Abstract
Understanding food choices made for meals in overweight and obese individuals may aid strategies for weight loss tailored to their eating habits. However, limited studies have explored food choices at meal occasions. The aim of this study was to identify the usual food choices for meals of overweight and obese volunteers for a weight-loss trial. A cross-sectional analysis was performed using screening diet history data from a 12-month weight-loss trial (the HealthTrack study). A descriptive data mining tool, the Apriori algorithm of association rules, was applied to identify food choices at meal occasions using a nested hierarchical food group classification system. Overall, 432 breakfasts, 428 lunches, 432 dinners and 433 others (meals) were identified from the intake data (n 433 participants). A total of 142 items of closely related food clusters were identified at three food group levels. At the first sub-food group level, bread emerged as central to food combinations at lunch, but unprocessed meat appeared for this at dinner. The dinner meal was characterised by more varieties of vegetables and of foods in general. The definitions of food groups played a pivotal role in identifying food choice patterns at main meals. Given the large number of foods available, having an understanding of eating patterns in which key foods drive overall meal content can help translate and develop novel dietary strategies for weight loss at the individual level.