JMIR Publications

ABSTRACT

Background:

Current food consumption patterns are driving the steep increase of obesity and non-communicable diseases, and exacerbate environmental degradation. Recent research is increasingly exploring the potential of digital media to encourage positive changes in eating habits. Despite this growing interest, there is considerable uncertainty about the effectiveness of digital interventions in improving dietary behavior, and a comprehensive review comparing intervention approaches is lacking.

Objective:

The aim of this systematic review and meta-analysis is to evaluate the effectiveness of digital interventions in improving healthy and sustainable food consumption and explore differences in effects due to intervention characteristics, particularly the digital medium and behavior change techniques employed.

Methods:

A systematic search strategy was conducted across Web of Science, Embase and Scopus, supplemented by forward and backward reference searching. Eligibility screening led to the inclusion of 46 articles encompassing 62 interventions. Each intervention was coded for behavior change techniques using the Behavior Change Taxonomy v1. A meta-analysis with robust variance estimation was performed, and we conducted moderator analyses to evaluate the intervention characteristics associated with change in eating behavior.

Results:

A meta-analysis with robust variance estimation revealed a small positive effect of digital interventions on food consumption outcomes (d=0.33, P<.001), however substantial heterogeneity indicates considerable variation in effectiveness across intervention characteristics. Moderator analyses indicated that intervention effectiveness differed by platform, with social media interventions yielding significantly stronger effects than other modalities. Incorporating prompts or cues significantly enhanced the effectiveness of digital interventions. Additionally, interventions including social support and comparison of behavior yielded larger effects (although not statistically significant), and the narrative synthesis of findings highlighted the effectiveness of social norm communication. Interventions aimed at reducing meat consumption and promoting plant-based eating showed comparable effects. Although digital interventions appeared most effective among young adults, age-related differences were not statistically significant. Few studies included adolescents or individuals from lower socioeconomic backgrounds.

Conclusions:

This review underscores the potential of digital interventions in improving eating behavior, though their success varies by intervention design and population targeted. There is a lack of research on adolescents, and despite our aim to explore effects specifically among low SES groups, limited available research restricted our ability to conduct subgroup analyses for this population. Clinical Trial: PROSPERO CRD42023487955; https://www.crd.york.ac.uk/PROSPERO/view/CRD42023487955