Effectiveness of a universal digital–human parenting intervention in promoting early childhood development and protection: A pragmatic cluster randomized controlled trial

Abstract

Delayed early childhood development and violence against children are major global challenges, particularly in low-resource settings. Universal digital–human parenting interventions may offer a scalable solution by overcoming barriers associated with traditional in-person programs. This study reports the first pragmatic randomized controlled trial evaluating a blended chatbot-based parenting intervention delivered within the Chinese preschool system. The trial was conducted in a lower-middle-income city in central China. Twenty-one preschool classes were cluster-randomized to a treatment group (n = 10) or waitlist control (n = 11). Primary caregivers of enrolled children participated following informed consent. The intervention comprised a 2.5-month chatbot-led digital parenting program supported by weekly or twice-weekly online group sessions facilitated by headteachers and social workers. Data were collected at baseline, post-intervention, and at 6- and 12-month follow-ups. Primary outcomes were caregiver-provided early learning and stimulation, and caregiver-perpetrated violence. Analyses followed the intention-to-treat principle using multilevel regression models. Equity effects related to caregiver and child disability were explored through moderation and subgroup analyses. Sustainability of impacts was assessed, and complier average causal effects examined the role of intervention completion. Between March 2024 and June 2025, 541 caregivers of children aged 3–6 years were enrolled (treatment: n = 272; control: n = 269), of whom 25.2% were male. Overall, 60.3% completed all chatbot modules. At post-intervention, the program significantly improved early learning and stimulation (β = 1.79, 95% CI [0.24, 3.34]) and reduced caregiver-perpetrated violence (IRR=0.87, 95% CI [0.80, 0.96]). The intervention showed potential to advance equity for families affected by disability, with some effects sustained at follow-up. Complier analyses indicated reduced endorsement of corporal punishment and lower parental anxiety among participants completing at least 30 modules. Universal digital–human parenting interventions embedded in preschool systems can enhance early childhood development and reduce violence, highlighting the importance of human support and cultural adaptation to optimize engagement and outcomes.