AI Model Flags Alzheimer's Risk Years Before Symptoms

Validation across more than 1.6 million patient records shows routine blood tests and records can signal Alzheimer's risk early.

Updated on Nov. 16, 2026 in Alzheimer’s

Photograph of a rack of clear glass sample vials holding pale amber liquid on a steel laboratory bench under cool clinical light.
Researchers at Israel's Clalit Health Services and Ariel University validated an AI model that flags Alzheimer's risk years before symptoms, using 1.6 million patient records. AI Illustration. Upload story photo >

Live Poll

Would you get a checkup if it could flag your Alzheimer's risk early?

ALZAI Health completed large-scale validation of an AI model that identifies Alzheimer's risk years before symptoms appear, using more than 1.6 million patient records from Israel and the United States. The model analyzes routine blood test results, existing diagnoses and demographic information without initially requiring brain scans or invasive procedures.

Why it matters

Alzheimer's disease can develop 15 to 20 years before noticeable symptoms, and early diagnosis creates an opportunity to slow its progression. Low-cost screening built on existing medical data could widen early detection as dementia cases are projected to exceed 150 million by 2050.

The validation drew on more than 1.6 million patient records from Israel and the United States, and Alzheimer's can develop 15 to 20 years before symptoms appear. The reported figures do not yet indicate predictive accuracy or whether earlier detection changes disease outcomes.

The players

ALZAI Health

A company established through a partnership involving Mor Research Applications, Ariel Scientific Innovation and its founders, listed on the TSX Venture Exchange in Toronto.

Clalit Health Services

Israel's largest healthcare provider and one of the country's four health funds, whose research infrastructure supported the study.

The details

The research originated at Clalit Health Services' Beilinson Hospital in Petah Tikva in collaboration with Ariel University, using Clalit's research infrastructure to examine hundreds of thousands of medical records. The machine-learning models were trained on extensive datasets to identify patterns linked to later Alzheimer's risk.

Timeline

  1. Alzheimer's biological changes may already be underway 15 to 20 years before symptoms appear.

  2. ALZAI Health was listed on the TSX Venture Exchange approximately two months ago.

  3. Dementia cases are projected to exceed 150 million by 2050.

Big Picture

The study extends Clalit Health Services' records-based research program at Beilinson Hospital, which uses routine health-system data for clinical research. Applying machine learning to that data to flag Alzheimer's risk years before symptoms marks a shift toward screening built on existing medical records.

Routine blood tests and existing medical records could one day be used to flag early Alzheimer's risk without brain scans. Readers should discuss family history and cognitive concerns with a clinician, since the tool's availability and accuracy for individuals are not yet established.

The takeaway

Early Alzheimer's detection may become possible through data already collected in routine care. Tracking changes in memory and thinking with a doctor remains the most practical step while such tools are still being validated.

Further reading

Explore more on Alzheimer's research and early detection efforts.

Live Poll

Would you get a checkup if it could flag your Alzheimer's risk early?