Family influences on children's physical activity and fruit and vegetable consumption.

Background: There is evidence of a clustering of healthy dietary patterns and physical activity among young people and also of unhealthy behaviours. The identification of influences on children's health behaviors, particularly clustered health behaviors, at the time at which they develop is imperative for the design of interventions. This study examines associations between parental modelling and support and children's physical activity (PA) and consumption of fruit and vegetables (FV), and combinations of these behaviours.

Methods: In 2002/3 parents of 775 Australian children aged 10-12 years reported how frequently their child ate a variety of fruits and vegetables in the last week. Children wore accelerometers for eight days during waking hours. Parental modelling and parental support (financial and transport) were self-reported. Binary logistic and multinomial logistic regression analyses examined the likelihood of achieving >/= 2 hours of PA per day (high PA) and of consuming >/= 5 portions of FV per day (high FV) and combinations of these behaviors (e.g. high PA/low FV), according to parental modelling and support.

Results: Items of parental modelling and support were differentially associated with child behaviours. For example, girls whose parents reported high PA modelling had higher odds of consuming >/= 5 portions of FV/day (OR = 1.95, 95% CI = 1.32-2.87, p < 0.001). Boys whose parents reported high financial support for snacks/fast foods had higher odds of having 'high PA/low FV' (OR = 2.0, 95% CI = 1.1-3.7).

Conclusion: Parental modelling of and support for physical activity and fruit and vegetable consumption were differentially associated with these behaviours in children across behavioural domains and with combinations of these behaviours. Promoting parents' own healthy eating and physical activity behaviours as well encouraging parental modelling and support of these behaviours in their children may be important strategies to test in future research.