The results from the Icelandic part of the Pro Children survey shows that a large proportion of 11-year-old children's fruit and vegetable intake can be explained by environmental and personal factors. Environmental factors appear to be more important for children's vegetable intake than for their fruit intake. The personal factors found to be associated with fruit and vegetable intake were self-efficacy, preferences, liking and knowledge, suggesting the importance of addressing these variables in interventions.
All independent variables were significantly correlated to children's fruit and vegetable intake, with the exception of availability at school. In the present study, however, 31% of the children reported having eaten fruit in school the previous day, according to the 24-hour recall. The actual availability of fruit and vegetables at schools is low in Iceland, but as many children bring fruit with them to school, the availability at home appears to be a more important determinant of fruit and vegetable intake in Iceland. However, this could be changed by increasing the availability of fruits and vegetables at schools. In a Norwegian study it was shown that free subscription of fruits and vegetables to all pupils at school, at no cost to their parents, is an effective strategy to increase overall fruit and vegetable consumption . Paid subscription to a fruit and vegetable program, however, had a limited effect on the overall fruit and vegetable intake in the Norwegian study, but a Danish subscription study showed an increase of the same magnitude of fruit intake among both non-subscribers and subscribers . In the Danish study the fruit was given at school during a "fruit break", which seems to offer a positive setting for eating fruit and vegetables.
In the regression analysis, respectively 31% and 39% of the variance in children's fruit and vegetable intake was explained by the model. The predictiveness of the present model is considered good compared to similar studies. In a Norwegian study among school children (6th and 7th graders), 34% of the variance of the children's reported fruit and vegetable intake was explained by the measured factors . In an American study among adolescents, where a comprehensive model, including numerous socio-environmental, personal and behavioural factors were tested, only 13% of the variance in fruit and vegetable intake was explained . A review of the literature on models with psychosocial variables predicting dietary fat and fruit and vegetable consumption generally revealed low predictiveness of the models as they explained less than 30% of the variance, and among children and adolescents they were even less predictive .
The background variables of gender, residence and socioeconomic status (SES) were included in the model as these variables can be confounders . Gender and the father's socioeconomic status contributed significantly to the explanation of variance of vegetable intake, while parents' socioeconomic status did not affect children's fruit intake. Gender became non-significant for fruit intake when personal variables were added to the model, which could be explained by the clear gender differences in personal factors. Social and personal variables also explain more of the variance among girls than boys; however, among boys background and environmental variables seem to explain more of the variance. Studies from other countries have shown that children and adolescents from lower socioeconomic background consume less fresh fruit and vegetables [24–27]. Socioeconomic differences have been small in Iceland, but in the last decade, in an upswing of the Icelandic economy, inequality has grown . The most effective and efficient way to reach a large segment of the population is through elementary schools . While the efforts to increase fruit and vegetable intake need to reach all children, they need in particular to be suitable to those at highest risk for inadequate intake. Gender must therefore to be taken into consideration, and the association between SES and fruit and vegetable consumption should be followed over time to prevent any unfortunate growing inequality in intake among low SES groups.
Availability at home was found to be one of the most important determinants of children's vegetable intake; it seems, however, less important for children's fruit intake, although still important. This is consistent with studies from other countries that have found that those having high availability/accessibility to fruit and vegetables eat more than those with lower availability/accessibility [20, 21, 30]. It must be noted that the environmental factors in the present study are perceived environmental factors as the data was self-reported by the children. Thus perceived availability might be increase by making fruit and vegetable more visible, for example keeping fruit in a bowl on the table instead of in the refrigerator. The availability/accessibility of fruits and vegetables has been proposed as one of the most important determinants for children's and adolescent's intake of fruits and vegetables. Higher intake of fruits and vegetables could thus be reached, by making fruit and vegetable as easily available as possible.
Social factors seem to have a similarly strong effect on both fruit and vegetable intake. Modelling or subjective norm, i.e., the perceived fruit and vegetable intake of parents and friends, seems to have a stronger effect on girls than boys. Modelling may work through personal factors for instance increasing self-efficacy, which might be the reason for modelling becoming non-significant for fruit intake when the personal factors were added to the model.
Personal factors seem to have stronger effects on fruit intake than on vegetable intake. Self-efficacy was the strongest determinant of fruit intake, and somewhat weaker effect for vegetables. Children seem confident that they can eat a sufficient amount of fruit, if they want to. They feel a bit less confident, when it comes to eating sufficient vegetables. Vegetables are more often eaten as part of a meal than between meals. It is thus more in the hands of the parents than the kids themselves and may demand more cooking skills or preparation time, although Icelandic children at this age seem to prefer raw rather than cooked vegetables . It could be suggested that raw vegetable is preferred by children in the Nordic countries because they are more familiar to it. The consumption has traditionally been highest in the autumn when the vegetable is harvested, often eaten raw. In the Nordic countries there is not the same variety of vegetables as in the warmer southern countries, and in Iceland (at least) there is not the same tradition of cooked vegetable recipes (as the soup in Spain and Portugal for example). Knowledge of recommendations was also found positively related to intake. Increasing nutritional education, including skills necessary for preparing fruit and vegetable for consumption, might increase the frequency of fruit and vegetable intake.
Preferences and liking were the strongest personal predictors of vegetable intake. These factors were also predictors of fruit intake although not as strong. Preferences and liking are similar factors; however, they are measured differently. Liking was assessed by asking if they liked fruit and vegetables in general, while preferences were assessed by asking children how much they liked 12 frequently consumed fruits and vegetables. Preferences have been strongly correlated with fruit and vegetable intake in studies from other countries [20, 21, 30, 31]. Preferences were also found to be moderators of the relationship between availability and intake . Availability and accessibility, on the other hand, may be necessary for acceptance of fruits and vegetables [11, 21, 32]. Study using mere exposure of vegetable has shown increased liking in young children suggesting that repeatedly inviting child to taste small amount, without emphasis on how much they eat, is a good strategy to promoting liking , but it must be important that the quality of the vegetable is good. Reward on the other hand might be necessary to encourage children that simply refuse to taste the food at all, since the food must be tasted for the exposure to be effective . The interaction between availability and preferences supports the implementation of multi-component interventions with a strong environmental component; this might be especially important in low consumption groups.
The strength of this study is the large and representative sample, with a high participation rate. The data are self-reported; therefore, all measures are perceived measures, but thorough validity and reliability studies have shown measures to be valid and reliable [13, 15]. As this was a cross-sectional study, it cannot express causality between the determinants and frequency of fruit and vegetable intake. Further limitations of the study were that there were few items per scale, but in questionnaire development there is always a trade-off between precision and extensiveness .