Introduction: Minimally important differences (MID?s) for common outcomes of pulmonary rehabilitation are established in COPD. It is not known whether MID?s differ based on COPD disease characteristics.

Aim: To estimate MID?s for clinical outcomes of pulmonary rehabilitation dependent upon baseline characteristics.

Methods: 2791 people with COPD were split into derivation (n=2245; age 66±9 years; 50% males; FEV1 47±20% predicted) and comparator (n=546; age 66±9 years; 47% males; FEV1 46±21% predicted) cohorts. MID?s were estimated using 0.5xSD (symmetrically distributed) or 0.5xIQR (non-symmetrically distributed) for: 6-minute walk test (6MWT), constant work rate test (CWRT), CAT, SGRQ, HADS, and fat-free mass index (FFMI). MID?s were estimated based on baseline outcome scores, lung function, mMRC grade and FFMI.

Results: MID?s were comparable to previous values. MID?s for SGRQ-S (8.7 points), SGRQ-A (7.1 points), SGRQ-I (8.1 points), and FFMI (0.36kg/m2) were produced. There was greater variation of change in 6MWT, SGRQ-A, SGRQ-I, HADS and FFMI on which MID?s were determined when categorising for baseline values (p<0.05). Greater variation of change in 6MWT on which MID?s were determined was evident with COPD disease severity (p<0.05). Changes in 6MWT, CAT, CWRT, SGRQ-A, and FFMI depending on baseline mMRC score resulted in greater variation on which MID?s were determined (p<0.05). Baseline stratification for FFMI resulted in greater variation of change in CWRT (p<0.001) and HADS-D (p=0.043) on which MID?s were determined.

Conclusions: Findings suggest that baseline presentation should be considered for people with COPD when assessing the efficacy of pulmonary rehabilitation.