The Wearables Guide: What to Track, What to Ignore, and How to Choose
The consumer health wearable market has exploded — Oura Ring, Whoop, Garmin, Apple Watch, Continuous Glucose Monitors, smart scales, and dozens of other devices all promise to transform health through data. The reality: some of this data is genuinely valuable; some is noise marketed as signal; and some devices provide information that, without the right interpretive framework, can worsen rather than improve health decisions.
- The highest-value wearable data for longevity is: overnight HRV (autonomic nervous system balance, recovery quality), resting heart rate trends (cardiovascular fitness trajectory), sleep consistency (total sleep time and wake time regularity), and daily step count and activity level. These metrics are reasonably accurate and useful as trends. Overnight blood oxygen readings can hint at breathing problems but cannot diagnose sleep apnea.
- The lowest-value wearable data (despite high marketing prominence): calorie expenditure estimates (in one test of seven devices, none had an error below 20 percent), sleep stage classification (roughly half agreement with lab studies for deep sleep), and stress scores derived from HRV algorithms (useful as trend indicators, not absolute measures). Do not make specific behavioral changes based on single-day values of these metrics.
- A continuous glucose monitor (CGM) shows how your glucose responds to meals, exercise and sleep, which no other consumer device can capture. It is essential for many people with diabetes; for people without diabetes, the insights can be interesting, but there is little evidence that using one improves health outcomes.
- Choosing a wearable: prioritize battery life (daily charging interrupts sleep monitoring consistency), sensor accuracy validation (published peer-reviewed validation against clinical gold standards), and whether the device measures what you actually want to track. Oura Ring and Whoop have the most published academic validation for sleep metrics; Garmin and Apple Watch for activity and VO2 max estimation.
- The data without action is not valuable. The most important wearable habit is weekly review of trends — not daily checking of scores. A 4-week average HRV trend, a monthly resting heart rate trajectory, and a weekly sleep consistency score are more actionable than daily fluctuations that primarily reflect normal biological variation.
Consumer health wearables have fundamentally changed what is possible in personal health monitoring. For the first time, individuals can track continuous physiological data — heart rate variability, sleep staging, blood oxygen, activity, and in some devices glucose — outside of clinical settings, across days and weeks, in the context of real life. This is genuinely transformative for longevity medicine. It is also generating a new category of health anxiety driven by imprecise metrics, and a market that is not uniformly incentivized to accurately represent what its devices can and cannot measure.1
What Wearables Measure Well
Resting heart rate: Accurately measured from optical HR sensors, and a useful marker of fitness: lower resting heart rate is linked to lower mortality in population studies. Declining resting heart rate over months of aerobic training is one of the clearest indicators of improving cardiovascular fitness. A resting heart rate consistently below 60 BPM reflects high parasympathetic tone and cardiorespiratory fitness. Heart rate variability (RMSSD): Reasonably accurate from good-quality optical sensors measured overnight. The most valuable wearable metric for recovery monitoring and autonomic nervous system (see our heart rate variability article). Sleep/wake detection: Good at detecting sleep but much weaker at detecting time awake in bed, so total sleep can be overestimated. Reliable for tracking sleep timing consistency. Step count and activity: Reasonably accurate from accelerometers. Daily step counts and activity minutes are reliable and directly actionable metrics with strong epidemiological longevity associations.2
What Wearables Measure Poorly
Sleep staging (deep sleep, REM): As covered in our sleep tracking article, consumer devices agree with lab sleep studies only about half the time for deep sleep in validation studies. Individual night deep sleep values are unreliable. Weekly averages are more stable. Calorie expenditure: In a test of seven wrist-worn devices, heart rate was generally accurate (six devices had median errors below 5% during cycling), but no device estimated energy expenditure with less than 20% error, and errors were larger for men, people with higher BMI, people with darker skin and during walking. That is enough to mislead dietary decisions. Do not eat according to wearable calorie estimates. Stress scores: Composite stress algorithms from HRV and accelerometry are directionally useful for trend identification but not precise enough for day-to-day behavioral decisions. Blood glucose (non-invasive): No current consumer wearable accurately measures blood glucose from the wrist or similar optical approaches; all consumer glucose monitoring still requires a subcutaneous sensor (CGM).
The CGM: A Tier Above Other Wearables
The CGM is different from other consumer devices because it provides information no other consumer device can: how your glucose responds to specific meals, sleep and exercise. People’s blood sugar responses to identical meals vary widely.3 Food order effects, post-meal glucose responses, sleep quality effects on fasting glucose, and exercise timing effects are all invisible to standard bloodwork and fully visible to CGM. For people without diabetes, though, there is little evidence yet that wearing a CGM improves long-term health, and standard tests such as fasting glucose and HbA1c remain the basis for diagnosis.3
References
- 1Piwek L, et al. "The Rise of Consumer Health Wearables: Promises and Barriers." PLoS Med. 2016;13(2):e1001953. PubMed · DOI
- 2Shcherbina A, et al. "Accuracy in Wrist-Worn, Sensor-Based Measurements of Heart Rate and Energy Expenditure in a Diverse Cohort." J Pers Med. 2017;7(2). PubMed · DOI
- 3Zeevi D, et al. "Personalized Nutrition by Prediction of Glycemic Responses." Cell. 2015;163(5):1079-1094. PubMed · DOI
