Validity, Responsiveness, Long-Term Stability, and Minimal Detectable Change in the Six-Minute Walk Test in...

Background: Reduced physical fitness in individuals with schizophrenia contributes to disability and health disparities, highlighting the need for valid and responsive field measures in exercise interventions. Objective: This study examined the long-term stability, construct validity, and responsiveness of the Six-Minute Walk Test (6MWT) in individuals with schizophrenia undergoing structured exercise interventions. Methods: This secondary analysis included 70 participants from two randomized controlled trials of 12-week Baduanjin or brisk walking interventions. The cohorts were pooled after confirming comparable baseline characteristics and no significant group-by-time interaction in 6MWT change. Baseline and post-intervention assessments included the 6MWT, the Timed Up-and-Go (TUG), the cognitive dual-task TUG (TUGcognitive), and the 30-Second Chair Stand Test (30CST). Long-term stability of 6MWT scores was assessed using the intraclass correlation coefficient (ICC), and measurement error was estimated using the standard error of measurement (SEM) and minimal detectable change at the 95% confidence level (MDC₉₅). Construct validity was examined using Spearman correlations (ρ), and responsiveness was examined using the standardized response mean (SRM) and Cohen's d. Results: The 6MWT showed excellent long-term stability (ICC = 0.96, 95% CI: 0.93-0.97), with SEM = 13.41 m and MDC₉₅ = 37.19 m. Construct validity was supported by correlations with the TUG (ρ = -0.54, p < 0.001), 30CST (ρ = 0.45, p < 0.001), and TUGcognitive (ρ = -0.36, p = 0.002). Responsiveness was large (SRM = 1.08, 95% CI: 0.84-1.31), with a complementary effect size of d = 0.32 (95% CI = 0.24 to 0.42), and 24.29% of participants exceeded the MDC₉₅ threshold. Conclusions: The 6MWT demonstrated appropriate long-term stability, construct validity, and responsiveness as a measure of functional exercise capacity in schizophrenia, supporting its use for evaluating exercise outcomes in psychiatric rehabilitation. Because the present psychometric analysis did not include a non-exercise comparison group, the responsiveness estimates reflect within-intervention change and should not be interpreted as causal treatment effects of exercise. Larger studies are needed to identify responder characteristics and trajectories of changes in test performance.