Use of Pulmonary Rehabilitation after Chronic Obstructive Pulmonary Disease Hospitalization: An Analysis of Statewide Patient and Hospital Data

<b>Rationale:</b> Pulmonary rehabilitation (PR) is a clinically effective and cost-effective outpatient treatment for chronic obstructive pulmonary disease (COPD) that remains highly underused. Existing analyses of PR use patterns have been focused largely on patient characteristics, but hospital-level analysis is lacking and is needed to inform interventions aimed at improving use after COPD hospitalization. <b>Objectives:</b> To evaluate PR use across hospitals after COPD hospitalization in the state of Michigan, with the goal of characterizing hospital-level variation and identifying the characteristics of high-performing hospitals. <b>Methods:</b> This is a retrospective study of patients with COPD hospitalizations between January 1, 2018, and December 31, 2021, using claims data from the Michigan Value Collaborative and hospital data from the American Hospital Association annual survey. Our primary outcome was the initiation of PR within 30 days of discharge. Chi-square tests and analysis of variance were used to test for differences in patient and hospital covariates. Multilevel logistic regression was used to analyze associations between patient covariates and the primary outcome and to characterize hospital-level variation. <b>Results:</b> A total of 36,389 patients and 99 hospitals were included in the analysis. The majority of patients were older than 65 years of age, female, White, and Medicare fee-for-service insured. The rate of PR initiation within 30 days after hospitalization was 0.8%. Adjusted rates of PR initiation by hospital ranged from 0.4% to 2.0%. Compared with the set reference groups, being female, in the fifth Distressed Community Index quintile, and older than 85 years of age independently decreased the odds of initiating PR. Some variation in initiation rate was attributed to the hospital level (7%; intraclass correlation coefficient = 0.07 [95% confidence interval, 0.03-0.15]). The median odds ratio was 1.6 for PR initiation by hospital. <b>Conclusions:</b> Rates of PR initiation after COPD hospitalization are universally low across all hospitals, though there is some variation. Interventions targeted at patients alone are not sufficient to improve use. Hospital-based strategies to improve PR use after discharge, adapted from those being successfully used with cardiac rehabilitation, should be further explored.