Extracting relevant predictive variables for COVID-19 severity prognosis: An exhaustive comparison of feature selection techniques.

With the COVID-19 pandemic having caused unprecedented numbers of infections and deaths, large research efforts have been undertaken to increase our understanding of the disease and the factors which determine diverse clinical evolutions. Here we focused on a fully data-driven exploration regarding which factors (clinical or otherwise) were most informative for SARS-CoV-2 pneumonia severity prediction via machine learning (ML). In particular, feature selection techniques (FS), designed to reduce the dimensionality of data, allowed us to characterize which of our variables were the most useful for ML prognosis. We conducted a multi-centre clinical study, enrolling n = 1548 patients hospitalized due to SARS-CoV-2 pneumonia: where 792, 238, and 598 patients experienced low, medium and high-severity evolutions, respectively. Up to 106 patient-specific clinical variables were collected at admission, although 14 of them had to be discarded for containing ⩾60% missing values. Alongside 7 socioeconomic attributes and 32 exposures to air pollution (chronic and acute), these became d = 148 features after variable encoding. We addressed this ordinal classification problem both as a ML classification and regression task. Two imputation techniques for missing data were explored, along with a total of 166 unique FS algorithm configurations: 46 filters, 100 wrappers and 20 embeddeds. Of these, 21 setups achieved satisfactory bootstrap stability (⩾0.70) with reasonable computation times: 16 filters, 2 wrappers, and 3 embeddeds. The subsets of features selected by each technique showed modest Jaccard similarities across them. However, they consistently pointed out the importance of certain explanatory variables. Namely: patient's C-reactive protein (CRP), pneumonia severity index (PSI), respiratory rate (RR) and oxygen levels -saturation Sp O2, quotients Sp O2/RR and arterial Sat O2/Fi O2-, the neutrophil-to-lymphocyte ratio (NLR) -to certain extent, also neutrophil and lymphocyte counts separately-, lactate dehydrogenase (LDH), and procalcitonin (PCT) levels in blood. A remarkable agreement has been found a posteriori between our strategy and independent clinical research works investigating risk factors for COVID-19 severity. Hence, these findings stress the suitability of this type of fully data-driven approaches for knowledge extraction, as a complementary to clinical perspectives.

Utility of Differential White Cell Count and Cell Population Data for Ruling Out COVID-19 Infection in Patients With Community-Acquired Pneumonia.

INTRODUCTION: The main aim of this study was to assess the utility of differential white cell count and cell population data (CPD) for the detection of COVID-19 in patients admitted for community-acquired pneumonia (CAP) of different etiologies. METHODS: This was a multicenter, observational, prospective study of adults aged ≥18 years admitted to three teaching hospitals in Spain from November 2019 to November 2021 with a diagnosis of CAP. At baseline, a Sysmex XN-20 analyzer was used to obtain detailed information related to the activation status and functional activity of white cells. RESULTS: The sample was split into derivation and validation cohorts of 1065 and 717 patients, respectively. In the derivation cohort, COVID-19 was confirmed in 791 patients and ruled out in 274 patients, with mean ages of 62.13 (14.37) and 65.42 (16.62) years, respectively (p<0.001). There were significant differences in all CPD parameters except MO-Y. The multivariate prediction model showed that lower NE-X, NE-WY, LY-Z, LY-WY, MO-WX, MO-WY, and MO-Z values and neutrophil-to-lymphocyte ratio were related to COVID-19 etiology with an AUC of 0.819 (0.790, 0.846). No significant differences were found comparing this model to another including biomarkers (p=0.18). CONCLUSIONS: Abnormalities in white blood cell morphology based on a few cell population data values as well as NLR were able to accurately identify COVID-19 etiology. Moreover, systemic inflammation biomarkers currently used were unable to improve the predictive ability. We conclude that new peripheral blood biomarkers can help determine the etiology of CAP fast and inexpensively.

Machine learning-based model for prediction of clinical deterioration in hospitalized patients by COVID 19.

Despite the publication of great number of tools to aid decisions in COVID-19 patients, there is a lack of good instruments to predict clinical deterioration. COVID19-Osakidetza is a prospective cohort study recruiting COVID-19 patients. We collected information from baseline to discharge on: sociodemographic characteristics, comorbidities and associated medications, vital signs, treatment received and lab test results. Outcome was need for intensive ventilatory support (with at least standard high-flow oxygen face mask with a reservoir bag for at least 6 h and need for more intensive therapy afterwards or Optiflow high-flow nasal cannula or noninvasive or invasive mechanical ventilation) and/or admission to a critical care unit and/or death during hospitalization. We developed a Catboost model summarizing the findings using Shapley Additive Explanations. Performance of the model was assessed using area under the receiver operating characteristic and prediction recall curves (AUROC and AUPRC respectively) and calibrated using the Hosmer-Lemeshow test. Overall, 1568 patients were included in the derivation cohort and 956 in the (external) validation cohort. The percentages of patients who reached the composite endpoint were 23.3% vs 20% respectively. The strongest predictors of clinical deterioration were arterial blood oxygen pressure, followed by age, levels of several markers of inflammation (procalcitonin, LDH, CRP) and alterations in blood count and coagulation. Some medications, namely, ATC AO2 (antiacids) and N05 (neuroleptics) were also among the group of main predictors, together with C03 (diuretics). In the validation set, the CatBoost AUROC was 0.79, AUPRC 0.21 and Hosmer-Lemeshow test statistic 0.36. We present a machine learning-based prediction model with excellent performance properties to implement in EHRs. Our main goal was to predict progression to a score of 5 or higher on the WHO Clinical Progression Scale before patients required mechanical ventilation. Future steps are to externally validate the model in other settings and in a cohort from a different period and to apply the algorithm in clinical practice.Registration: ClinicalTrials.gov Identifier: NCT04463706.

COPD in SARS-CoV-2 pandemic. baseline characteristics related to hospital admissions.

OBJECTIVE: To develop a predictive model for COPD patients admitted for COVID-19 to support clinical decision-making. METHOD: Retrospective cohort study of 1313 COPD patients with microbiological confirmation of SARS-CoV-2 infection. The sample was randomly divided into two subsamples, for the purposes of derivation and validation of the prediction rule (60% and 40%,respectively). Data collected for this study included sociodemographic characteristics, baseline comorbidities, baseline treatments, and other background data. Multivariable logistic regression analysis was used to develop the predictive model. RESULTS: Male sex, older age, hospital admissions in the previous year, flu vaccination in the previous season, a Charlson Index>3 and a prescription of renin-angiotensin aldosterone system inhibitors at baseline were the main risk factors for hospital admission. The AUC of the categorized risk score was 0.72 and 0.69 in the derivation and validation samples, respectively. Based on the risk score, four groups were identified with a risk of hospital admission ranging from 21% to 80%. CONCLUSIONS: We propose a classification system to identify COPD people with COVID-19 with a higher risk of hospitalization, and indirectly, more severe disease, that is easy to use in primary care, as well as hospital emergency room settings to help clinical decision-making. CLINICALTRIALS.GOV IDENTIFIER: NCT04463706.

Predictors of mortality of COVID-19 in the general population and nursing homes.

The factors that predispose an individual to a higher risk of death from COVID-19 are poorly understood. The goal of the study was to identify factors associated with risk of death among patients with COVID-19. This is a retrospective cohort study of people with laboratory-confirmed SARS-CoV-2 infection from February to May 22, 2020. Data retrieved for this study included patient sociodemographic data, baseline comorbidities, baseline treatments, other background data on care provided in hospital or primary care settings, and vital status. Main outcome was deaths until June 29, 2020. In the multivariable model based on nursing home residents, predictors of mortality were being male, older than 80 years, admitted to a hospital for COVID-19, and having cardiovascular disease, kidney disease or dementia while taking anticoagulants or lipid-lowering drugs at baseline was protective. The AUC was 0.754 for the risk score based on this model and 0.717 in the validation subsample. Predictors of death among people from the general population were being male and/or older than 60 years, having been hospitalized in the month before admission for COVID-19, being admitted to a hospital for COVID-19, having cardiovascular disease, dementia, respiratory disease, liver disease, diabetes with organ damage, or cancer while being on anticoagulants was protective. The AUC was 0.941 for this model's risk score and 0.938 in the validation subsample. Our risk scores could help physicians identify high-risk groups and establish preventive measures and better follow-up for patients at high risk of dying.ClinicalTrials.gov Identifier: NCT04463706.