Vanderbilt Health researchers have received a grant from Eli Lilly and Company to build a new AI tool. The tool will live inside the hospital's electronic health record system.
The goal is simple. Get Alzheimer's patients to treatment faster.
New monoclonal antibody drugs can slow early cognitive decline in patients who test positive for amyloid. These drugs have opened new treatment paths for early stage Alzheimer's disease.
But there's a catch. Patients who qualify for these drugs often wait weeks before their first infusion.
That wait includes referrals, specialist evaluations, brain imaging, and insurance approval. Each step adds time.
To fix this, the Vanderbilt team is building an AI triage agent. The project runs for 18 months and costs $600,000.
The tool will focus on one specific handoff. That's the referral from primary care or geriatrics doctors to neurology specialists.
How the AI Tool Will Work
When a doctor enters a referral for a cognitive concern, the AI will step in. It will summarize the patient's relevant chart information right away.
The tool will also flag details that are missing. These are the kinds of gaps that usually slow down evaluations.
Then it will recommend how the case should be handled. Cases get marked as either priority or standard.
Doctors still have the final say. They can accept, change, or reject any suggestion the AI makes.
Dr. Amalia Peterson, a co-principal investigator on the project, described the problem clinicians face. "By the time a patient reaches our clinic, the clock has often been running for weeks," she said. She added that the tool won't replace clinical judgment, but it should cut down on delays.
Research Team and Next Steps
The project is led by You Chen, an associate professor of biomedical informatics. Peterson and geriatrician Sean Huang are serving as co-principal investigators.
Chen is also leading a separate $1 million Lilly-funded project. That one focuses on gaps in obesity care.
The Alzheimer's project starts with mapping where delays happen. Researchers will use AI to study care timelines from more than 5,300 Vanderbilt Health patients.
After that, the team will work with clinicians and staff to find the root causes of those delays. Then they'll test their AI agent in a real pilot program.
Success will be measured one way. Researchers will track how much time drops between diagnosis and a patient's first infusion.
Huang explained where the biggest need lies. "The pathway begins in primary care or geriatrics, where cognitive concerns are first recognized," he said. He noted the next steps can be hard for patients to navigate.
The team already has plans beyond this first project. A second AI agent could review MRI images and generate safety reports for radiologists. A third could handle insurance authorization paperwork for Alzheimer's therapies, tracking the process from start to finish.
