Cardiovascular
Weighted markers include ApoB, LDL-C, blood pressure, and resting heart rate. These have strong links to atherosclerotic and vascular risk over time.
LifeIndex LifeIndex is a longitudinal scoring framework that synthesizes selected bloodwork and wearable signals into one healthspan score on a 500-900 scale.
Weighted markers include ApoB, LDL-C, blood pressure, and resting heart rate. These have strong links to atherosclerotic and vascular risk over time.
Weighted markers include HbA1c, fasting glucose, and selected body composition context where available. These reflect insulin dynamics and metabolic strain.
Wearable recovery signals include resting heart rate trend, HRV trend, and sleep consistency proxies where available in standardized export fields.
Wearable activity signals include training volume and consistency proxies from recent periods to capture movement behavior and conditioning exposure.
Cardiovascular, metabolic, recovery, and activity domains capture most modifiable, behavior-linked longevity risk signals available in routine practice.
This structure aligns with practical preventive medicine models and enables a single framework that clinicians and patients can revisit every retest cycle.
Selected references:
The score range is 500-900, not 0-100, to avoid false precision at the extremes and to mirror realistic human variation in a health-data-rich population.
Domain scores are normalized and combined using a weighted average. Weights are set by signal reliability and expected longitudinal value in current model scope.
LifeIndex does not currently include every clinically relevant signal, and it does not diagnose disease states or predict specific events.
The score should not be used as a standalone medical decision engine. It is a structured measurement layer to support longitudinal interpretation.
Trajectory is prioritized over snapshot. A score that moves from 700 to 730 in six months is often more clinically meaningful than a static high score with no progress.
Peer percentile is provided for context, but the primary use case is change within the same individual across time windows.
Missing inputs reduce confidence and can suppress domain resolution. The model still returns a score when minimum viable fields are present.
Next versions will expand supported biomarkers, refine weights using larger cohorts, and improve handling for sparse or irregular data coverage.
We are evaluating additional signals now, but only promote them once they meet repeatability and interpretability standards.
Explore the clinic partnership track