How this is calculated
Health Age is a summary of four habits on an age scale, from published studies. It is calculated on your device, and the math is printed below in full. It is not a medical test. It does not look at any medical condition.
The formula
difference in years = (1 / G) × sum over the four habits of ( w × ln HR ), G = 0.1
Health Age = your age + difference in years, and never below 18
For each habit, HR is the ratio its study reports for your 90-day average against that habit's reference level, ln is the natural log, and w is the habit's weight. A habit's share in years is w × ln HR ÷ G. A negative total reads younger than your age; a positive one reads older.
The constants
| Habit | Weight w | Ratio HR | Where it does not move |
|---|---|---|---|
| Sleep, hours asleep a night | 0.8790 | 1.20 for each hour under 6, 1.34 for each hour over 9 | 6 to 9 hours |
| Resting heart rate, bpm | 0.7066 | 1.08 at 70 bpm and 1.33 at 90 bpm, in straight lines on the ln scale from 60, between and past them | 60 bpm or below |
| Cardio minutes a week | 0.6950 | 0.76 at 245, 0.71 at 418 and 0.68 at 500 minutes, in straight lines on the ln scale from 115 | Under 115, and 500 or more |
| Steps a day | 0.4773 | 0.88 for each 1,000 steps above 8,000, and 1 ÷ 0.88 for each 1,000 below | 9,000 or more |
The best possible total is 3.29 years younger: resting heart rate at 60 bpm or below, 9,000 steps a day or more, 6 to 9 hours of sleep and 500 cardio minutes a week or more. That is the cap the result shows. The study's own worked example, 60 bpm, 9,000 steps, 7 hours of sleep and 420 cardio minutes, comes to 2.9978 years younger, so an age of 29 reads 26.0.
Cardio minutes
One point for each minute of moderate effort and two for each minute of hard effort, added up over a week. Apple Watch Exercise minutes count every brisk minute once, so they are the floor of this number; hard minutes earn their second point on top.
The same math as code
This page runs these functions in JavaScript (health-age.js). The reference version, in Python with the natural log:
from math import log
W = {"sleep": 0.8790, "rhr": 0.7066, "cardio": 0.6950, "steps": 0.4773}
G = 0.1
def ln_hr_rhr(bpm): # resting heart rate, beats per minute
if bpm <= 60: return 0.0
if bpm <= 70: return (bpm - 60) / 10 * log(1.08)
return log(1.08) + (bpm - 70) / 20 * (log(1.33) - log(1.08))
def ln_hr_steps(steps): # mean steps per day; plateau at 9,000
return (min(steps, 9000) - 8000) / 1000 * log(0.88)
def ln_hr_sleep(hours): # mean hours asleep per night
if hours < 6: return (6 - hours) * log(1.20)
if hours <= 9: return 0.0
return (hours - 9) * log(1.34)
def ln_hr_cardio(minutes): # cardio minutes per week; plateau at 500
if minutes < 115: return 0.0
if minutes < 245: return (minutes - 115) / 130 * log(0.76)
if minutes < 418: return log(0.76) + (minutes - 245) / 173 * (log(0.71) - log(0.76))
if minutes < 500: return log(0.71) + (minutes - 418) / 82 * (log(0.68) - log(0.71))
return log(0.68)
def part_years(name, ln_hr): # one habit's share, in years
return W[name] * ln_hr / G
Sources
- Lee, Esmaeilpour et al., arXiv:2611.08568 (2026): the formula and the weights.
- Spiegelhalter, “How old are you, really?”, BMC Med Inform Decis Mak 16:104 (2016): the age-scale method.
- Zhang, Shen, Qi, CMAJ 188(3):E53 (2016): resting heart rate.
- Master et al., Nature Medicine 28:2301 (2022): steps.
- Daghlas et al., JACC 74(10):1304 (2019): sleep duration.
- Kany et al., Circulation 150(16):1236 (2024): activity minutes.