The minimal detectable difference for endurance shuttle walk test performance in people with COPD on completion of a program of high-intensity ground-based walking.

BACKGROUND:In people with moderate-to-severe chronic obstructive pulmonary disease (COPD), the minimal detectable difference (MDD) in endurance shuttle walk test (ESWT) performance following exercise training is unclear. We sought to determine the MDD for ESWT performance following supervised ground-based walking training using anchor- and distribution-based approaches and report whether these values exceeded random variation in test performance. METHODS:Participants with COPD trained for 30-45 min, 2-3 times weekly for 8-10 weeks. The ESWT was performed before and after the training period. Immediately after training, participants rated their change in walking ability using a Global Rating of Change scale. Receiver Operating Characteristic curves were used to derive the value that best separated those who perceived their improvement in walking ability to be at least 'a little' better from 'almost the same, hardly any change'. These values were compared with those calculated using a distribution-based method. Random variation in test performance was defined as the minimal detectable change (MDC), calculated using the standard error of measurement. RESULTS:78 participants (aged 70 ± 8 yr and FEV1 43 ± 15% predicted) completed the ESWT before and after training. The value that best separated those who perceived their walking ability as 'a little' better was 70 s. The 95% confidence intervals around this estimate traversed zero. The distribution-based estimate was 156 s. The MDC was 227 s. CONCLUSIONS:The MDD established using the anchor- and distribution-based approaches differed considerably. Large variation in test performance cautions against using the MDD to interpret changes in an individual. CLINICAL TRIALS REGISTRATION:Australian New Zealand Clinical Trials Registry (ACTRN12609000472279).