What affects the quality of diet of preschool children? A principal component analysis of maternal, child, and mealtime environment characteristics

Abstract

BackgroundPreschool children in the UK commonly have poor-quality diets. Research to identify influences on children's diets has focused on the immediate environment, identifying maternal and family determinants that seem important. However, these determinants are generally studied either in isolation or in conjunction with only a few others, which limits our understanding of their importance and whether they interact to determine a child's quality of diet. This study assessed a range of maternal, child, and home environmental factors and explored (i) how they inter-relate and (ii) whether patterns of these factors are associated with children's quality of diet. MethodsIn this cross-sectional survey, mothers enrolled in the Southampton Initiative for Health, a community-based intervention study, who had a child aged 2–5 years were invited to take part. Informed verbal consent was obtained. The study was approved by the University of Southampton, Faculty of Medicine ethics committee. 348 mother–child pairs were studied. Mothers completed questionnaires about maternal, child behavioural, and psychosocial factors, as well as characteristics of the child's mealtime environment. Quality of diet was assessed with a validated food frequency questionnaire from which a standardised diet Z score was calculated. A high score described children with high consumption of vegetables, fruit, water, and wholemeal bread. FindingsThe mothers' mean age was 32 years (SD 5·4), 134 (39%) had left education at age 16 years, and 198 (57%) had been employed since the birth of their child. The children's mean age was 3 years (SD 0·9) and 191 (55%) were the first born. A Spearman's correlation matrix showed inter-relationships between many of the maternal, child, and environmental factors. Principal components analysis was done to explore patterns of association. The first two components (labelled empowered mother and neophobic child) explained 19% and 11% of the variance in the data, respectively. A high score on component one described mothers who felt in control of their lives, had a strong sense of wellbeing, prioritised food, had high levels of self-efficacy, and who were food secure. Their children ate at table and not while watching television. Component two was independent of component one and described mothers with neophobic children, who battled about food, whose child did not eat with the family, and who tried to control their child's diet. Component scores were converted to Z scores and were entered into multiple regression models to predict a child's quality of diet. After adjustment for confounders, including mother's educational attainment, age, and employment status, and child's birth order the empowered mother scores were strongly positively associated with child's quality of diet (β 0·45, 95% CI 0·34–0·55; p<0·0001). The neophobic child scores were not associated with child's quality of diet (p=0·1). InterpretationAs far as we are aware, this is the first study to explore how patterns of maternal, child, and environmental factors determine preschool children's quality of diet. The most important predictors of child's quality of diet were having a mother with a strong sense of wellbeing, self-efficiency, food involvement, and control over life, who actively managed the child's food environment. Our data suggest that interventions to improve the quality of young children's diets should consider the importance of maternal psychosocial factors. The data were cross-sectional thus causality cannot be identified. Gathering dietary data with parental report incurs possible social desirability bias. Questionnaire data are subjective; however, all the assessments had been validated. The sample size for this study was small and we cannot assume that participants are representative of the wider population; however, mothers did represent a wide range of educational attainment and age. FundingMedical Research Council and the NIHR Southampton Biomedical Research Centre.