Long-term follow-up of physical activity behavior in older adults.

OBJECTIVE To examine the contribution of social-cognitive factors (self-efficacy and affect) in predicting long-term physical activity in a sample of older adults (N=174). DESIGN A prospective design assessed physical activity and psychosocial variables at 2 and 5 years following a 6-month randomized, controlled exercise trial. MAIN OUTCOME MEASURES The primary outcome variable was self-reported physical activity, with previous behavior, self-efficacy, and affect assessed as determinants of physical activity. RESULTS Covariance modeling analyses indicated that physical activity at Year 2 was the strongest predictor of physical activity at 5-year follow-up. Both self-efficacy and affect at Year 2 were also associated with physical activity at Year 5, as was original treatment condition. Variables accounted for 35% of the variance in Year 5 activity. CONCLUSION Older adults with higher levels of physical activity, more positive affect, and higher self-efficacy at Year 2 were more likely to continue to be active at Year 5. This study is one of the longest follow-ups of exercise behavior in older adults and has implications for structuring environments to maximize the maintenance of physical activity.