Exploring the Clinical Characteristics of COVID-19 Clusters Identified Using Factor Analysis of Mixed Data-Based Cluster Analysis

• Main Authors: Liang Han, Pan Shen, Jiahui Yan, Yao Huang, Xin Ba, Weiji Lin, Hui Wang, Ying Huang, Kai Qin, Yu Wang, Zhe Chen, Shenghao Tu
• Format: Article
• Language: English
• Published: Frontiers Media S.A. 2021-07-01
• Series: Frontiers in Medicine
• Subjects: COVID-19; cluster analysis; factor analysis of mixed data; symptoms; laboratory findings; support vector machine
• Online Access: https://www.frontiersin.org/articles/10.3389/fmed.2021.644724/full

The COVID-19 outbreak has brought great challenges to healthcare resources around the world. Patients with COVID-19 exhibit a broad spectrum of clinical characteristics. In this study, the Factor Analysis of Mixed Data (FAMD)-based cluster analysis was applied to demographic information, laboratory indicators at the time of admission, and symptoms presented before admission. Three COVID-19 clusters with distinct clinical features were identified by FAMD-based cluster analysis. The FAMD-based cluster analysis results indicated that the symptoms of COVID-19 were roughly consistent with the laboratory findings of COVID-19 patients. Furthermore, symptoms for mild patients were atypical. Different hospital stay durations and survival differences among the three clusters were also found, and the more severe the clinical characteristics were, the worse the prognosis. Our aims were to describe COVID-19 clusters with different clinical characteristics, and a classifier model according to the results of FAMD-based cluster analysis was constructed to help provide better individualized treatments for numerous COVID-19 patients in the future.