Photo: Kaylee Pugliese
In a study led by Honghuang Lin, PhD, professor of medicine and co-director of the Program in Digital Medicine at UMass Chan Medical School, researchers developed a dynamic prediction model for Alzheimer’s disease based on continually monitoring and updating information on cognitive functions.
Many risk prediction models rely on single-time measurements of risk factors. However, Alzheimer’s is a progressive neurodegenerative disorder, and single-time models may not effectively capture the dynamic changes in risk factors over time, which is critical for tailoring interventions as the disease progresses.
Drawing a parallel between a dynamic model for Alzheimer’s disease and existing models for predicting 10-year risk of cardiovascular disease, Dr. Lin explained, “Each year, you visit your doctor, and based on new data, your risk is updated, allowing you to adjust your exercise, diet or medication accordingly. Our project aims to determine whether incorporating additional cognitive measures can enable continuous updates to an individual’s lifetime risk of developing Alzheimer’s.”
The study, published in Alzheimer’s & Dementia, was supported by the National Institutes of Health, the Alzheimer’s Association and the American Heart Association.
Investigators studied participants in the Religious Orders Study (ROS) and Rush Memory Aging Project (MAP), collectively known as ROSMAP. These are ongoing longitudinal cohort studies initiated in 1994 and 1997, respectively. The analysis included 2,384 participants who exhibited no cognitive impairment at baseline and had at least one clinical evaluation.
The primary outcome of interest was the lifetime risk of developing Alzheimer’s disease. Clinical diagnostic assessments occur annually and involve a combination of cognitive testing, clinical evaluations and diagnostic classifications by clinicians adhering to accepted national criteria.
Cognitive assessment incorporated scores from five key domains: perceptual speed, visuospatial ability, episodic memory (word list recall, immediate and delayed recall, etc.), semantic memory, and working memory. Lin said that while these domains could be measured at different visits, regularly updating information across all domains over a person’s lifetime is important.
The model based on all five cognitive domains performed significantly better than models based on individual domains. Additionally, an increasing number of assessments further enhanced the model’s prediction power for identifying individuals at risk of developing Alzheimer’s before age 85 or 90.
And early intervention can make a difference.
“There are FDA-approved drugs that can help patients diagnosed in early stages of Alzheimer’s disease,” said Lin. “Usually the diagnosis of Alzheimer’s can take several months to a few years. But for people with high risk, they can benefit from more frequent medical evaluations and additional diagnostic testing to enable earlier detection and intervention.”
Enhancing disease prediction capabilities extends beyond Alzheimer’s, according to Lin. The model can also be applied to other scenarios, such as monitoring ICU patients for potential rapid changes in their condition by continually analyzing blood pressure, heart rate and other critical vital signs.
