Teun Remmers

Investigating determinants of children's outside play | 51 Background A natural and cheap way of increasing children’s physical activity (PA) is by stimulation of outside play, defined as play behavior without any given tasks or goals; unstructured free play. Child outside play (OP), as a specific type of PA behavior, has been shown to increase a child’s total PA level (1, 2) and children who spent more time outdoors were shown to be more active than children who spent less time outdoors (3, 4, 5). Recommendations from an expert committee on the prevention of childhood obesity acknowledge unstructured play as most appropriate to increase PA in young children (6). OP is also positively associated with children’s social skills as they learn to account for other children (7, 8, 9) and provides children with opportunities to acquire new motor skills such as climbing, jumping, hanging and sliding in a self-regulatory way (10). Studies have shown that relatively minor adjustments to school playgrounds lead to increases in PA, which makes OP relatively modifiable (11). In order to enable effective increasing of child’s engagement in this promising type of PA behavior, determinants of this behavior should be assessed. Previous studies mainly focused on determinants of PA in general. These studies found that gender (i.e. male), child’s PA enjoyment, and summer and spring season were consistently related with higher levels of PA in children (12). There may be specific determinants for OP as a specific type of children’s PA behavior, however studies focusing on determinants of OP specifically are scarce. Currently, two studies have specifically examined determinants of OP, also investigating attributes of the physical environment (PE). A cross-sectional study reported that lower parental education and the importance parents ascribe to OP were associated with more OP in 4- to 12-year-old children (13). A longitudinal study reported that children’s outdoor tendencies and parental encouragement were related to increases in OP across a time span of five years (14). Both studies indicate that family environment may play an important role in child’s duration of OP. Studies that incorporate a longitudinal design and use a broader range of environmental variables (e.g. parent perceived safety) to identify determinants of OP are warranted (14). Therefore, the present study examined whether the family environment and perceived PE are associated with the duration of children’s OP behavior, acknowledging the potential role of gender and seasonal variety and using a broad range of variables covering both the family environment and perceived PE. Methods Study design and population The present study was embedded in the “Be Active, Eat Right” study, a cluster RCT investigating the effect of an overweight prevention protocol described in detail elsewhere (9). The Medical Ethics Committee of the Erasmus University Rotterdam