Artificial intelligence can be utilized to detect early signs of mild cognitive impairment (MCI) and with future data, might more accurately predict the markers leading to it.
That’s according to a study published in The Journal of the Alzheimer’s Association. The study was conducted by a team of researchers from Geisinger, Indiana University-Purdue University Indianapolis, the Indiana Clinical Translational Science Institute and the Regenstrief Institute.
The researchers sought to test machine learning models on electronic medical records for the detection of Alzheimer’s Disease and Related Dementia (ADRD) three years early.
The process involved using a dataset of 1,379 individuals with MCI, the second stage of ADRD, and 5,516 individuals without MCI, and 209 words or phrases “extracted from clinical notes.” The AI programs were trained using a variety of passive digital markers, and four AI algorithms were utilized for the study.
Additional factors included age at prediction time, gender and race. The markers were “extracted from electronic medical records collected within two years before the prediction date.”
While the programs were developed to identify ADRD, they can also predict MCI one year and three years in advance with varying degrees of accuracy, the researchers found. One program in particular, called Adaboot, used a “gradient boosting algorithm” and outperformed the remaining three evaluated in the study.
Among the top markers for 1-year MCI predictions were cognitive functions, prescription procedure, ear structure, brain and antidepressive agents. For 3-year MCI predictions, the markers included antidepressive agents, wounds and injuries, accidents, cognitive functions and influenza.
In order to increase the accuracy of the models, the study’s authors wrote there will be further leveraging of “structured” EMR data.
Earlier detection of MCI could open up new opportunities for memory care providers, as many view early-stage memory care services as a growth avenue, particularly for reaching consumers earlier in their lives — even before they make the move into senior living communities.
“When we talk about early-stage memory care, we need to begin to think about those other product lines that are in front of that continuum,” Juliet Holt-Klinger, a dementia expert who formerly held a senior role at Brookdale Senior Living (NYSE: BKD), said at the 2022 Senior Housing News BRAIN event. “We need to be talking about all of this in our ILs and our ALs and our home care companies and hospitals and doctors offices.”