Is training speed an accurate predictor of who attains a minimal clinically important difference in the six-minute walk test in people with chronic stroke?

People with chronic stroke have significant impairments in their walking capacity. Minimal clinically important differences (MCIDs) can be used to interpret changes in patient outcomes following interventions. There is significant variability in the response to moderate-to-high walking interventions in people with chronic stroke. One reason for this response variability could be the lack of understanding of the threshold exercise dose needed to achieve an MCID. The purpose of this analysis was to determine the threshold of exercise training speed most predictive of a small (≥20 m) or moderate (≥50 m) clinically important difference in 6MWT in people with chronic stroke. Participants with chronic stroke with a walking speed of 0.3–1.0 m/s were randomized into a 12-week (1) fast-walking training or (2) fast-walking training and step-activity monitoring intervention. This analysis included participants (<i>n</i> = 129; age: 63.1 ± 12.5, 46% female) with complete pre- and post-intervention data. Exercise intensity was quantified as average training speed. Receiver operating characteristic curves analyzed whether training speed is predictive of attaining a clinically important difference in the 6MWT. Training speed had poor, non-significant accuracy of predicting a small (AUC [95% CI] = 0.584 [0.475–0.693], <i>p</i> = 0.131) or moderate (AUC [95% CI] = 0.597 [0.498–0.696], <i>p</i> = 0.056) change in 6MWT. The average walking training speed during this high-intensity walking intervention did not accurately predict which people with chronic stroke would attain a small or moderate clinically meaningful change in 6MWT distance.