Peer-reviewed article

Effect of an artificial intelligence-assisted community-based intervention on child restraint system practices in Shanghai, China.

Ning Gao, Xihui Wang, Shuna Gao, Juanjuan Peng, Yang Zheng et al. · 20 Apr 2026

Abstract

BACKGROUND: Child restraint systems (CRS) can effectively prevent injuries to children in road traffic crashes. However, in China, the rates of CRS ownership and use remain relatively low. METHODS: We conducted a community-based intervention trial with an integrated multimodal intervention from August 2024 to March 2025 in five communities in Huangpu District, Shanghai. A total of 501 participants were recruited (201 in the intervention group, 300 in the control group). The intervention included (1) distribution of CRS educational brochures and health education materials, (2) dissemination of online educational articles, and (3) Artificial intelligence-assisted (AI-assisted) voice calls delivering standardized reminders and safety education. The control group received only pamphlets on dietary nutrition. RESULTS: A total of 501 participants completed the study. After the intervention, the rate of CRS ownership showed no significant difference between the intervention and control groups (90.05% versus 91.33%, p = 0.74). However, the rate of consistent CRS use was significantly higher in the intervention group (50.75%) than in the control group (38.33%) (p < 0.01). The intervention group also showed fewer orientation-related installation errors than the control group (0.00% versus 12.55%, p < 0.01). Logistic regression with interaction terms indicated no significant intervention effect on ownership rate (OR = 0.97, 95% CI: 0.41-2.30), but a significant increase in consistent use rate (OR = 2.18, 95% CI: 1.25-3.82). Further analysis showed that ownership rate was associated with child age, parental education level, vehicle price, and travel frequency, while consistent use rate was associated with household registration, travel frequency, and travel distance. CONCLUSION: An integrated community-based intervention combining conventional health education with AI-assisted follow-up improved consistent CRS use and reduced orientation-related installation errors. This approach may be a useful complement to current CRS promotion strategies.