A Prospective Two-stage Cognitive and Blood-based Screening For Alzheimer’s Risk In A Community Population

Background: Early Alzheimer’s disease (AD) detection is limited by costly, specialist-dependent diagnostics. Scalable, community-based screening (CBS)
overseen by primary care physicians (PCPs) is needed to improve care pathway efficiencies. Previously, a two-stage model combining brief cognitive
assessment with targeted blood-based biomarker (BBM) testing demonstrated strong retrospective performance in the Bio-Hermes study.1 Here, we
prospectively evaluated the feasibility and utility of this approach in a real-world community setting.

Methods: The first 453 older adults (55+) in The Villages®, FL were enrolled in the CARE-AD study 1/12/26-3/26/26 through local PCP clinics. Participants
completed a self-report health/lifestyle survey, the Cognivue Clarity® digital cognitive assessment with integrated Cognivue Amyloid Risk Measure
(CARM),2 and the LucentAD® Complete multi-analyte blood test which combines pTau217, Ab42/40 ratio, NfL, and GFAP into a single score.3 A
prospective decision-tree model utilized cognitive performance as the initial triage, with BBM testing applied as a secondary adjudication step to refine
amyloid risk into actionable categories: Non-AD dementia, AD Dementia, Probable-Mixed Dementia, and Preclinical AD. Participant perceptions of utility
and value were also assessed.

Results: Mean age was 73.4. Using the 2-stage paradigm, 24.7% (n=112) were identified as cognitively impaired via Clarity. For this sub-cohort, the BBM
provided biological clarification: 41 (~37%) were double-positive, representing a high-priority group for potential expedited referral. Conversely, the BBM
identified 71 impaired individuals (63%) as amyloid-negative; given the 96% NPV, amyloid pathology could be ruled out, potentially reducing the
specialty referral burden by >60%. Although the intended clinical workflow targets only the impaired, this study’s design also identified 21.4%(n=97) as
Preclinical AD (cognitively normal/BBM-positive), offering a critical window for future intervention. Finally, 54.4% of participants received high-value
assurance as both cognitively and biomarker normal, contributing to an overall experience rating of 9.7/10.

Conclusion: This study demonstrates the feasibility of a two-stage CBS model to optimize resource allocation in a community screening context. By
providing biological clarification of cognitive impairment at the local level, the model prioritizes high-probability patients for potential specialty referral
while offering meaningful assurance to the majority of participants. This scalable approach addresses specialist backlogs and enables earlier intervention
when therapeutic impact is greatest.