ABSTRACT
Objective: To describe demographics, clinical, and respiratory mechanics (including ventilatory management details) of patients admitted to the Intensive Care Unit (ICU) with severe COVID-19 and to evaluate the effectiveness of gas exchange variables, ventilatory parameters, and ICU illness severity scores in predicting 28-day mortality. Design: Single-center retrospective cohort study. Setting: Portuguese medical-surgical ICU. Patients: Adults sequentially admitted to the ICU, from March 18 to May 12, 2020, with critical COVID-19 requiring invasive mechanical ventilation (IMV) for over 48 hours. Interventions: None, due to study design. Measurements and results: Data regarding positioning, positive end-expiratory pressure (PEEP), driving pressure, static lung compliance, and lowest daily arterial oxygen partial pressure to fractional inspired oxygen (PaO2/FiO2) ratio throughout the first 5 days of . ICU admission were collected from daily ventilatory assessment charts. The median ICU length of stay was 11.3 days and median IMV duration was 9.5 days. The 28-day mortality was 12.1%. When comparing non-survivors and survivors, significant differences were found regarding Simplified Acute Physiology Score (SAPS) II (48.5, IQR 14.0 vs. 32.0, IQR 11.0, p=0.004), PaO2/FiO2 ratio before endotracheal intubation (101.3, IQR 22.5 vs. 174.1, IQR 9.5, p=0.01) and throughout ICU stay. Over 90% of patients were submitted to prone positioning. Use of low PEEP levels and maintenance of low driving pressures in patients whose overall compliance was low as possible. Conclusions: Significant differences were found regarding SAPS II and PaO2/FiO2 ratios between survivors and non-survivors, eliciting further investigation as potential mortality predictors. With the second wave of the pandemic taking shape, sharing previous experience is crucial to further coordinate efforts internationally.
ABSTRACT
The advent of the Covid-19 pandemic has resulted in a global crisis making the health systems vulnerable, challenging the research community to find novel approaches to facilitate early detection of infections. This open-up a window of opportunity to exploit machine learning and artificial intelligence techniques to address some of the issues related to this disease. In this work, we address the classification of ten SARS-CoV-2 protein sequences related to Covid-19 using k-mer frequency as features and considering two objectives;classification performance and feature selection. The first set of experiments considered the objectives one at the time, four techniques were used for the feature selection and twelve well known machine learning methods, where three are neural network based for the classification. The second set of experiments considered a multi-objective approach where we tested a well known multi-objective approach Non-dominated Sorting Genetic Algorithm II (NSGA-II), and the Multi-dimensional Archive of Phenotypic Elites (MAP-Elites), which considers quality+diversity containers to guide the search through elite solutions. The experimental results shows that ResNet and PCA is the best combination using single objectives. Whereas, for the mulit-classification, NSGA-II outperforms ME with two out of three classifiers, while ME gets competitive results bringing more diverse set of solutions.