Surgeons face daily challenges in identifying key medical issues and optimising patient health before surgery, often navigating complex perioperative guidelines.
The rapid pace and frequent updates make consistent adherence difficult, risking suboptimal preparation and procedural delays, while increasing administrative burdens.
Now, PEACH (PErioperative AI Chatbot), a revolutionary clinical decision support system, is set to transform this landscape. By enabling physicians to input patient data from electronic health records seamlessly, PEACH performs critical tasks, promising to enhance surgical readiness and streamline workflows for better patient outcomes.
These include:
• Answering perioperative-related anaesthesia questions
• Drafting referral letters and patient instructions
• Summarising key anaesthesia-related issues
• Generating perioperative care plans.
Dr Ke Yuhe, associate consultant with the Department of Anaesthesiology at Singapore General Hospital, led the development of PEACH.
He said: ‘PEACH represents a significant step forward in how we manage preoperative assessments. By integrating our comprehensive perioperative guidelines with AI technology, we've created a tool that not only enhances patient safety but also helps doctors make more efficient and consistent decisions.’
Large Language Models (LLMs) are emerging as powerful healthcare tools, but their use with clinical data raises substantial concerns regarding data privacy and information security.
To address these issues, Dr Ke and her team integrated SGH’s perioperative guidelines into PAIR, an AI assistant created by Open Government Products (OGP), which is accessible only on hospital-issued, encrypted laptops.
The chatbot can also triage pre-surgery patients, recommending whether assessment should be conducted via phone, in-person at the pre-admission clinic or on surgery day.
This capability helps reduce waiting time at pre-admission clinic, which sees about 25,000 pre-surgery patients annually, allowing doctors to focus on complex cases and minimising unnecessary hospital visits.
Two comprehensive studies validated PEACH’s effectiveness, leading to Singapore’s Health Sciences Authority’s approval for its deployment in healthcare settings, including SGH.
The first study in November 2024, examining 240 PEACH interactions, demonstrated about 98% accuracy in making pre-surgery recommendations, with minimal hallucination (less than 2%) and guideline deviation rate (less than 1%).
The second study, conducted between January and February 2025 across more than 270 actual patient assessments, showed that doctors primarily used PEACH for complex cases.
They were also asked to rate the chatbot on a scale of 1 to 5, and it scored 4.94 for safety, 4.81 for explainability, and 4.72 for comprehensibility.
Human reviewers confirmed the accuracy of the output across the studied cases, with PEACH proving particularly effective in handling medium-complexity cases, reducing documentation time by nearly six minutes per patient.
Overall, PEACH implementation was projected to save over 600 junior doctor hours and nearly 60 senior doctor hours annually.
Associate Professor Hairil Rizal, Senior Consultant and Clinician Scientist, Department of Anaesthesiology, SGH, said: ‘What excites me most about PEACH is how it is transforming our clinical workflow at PAC. When you're seeing thousands of pre-surgery patients annually, every minute saved on administrative tasks is a minute gained for patient care.
‘Our studies show that doctors are confidently using PEACH for complex cases, and the system’s ability to triage patients effectively means we can better allocate our resources where they're needed most. This is about working smarter to deliver better care.’
