New data platform tracks the complex path to Alzheimer’s and could transform how its risk is predicted

A powerful new real-world data platform could transform how scientists predict and understand Alzheimer’s disease and Alzheimer’s disease-related dementias (AD/ADRD), reports a new study at Columbia University Mailman School of Public Health and collaborators at the Vagelos College of Physicians and Surgeons, the School of Nursing as well as the University of Miami and University of Chicago.

The project, known as the M3AD Study and Real-World Data Metaplatform, represents one of the most comprehensive efforts to date to use large-scale clinical data to advance precision aging research and accelerate discoveries in Alzheimer’s prevention and care. The study is published in the journal Alzheimer’s & Dementia.

Drawing on electronic health records from three U.S. cities, including approximately 60,000 individuals with AD/ADRD, the platform allows researchers to track how multiple chronic diseases, behaviors, and social conditions interact over time to shape dementia risk, creating one of the largest and most comprehensive datasets ever assembled to study dementia.

Unlike traditional studies that examine a limited set of individual diseases, the platform analyzes interacting health conditions, behaviors, and social factors simultaneously and over time, so that the full complexity of aging and multimorbidity is captured and predictions of dementia risk and progression are improved.

“As people live longer, chronic diseases increasingly occur together, creating complex health trajectories that traditional disease-by-disease research does not easily capture,” says Moise Desvarieux, MD, Ph.D., associate professor of Epidemiology and corresponding author.

“The new platform addresses this challenge by integrating and harmonizing electronic health records (EHRs) from nearly 10 million patients across three major health systems in New York City, Chicago, and Miami, including the roughly 60,000 patients with Alzheimer’s disease and related dementias.”

The research is motivated by the growing scale of dementia in the United States. More than 7.2 million older Americans are living with Alzheimer’s disease, including roughly 35% of people aged 85 and older.

At the same time, multimorbidity—having two or more chronic diseases—affects nearly 90% of adults over age 60, complicating diagnosis, treatment, and care. The initiative challenges the long-standing view of dementia as a single, isolated disease,” continues Desvarieux. “Alzheimer’s and related dementias do not occur in isolation,” he notes.

“They emerge from complex interactions among multiple diseases, behaviors, epigenetics and life circumstances that influence one another over time.”

“Electronic health records make it possible to analyze these interacting trajectories across decades of care. By examining longitudinal clinical histories, the platform may help identify previously unrecognized early warning signs of dementia, while capturing the broader context of patients’ lives,” says co-author George Hripcsak, MD, MS, Vivian Beaumont Allen professor of biomedical informatics at Columbia.

The three-city consortium includes: NewYork‑Presbyterian Hospital Clinical Data Warehouse, with 32 years of data from roughly 6 million patients, including 33,000 with AD/ADRD; University of Chicago Clinical Research Data Warehouse, capturing data from more than 2 million patients, including 11,000 with AD/ADRD; and University of Miami Health System, with about 1.4 million patients, including 13,000 with AD/ADRD.

Together, these datasets form a multiethnic population spanning whites, Blacks, Hispanics, and Asians, enabling researchers to study dementia risk across diverse populations.

The project also incorporates advanced analytical approaches, including machine-learning models and a federated platform that allows institutions to analyze data collaboratively while preserving privacy.

“These models can be expanded to integrate additional real-world data sources, including imaging, genetic information, and novel biomarkers as they are being identified,” adds co-author Habibul Ahsan, MD, MMedSc, Louis Block Distinguished Service professor of Public Health Sciences, Family Medicine, and Human Genetics and director, Institute for Population and Precision Health at the University of Chicago.

“Medicine has traditionally focused on treating individual organs and diseases,” co-author Tatjana Rundek, professor of Neurology and director of the Evelyn F. McKnight Brain Institute at the University of Miami, said.

“But aging is not just about one condition, it is about multiple health issues interacting over time. To truly understand and prevent dementia, we need to look at the whole person, even before symptoms appear,” continues Rundek.

“This groundbreaking study gives us a unique opportunity to uncover early signs of dementia, offering insights that could transform real-world care, especially for individuals managing multiple health challenges at once.”

The new platform also integrates predictive tools such as the Electronic Health Record Risk of Alzheimer’s and Dementia Assessment Rule (eRADAR), an algorithm that uses routine EHR data to identify individuals who may have undiagnosed dementia and should receive further clinical evaluation.

Beyond prediction, the platform will allow researchers to test prevention hypotheses in real-world populations—for example, how smoking cessation, healthy weight, and blood-pressure control during middle age influence later cognitive decline.

“We are also embedding these clinical data within neighborhood census-tract information, allowing us to examine how social and environmental conditions shape Alzheimer’s risk and progression over time,” adds co-author Allison Aïello, Ph.D., James S. Jackson Healthy Longevity professor of Epidemiology and interim director of the Columbia Aging Center.

“Our multidisciplinary approach creates a dynamic platform to improve risk prediction, guide complex clinical management, and evaluate future treatments for Alzheimer’s disease in real-world settings,” the researchers said.