Co-creating child-inclusive food group categories and illustrations for dietary assessment: a mixed methods study

Abstract

Accurate dietary assessment is essential for capturing detailed information on food and nutrient consumption, with applications in nutrition research, public health, and clinical practice. However, developmental and cognitive factors, including limited food knowledge and literacy skills, complicate dietary assessment in children. From approximately 8 years of age, children become increasingly capable of (partially) reporting dietary intake, provided that tools are adapted to their cognitive abilities and experiences. Co-creating dietary assessment tool features with children may support this process by aligning design with children’s cognitive and practical capacities. This study therefore aimed to co-create food group categorizations and illustrations to support food identification and selection within a self-reported dietary assessment tool, using a sequential exploratory mixed methods approach. Food items ( n  = 63) were selected based on the 50 most frequently consumed foods among children aged 8–10 years from Dutch National Food Consumption Survey (DNFCS) data (2019–2021), supplemented by two caregiver focus groups to ensure coverage across 14 GloboDiet food categories. Children categorized these items into self-defined food groups, assigned labels, and created illustrations to represent groups. Qualitative data (i.e., interview transcripts, label, drawings) were analyzed using thematic analysis with inductive and deductive coding approaches. Quantitative analyses (i.e., card sort) included Jaccard similarity matrices and hierarchical cluster analysis to identify clusters of food groups. On average, children ( N  = 43) created 10.0 (SD = 4.6) piles, yielding fifteen distinct clusters: “vegetables”, “fruit”, “chocolate”, “other”, “potatoes”, “sauce”, “spreads”, “dairy”, “rice & pasta”, “eggs”, “beans”, “meat & fish”, “breakfast”, “unhealthy”, and “drinks”. From 220 drawings (M = 5.1, SD = 3.0 per child), design strategies to support recognition included the use of utensils (48.6%), examples (48.6%), symbols (23.2%), human features (13.6%), and contextual elements (9.5%), with color (48.9%) and language (41.9%). Children’s perspectives on food categorization and visual recognition, combined with DNFCS and caregiver input, resulted in fifteen food groups and child-derived icon design strategies that may enhance the usability and accuracy of self-reported dietary assessment in children; further translation into a functional food search interface through multidisciplinary collaboration and empirical evaluation of effectiveness and efficiency is warranted.