Machine learning algorithms applied to the diagnosis of COVID-19 based on epidemiological, clinical, and laboratory data

Jornal Brasileiro de Pneumologia, Journal Year: 2025, Volume and Issue: unknown, P. e20240385 - e20240385

Published: Feb. 4, 2025

Objective: To predict COVID-19 in hospitalized patients with SARS a city southern Brazil by using machine learning algorithms. Methods: The study sample consisted of = 18 years age admitted to the emergency department and Hospital Escola - Universidade Federal de Pelotas between March December 2020. Epidemiological, clinical, laboratory data were processed algorithms order identify patterns. Mean AUC values calculated for each combination model oversampling/undersampling techniques during cross-validation. Results: Of total 100 SARS, 78 had information RT-PCR testing SARS-CoV-2 infection therefore included analysis. Most (58%) female, mean was 61.4 ± 15.8 years. Regarding models, random forest slightly higher median performance when compared other models tested adopted. most important features diagnose leukocyte count, PaCO2, troponin levels, duration symptoms days, platelet multimorbidity, presence band forms, urea age, D-dimer an 87%. Conclusions: Artificial intelligence represent efficient strategy high clinical suspicion, particularly situations which health care systems face intense strain, such as pandemic.

Language: Английский

---

On the Maintenance Oversight of the Healthcare Sector Based on Artificial Intelligence

Studies in systems, decision and control, Journal Year: 2025, Volume and Issue: unknown, P. 395 - 425

Published: Jan. 1, 2025

Language: Английский

---

Informatics-Driven Unsupervised Learning of Comorbidity Clusters for COVID-19 Reinfection Risk: A Finite Mixture Modeling Approach

Informatics in Medicine Unlocked, Journal Year: 2025, Volume and Issue: unknown, P. 101649 - 101649

Published: May 1, 2025

Language: Английский

---

Developing a Framework for Harnessing Disruptive Emerging Technologies in Health for Society 5.0 in a Developing Context

Published: May 3, 2025

Language: Английский

---

Harnessing Machine Learning in Early COVID-19 Detection and Prognosis: A Comprehensive Systematic Review

Cureus, Journal Year: 2023, Volume and Issue: unknown

Published: May 1, 2023

During the early phase of COVID-19 pandemic, reverse transcriptase-polymerase chain reaction (RT-PCR) testing faced limitations, prompting exploration machine learning (ML) alternatives for diagnosis and prognosis. Providing a comprehensive appraisal such decision support systems their use in management can aid medical community making informed decisions during risk assessment patients, especially low-resource settings. Therefore, objective this study was to systematically review studies that predicted or severity disease using ML. Following Preferred Reporting Items Systematic Reviews Meta-Analysis (PRISMA), we conducted literature search MEDLINE (OVID), Scopus, EMBASE, IEEE Xplore from January 1 June 31, 2020. The outcomes were prognostic measures as death, need mechanical ventilation, admission, acute respiratory distress syndrome. We included peer-reviewed observational studies, clinical trials, research letters, case series, reports. extracted data about study's country, setting, sample size, source, dataset, diagnostic outcomes, prediction measures, type ML model, accuracy. Bias assessed Prediction model Risk Of ASsessment Tool (PROBAST). This registered International Prospective Register (PROSPERO), with number CRD42020197109. final records extraction 66. Forty-three (64%) used secondary data. majority Chinese authors (30%). Most (79%) relied on chest imaging prediction, while remainder various laboratory indicators, including hematological, biochemical, immunological markers. Thirteen explored predicting severity, rest diagnosis. Seventy percent articles deep models, 30% traditional algorithms. reported high sensitivity, specificity, accuracy models (exceeding 90%). overall concern bias "unclear" 56% studies. mainly due concerns selection bias. may help identify patients particularly context imaging. Although these reflect exhibit accuracy, novelty biases dataset make them replacement clinicians' cognitive decision-making questionable. Continued is needed enhance robustness reliability

Language: Английский

---

Ultrasound-based AI for COVID-19 detection: a comprehensive review of public and private lung ultrasound datasets and studies

Multimedia Tools and Applications, Journal Year: 2025, Volume and Issue: unknown

Published: April 5, 2025

Language: Английский

---

An overview of reviews on digital health interventions during COVID- 19 era: insights and lessons for future pandemics

Archives of Public Health, Journal Year: 2025, Volume and Issue: 83(1)

Published: May 9, 2025

Language: Английский

---

Glucocorticoid Therapy in COVID-19

Seminars in Respiratory and Critical Care Medicine, Journal Year: 2023, Volume and Issue: 44(01), P. 100 - 117

Published: Jan. 16, 2023

Abstract Coronavirus disease 2019 (COVID-19) pneumonia caused by severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) has resulted in significant mortality pandemic proportions. Inflammation response to the infection contributes pathogenesis of pneumonia. This review will discuss prior studies on use glucocorticoids treat infections, rationale for COVID-19, and existing data. We also highlight outstanding research questions future studies.

Language: Английский

---

Machine and deep learning methods in identifying malaria through microscopic blood smear: A systematic review

Engineering Applications of Artificial Intelligence, Journal Year: 2024, Volume and Issue: 133, P. 108529 - 108529

Published: May 2, 2024

Language: Английский

---

Artificial neural network based prediction of the lung tissue involvement as an independent in‐hospital mortality and mechanical ventilation risk factor in COVID‐19

Journal of Medical Virology, Journal Year: 2023, Volume and Issue: 95(5)

Published: May 1, 2023

During COVID-19 pandemic, artificial neural network (ANN) systems have been providing aid for clinical decisions. However, to achieve optimal results, these models should link multiple data points simple models. This study aimed model the in-hospital mortality and mechanical ventilation risk using a two step approach combining variables ANN-analyzed lung inflammation data.A set of 4317 hospitalized patients, including 266 patients requiring ventilation, was analyzed. Demographic (including length hospital stay mortality) chest computed tomography (CT) were collected. Lung involvement analyzed trained ANN. The combined then unadjusted multivariate Cox proportional hazards models.Overall associated with ANN-assigned percentage (hazard ratio [HR]: 5.72, 95% confidence interval [CI]: 4.4-7.43, p < 0.001 >50% tissue affected by pneumonia), age category (HR: 5.34, CI: 3.32-8.59 cases >80 years, 0.001), procalcitonin 2.1, 1.59-2.76, 0.001, C-reactive protein level (CRP) 2.11, 1.25-3.56, = 0.004), glomerular filtration rate (eGFR) 1.82, 1.37-2.42, 0.001) troponin 2.14, 1.69-2.72, 0.001). Furthermore, is also ANN-based 13.2, 8.65-20.4, involvement), age, 1.91, 1.14-3.2, 0.14, eGFR 1.2-2.74, 0.004) variables, diabetes 2.5, 1.91-3.27, cardiovascular cerebrovascular disease 3.16, 2.38-4.2, chronic pulmonary 2.31, 1.44-3.7, 0.001).ANN-based strongest predictor unfavorable outcomes in represents valuable support tool

Language: Английский

---