DMA Clinical Pilates Directional-Bias Assessment: Reliability and Predictive Validity

Randomized, repeated-measures crossover design. To determine the interrater reliability of directional-bias assessment and to investigate its validity for predicting immediate changes in dynamic postural stability and muscle performance following directionally biased exercises. Directional bias in dynamic postural stability deficits may be associated with outcome following intervention. Two researchers independently assessed 33 participants, each with a history of more than 1 unilateral lower-limb injury, for directional bias. Interrater reliability was evaluated with the kappa coefficient and a prevalence-adjusted and bias-adjusted kappa coefficient. Participants were randomly allocated to perform matched-bias (MB) or unmatched-bias (UB) exercises first, in 2 crossover groups. Two outcome measures, time to stabilization and rebound hopping, were assessed before and following each exercise intervention, using a force plate. Crossover trial data were analyzed by t tests for period, interaction, and treatment effects, and repeated-measures analyses of variance were used to investigate differences between baseline, MB, and UB. Interrater reliability of directional-bias assessment was substantial (κ = 0.75; prevalence-adjusted and bias-adjusted κ = 0.76). Following MB exercises, medial/lateral time to stabilization and time on the ground during rebound hopping were significantly shorter (P = .01 and P = .05, respectively) compared with UB exercises. Compared with baseline, pairwise change in anterior/posterior time to stabilization (P = .008) improved following MB, whereas time in the air decreased following UB (P = .036). Directional-bias assessment demonstrates substantial reliability, and outcomes suggest validity for predicting immediate improvements following matched directionally biased exercises.