Jiabin Shen

Objective Over 7,000 American children die from pedestrian injuries annually, and pedestrian injury ranks among the top 5 causes of unintentional child death. Prevention efforts are multifaceted. One strategy, use of virtual reality (VR) to teach children to cross streets, is of growing interest to public health practitioners. The present study is a systematic review and meta-analysis that examined the efficacy of using VR to teach children pedestrian safety. Methods Following PRISMA guidelines, searches among 7 databases were completed, followed by abstract/full-text screening and data extraction. Hedge’s g was computed for the effect sizes of 3 outcomes: pedestrian knowledge, pedestrian-relevant cognition (attention to traffic, time to contact, start delay), and pedestrian behaviors (safe crossings, unsafe crossings). Risk of bias was assessed using Cochrane guidelines. Meta-regression analyses and subgroup analyses were conducted including 3 moderators: age, sex, and immersion level. Results A total of 20 studies, reported in 24 articles, were included in the qualitative analysis. Meta-analysis of the 12 studies with sufficient quantitative data available showed a statistically significant medium effect of VR safety interventions on child pedestrian knowledge and behavior. Mixed results emerged for the effect of VR safety interventions on children’s pedestrian-relevant cognition. Age and sex moderated the effect of VR training on pedestrian knowledge. Conclusions This synthesis of the literature on pediatric VR pedestrian safety interventions suggests an overall beneficial impact of VR interventions to teach children how to cross streets safely. Efforts should continue to develop and disseminate effective VR interventions. Registration ID CRD42022309352