A groundbreaking study published in Gastroenterology shows that a machine learning model (random forest analysis) significantly outperforms conventional tools in identifying high-risk cirrhosis patients. Researchers analyzed data from 121 hospitals globally (via the CLEARED consortium), validating the model with U.S. Veterans’ data.
Key findings:
- Accurate across both high- and low-income settings.
- Maintained high performance even with just 15 variables.
- Patients were effectively triaged into high- and low-risk categories.
- Clinical scalability makes this AI tool practical and impactful.