Strata combines several recorded measures into a directional biological-age estimate. The headline number is only one part of the view. Read the contributing components, partial or missing inputs, and displayed uncertainty before treating a change as meaningful.
Start with the contributing inputs
The model has components for VO2 max, resting heart rate, body composition, sleep, activity, cognitive check-ins, emotional check-ins, HRV, and a bloodwork-based PhenoAge component when applicable. Availability depends on the records and the component's requirements.
An unavailable component is not a measurement of zero. A partial component may use incomplete inputs with reduced confidence. Open component detail to see what is present rather than assuming every area contributed equally.
Read completeness beside the estimate
The completeness view distinguishes contributing, partial, and missing components. A headline number assembled from fewer inputs needs that context. Adding records can change the model's output without demonstrating that your health changed by the same amount.
Strata's displayed interval expresses uncertainty within its model. It does not capture every possible source of error or prove clinical accuracy. A narrow interval should not be presented as a validated assessment of your true biological age.
Distinguish a research component from the combined model
The bloodwork component draws on PhenoAge research. The original work distinguishes a clinical phenotypic-age measure from a DNA-methylation-based measure. Strata's combined estimate should not be described as a DNA-methylation test simply because it includes a related clinical component.
Research supporting one component does not independently validate the entire combination used in an app. Read the model as a product estimate and preserve that boundary when discussing the result.
Reference: Levine et al. (2018): An epigenetic biomarker of aging for lifespan and healthspan
Review changes with their inputs
When a trajectory changes, inspect which records or components changed too. A newly available bloodwork panel, different coverage, or corrected measurement can affect the estimate. Separate those changes from claims about an intervention.
The feature is not shown for child profiles. For adults, use it as a directional way to explore recorded inputs, not a diagnosis, lifespan prediction, or reason to alter treatment.
- Read the estimate and displayed uncertainty.
- Inspect completeness and partial inputs.
- Open the contributing component details.
- Review input changes alongside the trajectory.
An illustrative example
A number needs its inputs
The mockup shows a deliberately fictional estimate and a mixture of contributing, partial, and missing components. It is not a computed result, a validated model output, or a claim that the example person has become younger.
Keep this in mind
The components and their limitations are part of the estimate, not fine print beneath it.