Reliability of gait and dual-task measures in multiple sclerosis

Background: Single-task (ST) and dual-task (DT) assessments are commonly used to evaluate motor and cognitive impairment in people with multiple sclerosis (MS). Although variability can influence repeated DT testing measures, the reliability of several DT variables over time has not been adequately explored. For instance, a third testing session has never been included to observe whether DT has a learning effect. DT cognition rate reliability has not yet been examined and dual-task cost (DTC), a widely used calculation for DT interpretation, has not been proven reliable.

Research question: To evaluate the reliability of ST and DT measures of gait and cognition over three test sessions.

Methods: This was a cross-sectional study involving 18 people with MS and 12 controls. Participants attended three test sessions, each one week apart. ST and DT (serial seven subtraction) gait variables, DTC, coefficient of variability (CV), and cognition rate were extracted and calculated using an instrumented walkway. Reliability was assessed using intraclass correlation coefficients (ICC) or Kendall's coefficient of concordance (KCC; nonparametric test) and minimum detectable change (MDC); between-session learning effect was assessed using repeated measures ANOVA.

Results: ICC/KCC values for ST and DT gait variables ranged from moderate to excellent (0.50-0.99). However, reliability for DT stride width and cognition rate was lower in controls. In general, DTC and CV variables had poor ICCs and high MDC values (49.19-1478.67 %), although some DTC variables had moderate or higher reliability in controls. Cognition rate was reliable in both MS (ICC 0.91) and controls (ICC 0.84). A learning effect between sessions was observed for DT velocity in both groups and for DTC cadence in people with MS.

Significance: ST and DT gait measures as well as DT cognition rate are reliable outcomes for repeated testing, while DTC and CV variables may not be suitable for long-term monitoring.