Could be that the healthier are more likelier to run. Correlation is not causation.
Exactly. Confirmation bias.