NUH Uses AI-Assisted Evidence Tools to Support Clinical Decision Making, Clinical Education and Training
Healthcare systems are facing an unprecedented expansion in medical knowledge. While authoritative guidelines and textbooks remain foundational, clinicians increasingly rely on a combination of structured references and AI-assisted tools to navigate day to day clinical questions and in academic hospital environments like NUH, these tools are also reshaping how medical education is delivered.
Three categories of AI-assisted tools are commonly used together across both clinical and educational settings.
ClinicalKey Physician serves as a core clinical reference library, providing access to textbooks, practice guidelines, and peer reviewed content. Clinicians turn to it when they need structured, domain-specific information, for example, reviewing standard approaches to disease management or refreshing knowledge in less frequently encountered conditions. In formal education pathways, our educators use it to anchor teaching in recognised reference materials, giving our trainees a reliable starting point for understanding conditions, therapies, and guidelines.
ClinicalKey AI builds on this content base by enabling natural language search and summarisation, with direct links back to source material. In practice, clinicians often use it to rapidly explore questions such as:
• What does recent literature say about treatment options for a specific patient profile?
• How do different guidelines approach a clinical scenario?
The emphasis remains on traceability as answers are reviewed alongside the underlying references rather than taken at face value.
In case-based learning, our trainees use ClinicalKey AI in pretty much the same way as our clinicians, posing natural language questions that arise during ward discussions or tutorials, then tracing answers back to original sources. This reinforces habits of verification and critical reading, rather than shortcutting them.
OpenEvidence is used throughout the day like ClinicalKey AI. As an open and lightweight evidence synthesis tool, it supports tasks such as quick literature scanning, comparison of trial findings, or exploratory questioning during care, teaching and discussion.
In educational settings, OpenEvidence has found a particular niche in generating discussion prompts for tutorials, exploring how evidence has evolved over time, and supporting exam question design or reflective learning. Because it is not tied to institutional workflows, its use tends to be flexible and learner-driven.
Across all three AI-assisted tools, the common principle is augmentation rather than replacement. Medical education increasingly takes place in parallel with service delivery, leaving limited time for traditional literature review and teaching methods, and these tools help reduce friction in accessing evidence. Within this reality, AI-assisted evidence tools have started to play a supporting role in clinical learning environments.
Clinical judgement, peer discussion, and institutional protocols continue to guide decisions. The AI-assisted tools do not replace appraisal skills. Instead, they serve as assistive scaffolding, helping both clinicians and trainees engage with evidence more efficiently while maintaining academic rigour.
Topic: Patient Safety Technologies, Artificial intelligence, Machine learning, Technological innovation, Virtualization
