SIMFONI Pipe Cleaner 1 (Fine-tuning in enTRUST with Public Data Assets)
Main Applicant – Dr Mukkesh Kumar, Head of Data Management Platform, A*STAR

Proactive monitoring before surgery could help doctors respond sooner, reduce complications, and support faster recovery. To address this, the SurgiSentinel project developed an AI agent that helps doctors identify early signs of post-surgical risk in patients preparing for elective surgery. The system was trained with anonymised patient data from Seoul National University Hospital. This earlier work showed that AI could support more proactive care, and it was recognised with the IMDA Agentic AI Special Award at IMAGINE AI 2024.

The SIMFONI programme now aims to explore whether a similar AI agent can be successfully adapted. As the first step, SIMFONI Pipe Cleaner 1 (PC1) will attempt to reproduce the SurgiSentinel training process within the enTRUST platform. SurgiSentinel was chosen for this study because it provides a clear and practical real-world example of how AI can be used to improve patient safety after surgery.

PC1 will help lay the foundation for later work in SIMFONI, including Pipe Cleaner 2 (PC2) with the national health clusters (NHG, NUH, SHS). If successful, this could help enable faster intervention, fewer complications, and shorter hospital stays for surgical patients, while also demonstrating how AI tools can be responsibly introduced into Singapore’s healthcare data analytics environment.