Derivación y validación de un nuevo puntaje predictivo de mortalidad por covid-19 en la altura

Objective: To develop a mortality predictive score for covid-19 patients. Materials and Methods: A twophase retrospective, analytical, observational, and cross-sectional study was conducted. Six hundred twenty medical records were reviewed, with 320 patients in the derivation cohort and 300 patients in the validation cohort. Analysis of variance, Pearson’s chi-square test, and multivariate analysis with binary regression were used to determine sensitivity, specificity, and negative and positive predictive values. Receiver operating characteristic curves were used to compare the scores with the news and hews scores. Results: The two scores included age, lymphocyte count, SatO₂/FiO₂, leukocytes, platelets, absence of symptoms, arterial hypertension, dild, and dhl. The dhl score had an area under the curve (auc) of 0.838, with 100 % mortality for 7.75 points or above, whereas the first score had an auc of 0.826. The auc for the first score in the validation cohort was 0.831, whereas the auc for the dhl score was 0.855. The auc for the hews score was 0.451, whereas the auc for the news score was 0.396. Conclusions: Two clinical measures with high discrimination power, superior to the British hews and news scores, were developed to predict mortality in covid-19 patients.