Exploring novel determinants of exercise behavior: a lagged exposure-wide approach

Many middle-aged to older adults do not engage in regular exercise at all, despite its importance for healthy aging. Extensive research grounded in behavioral and social science theories has identified numerous determinants of exercise. However, few studies used an exposure-wide approach, a data-driven exploratory method particularly useful for identifying novel determinants.

We used data from 13 771 participants in the Health and Retirement Study, a diverse, national panel study of adults aged >50 years in the United States, to evaluate 62 candidate determinants of exercise participation. Candidate predictors were drawn from the following domains: health behaviors, physical health, psychological well-being, psychological distress, social factors, and work. We used Poisson regression with robust error variance to individually regress exercise in the outcome wave (t2: 2014/2016) on baseline candidate predictors (at t1: 2010/2012) controlling for all covariates in the previous wave (t0: 2006/2008).

Some physical health conditions (eg, physical functioning limitations and lung disease), psychological factors (eg, health mastery, purpose in life, and positive affect), and social factors (eg, helping others, religious service attendance, and volunteering) were robustly associated with increased subsequent exercise. Among factors related to psychological distress, perceived constraints stood out as a factor in reducing exercise.

We identified potentially novel exercise determinants, such as helping friends/neighbors/relatives, religious attendance, and volunteering, that have not been captured using a theory-driven approach. Future studies validating these findings experimentally in midlife and older adults are needed.