Patterns and Clinical Outcomes of Physical Activity and Sedentary Behavior Across 20 Million Days of Wearable Monitoring in U.S. Adults

BackgroundPhysical activity is a key modifiable determinant of health, yet current guidelines primarily emphasize moderate-to-vigorous physical activity and provide limited guidance on other actionable dimensions such as daily step count and sedentary time. The growing availability of wearables enables high-resolution measurement of these metrics and assessment of their associations with clinical outcomes. MethodsMinute-level wearable data from adult participants of All of Us between 2015-2023 were used to calculate step count and sedentary time. Patterns of step count and sedentary behavior were assessed across temporal scales, U.S. states, and sociodemographic groups. Phenome-wide association (PheWAS) and dose-response analyses were conducted to evaluate associations with incident disease. ResultsAmong 50,300 participants (68% female, 71% White, median age 55 years), 19,845,612 person-days of wearable monitoring over a median follow-up of 13 months were included in this analysis. Step counts peaked at midday with weekly highs on Saturdays and seasonal highs in May, while sedentary time was highest in midafternoon and winter. Higher step counts and lower sedentary time were observed among males, individuals with higher income and education, and residents of the Northeast, upper Midwest, and West Coast. In PheWAS (n=31,446), higher step count and lower sedentary time were associated with lower risk of multiple diseases, including obesity, cardiometabolic risk factors, and mood and anxiety disorders. Dose response analyses demonstrated heterogeneous relationships across disease groups, with cardiovascular benefits plateauing beyond 9,000-10,000 steps/day and below 600-640 minutes/day of sedentary time. ConclusionsIn this nationwide cohort of U.S. adults, wearable-derived physical activity and sedentary behavior showed distinct temporal and geographic patterns with significant disease-specific associations with clinical outcomes. These findings support the use of longitudinal high-resolution wearable data for advancing personalized guidelines of physical activity and sedentary behavior.