Le comportement alimentaire représente un enjeu majeur pour la santé publique du fait de son impact sur les pathologies chroniques et plus spécifiquement l’obésité. Les préférences temporelles sont des facteurs psychologiques liées aux capacités d’autorégulation et pourraient être associées aux comportements alimentaires et à l’obésité. L’objectif principal de cette thèse était d’étudier la relation entre les préférences temporelles, le comportement alimentaire et l’obésité en population générale. Les préférences temporelles ont été évaluées à partir de questionnaires mesurant l’impulsivité et la considération des conséquences futures. Ce travail a été réalisé au sein de la cohorte NutriNet-Santé basée sur internet. L’impulsivité était associée à une moins bonne qualité de l’alimentation, à un grignotage plus fréquent et aux troubles des conduites alimentaires. Une considération pour les conséquences futures plus importante était associée à une plus forte motivation envers les facteurs santé et environnement lors des choix alimentaires, à une meilleure qualité de l’alimentation et à une part plus importante d’aliments issus de l’agriculture biologique dans le régime alimentaire. L’impulsivité et une faible considération pour les conséquences futures étaient associées positivement à l’obésité et modéraient de façon quantitative la relation positive entre alimentation émotionnelle et obésité. Ces résultats suggèrent une influence des préférences temporelles sur les comportements alimentaires et l’obésité, et apportent de nouveaux arguments en faveur de la prise en compte des déterminants psychologiques dans les stratégies de prévention de santé publique. / Eating behavior represents a major stake regarding public health because of its impact on noncommunicable diseases and more specifically obesity. Time preferences are psychological factors linked with self-regulation behaviors and could be associated with eating behavior and obesity. The main objective of this thesis was to study the relationship between time preferences, eating behavior, and obesity in a general population. Time preferences were assessed with questionnaires measuring impulsivity and consideration of future consequences. This work was conducted as part of the NutriNet-Santé study, which is a large web-based cohort. Impulsivity was associated with a lower diet quality, a higher snacking frequency, and eating disorders. A high level of consideration of future consequences was associated with greater concern for health and environment when choosing food, with a better diet quality and a higher contribution of organic foods in the diet. Impulsivity and a low consideration of future consequences were positively associated with obesity and quantitatively moderated the relationship between emotional eating and obesity. These results suggest that time preferences influence eating behavior and obesity, and bring new arguments in favor of the consideration of psychological determinants in public health prevention strategies

Background: The alcohol-attributable burden of disease is high among socially disadvantaged individuals. Interventional efforts intending to have a public health impact should also address the reduction of social inequalities due to alcohol. Objective: The aim was to test the moderating role of educational background on the efficacy of a computer-based brief intervention addressing the full spectrum of alcohol use. Methods: We recruited 1646 adults from the general population aged 18 to 64 years (920 women, 55.9%; mean age 31 years; 574 with less than 12 years of school education, 34.9%) who reported alcohol use in the past year. The participants were randomly assigned a brief alcohol intervention or to assessment only (participation rate, 66.9%, 1646/2463 eligible persons). Recruitment took place in a municipal registry office in one German city. All participants filled out a self-administered, tablet-based survey during the recruitment process and were assessed 3, 6, and 12 months later by study assistants via computer-assisted telephone interviews. The intervention consisted of 3 computer-generated and individualized feedback letters that were sent via mail at baseline, month 3, and month 6. The intervention was based on the transtheoretical model of behavior change and expert system software that generated the feedback letters automatically according to previously defined decision rules. The outcome was self-reported change in number of alcoholic drinks per week over 12 months. The moderator was school education according to highest general educational degree (less than 12 years of education vs 12 years or more). Covariates were sex, age, employment, smoking, and alcohol-related risk level. Results: Latent growth modeling revealed that the intervention effect after 12 months was moderated by educational background (incidence rate ratio 1.38, 95% CI 1.08-1.76). Individuals with less than 12 years of school education increased their weekly alcohol use to a lesser extent when they received the intervention compared to assessment only (incidence rate ratio 1.30, 95% CI 1.05-1.62; Bayes factor 3.82). No difference was found between groups (incidence rate ratio 0.95, 95% CI 0.84-1.07; Bayes factor 0.30) among those with 12 or more years of school education. Conclusions: The efficacy of an individualized brief alcohol intervention was moderated by the participants’ educational background. Alcohol users with less than 12 years of school education benefited, whereas those with 12 or more years did not. People with lower levels of education might be more receptive to the behavior change mechanisms used by brief alcohol interventions. The intervention approach may support the reduction of health inequalities in the population at large if individuals with low or medium education can be reached.