Sign in Subscribe
Diseases

Novel Predictors of Alzheimer's Disease in Down Syndrome Identified Using Machine Learning

Strategic Insights

— 1 min read

Researchers at Boston University School of Public Health have identified new predictors of Alzheimer's disease in individuals with Down syndrome using machine learning techniques. The study examined a cohort of adults with Down syndrome enrolled in Medicaid and/or Medicare between 2011 and 2019, and found that certain co-occurring conditions, such as epilepsy and hypothyroidism, significantly increased the risk of developing Alzheimer's disease. The machine learning model used in the study had a high area under the curve of 0.86 and a positive predictive value, indicating its ability to accurately identify individuals at risk of developing Alzheimer's disease.

Key Takeaways:

  • The study used machine learning techniques to identify predictors of Alzheimer's disease in individuals with Down syndrome.
  • The cohort had a mean age at entry of 44.6 years, with 46.2% of participants being male and 73.7% being white non-Hispanic.
  • 16,398 participants had incident Alzheimer's disease diagnoses over the study period.
  • The machine learning model had an area under the curve of 0.86 and a high positive predictive value.
  • Strongest predictors of increased probability of Alzheimer's disease were age, dual Medicaid/Medicare enrollment, incident epilepsy or incident ulcer three years before index date, any hypothyroidism, schizophrenia, or hyperlipidemia.
  • The study found a synergistic interaction between epilepsy and enrollment by age.
  • The predictors identified in the study align with known predictors in the general population and signal Alzheimer's disease symptom onset.
  • The research highlights areas for further etiologic inquiry and intervention.

Statistics:

  • 44.6 years: mean age at entry of the cohort.
  • 46.2%: proportion of males in the cohort.
  • 73.7%: proportion of white non-Hispanic participants in the cohort.
  • 16,398: number of participants with incident Alzheimer's disease diagnoses over the study period.
  • 0.86: area under the curve of the machine learning model.
  • 0.???: positive predictive value of the machine learning model.

Sources:

  • NewsRx. New Down Syndrome Study Results Reported from Boston University School of Public Health (Novel predictors of Alzheimer's disease in Down syndrome identified using machine learning). Managed Care Weekly Digest. October 27, 2025; p 57.
  • Journal of Alzheimer's Disease. Novel predictors of Alzheimer's disease in Down syndrome identified using machine learning. 2025:13872877251385423.

Read more