A comparison of dietary patterns derived by cluster and principal components analysis in a UK cohort of children

Andrew D a C Smith (1) , Pauline Emmett (1) , Pk Newby (1) , Kate Northstone (1)

1 School of Social and Community Medicine

BACKGROUND/OBJECTIVES To identify dietary patterns in a cohort of 7-year-old children through cluster analysis, compare with patterns derived by principal components analysis (PCA), and investigate associations with socio-demographic variables. SUBJECTS/METHODS The main caregivers in the Avon Longitudinal Study of Parents and Children (ALSPAC) recorded dietary intakes of their children (8 279 subjects) using a 94-item food frequency questionnaire. Items were then collapsed into 57 food groups. Dietary patterns were identified using k-means cluster analysis and associations with socio-demographic variables examined using multinomial logistic regression. Clusters were compared with patterns previously derived using PCA. RESULTS Three distinct clusters were derived: Processed (4 177 subjects), associated with higher consumption of processed foods and white bread, Plant-based (2 065 subjects), characterized by higher consumption of fruit, vegetables and non-white bread, and Traditional British (2 037 subjects), associated with higher consumption of meat, vegetables and full-fat milk. Membership of the Processed cluster was positively associated with girls, younger mothers, snacking, and older siblings. Membership of the Plant-based cluster was associated with higher educated mothers and vegetarians. The Traditional British cluster was associated with council housing and younger siblings. The three clusters were similar to the three dietary patterns obtained through PCA; each principal component score being higher on average in the corresponding cluster. CONCLUSIONS Both cluster analysis and PCA identified three dietary patterns very similar both in the foods associated with them and socio-demographic characteristics. Both methods are useful for deriving meaningful dietary patterns. KEYWORDS Dietary patterns, children, cluster analysis, principal components analysis, ALSPAC

Mots clés: Dietary patterns; children; cluster analysis; principal components analysis; ALSPAC

Andrew D a C Smith, Pauline Emmett, Pk Newby, Kate Northstone. A comparison of dietary patterns derived by cluster and principal components analysis in a UK cohort of children. European Journal of Clinical Nutrition, 2011, ⟨10.1038/ejcn.2011.96⟩. ⟨hal-00644720⟩