Apple’s own validation reports a mean VO₂max error of +1.4 ± 4.7 mL/kg/min, against a projected reference
What does Apple say about Cardio Fitness accuracy?
The finding
Validity error was 1.2 ± 4.4 mL/kg/min in the design set and 1.4 ± 4.7 in the validation set. Test-retest reliability ICC was 0.86 (0.80–0.90).
- Authors
- Apple Inc.
- Published
- Using Apple Watch to Estimate Cardio Fitness with VO₂ max (white paper) · 2021
- Evidence
- Read in the primary source; numbers checked.
- Source
- www.apple.com/healthcare/docs/site/Using_Apple_Watch_to_Estimate_Cardio_Fitness_with_VO2_max.pdf
How it was studied
Manufacturer study on Apple Watch Series 4: design set of 534, validation set of 221 (mean age 55 ± 17; 38% over 65). The reference was a linear projection of submaximal heart rate and VO₂ to age-predicted maximum heart rate, not a measured maximum.
The numbers
Estimates range from 14 to 60 mL/kg/min and need an outdoor walk, run or hike on a grade under 5%. Apple’s Low Cardio Fitness alerts use FRIEND’s lowest quintile for ages 20–59, and 18 (men) or 15 (women) mL/kg/min from age 60.
What it does not show
Written by the vendor, and the reference is itself an estimate rather than a measured VO₂max.
What Longstanding does with it
Apple scores against the measured FRIEND norms; the app scores against the Cooper table. The same watch value can therefore read as low fitness in one place and fine in the other. We name the table on every result so you can tell which.