A team of Georgetown researchers, including health economists, a neurologist and a computer scientist, are teaming up to develop machine learning algorithms that they hope will help identify people with early Alzheimer’s or other types of dementia based on their financial decision-making and behavior.
“Earlier diagnosis yields many benefits for affected individuals and their families — like facilitating planning and ensuring early access to support systems — but wide-scale screening tools are lacking,” says lead study investigator and health economist Carole Roan Gresenz, a professor in Georgetown’s School of Health and McCourt School of Public Policy. “We want to know if financial red flags years prior to diagnosis might help identify individuals who would benefit from further clinical evaluation, supporting earlier diagnosis and allowing families to take steps to safeguard their financial future.”
Using Financial Data to Identify Alzheimer’s and Dementia
With a $2.8 million, four-year grant from the National Institute on Aging at the National Institutes of Health, Gresenz and her colleagues will build on previous research. That work, which surveyed data on roughly 10,000 households, showed that prior to an Alzheimer’s diagnosis, a person in the early stages of the disease faces a heightened risk of adverse financial outcomes — a likely consequence of compromised decision-making when managing money, in addition to exploitation and fraud by others. The research was published in 2019 in the journal, Health Economics.
“We want to understand more about the specific types of financial decisions and choices that underlie these findings,” Gresenz explains. “We will explore whether financial information offers the potential for early identification of individuals who are in the initial stages of Alzheimer’s disease and who should be prioritized for clinical screening.”
Gresenz says over the next several years, the team will use data collected by financial bureaus and Medicare to explore how a range of different financial markers captured in credit data change in the years prior to a dementia diagnosis.
They’ll do that with the help of a massive database created in 2022 with a previous grant from the National Institute on Aging. The database merged Medicare data with credit data in collaboration with the Federal Reserve Bank of New York. After updating the database with new and more recent information, the team plans to develop prediction algorithms, which involve a continual process of developing, testing, modifying and re-testing algorithms.
Collaboration Across Georgetown and Beyond
The project will include partnerships across disciplines at Georgetown, including the Massive Data Institute, and other external institutions and partners.
“This is a substantial undertaking that requires partnership across multiple entities and institutions, a diverse interdisciplinary team, and resources for data creation and analysis,” says health economist Jean M. Mitchell, a professor at the McCourt School of Public Policy and an investigator on the project.
In addition to Gresenz and Mitchell, Georgetown investigators on the project include R. Scott Turner, director of Georgetown’s Memory Disorders Program and a professor of neurology in the School of Medicine; and Lisa Singh, a professor of computer science in the College of Arts & Science and director of the Massive Data Institute at the McCourt School of Public Policy. Additional collaborators include Wilbert van der Klauuw, senior vice president and director of the Federal Reserve Bank of New York’s Center for Microeconomic Data Research and Statistics, and programmers and graduate students in the Massive Data Institute.
A Personal Connection to Alzheimer’s Research
For Gresenz, her research was inspired in part by her own experiences of helping her mother, who had Parkinson’s disease and dementia.
“I look back and wonder if some of the changes in her financial behavior were harbingers that I could and should have paid more attention to,” she says.
“As I have talked with more and more people about this research, I have become the steward of many stories about loved ones and the toll of financial distress prior to an Alzheimer’s diagnosis.”
Gresenz says she hopes this work will provide a new tool to help families avoid financial difficulties.
“Credit data provides an opportunity to move from anecdote to evidence as it relates to the financial lives of those with Alzheimer’s or a related dementia and the disease’s progression,” she says.
Research described in this publication is supported by the National Institute on Aging of the National Institutes of Health (R01AG080623).
