KKH Nurses Launch AI Sidekick for Pediatric Care
- •Nurses at KKH in Singapore built a chatbot to automate pediatric drug dosage calculations.
- •The Pediatric Nursing Sidekick cuts calculation time to seconds using the government-secure Pair platform.
- •The team plans a cluster-wide rollout after reaching 200 conversations in two months of use.
Nurses at Singapore's KK Women’s and Children’s Hospital (KKH) have developed the "Pediatric Nursing Sidekick," an AI-powered chatbot designed to streamline clinical workflows for ward staff. By using the government-secure AI platform Pair, nurses created a tool that automates complex weight-based drug dose calculations and answers ward protocol questions. This implementation reduced the time required for medication calculations from minutes to seconds, while also providing safety alerts when dosage levels exceed designated thresholds.
The project emerged after previous attempts using third-party low-code platforms and student collaborations failed due to data security regulations, technical maintenance requirements, and hardware restrictions. Pair, the government-backed tool, addressed these barriers by eliminating server management and allowing direct access via existing ward laptops. According to Advanced Practice Nurse (APN) Joanne Cheng, the system logged over 200 conversations within the hospital in the first two months.
The development process relied on nurses using natural language prompts to generate Python code for the chatbot, iterating based on feedback from clinical staff. This bottom-up approach proved highly effective, as the tool directly addressed frontline pain points. The team’s success prompted the pharmacy division to contribute additional data, and discussions are currently underway for a cluster-wide rollout across the SingHealth network.
The team also participated in a nursing hackathon as part of the annual SingHealth Nursing & Research Innovation event, which provided a platform to refine their prompts and pitch the tool to stakeholders. The nurses emphasized that the project succeeded through extensive networking and openness to sharing partial solutions, rather than a top-down mandate. The tool's ability to evolve based on constant feedback from users served as the primary driver for its adoption and eventual consideration for larger-scale implementation.