Innovative AI Platform Predicts Disease Severity and Hospitalization Length During Viral Outbreaks
An innovative patient triage platform powered by artificial intelligence (AI) has been developed to predict patient disease severity and length of hospitalization during viral outbreaks, according to researchers. This platform, which utilizes machine learning and metabolomics data, aims to improve patient management and assist healthcare providers in efficiently allocating resources during severe viral outbreaks that can overwhelm local healthcare systems.
The study, published in the journal Human Genomics, was led by Vasilis Vasiliou, a professor of epidemiology at Yale University School of Public Health. The researchers developed the AI-powered patient triage platform using COVID-19 as a disease model, integrating routine clinical data, patient comorbidity information, and untargeted plasma metabolomics data to drive its predictions.
Lead author Georgia Charkoftaki, an associate research scientist in the environmental health sciences department, stated that the platform distinguishes itself from typical COVID-19 AI prediction models and serves as the foundation for a proactive and methodical approach to address future viral outbreaks.
Using machine learning, the researchers built a model of COVID-19 severity and prediction of hospitalization based on clinical data and metabolic profiles collected from 111 COVID-19 patients admitted to Yale New Haven Hospital during a two-month period in 2020, as well as 342 healthy individuals who served as controls.
The study identified several elevated metabolites in plasma that had a distinct correlation with COVID-19 severity, including allantoin, 5-hydroxy tryptophan, and glucuronic acid. Notably, patients with elevated blood eosinophil levels were found to have a worse disease prognosis, indicating a potential new biomarker for COVID-19 severity. The researchers also observed that patients requiring positive airway pressure or intubation exhibited decreased plasma serotonin levels, which warrants further research.
The AI-assisted patient triage platform has three essential components:
1. Predicting disease severity and length of hospitalization
2. Classifying patients’ conditions upon arrival at the emergency department
3. Providing pre-hospital patient management through user-friendly software called COVID Severity by Metabolomic and Clinical Study (CSMC)
While the study acknowledges limitations, such as the absence of COVID-19 vaccines and specific treatments in the sample collection period, the researchers believe their findings contribute to a more effective and data-driven public health response against COVID-19 and future viral outbreaks.
Collaborating institutions in the research include the Laboratory of Analytical Chemistry at the National and Kapodistrian University of Athens, Greece; Imperial College London; and the São Carlos Institute of Chemistry at the University of São Palo, Brazil.
This innovative AI platform shows promising potential in predicting disease severity and length of hospitalization during viral outbreaks, enabling healthcare providers to optimize patient health outcomes and allocate hospital resources efficiently. With further development and refinement, this technology could revolutionize patient triage and improve overall healthcare delivery during times of crisis.
