Asociación del tiempo de hospitalización frente a variables sociodemográficas, clínicas y paraclínicas de pacientes pediátricos con infección por virus Epstein Barr mediante modelos de regresión

Recently, some studies are researching the veracity of the American literature, on which the vast majority of medical schools in Latin America are based, with the diagnosis and evolution of diseases in cohorts from different countries. One example is the work of Moreno (2020) which characterizes and differs in certain diagnoses of the disease caused by the Epstein Barr virus, in a pediatric population of a clinic in Bogotá, Colombia, between the years 2015 and 2019. With the previous work, a possible fault in the diagnosis was identified due to these differences with the teaching parameters, which generates inefficiency in the hospitalization times of the patients. Therefore, a comparison of regression models that explain the association of the sociodemographic, clinical and paraclinical variables of the patients with the number of hospitalized days in the studied cohort was carried out. Models were made with a frequentist and Bayesian approach, supported by the selection of variables by Step AIC methods, evaluation of importance by Random Forest, or probability of inclusion for handling overfitting. Variables such as age, presence of myalgia, and thrombocytosis, among others, that explain the hospitalization time of pediatric patients with Epstein Barr virus infection in the studied cohort were identified. After discussing the results obtained, it was concluded that all the variables generated from the different proposed models would be used since, on the one hand, possible shortcomings of some models are complemented with the others and, on the other hand, they will be the basis argued of the following study with a representative sample of the local cohort.

• Autores: Baquero Sánchez, Jorge Arturo
• Tipo de recurso: Trabajo de grado de pregrado
• Fecha de publicación: 2022
• Institución: Universidad El Bosque
• Repositorio: Repositorio U. El Bosque
• Idioma: spa
• OAI Identifier: oai:repositorio.unbosque.edu.co:20.500.12495/8544
• Acceso en línea: http://hdl.handle.net/20.500.12495/8544
• Palabra clave: Aprendizaje de máquina; Virus Epstein Barr; Tiempo de hospitalización; Modelos lineales generalizados; Modelos de regresión bayesianos; 519.5; Epstein Barr Virus; Hospitalization time; Generalized linear model; Machine learning; Bayesian regression model
• Rights: openAccess
• License: Atribución-NoComercial-CompartirIgual 4.0 Internacional