ABSTRACT
The majority of early prediction scores and methods to predict COVID-19 mortality are bound by methodological flaws and technological limitations (e.g., the use of a single prediction model). Our aim is to provide a thorough comparative study that tackles those methodological issues, considering multiple techniques to build mortality prediction models, including modern machine learning (neural) algorithms and traditional statistical techniques, as well as meta-learning (ensemble) approaches. This study used a dataset from a multicenter cohort of 10,897 adult Brazilian COVID-19 patients, admitted from March/2020 to November/2021, including patients [median age 60 (interquartile range 48-71), 46% women]. We also proposed new original population-based meta-features that have not been devised in the literature. Stacking has shown to achieve the best results reported in the literature for the death prediction task, improving over previous state-of-the-art by more than 46% in Recall for predicting death, with AUROC 0.826 and MacroF1 of 65.4%. The newly proposed meta-features were highly discriminative of death, but fell short in producing large improvements in final prediction performance, demonstrating that we are possibly on the limits of the prediction capabilities that can be achieved with the current set of ML techniques and (meta-)features. Finally, we investigated how the trained models perform on different hospitals, showing that there are indeed large differences in classifier performance between different hospitals, further making the case that errors are produced by factors that cannot be modeled with the current predictors.
Subject(s)
COVID-19 , Adult , Humans , Female , Middle Aged , Male , Brazil , Hospitals , Hospitalization , Machine LearningABSTRACT
Introduction: Neurological manifestations have been associated with a poorer prognosis in COVID-19. However, data regarding their incidence according to sex and age groups is still lacking. Methods: This retrospective multicentric cohort collected data from 39 Brazilian hospitals from 17 cities, from adult COVID-19 admitted from March 2020 to January 2022. Neurological manifestations presented at hospital admission were assessed according to incidence by sex and age group. Results: From 13,603 COVID-19 patients, median age was 60 years old and 53.0% were men. Women were more likely to present with headaches (22.4% vs. 17.7%, p < 0.001; OR 1.36, 95% confidence interval [CI] 1.22-1.52) than men and also presented a lower risk of having seizures (OR 0.43, 95% CI 0.20-0.94). Although delirium was more frequent in women (6.6% vs. 5.7%, p = 0.020), sex was not associated with delirium in the multivariable logistc regresssion analysis. Delirium, syncope and coma increased with age (1.5% [18-39 years] vs. 22.4% [80 years or over], p < 0.001, OR 1.07, 95% CI 1.06-1.07; 0.7% vs. 1.7%, p = 0.002, OR 1.01, 95% CI 1.00-1.02; 0.2% vs. 1.3% p < 0.001, OR 1.04, 95% CI 1.02-1.06), while, headache (26.5% vs. 7.1%, OR 0.98, 95% CI 0.98-0.99), anosmia (11.4% vs. 3.3%, OR 0.99, 95% CI] 0.98-0.99 and ageusia (13.1% vs. 3.5%, OR 0.99, CI 0.98-0.99) decreased (p < 0.001 for all). Conclusion: Older COVID-19 patients were more likely to present delirium, syncope and coma, while the incidence of anosmia, ageusia and headaches decreased with age. Women were more likely to present headache, and less likely to present seizures.