A top 10 pharma company used AI/ML to predict asthma patients at high risk of acute exacerbations, aiming to improve care and trial efficiency. The predictive model, built with 681 feature groups, identified the top 5% of patients as 188 times more likely to have an exacerbation within three months. Precision-recall analysis showed that 20% of future exacerbations could be predicted with 46% accuracy, highlighting the model’s clinical utility. Subgroup analysis revealed five distinct patient clusters, enabling tailored interventions, faster recruitment in clinical studies, and targeted education for providers and payors to address unmet needs.

Join for free to read