Duke Initiative Aims To Bring Dementia AI Tools Into Clinical Practice

TL;DR

For listenersOffer from Amazon

Turn quiet afternoons into listening time

• Thousands of audiobooks, podcasts and originals
• Listen on your phone, tablet or Echo — also offline
• Cancel anytime

Try Audible free →Free trial for new members

As an affiliate, we earn on qualifying purchases.Hide

Duke will receive nearly $25 million over five years as one of five National Institute on Aging AI and Technology Collaboratories. The center plans to fund research nationwide and develop shared data, evaluation and implementation resources to help assess whether AI tools can be used safely and effectively in aging and dementia care.

Duke University has been selected as one of five centers in a National Institute on Aging program intended to help move artificial intelligence tools for aging and dementia from research toward clinical use. The Duke Artificial Intelligence and Technology Collaboratory for Aging: Governance and Evaluation will receive nearly $25 million over five years to support research nationwide and build shared systems for evaluating, implementing and monitoring AI tools.

The Duke center, known as Duke AITC-AIGE, will receive about $5 million a year, with roughly $4 million annually allocated to research projects nationwide, according to the report. At least 80% of those research funds will support Alzheimer’s disease and related dementias; the remainder will go to broader aging research, said Sheng Luo, a Duke professor of biostatistics and bioinformatics and a principal investigator.

Based within the Duke Clinical and Translational Science Institute, the center plans to provide pilot grants, data-science infrastructure, implementation assistance and safeguards for responsible AI. Its resources are also expected to include evaluation methods, safety-monitoring approaches and standardized “model cards” describing an AI tool’s intended purpose, training data, test performance, limitations and possible biases.

The initiative is organized into three cores. Juliessa Pavon, an associate professor in geriatrics, leads the Administrative Core; Rishi Kamaleswaran, an associate professor of surgery and anesthesiology, leads the Clinical Implementation & Data Science Core; and Luo leads the Research Innovation Core. Duke said the center is the only one of the five led from the Southeast.

Building Evidence Before Clinical Use

The initiative addresses a gap between developing AI algorithms and establishing whether they are dependable and useful in routine care. Funding and shared infrastructure could help researchers test tools in real-world settings, rather than relying only on research-stage performance, and give health systems clearer information before deciding whether a tool is appropriate for patients.

Potential uses described by Duke include analyzing speech, behavior and wearable-device data to identify early cognitive changes, assessing risks such as falls or wandering, supporting medication review and helping caregivers access personalized guidance. These are possible applications, not evidence that the center has already produced or clinically validated such tools. The stated goal is to assess which approaches work and whether they improve outcomes or quality of life.

The center’s planned model cards are intended to make tool limitations and suitable uses easier to understand for clinicians, researchers, health systems and patients. That information may help users judge whether an AI system fits a particular setting, while evaluation and monitoring work could reveal when performance is unreliable or changes over time.

From Research to Routine Care

The National Institute on Aging’s Artificial Intelligence and Technology Collaboratories program selected five centers to create resources for advancing AI in aging and Alzheimer’s disease and related dementias. Duke’s award is one part of that national effort; the source report does not provide the names or funding details of the other centers.

Pavon said the problem is not a lack of algorithms, but a shortage of solutions that have been rigorously evaluated, safely implemented, monitored over time and shown to improve real-world outcomes for older adults. Duke’s project is designed to build a pathway from promising research to tools that can be considered for practice, with attention to governance and evaluation as well as technical development.

One planned resource, NeuroAI, is described as a secure platform for analytics and evaluation. Researchers will be able to bring together different kinds of health information—including medical records, brain images, laboratory results and genetic data—to study aging and dementia and develop and assess AI systems. The source does not specify when the platform will become available or which datasets will be included.

“We do not have a shortage of AI algorithms. We have a shortage of AI solutions that have been rigorously evaluated, safely implemented, continuously monitored – and shown that it really improves real-world outcomes for older adults.”

— Juliessa Pavon, Duke associate professor of geriatrics and principal investigator

Clinical Impact Still Unproven

The announcement describes a five-year initiative and its planned resources; it does not identify any specific AI tool that has been approved for clinical use or shown through this center to improve patient outcomes. Which tools will be funded, how they will perform in different care settings and whether they will change clinical decisions remain open questions.

The source also does not give a timeline for the first funding awards, the launch of NeuroAI, or completion of particular evaluations. While the plans include safety monitoring and tools to describe limitations, the report does not set out specific standards, oversight procedures or measures for determining whether an AI system is ready for routine use.

Funding and Evaluation Ahead

Over the award’s five-year period, Duke says it will operate a national pilot-funding program and provide investigators with support in AI evaluation, data science, implementation, ethics, governance and stakeholder engagement. The center also plans to develop shared data and evaluation resources, including model cards and safety-monitoring approaches.

The next evidence to watch will be details about the first funded projects, access to the platform and results from testing tools in real-world settings. Those findings will help show whether the initiative can move beyond infrastructure-building to identify AI applications that are safe, reliable and useful for older adults and people living with dementia.

Key Questions

What is Duke AITC-AIGE?

It is Duke’s Artificial Intelligence and Technology Collaboratory for Aging: Governance and Evaluation, one of five centers selected for the National Institute on Aging’s AI and Technology Collaboratories program.

How much funding will the Duke center receive?

Duke is set to receive nearly $25 million over five years, or about $5 million annually. The report says roughly $4 million a year will support nationwide research projects.

Will the initiative immediately bring dementia AI tools into clinics?

No. The announcement concerns funding and plans to build research, evaluation and implementation resources. It does not say that a particular tool is ready or approved for clinical use.

What kinds of AI applications could researchers study?

Possible areas include detecting cognitive changes using speech, behavior or wearable data; assessing risks such as falls or wandering; supporting medication review; and helping caregivers. These are research possibilities, not confirmed clinical benefits.

What are model cards?

They are planned standardized summaries of an AI tool’s purpose, training data, test performance, limitations, potential biases and appropriate uses, intended to help users judge whether it fits a particular situation.

Source: rss