Associations between socioeconomic status, home food availability, parental role-modeling, and children's fruit and vegetable consumption : a mediation analysis

Background: Recent literature has suggested that associations and interactions between family socioeconomic status (SES) and home food environment influence children's diet, but little is known about the mediation roles of parental role-modeling and food availability in the socioeconomic inequalities of children's diet. This study aimed to determine the associations between family SES and children's fruit and vegetable (FV) consumption and to assess the mediation roles of parental role-modeling and food availability in the above associations. Methods: Cross-sectional data of 574 Finnish children (aged 3 to 6) were analyzed. Parents completed an FFQ assessing their children's FV consumption frequency and a questionnaire assessing SES and home food environment. Two exposure variables: parental educational level ("low", "middle", and "high") and the relative family income tertiles of the family were used. The frequencies of parental role-modeling of FV and sugary food and drink (SFD) consumption, and the availability of FV and SFD at home were calculated. Single- and multiple-mediator models were created using IBM SPSS 27.0. Results: The positive association between high parental educational level and children's FV consumption (direct effect coefficient: 2.76, 95% CI: 0.51-4.86) was partially mediated by more frequent parental role-modeling of FV consumption (indirect effect coefficient: 0.89, 95% CI: 0.10-1.76), higher availability of FV (indirect effect coefficient: 1.00, 95% CI: 0.35-1.77), and lower availability of SFD (indirect effect coefficient: -0.30, 95% CI: -0.72 - -0.01). The relative family income was not directly associated with the outcome. However, the higher relative family income level indirectly predicted the Children's FV consumption (full mediation) through more frequent parental role-modeling of FV consumption (indirect effect coefficient: 0.91, 95% CI: 0.06-1.83) and higher availability of FV (indirect effect coefficient: 0.98, 95% CI: 0.40-1.67). Parental role-modeling on SFD consumption did not mediate any of the above associations. Conclusions: Parental educational level showed more associations with children's FV consumption than relative family income. Our findings suggest that reducing the availability of SFD is as important as increasing the availability of FV to enhance children's FV consumption. Future interventions to improve children's dietary behaviors should pay greater attention to the lower SES segments of society. Longitudinal studies and intervention studies supporting these findings are needed for making meaningful recommendations for health promotion