[brandonb]

Yep—resting heart rate is a good proxy, and readily measurable from Apple Watch.

A definitive study would randomly assign people to an exercise routine, follow them for 20-30 years, and then measure hard outcomes like heart attacks, strokes, and all-cause mortality.

Since that's pretty challenging, all of the studies on exercise (and nutrition, for that matter) make two types of approximation:

1. analyze correlations in data rather than randomize, and/or

2. analyze proxy metrics like resting heart rate or blood pressure rather than hard endpoints.

That's part of what makes research on health so complex!