Principal component analysis in covid-19 clinical variable clustering in santiago de cuba

ABSTRACT

Introduction: Since March 2020, Cuba has been affected by SARS-CoV-2, a highly infectious coronavirus that causes COVID-19. In COVID-19 a set of associated symptoms is presented and its evolution can be influenced by the presence of certain personal pathological antecedents in the host. Objective: To identify through principal components the grouping of clinical variables in cases with COVID-19 in Santiago de Cuba province, Cuba. Methods: We conducted an observational, descriptive and transversal study. The study population consisted of the 49 confirmed cases with COVID-19 in the province of Santiago de Cuba. Ten clinical variables were selected nine related to symptoms and personal pathological history, and one to the state “deceased”. Principal component analysis was applied as a statistical technique. Results: Variables were represented at the level of the first two principal components. The first component was associated to symptoms and the second component to personal pathological antecedents not associated to the respiratory system. This representation revealed that variables leading to an unfavorable evolution of cases were located in the first and fourth quadrants of the plane, being remarkable for those located in the fourth quadrant. The second and third quadrants were indicators of the favorable evolution, being marked in the second quadrant. Conclusions: The principal component analysis groups the clinical variables and corroborates that personal pathological antecedents have an essential role in the unfavorable evolution of patients with COVID-19. © 2021, Editorial Ciencias Medicas. All rights reserved.