YEAR IN REVIEW
2025 marked a year of significant transformation for the initiative formerly known as Precision Health. In January, we completely revamped our brand, taking on the name of “AI & Digital Health Innovation” (AI&DHI) to better signify our expanded services, interactive communities, and our commitment to advancing research at the intersection of AI and health at U-M.
A new year, a new Precision Health us.
Our Mission
Harness data and expertise to power innovations that improve people’s health.
Our Vision
Support and accelerate the research, development, and testing of new approaches, tools and devices to evolve healthcare.
The months that followed brought a surge of innovation and collaboration. The highlights that follow capture the impact of this first year under our new banner, showcasing the achievements, insights and growth that shaped 2025.
Jump to Section
AI&DHI by the Numbers / Research Implementation / Data Solutions / Michigan Genomics Initiative / Events & Outreach
AI&DHI: History by the Numbers
We provide researchers with multimodal deidentified health data on over 5.5 million U-M Health patients, genetic data, mobile health data, compute resources, and research implementation services. These resources support grant funding, publications, and health AI model pilots at Michigan Medicine.
This year, we compiled data from 2019-2024 and determined:
researchers and community members have used our services.
unique users
have leveraged our Data Platform.
research projects
have been supported by AI&DHI.
unique users have leveraged our compute resources (at no cost
to them!)
terabytes of AI&DHI storage have been used by researchers.
Michigan Genomics Initiative participants have been genotyped.
publications have been supported by AI&DHI.
in total grant funding has been awarded to AI&DHI researchers.
Research Implementation
In 2025, we facilitated the implementation of 4 models for piloting at Michigan Medicine:
PICTURE Pediatric Model
Researchers developed these models and turned to AI&DHI to provide the guidance needed to:
Develop appropriate workflows to maximize clinical impact
Submit for IRB approval
Present to clinical, operational, and governance committees
Plan for design sessions with clinicians, nurses, and support staff
Explore options to display and alert in MiChart
Data Solutions
research community members utilized our Data Platform in 2025.
unique users utilized our Data Platform in 2025.
research projects utilized AI&DHI data and resources in 2025.
unique users leveraged our compute resources in 2025.
Major Data Initiatives from 2025
Expanding Access to Deidentified Health Data
We implemented a robust, layered framework for deidentifying Michigan Medicine’s clinical reports and notes in accordance with Safe Harbor principles, ensuring researchers can responsibly access this vast data resource while maintaining patient privacy.
In CY25, we released the deidentified Radiology reports for the entire Michigan Medicine patient cohort (~5.5M strong), for a total of about 9 million narratives and 7.5 million impressions. We also released deidentified Pathology and Microbiology reports for the entire Michigan Medicine patient population. These notes are all linkable to all deidentified EHR data available in our Data Platform.
Additions to CMS Claims / Michigan Medicine EHR Data Repository
In CY25, we enriched our CMS claims/Michigan Medicine EHR Data Repository to include years 2022 and 2023. The Data Repository combines CMS claims data with the health data of Michigan Medicine Medicare patients, and it enables researchers to access all of the patients’ imaging studies, procedures, or other billed services, even when they were not performed at U-M.
Enrichment of AI&DHI’s Deidentified Chest X-ray Repository
The team augmented the existing imaging repository with deidentified CXRs for the years 2013, 2014, and 2024. These images are all linked to the deidentified EHR data in the AI&DHI Data Platform.
Michigan Genomics Initiative (MGI)
The Michigan Genomics Initiative (MGI) community is built around shared data and resources with a focus on lowering barriers of entry and allowing a wide range of researchers and clinicians to become involved in genetics research. The goal of MGI is to enable the adoption of genetic understanding in clinical practice.
In March of 2025, MGI completed genotype calling, quality control, and genotype imputation for its Freeze 7 dataset. Significant improvements and additions in 2025 include:
variants per participant are now available.
additional Genome-Wide Association Study (GWAS) participants were added in 2025.
new Whole Genome Sequenced (WGS) participants were added in 2025.
Major MGI Initiatives from 2025
Whole Genome Sequencing (WGS) Pilot
Building on the 100,000+ patient GWAS data set from MGI, and in collaboration with the MM Dept of Pathology, AI&DHI initiated a whole genome sequencing (WGS) pilot to provide researchers a more comprehensive view of the genetics from subsets of the MGI cohort. The first cohort of samples (~2000 patients) have been processed demonstrating a robust workflow across the 4 key technical units and excellent data quality. Additional sequencing is underway. This effort serves dual purposes providing researchers with state-of-the-art genetic data while supporting clinical pathology efforts to build-out the internal infrastructure that will support WGS return of results within MM health system operations.
Additional sequencing will expand this cohort to 10,000 samples in 2026.
Events & Outreach
AI&DHI Kickoff Events
Introducing the new AI&DHI and our services to the broader U-M research community.
Feb 26
School of Public Health
ATTENDEES
including many students
April 24
College of Engineering
ATTENDEES
with faculty engaging
from across campus
May 21
NCRC
ATTENDEES
marking a strong
turnout
AI&DHI Roadshows
Highlighting our data resources/services at UM Department meetings
(Michigan Medicine, School of Public Health, Pharmacy, MIDAS)
ATTENDEES
2025 Hands-On Workshop Events
Engaging directly with the U-M research community to share resources and expertise
”Best Practices for Code Sharing”
”Access Health Research Data from Your Laptop”
Annual Symposiums
Uniting and celebrating research efforts in AI and health at U-M.
September 12
AI & Health Symposium
ATTENDEES
September 19
MGI Symposium
ATTENDEES
Outreach
researchers and community members
are subscribed to our newsletter.
researchers and community members
follow us over LinkedIn.
Background Photo: Audience members at the 2025 AI & Health Symposium listen to a keynote presentation by Dr. Marinka Zitnik, Associate Professor of Biomedical Informatics at Harvard Medical School. The Symposium was hosted jointly by AI&DHI and e-HAIL with additional support from the Institute for Healthcare Policy and Innovation (IHPI).