Development of a Risk Stratification and Priority Scheme for Active Chronic Obstructive Pulmonary Disease Screening among Inpatients in County-level Hospitals postprint

Abstract: Background　As chronic obstructive pulmonary disease (COPD) has been officially incorporated into the National Basic Public Health Service Program, early identification and standardized case finding have become the core of disease management. Although county-level medical institutions serve as the frontline of primary prevention and control, they face the challenge of insufficient screening capacity, and universal undifferentiated screening inevitably leads to inefficient dissipation of resources. Objective　To explore a priority stratification scheme for active screening of COPD among inpatients in county-level hospitals, aiming to improve screening efficiency and conserve screening resources. Methods　A total of 2 276 hospitalized patients from a county-level hospital (Baiyun Hospital of Guangzhou First People's Hospital) affiliated with the Western Medical Group of Baiyun District, Guangzhou, between December 2023 and October 2024 were enrolled as study subjects. Upon admission, patients routinely underwent COPD-SQ screening. Those with a positive COPD-SQ score ( ≥ 16 points) and/or COPD risk factors were referred for pulmonary function testing to further confirm the COPD diagnosis. General patient characteristics, disease-related data, and active screening data were collected through the Hospital Information System (HIS). Eight variables were included for unsupervised clustering analysis: age, COPD-SQ score, smoking history, positive COPD-SQ screening result, presence of chronic respiratory symptoms, inclusion of common comorbidities in the complete diagnosis, completion of pulmonary function testing, and confirmed COPD diagnosis via pulmonary function testing. Clinical data were compared across different patient clusters to identify potential subgroup patterns. Subsequently, supervised classification analysis was performed based on the labels derived from clustering, and the active COPD screening outcomes were compared among groups to evaluate the grouping performance. Results　Among the 2,276 hospitalized patients, 550 (24.17%) had positive COPD-SQ screens, 782 (34.36%) had a smoking history, 725 (31.85%) had chronic respiratory symptoms, 246 (10.81%) underwent pulmonary function testing, and 123 (5.40%) were confirmed with COPD by pulmonary function testing. Based on the unsupervised clustering results and comprehensive consideration of methodological logic, etiological rationale, and grassroots operability, smoking history and presence of chronic respiratory symptoms were selected as the labels for supervised classification. The hospitalized patients were thus divided into a priority screening group (n=311), a recommended screening group (n=885), and a deferred screening group (n=1 081). The priority screening group had the highest proportions of positive COPD-SQ screening, completion of pulmonary function testing, and confirmed COPD diagnosis via pulmonary function testing, followed by the recommended screening group, with the deferred screening group showing the lowest proportions (P<0.017). Conclusion　Priority stratification for active COPD screening among county-level hospital inpatients can be effectively based on smoking history and chronic respiratory disease symptoms, allowing for tailored screening strategies. However, the applicability of this scheme requires further validation in diverse clinical settings.