UCLA researchers developed an AI model using EHR data to better detect undiagnosed Alzheimer's disease, especially in underrepresented groups.
Key Details
- 1The AI leverages semi-supervised positive unlabeled learning to identify probable Alzheimer's cases from electronic health records (EHRs).
- 2Tested on data from over 97,000 UCLA Health patients across multiple ethnic groups.
- 3Achieved 77–81% sensitivity in various populations compared to 39–53% for traditional models.
- 4Model validation included polygenic risk scores (APOE ε4 allele), supporting accuracy of predictions.
- 5Tool addresses underdiagnosis, which especially affects African American and Hispanic/Latino communities.
- 6Planned prospective validation in other health systems; patent application filed for framework.
Why It Matters
AI-driven detection of undiagnosed Alzheimer's may support earlier intervention, which is critical as new treatments emerge. Importantly, this model addresses equity concerns by reducing diagnostic bias across diverse populations, a key consideration for both clinical and AI application in healthcare.
Source
EurekAlert
Related News
•EurekAlert
AI Accelerates Radiopharmaceuticals, Boosts Personalized Dosimetry in Cancer
Machine learning is driving advancements in radiopharmaceutical drug discovery and optimizing patient-specific dosimetry for precision cancer therapy.
•EurekAlert
Physicians Overly Trust Erroneous AI, Ignore Contradictory Evidence
Physicians tend to trust incorrect AI advice, even when evidence contradicts it, suggesting risks in clinical decision-making with AI tools.
•EurekAlert
Concerns Raised Over Unverified Datasets in AI Health Prediction Models
A new study finds widely used AI health prediction models are built on datasets with unverifiable origins, raising safety and validity concerns.