Possible Reverse Causation and Confounding in Study of the Association of Sedentary Behavior With Cancer Mortality.