An AI-based chatbot to support health-related social needs among pediatric primary care population: Protocol for a pilot randomized controlled trial

Clinical Snapshot

Key Findings

• P Value: Not applicable - protocol paper

• Effect Size: Not applicable - protocol paper

• Primary Outcome: Study feasibility, acceptability, and usability measures

• Nnt Or Sensitivity: Not applicable - feasibility study protocol

• Confidence Interval: Not applicable - protocol paper

Clinical Application

Requires technological infrastructure and staff training; integration with existing electronic health records needed Would require adaptation to Australian social services landscape, Medicare system integration, and compliance with Australian privacy regulations. Relevance to RACGP guidelines on social determinants of health screening Caregivers of young children in primary care settings with potential unmet social needs

Abstract

Unmet health-related social needs (HRSNs) are major drivers of poor health outcomes in early childhood. Children with unmet HRSNs are at greater risk for developmental delays, caregiver stress, and increased healthcare utilization, yet current screening approaches in pediatric primary care are resource-intensive and inconsistently implemented. AI-powered chatbots (conversational agents or virtual assistants) may offer a private, secure, scalable, and cost-effective alternative for identifying unmet needs and connecting families to services. This protocol describes a pilot randomized controlled trial designed to evaluate the feasibility, acceptability, and usability of DAPHNE, an AI-driven chatbot developed to facilitate the identification of unmet HRSNs and provide personalized community resource referrals. One hundred caregivers of children under two years of age will be recruited from Nationwide Children's Hospital pediatric primary care clinics and randomized to either the standard care (control) group or DAPHNE+ Standard care (intervention) group (n = 50 each arm). Caregivers will complete surveys at baseline, 1 month, 3 months, and 6 months post-intervention (depending on the measure). For the intervention group, participants will receive weekly chatbot prompts and on-demand access throughout the 6-month study period. Primary outcomes include study feasibility (recruitment, retention, and survey completion across both arms), acceptability (caregiver-reported ratings in both arms and intervention-specific ratings), and usability of the DAPHNE chatbot (System Usability Scale among intervention participants). Secondary outcomes includes caregiver-reported outcome measures (caregiver stress, self-efficacy, satisfaction with resource access, quality of life), and electronic health record-derived measures (including documentation of HRSN screening and referrals, adherence to well-child visits, missed appointments, emergency department utilization, and estimated healthcare costs). In addition, ten primary care providers will also participate to assess workflow integration and report on current HRSN practices. Mixed-methods analyses will integrate survey data, chatbot engagement metrics, and qualitative interviews to refine both the intervention and the study protocol. The results of this study will inform the design of a future multi-site trial to evaluate the efficacy and implementation of DAPHNE for addressing HRSNs in pediatric primary care. Trial registration: NCT07168382.

References

1. 1.Sezgin, E., Clarkson, E., Logan, F., Jackson, D. I., Hussain, S. A., Stokes, J., Bunger, A., Brock, G., Fosler-Lussier, E., Kemper, A. R., & Pai, A. L. (2026). An AI-based chatbot to support health-related social needs among pediatric primary care population: Protocol for a pilot randomized controlled trial. PLoS One. https://doi.org/10.1371/journal.pone.0337868