Which psychological, social and physical environmental characteristics predict changes in physical activity and sedentary behaviors during early retirement? A longitudinal study

BACKGROUND: In the context of healthy ageing, it is necessary to identify opportunities to implement health interventions in order to develop an active lifestyle with sufficient physical activity and limited sedentary time in middle-aged and older adults. The transition to retirement is such an opportunity, as individuals tend to establish new routines at the start of retirement. Before health interventions can be developed, the psychological, social and physical environmental determinants of physical activity and sedentary behaviors during early retirement should be identified, ideally with longitudinal studies. The aim of this paper was first to examine whether psychological, social and physical environmental factors at the start of retirement predict longitudinal changes in physical activity and sedentary behaviors during the first years of retirement. Second, moderating effects of gender and educational levels were examined. METHODS: This longitudinal study was conducted in Flanders, Belgium. In total, 180 recently retired (>1 month, <2 years at baseline) adults completed a postal questionnaire twice (in 2012–2013 and two years later in 2014–2015). The validated questionnaire assessed socio-demographic information, physical activity, sedentary behaviors, and psychological, social and physical environmental characteristics. Multiple moderated hierarchic regression analyses were conducted in SPSS 22.0. RESULTS: Higher perceived residential density (p < 0.001) and lower aesthetics (p = 0.08) predicted an increase in active transportation (adjusted R(2) = 0.18). Higher baseline self-efficacy was associated with an increase in leisure-time physical activity (p = 0.001, adjusted R(2) = 0.13). A more positive perception of old age (p = 0.04) and perceiving less street connectivity (p = 0.001) were associated with an increase in screen time (adjusted R(2) = 0.06). Finally, higher baseline levels of modeling from friends (p = 0.06) and lower perceived land use mix access (p = 0.09) predicted an increase in car use (adjusted R(2) = 0.06). A few moderating effects, mainly of educational level, were found. DISCUSSION: Walkability characteristics (perceived residential density) and self-efficacy at the start of retirement are the most important predictors of longitudinal changes in active transportation and leisure-time physical activity. Few moderating effects were found, so health interventions at the start of retirement focusing on self-efficacy and specific walkability characteristics could be effective to increase physical activity in recently retired adults. No firm conclusions can be drawn on the importance of the examined predictors to explain change in car use and screen time, possibly other factors like the home environment, or automatic processes and habit strength are more important to explain sedentary behaviors.