Este artigo é um Preprint
Preprints são relatos preliminares de pesquisa que não foram certificados pela revisão por pares. Eles não devem ser considerados para orientar a prática clínica ou comportamentos relacionados à saúde e não devem ser publicados na mídia como informação estabelecida.
Preprints publicados online permitem que os autores recebam feedback rápido, e toda a comunidade científica pode avaliar o trabalho independentemente e responder adequadamente. Estes comentários são publicados juntamente com os preprints para qualquer pessoa ler e servir como uma avaliação pós-publicação.
Early prediction of mortality risk among severe COVID-19 patients using machine learning
Preprint
em En
| PREPRINT-MEDRXIV
| ID: ppmedrxiv-20064329
ABSTRACT
BackgroundCoronavirus disease 2019 (COVID-19) caused by severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) infection has been spreading globally. The number of deaths has increased with the increase in the number of infected patients. We aimed to develop a clinical model to predict the outcome of severe COVID-19 patients early. MethodsEpidemiological, clinical, and first laboratory findings after admission of 183 severe COVID-19 patients (115 survivors and 68 nonsurvivors) from the Sino-French New City Branch of Tongji Hospital were used to develop the predictive models. Five machine learning approaches (logistic regression, partial least squares regression, elastic net, random forest, and bagged flexible discriminant analysis) were used to select the features and predict the patients outcomes. The area under the receiver operating characteristic curve (AUROC) was applied to compare the models performance. Sixty-four severe COVID-19 patients from the Optical Valley Branch of Tongji Hospital were used to externally validate the final predictive model. ResultsThe baseline characteristics and laboratory tests were significantly different between the survivors and nonsurvivors. Four variables (age, high-sensitivity C-reactive protein level, lymphocyte count, and d-dimer level) were selected by all five models. Given the similar performance among the models, the logistic regression model was selected as the final predictive model because of its simplicity and interpretability. The AUROCs of the derivation and external validation sets were 0.895 and 0.881, respectively. The sensitivity and specificity were 0.892 and 0.687 for the derivation set and 0.839 and 0.794 for the validation set, respectively, when using a probability of death of 50% as the cutoff. The individual risk score based on the four selected variables and the corresponding probability of death can serve as indexes to assess the mortality risk of COVID-19 patients. The predictive model is freely available at https//phenomics.fudan.edu.cn/risk_scores/. ConclusionsAge, high-sensitivity C-reactive protein level, lymphocyte count, and d-dimer level of COVID-19 patients at admission are informative for the patients outcomes.
cc_by_nc_nd
Texto completo:
1
Coleções:
09-preprints
Base de dados:
PREPRINT-MEDRXIV
Tipo de estudo:
Diagnostic_studies
/
Prognostic_studies
/
Rct
Idioma:
En
Ano de publicação:
2020
Tipo de documento:
Preprint