Chronic Obstructive Pulmonary Disease is a disease of multiple heterogeneous conditions changing over time: a Bayesian probability model-based meta-analysis of the ECLIPSE data

Since 2011, patients with chronic obstructive pulmonary disease (COPD) have been classified into four categories (A to D). Investigators of the Evaluation of COPD Longitudinally to Identify Predictive Surrogate End-points (ECLIPSE) study reported stability and changeover time in the distribution of A-B-C-D categories of ECLIPSE cohorts. The rule of change, however, in the distribution seems different among different cohorts, and remains to be clearly defined. We analyzed the ECLIPSE data by Bayes probability model and obtained a mathematical expression of a set of matrices for the rule of change to define the rule of change in the distribution of A-B-C-D categories. Applying the matrices to each category group in the ECLIPSE revealed that there are common distributions of categories among the category groups. We extended this result to each COPD patient as the distribution of A-B-C-D categories of changing conditions over time. Our proposal is optimizing the care and treatment for each COPD patient by targeting the current condition classified with the A-B-C-D categories..