A tailored machine learning approach for mortality prediction in severe COVID-19 treated with glucocorticoids.

Journal: The international journal of tuberculosis and lung disease : the official journal of the International Union against Tuberculosis and Lung Disease
PMID: 39187998

Abstract

BACKGROUNDThe impact of severe COVID-19 pneumonia on healthcare systems highlighted the need for accurate predictions to improve patient outcomes. Despite the established efficacy of glucocorticoids (GCs), variable patient responses are observed, and the existing clinical scores are limited in predicting non-responders. We propose a machine learning (ML) based approach to predict mortality in COVID-19 pneumonia treated with GCs.METHODSThis is an ML-driven retrospective analysis involving 825 patients. We leveraged XGBoost to select the most appropriate features from the initial 52, including clinical and laboratory data. Six different ML techniques were compared. Shapley additive explanation (SHAP) values were used to describe the influence of each feature on classification. Internal validation was performed.RESULTSNine key predictors of death were identified: increasing C-reactive protein (CRP), decreasing arterial partial pressure of oxygen to fraction of inspired oxygen ratio (PaO₂/FiO₂), age, coronary artery disease, invasive mechanical ventilation, acute renal failure, chronic heart failure, PaO₂/FiO₂ earliest value, and body mass index. Random forest achieved the highest test area under the receiver operating characteristic curve at 0.938 (95% CI 0.903-0.969). SHAP values highlighted age and PaO₂/FiO₂ improvement as the most influential features; the latter showed a higher impact than CRP reduction over time.CONCLUSIONThe proposed ML algorithm effectively predicts the risk of hospital death in COVID-19 pneumonia patients undergoing GCs. This approach can be adapted to datasets measuring similar clinical variables..

Authors

  • F Salton
    Pulmonology Unit, University Hospital of Cattinara, Trieste, Italy;, Department of Medical Surgical and Health Sciences, University of Trieste, Trieste, Italy.
  • M Rispoli
    Department of Mathematics, Informatics and Geosciences, University of Trieste, Trieste, Italy.
  • P Confalonieri
    Pulmonology Unit, University Hospital of Cattinara, Trieste, Italy;, Department of Medical Surgical and Health Sciences, University of Trieste, Trieste, Italy.
  • A De Nes
    Pulmonology Unit, University Hospital of Cattinara, Trieste, Italy;, Department of Medical Surgical and Health Sciences, University of Trieste, Trieste, Italy.
  • E Spagnol
    Department of Medical Surgical and Health Sciences, University of Trieste, Trieste, Italy.
  • A Salotti
    Pulmonology Unit, University Hospital of Cattinara, Trieste, Italy;, Department of Medical Surgical and Health Sciences, University of Trieste, Trieste, Italy.
  • B Ruaro
    Pulmonology Unit, University Hospital of Cattinara, Trieste, Italy;, Department of Medical Surgical and Health Sciences, University of Trieste, Trieste, Italy.
  • S Harari
    Department of Clinical Sciences and Community Health, University of Milan, Milan, Italy;, Division of Internal Medicine, Multimedica Istituto di Ricovero e Cura a Carattere Scientifico, Milan, Italy.
  • A Rocca
    Department of Medical Surgical and Health Sciences, University of Trieste, Trieste, Italy.
  • L Manzoni
    Department of Mathematics, Informatics and Geosciences, University of Trieste, Trieste, Italy.
  • M Confalonieri
    Pulmonology Unit, University Hospital of Cattinara, Trieste, Italy;, Department of Medical Surgical and Health Sciences, University of Trieste, Trieste, Italy.

Keywords

External Resources

Popular Topics
Recent Journals