Changes in Social Integration Predict Changes in Physical Activity: A 25-Year Prospective Study

Background: Social relationships are among the strongest predictors of health. The extent to which one is embedded in social networks, also referred to as social integration, has been associated with physical activity in short-term longitudinal and cross-sectional studies. How changes in social integration impact physical activity over longer periods of time is not well understood. Methods: Longitudinal data from 5 waves of the Americans’ Changing Lives study were used (wave 1: n = 3617; wave 5: n = 1427). Data were modeled using latent growth curves to understand the trajectories of social integration and physical activity, separately. Latent interlocking growth curve methods were used to understand if and how changes in social integration predicted changes in physical activity. Results: Physical activity did not change over the 25-year period (P = .68). Social integration significantly decreased (P = .025). Changes in social integration predicted changes in physical activity (β = 0.12, P < .05). Conclusions: Changes in how often adults speak to family and friends predict changes in physical activity across a 25-year period. Group- and social network–focused physical activity research, advocacy, and interventions are warranted.