BackgroundDigital interventions may help address low vegetable intake in adults, however there is limited understanding of the features that make them effective. We systematically reviewed digital interventions to increase vegetable intake to 1) describe the effectiveness of the interventions; 2) examine links between effectiveness and use of co-design, personalisation, behavioural theories, and/or a policy framework; and 3) identify other features that contribute to effectiveness.MethodsA systematic search strategy was used to identify eligible studies from MEDLINE, Embase, PsycINFO, Scopus, CINAHL, Cochrane Library, INFORMIT, IEEE Xplore and Clinical Trial Registries, published between January 2000 and August 2022. Digital interventions to increase vegetable intake were included, with effective interventions identified based on statistically significant improvement in vegetable intake. To identify policy-action gaps, studies were mapped across the three domains of the NOURISHING framework (i.e., behaviour change communication, food environment, and food system). Risk of bias was assessed using Cochrane tools for randomized, cluster randomized and non-randomized trials.ResultsOf the 1,347 records identified, 30 studies were included. Risk of bias was high or serious in most studies (n = 25/30; 83%). Approximately one quarter of the included interventions (n = 8) were effective at improving vegetable intake. While the features of effective and ineffective interventions were similar, embedding of behaviour change theories (89% vs 61%) and inclusion of stakeholders in the design of the intervention (50% vs 38%) were more common among effective interventions. Only one (ineffective) intervention used true co-design. Although fewer effective interventions included personalisation (67% vs 81%), the degree of personalisation varied considerably between studies. All interventions mapped across the NOURISHING framework behaviour change communication domain, with one ineffective intervention also mapping across the food environment domain.ConclusionFew digital interventions identified in this review were effective for increasing vegetable intake. Embedding behaviour change theories and involving stakeholders in intervention design may increase the likelihood of success. The under-utilisation of comprehensive co-design methods presents an opportunity to ensure that personalisation approaches better meet the needs of target populations. Moreover, future digital interventions should address both behaviour change and food environment influences on vegetable intake.