Validity and reproducibility of C-Mill walking-adaptability assessment in polio survivors

Abstract

BackgroundThe C-Mill interactive treadmill allows for a safe walking-adaptability assessment, unveiling reduced walking adaptability in polio survivors compared to healthy individuals, possibly related to their high fall rate. However, evidence on its validity and reproducibility is scarce. Research questionWhat is the validity and reproducibility of C-Mill walking-adaptability assessment in polio survivors? MethodsPolio survivors with a history and/or fear of falling (n = 46) performed two walking-adaptability assessments, 1–2 weeks apart, including target-stepping tests (with 0%, 20% and 30% inter-target variance) and obstacle-avoidance tests (anticipatory and reactive). We examined (1) face validity by determining Group effects (for subgroups stratified for fall frequency, fear of falling and leg muscle weakness) and Condition effects (for difficulty level) on walking-adaptability outcomes, (2) construct validity by correlating walking-adaptability and balance outcomes, and (3) content validity by establishing possible ceiling effects. We determined whether face-validity findings were reproducible over test occasions and calculated Intraclass Correlation Coefficients (ICC) and the 95% Limits of Agreement (LoA) for walking-adaptability outcomes. ResultsWalking-adaptability outcomes differed in to-be-expected directions for subgroups stratified for fall frequency and leg muscle weakness and for difficulty levels, all reproducible over test occasions. Correlations between walking-adaptability and balance outcomes were mainly low (r < 0.587). Ceiling effects were present for anticipatory obstacle-avoidance and balance outcomes, but not for reactive obstacle avoidance. ICCs [95% confidence intervals] were good for the challenging 20% (0.80[0.67–0.88]) and 30% target-stepping conditions (0.74[0.57–0.85]) and for the reactive obstacle-avoidance (0.76[0.59–0.86]) condition, but not for 0% target-stepping and anticipatory obstacle–avoidance (ICC<0.62) conditions. Likewise, the narrowest LoA were observed for the 20% and 30% target-stepping conditions. SignificanceWe proved face, construct and content validity of C-Mill walking-adaptability assessment in polio survivors with a history of falls and/or fear of falling. Adding walking-adaptability assessment, particularly the more challenging tests given their superior reproducibility, to currently used clinical tests could improve fall-risk evaluation in this population.