Introduction: Leisure time physical activity in compliance with recommended levels is associated with improved health and lower mortality, but little is known on whether these physical activity habits are stable among adults and what characteristics predict physical activity changes. Our objective was to determine change in the levels of leisure time physical activity among adults during a period of 10 yr.
Methods: Detailed information on time spent on cycling, gardening, doing odd jobs, and sports from three measurement periods (1993-1997, 1998-2002, and 2003-2007) of the population-based Doetinchem Cohort Study was used to define being active: spending at least 3.5 h·wk−1 on moderate to vigorous physical activities, an approximation of the Dutch recommended level.
Results: Almost one-third (31.4%) of the population were active at all three points in time, 3.6% were inactive, and 45.0% of the participants changed their level of physical activity, almost equally distributed over decreasers, increasers, and varying. Not smoking (odds ratio (OR) = 1.47, 95% confidence limits (CL) = 1.14-1.89) and high socioeconomic status (OR = 1.43, 95% CL = 1.07-1.92) were associated with staying active. Inactive men (OR = 0.73, 95% CL = 0.57-0.94) had the highest risk of staying inactive, whereas good perceived health was associated with becoming active (OR = 1.49, 95% CL = 1.09-2.03).
Conclusions: The finding that, in a decade, almost half of the population changed from active to inactive or vice versa affects the interpretation of the long-term health effects of physical activity measured only once, and it stresses the importance of interventions not only in increasing physical activity levels but also in maintaining a physically active lifestyle.