Eva Vanbrabant¹, Jikke Hesen¹, Suzan Jordan¹, Codrin Mironiuc¹, Lotte HJM Lemmens¹, Yuliya Shapovalova³, Gijs H. Goossens², Anne Roefs¹

1.Faculty of Psychology and Neuroscience, Maastricht University, Maastricht, the Netherlands, 2.Department of Human Biology, Institute of Nutrition and Translational Research in Metabolism (NUTRIM), Maastricht University Medical Centre+, 3.Institute for Computing and Information Sciences, Radboud University Nijmegen, Nijmegen, the Netherlands

eva.vanbrabant@maastrichtuniversity.nl

Background:

Obesity is a biopsychosocial disease and a major global health issue. An intensive lifestyle intervention is on average moderately effective, but effects vary highly across people. This study characterized individuals with overweight and obesity on personal, environmental, biological, psychological and behavioral factors that on average are related to body weight. Aiming to test if participants can be clustered according to similarity, and if treatment effectiveness is moderated by these variables.

Methods:

191 individuals with overweight and 204 with obesity were randomized to a six-month intensive lifestyle intervention (ILI) or an information (control) condition (INFO). Body weight indicators were obtained at pre- and posttest. A baseline profile was created, including biological (e.g., hormones), psychological (e.g., depression), environmental (e.g.,walkability), behavioral (e.g., diet quality) and personal (e.g., gender) variables.

Results:

Clustering analyses were done for the psychological and biological variables separately. In each, the optimal clustering solution revealed a favourable and non-favourable cluster. Next, hierarchical regression models were estimated. A model including the psychological variables, gender, and treatment type revealed that ILI was more effective (mean BMI decrease: ILI = 1.42 kg/m², INFO = 0.42 kg/m²; p < 0.001), and higher psychopathology and being male predicted more weight loss, with the model explaining 16% of variance in weight loss. Furthermore, a regression model including intervention type and two biological clusters confirmed the superiority of ILI, but effectiveness was not significantly moderated by biological cluster membership (mean BMI decrease: non-favorable cluster = 0.991 kg/m², favorable cluster = 0.873 kg/m²; p = 0.624), the model explained 11,3% of variance in BMI change because the significant effect of the intervention,as found previously. In ongoing analyses, we estimate hierarchical regression models including all relevant predictors of weight change. These will be presented at the conference.

Discussion/conclusion:

Predicting intervention effectiveness or weight change based on variables collected at baseline can significantly contribute to more personalized approaches in the management of overweight and obesity and related complications.