Mapping Digital Nudges and Recommender Systems for Obesity Prevention: Scoping Review

Background: Recommender systems are pivotal in organizing information to enhance noticeability, reduce overload, and streamline decision-making. They can be even more effective if combined with digital nudges. Digital nudging is a subtle approach that combines design, information, and interaction elements to create a choice architecture that can guide user behavior in digital environments. While promising in many fields, there is a notable gap in health promotion, particularly because digital nudges and recommender systems can encourage and support sustained healthier choices in nutrition, physical activity (PA), and sedentary behavior reduction to prevent overweight and obesity. Objective: This scoping review addresses these gaps by exploring how digital nudges and recommender systems are used in obesity prevention. Methods: We prospectively published the scoping review protocol and adhered to the PRISMA-ScR (Preferred Reporting Items for Systematic Reviews and Meta-Analyses extension for Scoping Reviews) guidelines. Eligibility was defined using the PCC (Population, Concept, Context) framework. We searched 7 bibliographic databases (MEDLINE and PsycINFO via Ovid, Web of Science, CINAHL via EBSCO, Scopus, ACM Digital Library, and IEEE Xplore) up to October 2023. Following a 2-stage screening by independent reviewers, we selected 68 articles that included 94 user evaluations. Results: Most articles (36/68, 53%) report on recommender systems focused on nutrition, with fewer (16/68, 23%) aiming to promote PA. Most studies on digital nudges (11/68, 16%) targeted nutrition-related nudges for shopping and meal selection (8/68, 2%). Articles address PA and sedentary behavior less frequently (3/68, 4%). Three out of 68 (4%) articles report on recommender systems in combination with games, and 2 out of 68 (3%) articles report on recommender systems and digital nudges. Approaches to item retrieval vary widely, with 31 out of 68 (46%) articles failing to describe their methods. In the scoping review, we found a discrepancy between the target group for which the system was developed and the group with which the evaluation was conducted. Sixty-eight evaluations report positive results, while 26 studies report mixed, negative, or no-difference results. Conclusions: Integrating digital nudges and recommender systems might hold potential in overweight and obesity prevention by subtly encouraging healthier lifestyle choices. However, the heterogeneity in study designs, outcome measures, and reporting quality limits the comparability of findings and makes it difficult to draw robust conclusions about effectiveness. Future work should include detailed definitions, mechanism descriptions, broader geographic representation, and rigorous intervention testing and user evaluations to fully leverage these systems for improved health outcomes and to support sustainability and well-being objectives.