Bloodwork-free Early Screening for Alzheimer's Disease via Comorbid Pattern Recognition in Electronic Health Records

Early identification of Alzheimer's disease and related dementias (ADRD) remains limited by the need for specialized tests and late-stage diagnosis. The Zero-burden Risk Assessment (ZeBRA) is a AI-driven score that predicts incident ADRD up to a decade before diagnosis, using only routine electronic health record (EHR) data, without laboratory tests, imaging, or questionnaires. Trained on 487,989 cases and 12,483,718 controls from nationwide U.S. insurance claims and validated on held-back samples, and two independent cohorts, ZeBRA achieved AUC = 0.93 and 0.83 for predicting out to 1-year and 10-year horizons respectively, maintaining positive likelihood ratios (>10) at 95% specificity and stable discrimination over time (AUC drop ≈ 1 to 1.3% per year). Performance was consistent across age, sex, race, and ethnicity subgroups. In a limited prospective pilot, higher ZeBRA scores correlated with lower Montreal Cognitive Assessment (MoCA) scores, indicating a greater degree of cognitive impairment ( $R=-0.78$ ). Compared with prior EHR-based models, ZeBRA provides superior accuracy, cross-site generalizability, and demonstrates noise-corrected interpretability via our novel Λ-OR attrubution metric. Its scalability, low cost, and independence from specialized testing position ZeBRA as a practical tool for population-level early detection and presymptomatic trial enrichment.