Effectiveness of the Sleep Island WeChat Mini-Program for Promoting Sleep in Preschool Children

The Effectiveness of "Sleep Island": an LLM-enhanced Mini-Program to Promote Sleep in Preschool Children: a Randomized Controlled Trial

Behavioral sleep problems are common during the preschool years and are strongly influenced by parental sleep-related practices and the home environment. This randomized controlled trial aims to evaluate the effectiveness of Sleep Island, a parent-focused WeChat mini-program adapted from a theory- and evidence-based parent-child collaborative behavioral sleep promotion program.

A total of 122 families with preschool children will be randomly assigned in a 1:1 ratio to either the Sleep Island intervention plus routine sleep-health education or a waitlist control condition receiving routine sleep-health education. The intervention will be delivered over 4 weeks. Sleep Island incorporates electronic sleep monitoring, visualization of sleep patterns, individualized sleep-behavior feedback, sleep-health education, and personalized behavioral recommendations. Recommendations are generated using a large language model supported by retrieval-augmented generation from a curated pediatric sleep knowledge base and are reviewed by the research team before release.

Outcomes will be assessed at baseline, immediately after the intervention, and at 3- and 6-month follow-up. The primary outcome will be objectively measured child nocturnal sleep duration. Secondary outcomes will include other objective and parent-reported indicators of child sleep health, parental sleep-related practices, sleep environment, behavioral sleep problems, and parental knowledge, attitudes, self-efficacy, and beliefs regarding child sleep.

Behavioral sleep problems during early childhood are influenced by interactions among child characteristics, parental sleep-related practices, parent-child interactions, and the family environment. The original intervention was developed using the Intervention Mapping framework to identify modifiable determinants of preschool children's sleep and to link behavioral objectives with theory- and evidence-based intervention strategies.

Following development of the original program, the intervention was systematically adapted from primarily in-person delivery to a self-guided digital format delivered through a WeChat-accessible mini-program named Sleep Island. The digital adaptation was designed to preserve the principal behavioral targets and active intervention functions of the source program while reducing reliance on scheduled face-to-face contact. Core behavioral targets include establishing regular sleep-wake schedules, developing positive and predictable bedtime routines, improving parent-child bedtime interactions, optimizing the sleep environment, and using evidence-based responses to common behavioral sleep difficulties.

Sleep Island includes four principal intervention components: (1) electronic sleep assessment and monitoring; (2) graphical visualization of sleep patterns; (3) individualized sleep-behavior feedback; and (4) personalized sleep-promotion recommendations provided through the Sleep Silkbag. Additional sleep-health educational materials are available within the mini-program. Parents initially create a child profile and provide baseline sleep information. They then record their child's sleep using the electronic sleep diary. Recorded sleep information is summarized graphically and is used to generate individualized recommendations addressing sleep schedules, bedtime routines and activities, the sleep environment, and parent-child bedtime interactions.

Individualized recommendations are generated using a large language model operating within a retrieval-augmented generation architecture. The model is grounded in a domain-specific knowledge base developed from the original behavioral intervention protocol and pediatric sleep-health guidance. Generated recommendations are reviewed by the research team before being released to families to ensure that the recommendations are consistent with the intervention protocol and do not contain inappropriate or factually inaccurate advice.

In this randomized controlled trial, 122 eligible families will be randomized in a 1:1 ratio to the Sleep Island intervention plus routine health education or to a waitlist control group receiving routine health education. The intervention period will last approximately 4 weeks. Outcomes will be assessed at baseline, immediately after completion of the intervention, 3 months after the intervention, and 6 months after the intervention. The study will primarily evaluate whether the digitally delivered behavioral sleep-promotion program improves preschool children's sleep duration and will additionally examine changes in other dimensions of child sleep health, parental sleep-related practices, the home sleep environment, and parental sleep-related knowledge, attitudes, beliefs, and self-efficacy.

Inclusion Criteria:

• Parent or primary caregiver of a child aged 3 to 6 years.
• Child is generally healthy and living with the participating parent/caregiver.
• Parent/caregiver is responsible for the child's daily care and sleep-related routines.
• Parent/caregiver has access to a smartphone and is able to use WeChat and the Sleep Island mini-program.
• Parent/caregiver is willing to participate in the study procedures and follow-up assessments.
• Parent/caregiver provides informed consent for participation.

Exclusion Criteria:

• Child has a history of major perinatal complications, including significant birth resuscitation, asphyxia, hypoxic-ischemic encephalopathy, intracranial hemorrhage, or preterm birth.
• Child has a serious congenital or physiological disease that may substantially affect development or sleep.
• Child has a neurodevelopmental disorder or developmental delay.
• Child has a medical condition known to substantially affect sleep.
• Parent/caregiver has a severe mental disorder or other condition that would substantially impair their ability to report the child's sleep or participate in the intervention.
• Parent/caregiver is unwilling or unable to complete study procedures.