About

Two people waking with their backs turned

Targeting the gap between AI innovation and evidence

APEX-ADRD (Accelerating Platforms for Evaluating and eXecuting AI Lifecycles in ADRD) is an NIA-funded AI and Technology Collaboratory based at UCLA and built to serve investigators and technology developers nationwide. It targets the gap between AI innovation and evidence: few AI tools for older adults and people living with dementia have been tested in rigorous trials, and conventional validation takes 18 to 24 months, longer than many tools stay current. APEX-ADRD supports projects at every stage of the AI lifecycle, from co-design and AI-ready data through model development, real-world validation, implementation, and post-deployment monitoring, with particular depth in the later stages where promising tools most often stall.

Led by Drs. John Mafi, Joann Elmore, and Alex Bui, the Collaboratory combines health services research, geriatrics, cognitive neurology, biomedical informatics, biostatistics, causal inference, health economics, and AI ethics. 

Its Research Innovation Core offers awardees design studios, expertise matching, and consultation on study design, informatics, statistics, and economic evaluation. Through UCLA's INOVAi program, the team has run multiple rapid-cycle pragmatic trials of AI tools in live clinical settings, and brings that infrastructure and experience to pilot teams.

Awardees do not need to be at UCLA or hold UC data access to benefit. The Collaboratory can test external models behind the firewall on UC Health data (8.7 million patients, 152,000 with ADRD, six health systems) and report reproducibility, subgroup performance, and fairness back to the team. Every pilot receives standardized fairness and safety evaluation, and the Collaboratory is releasing open-source toolkits for AI evaluation, fairness, monitoring, and economic analysis for the national research community.

We especially welcome technology developers, including startups, seeking independent evidence for FDA review, health-system adoption, or follow-on NIH or SBIR/STTR funding, and teams working with rural and under-resourced populations. 

Robust AI Support

APEX-ADRD supports projects at every stage of the AI lifecycle, from co-design and AI-ready data through model development, real-world validation, implementation, and post-deployment monitoring, with particular depth in the later stages where promising tools most often stall.