Causal attributions for obesity and weight loss in the Healthy Weight Coaching program, a 12-month online lifestyle intervention

Introduction: We investigated how causal attributions—that is, individuals' beliefs about what caused their obesity—are associated with weight loss in a digital lifestyle intervention. Methods: Data were collected from Healthy Weight Coaching, a 12-month real-world online intervention. Weight and height were self-reported. Body mass index (BMI) was calculated using interpolated weights at 3, 6, 9, and 12 months. At baseline, participants completed an eight-item questionnaire on causal attributions, rating each as playing no role, some role, or a major role. TwoStep cluster analysis grouped individuals with similar response profiles. Results: Data were available for 1834 participants (1524 83.1% women, median age 52 years, median BMI 39.1 kg/m2). Participants most frequently attributed obesity to unhealthy dietary habits and lack of physical activity, whereas genetics, family lifestyle habits, and medication were least endorsed. Attributing obesity to stress (somewhat, standardized B=0.156 95% CI, 0.051–0.262; very much, B=0.204 95% CI, 0.100–0.307), fatigue (very much, B=0.154 95% CI, 0.073–0.235), adversities in life (somewhat, B=0.112 95% CI, 0.041–0.184; very much, B=0.100 95% CI, 0.024–0.176), or medication (somewhat, B=0.089 95% CI, 0.022–0.155; very much, B=0.108 95% CI, 0.025–0.192) were associated with lower weight loss compared with those not endorsing these attributions. However, only stress (“very much”) reached a small effect size according to Cohen’s criteria, whereas the remaining associations were negligible in magnitude despite statistical significance. Cluster analysis revealed four clusters. Compared to those Behaviourally focused, particularly endorsing dietary habits and physical activity, all other clusters showed higher endorsement of stress, fatigue, and adversities in life, with a cluster labelled Multifactorial fully endorsing all eight items. Relative to Behaviourally focused, all other clusters had lower odds of 5% weight loss and lost less weight (Broad lifestyle oriented, B=0.104 95% CI, 0.032–0.177, p=0.004; Psychologically burdened, B=0.144 95% CI, 0.063–0.225, p<0.001; Multifactorial, B=0.222 95% CI, 0.136–0.308, p<0.001). Only the Multifactorial cluster reached a small effect size. Conclusions: Strong endorsement of stress as a cause of obesity and membership in the Multifactorial cluster were associated with less favourable weight loss outcomes. However, effect sizes were small, suggesting limited standalone impact but potential relevance within broader behavioural and psychological treatment contexts.