Active transportation is associated with lower obesity risk: generalized structural equations model applied to...

This study aimed to identify latent (unobservable) dimensions representing specific physical activity-related behaviors and explore their potential effects on obesity burden and spatial distribution in Colombia. A cross-sectional study (n = 9,658) was conducted based on the Colombian National Survey of Nutritional Status. A generalized structural equations model was proposed, combining exposure and measurement models to define a disease model. Modeling identified latent dimensions of physical activity focused on screen time and means of transportation and estimated their direct and indirect effects on obesity occurrence. Mapping techniques were used to illustrate adherence to these dimensions. The latent dimensions identified were named "Screens use" and "Active transportation"; the latter was inversely associated with obesity occurrence (p = 0.004), with the use of bicycles being the dominant variable, contrasting with the use of motor vehicles. The mapping showed that departments with the highest adherence to the "Active transportation" construct have a lower prevalence of obesity. Bicycle use, as opposed to non-active transportation, represented a dimension of physical activity-related behaviors with a protective effect against obesity. This suggests that active transportation may be a crucial factor in the designing preventive interventions. Moreover, social inequalities may be contributing to the obesity epidemic and physical activity behaviors in Colombia, requiring equitable and multisectoral responses.