A Bayesian Model to Predict COVID-19 Severity in Children

Abstract

Background: We aimed to identify risk factors causing critical disease in hospitalized children with COVID-19 and to build a predictive model to anticipate the probability of need for critical care. Methods: We conducted a multicenter, prospective study of children with SARS-CoV-2 infection in 52 Spanish hospitals. The primary outcome was the need for critical care. We used a multivariable Bayesian model to estimate the probability of needing critical care. Results: The study enrolled 350 children from March 12, 2020, to July 1, 2020: 292 (83.4%) and 214 (73.7%) were considered to have relevant COVID-19, of whom 24.2% required critical care. Four major clinical syndromes of decreasing severity were identified: multi-inflammatory syndrome (MIS-C) (17.3%), bronchopulmonary (51.4%), gastrointestinal (11.6%), and mild syndrome (19.6%). Main risk factors were high C-reactive protein and creatinine concentration, lymphopenia, low platelets, anemia, tachycardia, age, neutrophilia, leukocytosis, and low oxygen saturation. These risk factors increased the risk of critical disease depending on the syndrome: the more severe the syndrome, the more risk the factors conferred. Based on our findings, we developed an online risk prediction tool (https://rserver.h12o.es/pediatria/EPICOAPP/, username: user, password: 0000). Conclusions: Risk factors for severe COVID-19 include inflammation, cytopenia, age, comorbidities, and organ dysfunction. The more severe the syndrome, the more the risk factor increases the risk of critical illness. Risk of severe disease can be predicted with a Bayesian model.