Relationships between the home environment and physical activity and dietary patterns of preschool children: a cross-sectional study

The prevalence of overweight and obesity in children have increased two and three-fold over the last decade in most developed countries [1–3]. Australian data suggest that one in five children are now above a healthy weight [3, 4]. Population data from South Australia have shown that these rapid increases have also occurred in children as young as four years [5].

The childhood obesity epidemic appears secondary to changes in modern society resulting in increased availability of energy dense foods and reduced opportunities for physical activity [6, 7]. There is growing evidence that many young children are consuming a diet inconsistent with recommendations and in addition, have inadequate levels of physical activity. For example, Australian dietary guidelines for children and adolescents recommend increasing the amount of fruit and vegetables consumed, reducing non-core snack foods, reducing consumption of sweetened beverages and choosing fat-reduced dairy from the age of 2 years [8]. Data from the Australian Food and Nutrition Monitoring Unit showed that children's total energy intake in a 24 hour period increased by 13% between 1985 and 1995, an increase largely consisting of non-core snack foods and sweetened drinks [9].

Australian physical activity guidelines recommend children spend a minimum of 60 minutes per day in moderate to vigorous physical activity and conversely spend a maximum of 120 minutes per day engaged in small screen entertainment [10]. A study of Australian 5–12 year olds showed that 15% of children this age did not fulfil the physical activity recommendations and 31% engaged in excessive electronic media use [11]. Consistent with this, a recent Australian study reported that 21% of 4 year olds spent greater than 2 hours engaged in TV viewing per day [12]. International studies report similar findings in preschool children. Reilly et al. used accelerometers to document physical activity of Scottish preschoolers. These young children spent only 20–25 minutes per day in moderate to vigorous physical activity and approximately 75% of the waking day engaged in sedentary activities [13]. Data from the US National Longitudinal Survey of Youth (1990–1998) showed that 41% of 2–3 year olds watched greater than 2 hours of television per day [14].

An ecological framework describes 'obesogenic' (or obesity promoting) environments in terms of macro and micro environments over which individuals have more or less control [15, 16]. Characteristics of such environments are hypothesised to be direct determinants of children's physical activity and dietary patterns. For children, micro-environments include the family home, along with the school and local neighbourhood. Macro-environments could include transport, food supplies, the built environs, health care services and a wide range of government services and policies. Although many children spend time in care away from their home, especially for preschool children, the family home remains one of the principal environments, with family rules and preferences largely determining food availability and opportunities for physical activity [17].

Relationships between the environment and children's physical activity have been summarised in a recent systematic review [18]. Three of the 33 studies reviewed included preschool children; Baranowski et al., reported that preschool children were least active outdoors during the hottest months, Burdette et al., reported that mothers' perceptions of neighbourhood safety were not associated with time children spent playing outdoors and Sallis et al., reported that the more play spaces within walking distance, the higher the physical activity of the child [19–21]. Conversely, Trost et al., reported that parental support, active toys, television watching, park visitation and child competence were not significantly associated with time spent in active play during preschool sessions [22]. Whilst studies have consistently shown that parental role-modeling of physical activity is an important predictor of older children's physical activity, few studies have reported on this relationship in younger children [23, 24].

Evidence supports positive relationships between greater numbers of televisions in the home and televisions in children's bedrooms, and greater amounts of time spent in sedentary activity [25]. A positive association has also been found between greater media use and children's body mass index (BMI), including a positive association between television viewing and preschool children's BMI [26, 27]. More complex relationships around screen time have also been reported. For example, Salmon et al., reported that low physical activity was associated with enjoyment of Internet and computer use, a preference for watching television, parents who use computers/electronic games and having pay TV and electronic games available at home in 10–12 year olds [28]. Yalcin et al., showed a positive relationship between viewing habits of siblings and parents and their preschool children and also that leaving the household TV switched on when not being watched was associated with preschool children having greater viewing times [29].

While Birch and colleagues have comprehensively investigated the role of parental child-feeding practices on dietary intake and thus energy balance of very young children, other aspects of the family food environment have not been explored [30, 31]. In a recent study of 5-to-6-year-old children, mother's who were less educated reported that the variety of fruits and vegetables available to them for purchase were limited and of poorer quality and that they were less likely to purchase fruit and vegetables because the family did not like them [32]. Hearn et al., assessed the presence of 10 varieties of fruit and 10 varieties of vegetables in the homes of 13 families who had school age children [33]. Weekend consumption of fruit and vegetables by children was associated with availability and access but weekday consumption was not.

In summary, previous research has predominantly focused on older children and most studies have assessed only one or two aspects of the home environment at any one time. The latter is problematic because conceptually it is possible that parents balance or compensate one aspect of the home environment with others. Assessing only one or two areas of the home environment does not give a comprehensive picture of what the home environment offers children. The aim of this study was to extend previous research conducted with older children by comprehensively describing characteristics of preschool children's home environments which may influence children's physical activity, sedentary behaviour and dietary patterns. Whilst we acknowledge the theoretical difficulty interpreting multiple comparisons, this approach gives a much broader and richer understanding of potential relationships between the home environment and children's physical activity and dietary patterns.

This study aimed to assess relationships between home environment characteristics and physical activity and dietary patterns of preschool children. Outdoor playtime, used as a proxy for physical activity, was positively associated with greater frequency of parental participation in physical activity, particularly mother's participation. The importance of parental modelling of physical activity has been demonstrated in other studies. Ferreira at al. in a review of correlates of children's physical activity, summarised 29 studies investigating parental physical activity and 29 and 31 studies investigating separately father's and mother's physical activity respectively [44]. In contrast to our study, Ferreira et al. concluded that father's level of physical activity had a stronger relationship with children' physical activity than mother's. This may have been because in this review, outcomes of studies of children aged 3–12 were combined. Mother's physical activity may be a stronger influence on the physical activity of younger children when children are of an age when mothers are more likely than fathers to spend time at home. Similarly, at this age it is plausible that having a parent accompany a child undertake outside activities is more important than children actually modelling parental behaviour. Regardless of the mechanism, encouraging parental physical activity appears to be an important and practical way of increasing both the physical activity of preschoolers and older children.

Children's physical activity was also positively associated with more outdoor play equipment and a larger backyard size. Whilst the size of these correlations were small, these findings are consistent with a number of other studies [44, 45]. A more interesting outdoor environment may encourage children to spend time in active outdoor play, but it is equally plausible that parents of sedentary children respond by investing less in such activities. Similarly, a larger backyard size may increase time children spend outdoors, but it is also plausible that more active families purchase homes with larger outdoor areas. In contrast, there was no association between presence of a children' playground close to the home and children's outdoor play, suggesting that the home rather than the neighbourhood environment is more important at this age. There was no significant association between amount of time spent in organised physical activities such as swimming lessons, dance, kindergym or gymnastics and children's outdoor playtime. These relatively costly activities generally take up only small amounts of time in the daily lives of preschool children (in our study 39% and 29% of children attended a swimming lesson or another form of organised physical activity once a week) and do not appear to influence the time children spend in free play at home. Finally, our inventory included an item about dog ownership. Dog ownership was not associated with outdoor playtime of preschoolers in our study, consistent with one other study, which showed that whilst dog ownership was associated with physical activity of older children, it was not associated with walking or cycling of 5–6 year old children [46].

Consistent associations were found between less screen time and greater parental monitoring of television use in the home and more rules about television viewing. In contrast to previous research, the number of televisions present in the family home was not associated with screen time [26, 29]. One plausible explanation is that the effect of this variable is much weaker now the majority of homes have multiple televisions (47% of families in our sample had three or more televisions compared to only 10% of families of preschoolers in the study of Yalcin et al. conducted in 1999) [29]. We also did not find an association between the presence of television in preschool children's bedrooms and screen time. Yalcin et al. also reported this negative finding [29]. This contrasts to the literature in older children where associations have been found between presence of television in children's bedrooms and both sedentary behaviour and obesity [12, 47]. Perhaps having a television in the bedroom is not such an important influence on younger children's viewing patterns compared with older children because younger children do not spend as much time involved in independent activities in their bedrooms. Another unexpected finding was that the presence of an Internet connection was significantly associated with lower levels of small screen entertainment time and a similar but non-significant association was found for presence of a computer in the home (p = 0.07). One plausible explanation for this finding is that preschool children do not use computer or the Internet for entertainment. Alternatively, we did not control for socio-economic status, and children of more affluent families may both have less screen time and be more likely to have computer and Internet available.

Previous research suggests that physical activity and sedentary behaviour represent different constructs [48]. Our study demonstrated that a different set of explanatory variables were associated with the amount of outdoor playtime of preschool children compared to time spent engaged in small screen entertainment consistent with this premise.

The most consistent nutritional variables associated with children's dietary patterns were the types and amounts of foods located within the family home. Other than the small study of Hearn et al., which focused solely on fruit and vegetable intake, this is the first time an association between children's intake and food availability in the home has been documented [33]. The association was found for both intake of core food groups (fruit and vegetables) and non-core food groups (sweet and salty snacks and sweetened beverages). A plentiful supply of fruit, vegetables and low fat dairy foods along with reducing the presence of fruit juice, sweetened beverages and non-core foods, were associated with higher intake of fruit and vegetables and lower intake of foods with less nutritional value. From a practical perspective, counselling parents regarding the types and amounts of foods stocked in the home may be more effective and less stigmatising than focusing on dietary patterns or body weights of individual children for either managing a child who is overweight or undertaking preventive education with parents.

Consistent with previous research, our study found significant associations between a variety of parental behaviours around food provision and children's dietary patterns [49]. Higher intake of fruit and vegetables was associated with parental behaviours restricting access to less healthy foods (fruit juice, high fat/sugar snacks) and restricting second helpings. In comparison, lower intakes of fruit and vegetables were associated with greater use of coercive parental behaviours such as offering food rewards to eat main meal and reminding child to 'eat up'. The causal direction of the relationship cannot be determined from this cross-sectional data. One plausible explanation is that parents of children who were poor consumers of fruit and vegetable from a younger age have responded by using these coercive techniques. Alternatively, Birch et al. argue from their experimental data that coercive behaviours can lead children to favour less healthy foods and specifically that restricting children's intake of less healthy foods can unintentionally lead to a preference for these foods [50, 51]. Our data suggests that coercive and restrictive parental behaviours should be considered separately. For example, we also showed that greater use of restrictive behaviours by parents was associated with lower intake of less healthy foods (sweetened beverages and non-core foods). Conversely, the use of food treats either by parent or other carers (a coercive behaviour), was associated with higher intake of less healthy foods. Thus parental restriction of access by children to less healthy foods appears to be associated with better dietary patterns whilst coercive behaviours are associated with poorer dietary patterns.

Finally, the positive relationships between allowing children to eat in front of the television and greater intake of less healthy food groups (sweetened beverages and non-core foods) and lower intake of more healthy food groups (fruit and vegetables) deserves mention. Post hoc analyses showed a small but significant positive correlation between the non-core food score and the small screen entertainment score (r = 0.14 and p = 0.02) and a small but significant negative Pearson correlation between the fruit and vegetable score and the small screen entertainment score (r = -0.17, p = 0.004). These findings suggest the relationship between television viewing and the development of obesity may well be partially mediated by dietary patterns in addition to reduced energy expenditure. This is supported by a recent randomised controlled trial of reducing television viewing to treat overweight 4–7 year olds where change in television viewing was related to the change in energy intake and not to changes in physical activity [52].

The study has a number of limitations. The data is cross-sectional and the temporal direction of associations cannot be determined. All analyses were univariable and did not take into account potential confounding variables such as socioeconomic status. The sample size was relatively small and parents were of higher socioeconomic status compared to the South Australian average. Whilst studies generally report an association between more healthy dietary patterns and higher socioeconomic status, recent studies fail to confirm an association between socioeconomic status and children's physical activity [32, 53].

The home environment inventory was analysed in terms of individual items rather than subscale scores and responses to individual items are inherently less stable than summed scores. This approach was taken because conceptually there was no theoretical basis on which to assume some items were related and could be summed. For example, just because parents may score high on one item (e.g. presence of non-core snack foods in the home) does not necessarily mean that they would score high on another item (e.g., presence of sweetened beverages in the home). The theoretical objection to presenting multiple comparisons (i.e., that a number of significant findings would be expected by chance) has been addressed by providing the total number of items (i.e., comparisons) in Additional file 1[42].

The relative newness of this area of research meant that we had to develop a questionnaire to measure the home environment, use a recently developed questionnaire to measure preschool children's dietary patterns and add supplementary questions to the physical activity questionnaire in order to measure sedentary activity. In addition, assessing the inter-rater reliability of the Physical and Nutritional Home Environment Inventory was beyond the scope of this study. Given the limited information regarding the validity of the questionnaires used, the results of this study need careful interpretation and ideally, replication.

The study has a number of strengths. This is one of few studies to measure elements of the home environment by direct observation. Other studies have relied on parental report regarding numbers of televisions, availability of outdoor play equipment or household food [19, 23, 28, 32]. In addition, the study included both assessments of dietary and physical activity patterns rather than focusing on only one side of the energy balance equation. Practical aspects of ensuring a healthy lifestyle for children by parents were examined and this has resulted in data that has real clinical utility for health professionals dealing with overweight children and their families. This is particularly true for the consumption of fruit and vegetables and non-core foods where our study found many significant associations with children's intake. Addressing these factors is particularly important because it is generally the over-consumption of non-core foods which result in excessive energy intake in children [54].

The key finding of this study was that many practical and potentially modifiable aspects of the home environment were associated with preschool children's physical activity and dietary patterns. Parents are important in terms of role-modelling physical activity, providing a safe and interesting backyard for children to play in, setting rules about how small screen entertainment is used in the home, behavioural approaches to family food consumption, and providing healthy food choices in the home. Preschool years are a key time as children spend a large proportion of time at home and this study supports focusing on the family home environment for obesity prevention and management.