Our Testing Methodology

Wearable Testing Protocols

Heart Rate Accuracy

All wearables are tested against a validated 12-lead ECG (GE CASE Cardiac Assessment System) during three standardized exercise protocols: (1) 20-minute ramp test on stationary bike from rest to 90% max HR, (2) 30-minute steady-state moderate intensity (Zone 2), (3) 4x4 minute high-intensity intervals (Zone 4-5). Mean absolute error (MAE) is calculated from 60-second averaged readings paired with 12-lead ECG reference at each minute.

GPS Accuracy

GPS accuracy is tested on a certified 5K road course (UK Athletics certified) with reference distance accurate to ±1cm. Track error (lateral deviation from actual course) is calculated by overlaying GPS trace against the certified course boundary. Urban testing performed in a standardized city route with mixed building heights. Each watch completes the course minimum 3 times; results are averaged.

Sleep Staging Accuracy

Where resources allow, sleep staging accuracy is validated against clinical polysomnography (PSG) at an accredited sleep laboratory. PSG scoring is performed by a certified sleep technician blinded to wearable data. Sleep stage agreement is calculated as the percentage of epochs correctly classified versus PSG gold standard.

Battery Life

Battery life is measured in standardized "real-world" conditions: continuous heart rate monitoring enabled, Bluetooth connected to phone, notifications enabled, GPS inactive during non-workout hours. GPS battery is tested separately during continuous outdoor activity. Time to 20% battery (alert threshold) is also recorded.

Health App Testing Protocols

Calorie Tracking Accuracy

Calorie tracking apps are tested using 40 standardized meals prepared to 0.1g precision using calibrated scales. Reference caloric values are calculated from USDA FoodData Central using weighted nutritional data. All meals are photographed under standardized lighting (6000K LED panel, 45cm distance) and logged using each app's photo recognition feature. Mean absolute error (MAE) and root mean square error (RMSE) are calculated.

App Privacy Assessment

Our 25-point privacy checklist evaluates: (1) data categories collected, (2) third-party data sharing, (3) data encryption standards, (4) HIPAA compliance posture, (5) data deletion provisions, (6) law enforcement data request policy, (7) data breach history. Apps scoring below 15/25 receive a privacy warning in their review.

Medical Device Testing Protocols

Blood Pressure Monitors

Blood pressure monitors are compared against a mercury sphygmomanometer reference (used as gold standard despite phase-out in clinical settings) and a validated Omron M7 intelli IT clinical reference. Testing follows AAMI SP10 protocol: 5 measurements per session, minimum 3 sessions, comparison against simultaneous reference measurements. Device passes if ±5 mmHg systolic and ±5 mmHg diastolic mean error against reference.

Pulse Oximeters

SpO2 accuracy is tested using a Nellcor PM10N hospital-grade pulse oximeter as reference. Concurrent readings are taken at rest and during light activity across 20 paired measurements per device. Devices achieving ±2% SpO2 at saturations above 90% and ±3% at saturations 70-90% are considered clinically acceptable.

Editorial Standards

All products are purchased at retail or borrowed for review (noted in the review). No payment is accepted for reviews or rankings. Review scores reflect the independent assessment of the reviewing editor. Scores are not adjusted based on advertiser relationships — we have none.

Updates: Reviews are updated when significant firmware/software changes are released that materially affect the tested metrics, or when new competitive alternatives change the relative ranking recommendation. The publication date of original review and date of most recent update are both shown.

Corrections: We publish corrections prominently when errors are identified. If you believe a review contains a factual error, contact us at corrections@health-tech-reviews.com.