Sleep, Activity, and Healthspan: What 4 Years of Wearable Data Reveal

Healthspan — the years lived in good function rather than mere survival — has become the central concept in modern longevity research. Two of its biggest levers are sleep and physical activity. A new analysis tells us how they interact.

The dataset

A Stanford-led consortium aggregated four years of de-identified wearable data from 1.2 million users, paired with electronic health records. The result is among the largest observational datasets ever assembled on consumer-grade physiological signals.

The key finding

The relationship between sleep and activity is not linear and not independent. Three loops emerged that predict 10-year cardiovascular outcomes more strongly than either variable alone:

1. Stable circadian-activity coupling — when bedtime and peak-activity windows are consistent week to week, predicted CV risk drops 18% versus chaotic patterns at the same total volume.
2. Activity timing — moderate activity 8–11 hours after waking is associated with deeper slow-wave sleep, which in turn predicts lower inflammatory markers.
3. Recovery balance — the ratio of high-intensity activity to deep-sleep minutes shows a U-shaped relationship with healthspan; both too little and too much carry risk.

What this changes

The clinical implication is that fitness advice optimized in isolation — "do X minutes of cardio" — leaves substantial benefit on the table. The same minutes timed differently, slept differently, can have markedly different long-term outcomes.

What the consumer can do

Two practical changes: stabilize your bedtime within a 30-minute window, and place your hardest activity in the late morning or early afternoon, not late evening. Both are within reach of most people without lifestyle upheaval.