Objectives: To examine the predictive validity, minimal important difference (MID) and responsiveness of the PRAISE tool. Design: Retrospective data analysis from HomeBase trial of home vs centre-based pulmonary rehabilitation. Setting: Tertiary health service. Participants: One hundred and sixty-six participants with COPD (100 men) with mean age 69 (SD 9) years, FEV 1 % predicted 50% (19). Interventions: Eight-week pulmonary rehabilitation program, conducted at the hospital or at home. Main outcome measures: The 15-item PRAISE tool comprising 10 general and five pulmonary rehabilitation-specific self-efficacy questions. Predictive validity was examined by exploring the relationship between baseline PRAISE score and objective change in physical activity following pulmonary rehabilitation using the SenseWear Armband. The MID was evaluated using anchor-based and distribution-based methods. Responsiveness was assessed with effect sizes. Results: A higher baseline PRAISE score (indicating better self-efficacy) was an independent predictor of reduced sedentary time following pulmonary rehabilitation (P = 0.03). A one point increase in PRAISE was associated with a decrease in sedentary time of 4 minutes/day (95% confidence interval −7.8 to −0.4 minutes/day). Anchor-based estimates of the MID were 0.5 to 1.5 points; however sensitivity and specificity were modest (area under the curve <0.70). Change in PRAISE score following pulmonary rehabilitation had an effect size of 0.21. Conclusions: The PRAISE tool has predictive validity and may be useful to identify those with high self-efficacy who are more likely to achieve important health behaviour changes with pulmonary rehabilitation. The small effect size suggests that the PRAISE tool was not responsive to changes following pulmonary rehabilitation. Trial registration number: NCT01423227, clinicaltrials.gov.
- Health behaviour
- Physical activity
- Pulmonary disease, chronic obstructive