HealthPartners Institute researchers have published new findings in Alzheimer’s & Dementia: Diagnosis, Assessment & Disease showing that artificial intelligence analytical methods can accurately identify predictors of Alzheimer’s disease prevalence at the county level.

Building on earlier research focused on individual-level risk factors, researchers examined the potential influence of population-level driving forces on disease prevalence. The AI model was trained using publicly available data, with 45 predictor variables that included culture, social, physical and economic environment, health risk factors and chronic disease.

The model efficiently reduced the number of predictor variables to 20 and explained 75% of the variance in Alzheimer’s disease. Researchers found that upstream factors, such as racial and ethnic minority status, sleep patterns, depression management and social association were the most important predictors for Alzheimer’s disease prevalence across communities.

“Our findings show both community conditions and personal circumstances can contribute to Alzheimer's risk,” said Nico Pronk, PhD, president and chief science office of HealthPartners Institute and lead investigator of the study. “Using AI research methodologies allows us to uncover new insight and directs us toward more effective prevention and healthier futures for our communities.”

Alzheimer’s disease is the leading cause of dementia in the US, impacting more than six million individuals. The disease has higher prevalence rates among underserved groups such as Black and Hispanic populations.