Mayo Clinic researchers have developed artificial intelligence tools that can predict which young children with asthma are at the highest risk for severe asthma attacks and serious respiratory infections, potentially as early as age 3. The findings, published in the Journal of Allergy and Clinical Immunology, represent a major step toward preventive, precision medicine in pediatric asthma. Asthma affects nearly 6 million children in the U.S. and is a leading cause of school absences, emergency room visits, and hospitalizations. Respiratory infections are the most common trigger for asthma flare-ups, but symptoms vary widely and evolve over time, making it difficult for doctors to identify high-risk patients early. The new AI tools aim to close this gap by analyzing vast amounts of patient data to detect patterns linked to severe outcomes. The study analyzed electronic health records from over 22,000 children born between 1997 and 2016 in southeastern Minnesota. Using machine learning and natural language processing, the AI tools extracted critical details from doctors’ notes—such as symptom history, family medical background, and allergic conditions—to apply two standard asthma diagnostic criteria: the Predetermined Asthma Criteria and the Asthma Predictive Index. Children who met both criteria formed a distinct subgroup with significantly higher risk. By age 3, this high-risk group had more than double the rate of pneumonia and nearly three times the rate of influenza compared to other children. They also experienced the highest rates of asthma exacerbations requiring steroids, emergency care, or hospitalization. Respiratory syncytial virus (RSV) infections were more frequent in this group during their first three years of life. These children were more likely to have a family history of asthma, eczema, allergic rhinitis, or food allergies. Laboratory data from prior studies revealed biological markers of allergic inflammation—such as elevated eosinophils, allergen-specific IgE, and periostin—along with early signs of reduced lung function. Together, these indicators point to a specific, high-risk asthma subtype driven by type 2 immune responses. The research is part of Mayo Clinic’s Precure initiative, which focuses on predicting and preventing serious diseases before they progress. The team plans to test the AI tools in broader, more diverse populations and healthcare systems. They also aim to integrate biological data, such as genetic and immune profiles, to refine asthma classification and improve early treatment. Additionally, the researchers are exploring a potential therapy to dampen overactive immune responses in asthma. Using lab-grown lung cell models, or organoids, they hope to accelerate the discovery of preventive treatments and develop new strategies for early intervention in childhood asthma.