Low versus high-resource Pulmonary Rehabilitation settings in COPD: a retrospective, propensity score-matched, non-inferiority study

Abstract

Background: Pulmonary rehabilitation (PR) is crucial for the management of people with chronic obstructive pulmonary disease (COPD) with well-established physical, psychological and social benefts. Access to PR is, however, limited. Implementing PR with minimal resources is being considered to increase its availability. Nonetheless, the effectiveness of this approach is unclear. This study aimed to assess whether PR for COPD in low-resource was non-inferior to high-resource settings. Methods: A retrospective study with people with COPD who participated in PR programs was conducted. Programs delivered at the Respiratory Research and Rehabilitation Laboratory, School of Health Sciences of the University of Aveiro (Lab3R-ESSUA) or at hospital outpatient departments were classifed as high-resource settings. PR programs conducted at city council facilities or primary healthcare centres were deemed low-resource settings. Outcomes of interest were change in functional exercise capacity assessed with the 6-minute walk test (6MWT), disease impact with the COPD assessment test (CAT) and health-related quality of life with the St. George Respiratory Questionnaire (SGRQ). The minimal clinically important differences for 6MWT (≥ 25m), CAT (≤ -2units) and SGRQ (≤ -4 points) were defned as the non-inferiority margins. Samples were matched using logistic regression-based propensity-score adjusted for sex, age, body mass index, lung function, smoking status, dyspnoea severity, and baseline values for all outcomes at a 1:1 ratio with nearest neighbour matching and a caliper of 0.2. Covariate balance was assessed using standardized mean differences (SMD). Settings were compared using chi-square test, Fisher’s exact test, independent t-test, or Mann–Whitney U test, as appropriate. A two-sided 95% confdence interval (CI) between-group differences was constructed to evaluate non-inferiority. Statistical analyses were performed in R software. Results: A total of 150 people with COPD completed PR. After matching, 102 people were included in the analysis, 51 in each setting. No signifcant differences were found in baseline characteristics (P>0.05) and all SMDs were less than 0.1, indicating good covariate balance. No signifcant differences were observed between settings in pre-post change for any of the outcomes (P>0.05). Non-inferiority of the low-resource settings compared with the high-resource settings was demonstrated only for the 6MWT [mean difference between low and high-resource settings (95%CI), 6.6 (-19.73; 32.93)], being inconclusive for CAT and SGRQ. Conclusion: PR with minimal resources seems to be non-inferior to high-resource PR in terms of functional exercise capacity in COPD. Other matching methods are being explored to minimise sample size loss and strengthen our results.