Exploration of specific domains within the COPD assessment test (CAT)

<bold>Objectives:</bold> CAT was developed as a measure of COPD health status impairment for clinical practice and research. It was designed to have a total score from the sum of 8 items, but subsequently investigators have explored the use of individual items. This study analysed the factor structure of CAT and tested the potential to reduce items and create domains. <bold>Methods:</bold> Data are from the baseline of COMPASS (Liang, et al. <italic>ERJ Open Res</italic> 2021;7:00201). Exploratory factor analysis (EFA) was used to identify factors with Eigenvalues >1.0. <bold>Results:</bold> CAT data from 1696 COPD patients were available, 89.6% male, mean age 65.4 years, mean post-bronchodilator FEV₁ 66.6% predicted. Using all 8 CAT items, 3 Factors were identified: Activity (items 3, 4, 5, 6), Cough & Sputum (items 1, 2) and Sleep & Energy (items 7, 8); items 3 and 8 did not load strongly and were dropped. Item 7 was then dropped as a single item domain, as it correlated poorly with other measures (<bold>Table</bold>). The % of total variance in the model explained by the 8 items was 56.3%, 6 items 53.9%, and 5 items 61.3%. Correlations between different CAT items with SGRQ, mMRC and FEV₁ are tabulated (<bold>Table</bold>). Notably, CAT 5-item correlated as strongly with SGRQ total score as CAT 8-item. <bold>Conclusion:</bold> These findings suggest CAT may be reduced to 5 items with 2 domain scores (Cough & Sputum and Activity). These scores show convergent validity with scores captured by other patient-centred outcomes. <bold>Funding</bold> GSK (208630) <fig><object-id>erj;64/suppl_68/PA2162/F1</object-id><object-id>F1</object-id><object-id>F1</object-id><graphic></graphic></fig>