BrainCheck® Population Analytics FAQ
What BrainCheck Population Analytics Does
What problem does Population Analytics solve that health systems can't solve today?
It systematically surfaces the patients in a panel who show research-derived data patterns associated with potential cognitive impairment but haven't yet been identified. Most systems today either don't proactively identify these patients or rely on ad hoc chart review. Population Analytics turns identification into a repeatable, prioritized population sweep for outreach planning.
What does the product do?
It analyzes structured EHR data to identify patients whose records contain research-derived patterns associated with potential cognitive impairment, helping health systems prioritize cognitive care outreach and clinical review. It supports population health management and care coordination.
Equally important, what does it not do?
It does not diagnose Alzheimer's, MCI, dementia, or any other condition. It does not generate a clinical risk score, recommend treatment, or replace clinician judgment. All diagnostic and treatment decisions remain the responsibility of the treating clinician.
What exactly does a health system receive?
A population-level dashboard report first, followed by a prioritized patient registry organized by outreach priority.
How BrainCheck Population Analytics Works
What structured EHR data is being analyzed?
Routinely collected structured fields already in the record: diagnosis codes, medications, encounter and utilization patterns, labs, and demographics including age. No new data collection or patient interaction is required, and the analysis does not use medical images, waveforms, or physiological signals.
How are patients organized into cohorts?
Patients are organized into transparent routing segments: patients with an existing dementia diagnosis (watchlist / care plan) and undiagnosed patients grouped by the data signals present in their records, ordered for outreach prioritization. A signal-count layer shows how many independent signals each patient's record contains. There is no numerical risk score and no clinical risk tier — grouping is signal-based by design.
Can clinicians understand why a specific patient appears in the registry?
Yes. Every patient's placement traces back to specific, named EHR data signals, so a clinician can independently review why a patient appears rather than relying on an opaque output. That signal-based transparency is a deliberate design choice — and, because there's no score to defer to, it's exactly what lets the tool remain a support for clinician judgment rather than a substitute for it.
How much implementation effort is required?
Very little on the customer's side. We ingest an existing structured EHR export; there's no new workflow to stand up on their side. The setup lift (data mapping, oversight) sits with us. Contrast this explicitly with the build-it-yourself path, which can take ~30 hours of setup, months of infrastructure, ~$300k+ in FTE + maintenance.
How is this different from a standard analytics dashboard?
The output ends in an action list, not just metrics. A dashboard shows the state of a population; Population Analytics hands the care team a specific list of patients to reach out to first, each connected to a concrete next step: outreach, then clinician-directed assessment, then care planning.
Validation & Clinical Use
How has the methodology been validated?
It's grounded in peer-reviewed longitudinal EHR research, including the Vanderbilt-led phenome-wide study (153M+ records across two independent datasets, identifying reproducible phenotypes that precede Alzheimer's diagnosis). Prospective pragmatic validation of the BrainCheck assessment is currently underway in a real-world clinical health system setting.
What evidence supports the approach?
Multiple peer-reviewed studies show routinely collected structured EHR data can identify previously unrecognized cognitive impairment. The Vanderbilt study identified 406 phenotypes (70+ replicated in an independent dataset); eRADAR is externally validated across large real-world systems with AUC up to 0.84 for previously undiagnosed dementia.
How should clinicians interpret the results?
As a population-health prioritization tool, not a diagnostic test or screening result. Flagged patients have EHR patterns that may warrant priority clinical consideration and must undergo appropriate clinical review before any diagnostic or treatment decision is made.
