The intra-rater reliability and agreement of compensatory stepping thresholds of healthy subjects.

The purpose of this study was to evaluate the test-retest, intra-rater reliability and agreement of compensatory stepping thresholds. A protocol was developed to establish anteroposterior single-stepping thresholds, anteroposterior multiple-stepping thresholds, and lateral single-stepping thresholds. Healthy, young subjects stood on a microprocessor-controlled treadmill, and responded to three series of progressively challenging surface translations. Subjects were instructed to "try not to step" when establishing single-stepping thresholds or "try to take only one step" when establishing multiple-stepping thresholds. Stepping thresholds were defined as the minimum disturbance magnitude that consistently elicited a single or second compensatory step. Thresholds were expressed as the ankle torque necessary to maintain upright posture. Thresholds studied included anterior single-stepping thresholds (τ = 273.0 ± 82.3 N m), posterior single-stepping, thresholds (τ = 235.5 ± 98.0 N m), anterior multiple-stepping thresholds (τ = 977.0 ± 416.3 N m), posterior multiple-stepping thresholds (τ = 701.9 ± 237.5 N m), stability-side lateral single-stepping thresholds (τ = 225.7 ± 77.7 Nm), and mobility-side lateral single-stepping thresholds (τ = 236.8 ± 85.4 N m). Based on intraclass correlation coefficients (ICC) and Bland-Altman plots, all thresholds demonstrated excellent reliability (ICC(2,1) = 0.87-0.97) and agreement. These results suggest that compensatory stepping thresholds have sufficient repeatability to be used in clinical and research-related assessments of fall-risk. Additional study is needed to determine the intra- and inter-rater reliabilities and validity of thresholds specific to the patient populations of interest.