Predicting clinical outcomes in the Machine Learning era: The Piacenza score a purely data driven approach for mortality prediction in COVID-19 Pneumonia

Background Several models have been developed to predict mortality in patients with COVID-19 pneumonia, but only few have demonstrated enough discriminatory capacity. Machine-learning(ML) algorithms represent a novel approach for data-driven prediction of clinical outcomes with advantages over statistical modelling. We developed the Piacenza score, a ML-based score, to predict 30-day mortality in patients with COVID-19 pneumonia. Methods 852 patients (mean age 70years, 70%males) were enrolled from February to November 2020. The dataset was randomly splitted into derivation and test. The Piacenza score was obtained through the Naïve Bayes classifier and externally validated on 86 patients. Using a forward-search algorithm the following six features were identified: age; mean corpuscular haemoglobin concentration; PaO2/FiO2 ratio; temperature; previous stroke; gender. In case one or more of the features are not available for a patient, the model can be re-trained using only the provided features. We also compared the Piacenza score with the 4C score and with a Naïve Bayes algorithm with 14 variables chosen a-priori. Results The Piacenza score showed an AUC of 0.78(95% CI 0.74-0.84, Brier-score 0.19) in the internal validation cohort and 0.79(95% CI 0.68-0.89, Brier-score 0.16) in the external validation cohort showing a comparable accuracy respect to the 4C score and to the Naïve Bayes model with a-priori chosen features, which achieved an AUC of 0.78(95% CI 0.73-0.83, Brier-score 0.26) and 0.80(95% CI 0.75-0.86, Brier-score 0.17) respectively. Conclusion A personalized ML-based score with a purely data driven features selection is feasible and effective to predict mortality in patients with COVID-19 pneumonia.