Wearable data can meaningfully estimate parts of health related to future healthspan—especially movement, fitness, sleep regularity, resting cardiac patterns, and resilience—but it cannot directly measure how many disease-free years a person has left. The strongest use is longitudinal risk and behavior context, not a personalized countdown.
Editorial note
Published 2026-07-16. Last reviewed 2026-07-16. Evidence, interpretation, and Metric’s implementation are identified separately below.
Define healthspan before estimating it
Healthspan usually means the period of life spent in good health or free from specified disease and disability. That definition requires an outcome and a time horizon. A watch observes current signals; it does not observe future diagnoses, future function, or the exact date health will change.
What wearables observe well
- Movement volume and pattern, including steps, walking cadence, active time, and sedentary periods.
- Workout frequency, duration, route, pace, and heart-rate response when a session is recorded.
- Resting and overnight heart trends, subject to sensor and sampling limits.
- Sleep timing and duration estimates, with less certainty for exact stage classification.
- Cardio-fitness estimates during eligible activities.
- Longitudinal consistency and recovery after disruption—signals that a one-time clinic measurement may miss.
What wearables usually miss
- Many laboratory, imaging, genetic, molecular, and organ-specific measurements.
- A complete medical history, examination, medication effect, symptom meaning, and diagnostic context.
- Social connection, environment, access to care, occupation, income, and other major determinants unless separately reported.
- Accurate nutrition intake without consistent logging or another data source.
- Future exposures, injuries, infections, and treatment changes.
What the evidence shows
Wearable activity contains real risk information
A model-development study using UK Biobank data and external NHANES evaluation found that device-derived daily steps and walking cadence produced modest improvements in five-year mortality prediction beyond traditional risk factors. The improvement was measurable but small overall, which is more informative than claiming that a step count determines longevity.
Patterns can be modeled as biological-age proxies
A 2021 study trained a model on week-long activity tracks and evaluated wearable-derived biological-age acceleration and resilience in large datasets. It showed that wearable behavior contains age- and morbidity-related structure. It did not turn a consumer device into a direct measure of all biological aging.
Individual metrics have outcome associations
Cardiorespiratory fitness, step patterns, and sleep regularity have each been associated with later outcomes in large cohorts. Those studies help prioritize useful trends, but observational associations can reflect baseline health and other confounding factors. They do not prove that a dashboard’s composite score is itself an intervention target.
Three reasonable levels of claim
- Measurement: “The device estimated 7,800 steps and 7 hours of sleep.”
- Pattern: “Activity has been lower and sleep timing less regular than this person’s recent baseline.”
- Risk proxy: “Patterns like these have been associated with future outcomes in defined cohorts.”
A direct forecast such as “you have 31 healthy years remaining” would require much stronger assumptions, a specified definition of health, a time-to-event model, broad clinical inputs, calibration, and an appropriate population. The precision of the sentence should never exceed the evidence underneath it.
How Metric uses healthspan language
Metric organizes authorized Apple Health and logged data around modifiable domains that matter to long-term health: cardio fitness, heart trends, activity, workouts, body composition, blood pressure, sleep, and nutrition. Its biological-age and aging-ratio views summarize selected inputs. They are not presented here as a prediction of disease-free years, lifespan, or a date of future decline.
How to use wearable data meaningfully
- Prioritize multi-week trends with stable wear and source settings.
- Choose metrics with an understandable measurement process and an action you can take safely.
- Use deviations to ask questions, not to diagnose their cause.
- Combine wearable trends with conventional measurements and appropriate care where relevant.
- Evaluate whether the behavior change improves the underlying metric and how you feel—not only a composite score.
Limitations
- Most outcome evidence is observational, so reverse causation and residual confounding remain possible.
- A week of device data may not represent a person’s usual year.
- Consumer algorithms, hardware, and firmware change over time.
- Cohort-level performance does not guarantee useful individual prediction.
- Wearable access and consistent wear can introduce selection bias into research and product models.
Sources and further reading
- Aging (2021): Deep longitudinal phenotyping of wearable sensor data reveals independent markers of longevity, stress, and resilienceDeveloped a wearable-activity-based biological-age-acceleration model and evaluated it in large observational datasets.
- The Lancet Healthy Longevity (2025): Predictive performance of wearable sensors for mortality risk in older adults: a model development and validation studyFound that step count and walking cadence produced modest improvements beyond traditional risk factors in two cohorts.
- JAMA Network Open (2018): Association of Cardiorespiratory Fitness With Long-term Mortality Among Adults Undergoing Exercise Treadmill TestingIn 122,007 referred adults, higher treadmill-measured cardiorespiratory fitness was associated with lower long-term all-cause mortality.
- Sleep (2024): Sleep regularity is a stronger predictor of mortality risk than sleep duration: A prospective cohort studyUsed accelerometer-derived sleep regularity in 60,977 UK Biobank participants and evaluated subsequent mortality associations.
- Medicine & Science in Sports & Exercise (2023): Development and Validation of a Machine Learning Wrist-worn Step Detection Algorithm with Deployment in the UK BiobankDeveloped a free-living wrist step algorithm and evaluated step-count associations with cardiovascular and all-cause mortality.
- Apple Support (2025): Track your cardio fitness levelsExplains when and how Apple Watch estimates VO2 max and the personal data and workout conditions used.
- Apple Support (2026): Get the most accurate measurements using your Apple WatchDescribes fit, motion, skin perfusion, tattoos, calibration, workout choice, and profile details that affect measurements.
Continue exploring
Health information disclaimer
Metric is a wellness product. This article is educational and does not provide medical advice, diagnosis, or treatment. Wearable measurements and app-generated scores are estimates; discuss symptoms, unusual readings, medication effects, and changes to a care or training plan with an appropriate healthcare professional.