Experience developing a pediatric medical chatbot in Singapore: a digital innovation for improved emergency care

Editorial reading guide · Collaboration

How would an automated pediatric advice service connect us to a clinician when a question exceeds the bot’s role?

Describes a service that joined automated information/triage with access to emergency doctors and nurses. The important collaboration is the handoff and the clinical team’s oversight of predefined advice. Most completed flows advised home monitoring, but the service could not link anonymous chats to actual care decisions or clinical outcomes.

Study context

Singapore children’s emergency service operational data: 57,302 chatbot conversations during 2022–2024 and 14,289 completed triage flows

Age / stage: Children served by a pediatric emergency department; chatbot users/caregivers not individually characterized

Read with three questions

1. Which parts of the service used automated clinical pathways and which used live staff?
2. What can a completed triage recommendation tell us, and what can it not tell us about what a family did next?
3. How do low feedback response rates and changing usage complicate satisfaction and impact claims?

Keep the limits in view

This operational case report had no comparator and no patient-linked follow-through, emergency-attendance, accuracy, or safety outcome validation. Feedback response rates were low and differed between chatbot and live chat. Explanations for declining use were speculative. Current GenAI integration was proposed, not evaluated in the reported service.