Exploring COPD Patient Clusters and Associations with Health-Related Quality of Life Using A Machine Learning Approach: A Nationwide Cross-Sectional Study

Abstract

Chronic obstructive pulmonary disease (COPD) is a complex condition marked by considerable interindividual heterogeneity. Comorbidities exacerbate this variability, worsening disease severity and reducing health-related quality of life (HRQoL). Despite the high prevalence of COPD in China, comorbidity clusters remain poorly characterized. This study aimed to identify and validate comorbidity clusters in Chinese patients with COPD using cluster analysis. This cross-sectional, multicenter cohort study used data from the Chinese Enjoying Breathing Program (2020–2023). HRQoL was measured using the EuroQol five dimension (EQ-5D). Dimension reduction was performed via multiple correspondence analysis on 31 variables, including indicators of 27 comorbidities and four socio-demographic characteristics. Unsupervised machine learning algorithms, K-means++, and hierarchical clustering identified distinct clusters. Robustness was assessed using random forest classification. Logistic regression evaluated the association between cluster membership and EQ-5D outcomes. Among 11 145 patients, 59.4% had comorbidities. Four clusters emerged: young male smokers, biomass-exposed females, respiratory comorbidity, and elderly multimorbid. The last two clusters had notably lower HRQoL. Cluster analysis identified four clinically meaningful COPD clusters based on comorbidities and risk profiles, each with distinct HRQoL outcomes. These findings support targeted public health interventions and integrated care strategies for COPD management.