Machine Learning Prediction for Supplemental Oxygen Requirement in Patients with COVID-19

Abstract

Background: The coronavirus disease poses an urgent threat to global public health and is characterized by rapid disease progression even in mild cases. In this study, we investigated whether machine learning can be used to predict which patients will have a deteriorated condition and require oxygenation in asymptomatic or mild cases of COVID-19.Methods: This single-center, retrospective, observational study included COVID-19 patients admitted to the hospital from February 1, 2020, to May 31, 2020, and who were either asymptomatic or presented with mild symptoms and did not require oxygen support on admission. Data on patient characteristics and vital signs were collected upon admission. We used seven machine learning algorithms, assessed their capability to predict exacerbation, and analyzed important influencing features using the best algorithm.Results: In total, 210 patients were included in the study. Among them, 43 (19%) required oxygen therapy. Of all the models, the logistic regression model had the highest accuracy and precision. Logistic regression analysis showed that the model had an accuracy of 0.900, precision of 0.893, and recall of 0.605. The most important parameter for 4 predictive capability was SpO2, followed by age, respiratory rate, and systolic blood pressure. Conclusion:In this study, we developed a machine learning model that can be used as a triage tool by clinicians to detect high-risk patients and disease progression earlier.Prospective validation studies are needed to verify the application of the tool in clinical practice.

Evaluating AI Proficiency in Nuclear Cardiology: Large Language Models take on the Board Preparation Exam

“…To our knowledge, previous studies have not yet examined the time-progressive performance of multiple multimodal compatible LLMs on medical subspecialty board exams, nor have they assessed GPT-4o. Previous research has explored GPT-4's ability to answer hybrid image and text-based questions on the Japanese Emergency Medicine Board Exam 12 and the American Shoulder and Elbow Surgeons Maintenance of Certification Exam 11 . GPT-4 with Vision (GPT-4V) was tested on the Japanese Otolaryngology Board Exam 10 , and GPT-4V Turbo was tested on the Japanese Diagnostic Radiology Board Exam 14 .…”

“…Particularly, the capacity of LLMs to analyze and interpret multimodal questions that include medical images-a critical component in cardiovascular imaging-remains underexplored. While some studies have begun to evaluate image analysis capabilities in individual LLMs, comprehensive comparisons across different models and over time are lacking [10][11][12][13][14] . This research gap underscores the need for more in-depth investigation into the image interpretation abilities of different LLMs, especially for diagnostics that integrate multimodal data, such as nuclear cardiology.…”

Machine Learning First Response to COVID-19: A Systematic Literature Review of Clinical Decision Assistance Approaches during Pandemic Years from 2020 to 2022

“…Similarly, Aslam [64] identified the impact of particular attributes on the prediction of mortality and mechanical ventilation support [66] in COVID-19 patients. Igarashi et al [71] introduced a model that can be implemented as a triage tool to detect the need for supplemental oxygen. Finally, understanding that COVID-19 hospitalization times are often long and may vary substantially from patient to patient, some works [56,70,76] aimed to develop a reliable prediction model of ED length of stay for COVID-19 patients and to identify clinical factors, such as age and comorbidities, associated with ED length of stay.…”

Performance of a Large Language Model on Japanese Emergency Medicine Board Certification Examinations

“…Artificial intelligence (AI) has the potential to reduce the burden of emergency physicians by fulfilling many expected roles [2][3][4] . Large language models (LLMs) are an advanced form of AI that learns from large quantities of textual information and engages in natural interactions with humans.…”