Singapore considers using AI tools to improve diagnosis in resource-limited medical settings

After cardiac arrest, families and doctors are often faced with uncertainty about the patient’s chances of recovery. This uncertainty is even greater in hospitals with limited resources, with limited access to advanced diagnostic tools and large datasets.

As an example of how artificial intelligence (AI) can help bridge this gap, researchers at Duke-NUS Medical School in Singapore and their collaborators adapted an advanced AI model to accurately predict neurological recovery after cardiac arrest in resource-limited settings.

Published in npj digital medicineIn this study, we applied transfer learning, an advanced AI approach that adapts a pre-trained model built on a large dataset to a new setting with limited local data. This method is particularly suitable for low- and middle-income countries because it improves performance in new environments without requiring extensive data collection.

AI tools have the potential to improve healthcare delivery, but appropriate governance frameworks are essential for safe and ethical implementation. Existing regulations for medical technology often do not address the inherent risks of AI, such as privacy concerns and model hallucinations, and do not clearly enforce accountability for the safe deployment and monitoring of new tools.

To address these gaps, researchers led by Duke-NUS proposed the creation of an international consortium, the Partnership for Oversight, Leadership, and Accountability in the Regulation of Intelligent System Generation Models in Healthcare (POLARIS-GM).

The consortium aims to establish practical best practice guidance to regulate new tools, monitor their impact, establish safety guardrails, and adapt them to resource-limited settings.