Identifying weight management clusters and examining differences in eating behaviour and psychological traits

Abstract

Previous research has suggested that differences in psychological traits and eating behaviours may characterise groups of individuals with similar weight management profiles, for example, individuals who have successfully maintained a significant weight loss compared to those who have not. However, studies have tended to rely on arbitrary thresholds to group individuals, which may have biased findings. Across two studies, we sought to identify and validate weight management profiles using a clustering approach (i.e., without pre-defined grouping thresholds) and to examine differences across groups in terms of psychological and eating behaviour traits. Data were collected using online questionnaires (Study 1 was a secondary data analysis and study 2 was a primary data analysis that also allowed for cluster validation). Cluster analysis was implemented with BMI, diet history, weight suppression (difference between highest and current weight) as primary grouping variables, and age and gender as covariates. Differences in psychological and eating behaviour traits (e.g., impulsivity and restraint) were explored across clusters. In study 1, 423 participants (27.21 ± 9.90 years) were grouped into 5 clusters: lean men, lean women, lean middle-aged women, successful and unsuccessful dieters. The cluster structure was broadly replicated in study 2 with 368 participants (34.41 ± 13.63 years). In both studies, unsuccessful dieters had higher restrained and emotional eating scores than lean individuals, and in study 1, they also had higher food addiction scores than successful dieters. Individuals could be meaningfully grouped in terms of their weight management profiles and differences in psychological and eating behaviour traits were evident across these groups. Findings are in line with existing theory (e.g., Lowe's three factor model of dieting) but longitudinal research is needed to determine directionality between weight management profiles and relevant traits in order to inform future targeted interventions.