Associations between socioeconomic, parental and home environment factors and fruit and vegetable consumption of children in grades five and six in British Columbia, Canada

Globally and in Canada, the prevalence of obesity and other nutrition-related risk factors for chronic disease have been climbing in both children and adults [1‐6]. The fruit and vegetable food group is widely recognized as crucial to promoting healthy weights and preventing chronic disease [7]. Epidemiological evidence suggests that regular fruit and vegetable consumption is associated with a significant reduction in the risk of chronic diseases such as cancer, cardiovascular disease, and obesity [8‐12]. Thus, understanding the factors that influence their consumption is vital to the health of the population.

A large body of evidence indicates that socio-economic status (SES) plays a significant role in the diet of adults [13‐19]. The results from research conducted on the relationship between SES (largely parental income and education) and the diet of children (pre-adolescents) and youth (adolescents) are less definitive. It could be expected that because most children and youth live at home with either a parent or a guardian, many of the disparities in health and nutrition that exist between low and high SES adults would translate to their children. However, some studies have failed to show these relationships while others have, although these effects are to a much smaller degree in these populations than in adults [20‐22].

For example, in the case of youth (adolescents) Riediger and colleagues conducted a cross-sectional study of Canadian adolescents using data from the Canadian Community Health Survey, Cycle 2.1. They found the proportion of adolescents consuming the recommended amounts of fruit and vegetables increased from 34.2% in the lowest categories of household income to 42.1% in the highest categories (Odds Ratio = 1.07, p < 0.001) [23]. As well, positive associations were found between household education level and fruit and vegetable consumption (Odds Ratio = 1.12, p < 0.001). Although studies conducted outside of Canada have shown a positive relationship between SES and nutrition in pre-adolescent children [24‐26], there is limited information on Canadian children. Researchers have identified both that social factors associated with fruit and vegetable consumption may differ for adolescents and children, and that there was a need for studies involving populations from different political contexts [27].

Beyond the socio-economic (income and education) determinants a number of other social factors have been associated with healthy eating in children. These include gender, culture, family (parental modeling, family structure, parenting style, meal structure – family meals, positive attitudes/parental support for healthy eating), exposure to media - in particular television, watching TV while eating, and home and school availability and access to healthy food choices [4, 28, 29]. Taylor, Evers and McKenna (2005) emphasized a need for studies on both the determinants of, and children’s eating behaviour, in Canada [4].

Thus the purpose of this research was to examine the relationship between key social determinants and fruit and vegetable consumption of pre-adolescent children (10-12 years of age) in British Columbia, Canada. Specifically, we examined the relationship between socio-economic status, parent and home environment factors and fruit and vegetable consumption in the past 24 hours.

In all, 312 boys and 353 girls completed the web based 24-hour dietary food recall at baseline. Those with high reported food group intakes (more than 3 standard deviations from the mean) were removed from analysis (FV >10.5, grain products >14.8, milk and alternatives > 6.8, meat and alternatives > 6.1, other > 11.2), resulting in a sample size of n = 597. Means and standard deviations for baseline intakes of food groups are presented in Table 1, which shows that mean servings were below recommendations for Vegetables and Fruit and for Milk and Alternatives, and approximated recommendations for Grain Products and Meat & Alternatives. The mean reported FAS was 5.98 (SD = 1.74) out of a maximum of 9. The mean neighbourhood environment score was 25.25 (SD = 3.26) out of a maximum of 32. Key demographic and other social factor descriptors for parents of participating children are displayed in Tables 2 and 3, respectively. Based on recommendations outlined in the 2007 Canadian food guide [32], 85.8% of participants failed to meet the guidelines of six daily servings of vegetables and fruit.

We set out to explore the associations between socio-economic and other social factors and fruit and vegetable consumption in pre-adolescent children in BC. First and foremost however, our findings echo previous research [23] that showed that irrespective of SES the majority of the children were failing to meet the recommended daily guidelines for fruit and vegetable consumption. The proportion below the recommendations in our study (85%) was even higher than the 62% of boys and 68% of girls who were below recommendations in the Canadian Community Health Survey, Cycle 2.2 (CCHS 2.2) [21]. This may be because the recommended minimum number of servings at the time the CCHS 2.2 was conducted was 5 servings per day, which was subsequently increased to 6 servings per day in the 2007 food guide [32].

Second our results provided little evidence to suggest that socio-economic status and other social determinants were associated with children’s fruit and vegetable consumption. Parent income is one of the primary indicators of socio-economic status, and it has been linked to higher intake of fruit and vegetables among children in some studies [12, 15, 23, 26]. Counter to these findings parent income was not significantly correlated with fruit and vegetable intake among children in our study. Our result is supported by studies among Australian adolescents [14], 9 and 10-year-old American girls [20] and 4-18 year old Canadian children and youth [21]. Interestingly, child-reported affluence, which is a surrogate measure for income, was associated, albeit very weakly. This may reflect subtle differences in the measures. The affluence measure is a relative measure, most closely associated with the concept of disposable income, while parent-reported income is an absolute measure, not adjusted for items like family size, housing costs, debt, etc. Nevertheless, the finding was in the trivial range in terms of effect size so the deviations between income and affluence may be negligible in terms of predictive utility.

Based on our correlation results however, it is possible that there is a developmental aspect to the relationships we explored. For instance parental income might be related to fruit and vegetable consumption through a set of behaviours that may aggregate and take time to fully manifest as measurable fruit and vegetable consumption effects. This could explain the significant relationship between SES and consumption found in adolescence and then more consistently in studies of adults. In our study for example, lower-income parents were more likely to report eating dinner in front of the television, which itself was significantly – although again, weakly -- correlated with decreased fruit and vegetable consumption in our study and others [28]. Similarly parents with a higher income were more likely to report good access to fruit and vegetables. This relationship between income and access to fresh produce has been highlighted previously [43]. However, reported access did not correlate with children’s consumption in our study, which may reflect the level of access measured. The lack of relationship may be because our measure was about neighbourhood access while the determinants literature identified access in the home [4, 27, 28]. We also used ‘perceived access’ rather than directly measured access and thus our measure might reflect parental awareness and attitudes rather than actual access.

Education is another commonly cited indicator of social economic status [4, 20, 23, 25, 30]. In fact, education was positively associated with parent income in this study. Evidence has indicated that lower parental education status is associated with poorer diet quality, including higher fat and lower micronutrient intake in children [4, 20]. In their study of Canadian adolescents, Riediger and colleagues reported a small but significant positive association between parental education and fruit and vegetable intake [23]. Our results did not support this direct relationship in pre-adolescent children. Instead, as with parent income, it appeared that relationship between parent education and fruit and vegetable intake had the potential to develop over time influenced by a set of lifestyle behaviours established during childhood. Specifically, lower parental education was associated with eating dinner at the table less frequently, and eating in front of the television more frequently. The latter was correlated with decreased fruit and vegetable intake in our study and in others [28].

The fact that parental income and education weren’t directly associated with fruit and vegetable consumption could also be related to measurement issues or other factors that play an important role in determining what children eat. Numerous studies have discussed factors associated with the dietary intake of children that are not directly related to socio-economic status [4, 28, 44, 45]. We found that parental perceptions about the ‘healthiness’ of their child’s diet related positively to fruit and vegetable consumption. The direction of influence can’t be elucidated in a cross-sectional survey. We also found that the perceptions of the neighbourhood environment score, which is primarily a measure of liking for the neighbourhood and access to safe physical activity opportunities, was positively associated with the perceived ‘healthiness’ of the child’s eating habits. Parents who rated their neighbourhood highly were more likely to indicate that they thought their child’s eating habits were good, encourage their children to eat healthy foods, and personally care about eating healthy foods. Notably, both income and education were positively associated with the neighbourhood composite score. However, these findings should be placed in context by noting that the variables explored failed to predict a meaningful amount of the variability in children’s fruit and vegetable intake (small to trivial effect sizes) and that intake was below recommended amounts for almost all children.

Our findings should be considered in light of a number of limitations. One such limitation, present in almost all studies of nutrient consumption in children, is that in order to collect data efficiently from a large number of subjects (both financially and practically speaking), it is often necessary to rely on children’s self-report of their dietary intake; dietary self-reports rely on memory, which is subject to error. The 24-hour recall method is a well-studied method in dietary recall testing, particularly when assessing mean intakes of population groups [30], however, the ability of children to accurately recall and report their intake has been called into question in a number of studies [46‐48]. We attempted to mitigate this limitation by measuring the oldest children in the schools.

Evidence indicates that while there are inherent difficulties in collecting nutritional data from large groups of children through the use of self-report surveys, if properly administered (i.e., the length of time between intake and report is minimized), such tests can give moderately accurate results, and are generally the most efficient and cost-effective method of collecting such data. However, this is likely true only if the children are ten years of age or older [49]. As the children in this study were between the ages of ten and twelve, with an average age of 11.3 years, it is probable that most of the participating children were of an age where they could accurately complete the surveys used in this study.

A second limitation, highlighted by the response rates and descriptive statistics, was a potential sampling bias. As with all studies we depended on volunteers and although it appears that the SES of the adult participants was close to the reported population norms in 2011 [50, 51] our data showed that most reported caring about eating healthy foods. Accordingly, they may have been more likely to engage in health promoting behaviours than a randomly selected population would be. Finally, there are limitations inherent to cross-sectional designs and the fact that intake was assessed on only a single day and might not reflect a pattern of nutrient consumption established over a period of time.

Canadian children in BC were not meeting Canada’s food guidelines for fruit and vegetable consumption irrespective of SES. Low fruit and vegetable intake appears to be a population wide concern. It is imperative that public health policy-makers and practitioners act to promote healthy eating, particularly of foods that are under consumed by large proportions of individuals, such as fruit and vegetables. This promotion, however, must be underpinned by a thorough understanding of the determinants of dietary behaviour. While socio-economic status is consistently associated with the nutritional status of adults, in BC we found only weak to trivial evidence that socio-economic factors relate to pre-adolescent children’s fruit and vegetable consumption. The home environment, parental perceptions and habits, were weakly associated with fruit and vegetable consumption and also with parental income and education and the influence of these and other factors deserves further exploration using longitudinal designs to explore the development of fruit and vegetable consumption behaviours over time.