sets differ by patient group, walking assessments should be specified according to expected walking ability to allow for targeted, cost-effective application of assessments.