AI-based Diagnostic Solutions

4 May 2023

NOTICE PAPER NO. 1885 
NOTICE OF QUESTION FOR WRITTEN ANSWER
FOR THE SITTING OF PARLIAMENT ON 8 MAY 2023

Name and Constituency of Member of Parliament
Assoc Prof Jamus Jerome Lim
MP for Sengkang GRC

Question No. 2944

To ask the Minister for Health (a) whether the Ministry has explored the credibility of AI-based diagnostic solutions; and (b) if so, what are the regulatory considerations that will be most important if there are more widespread adoption of such technologies.

Answer

1     Artificial Intelligence (AI) is ubiquitous, and being used in healthcare in Singapore.  For example, it is used in the JARVISDHL a programme by SingHealth and NUS to prevent Diabetes, Hypertension, hyperLipidemia.  It is the engine for SELENA+ (Singapore Eye LEsioN Analyzer plus) to diagnose diabetic retinopathy. It is also used for diagnosis and management of cardiovascular and chronic diseases.  It is deployed in health apps for preventive care, as well as in pharmaceutical research.  

2     MOH is open to the use of AI technologies in healthcare where they are proven to be safe, clinically efficacious, and cost-effective.  With the rapid advancement of AI technology, we believe many treatments using AI will emerge.  The challenge is to embrace and encourage them, while ensuring the safe adoption and deployment of AI in healthcare.  To do so, we have progressively put in place safeguards over the past few years.

3     MOH’s AI in Healthcare Guidelines guide AI-MD developers and implementers on the good practices they should adopt in rolling out such clinical services to patients.  

4     The Health Sciences Authority regulates multiple aspects of AI Medical Device (AI-MD) development as part of its medical device registration process to ensure the quality, safety, and efficacy of AI-MDs, and requires continued post-market surveillance to ensure continued real-world effectiveness.

5     MOH will be issuing an AI Governance Framework for healthcare that recommends detailed development and implementation-level good practices, such as risk prompts to help identify risks and controls for risk mitigation. 

6     The Bioethics Advisory Committee has also issued a draft advisory report and recommendations on the ethical use of big data and AI in human biomedical research on 2 May 2023 for public consultation.

7     MOH will continue to study international best practices, consolidate and develop our AI governance framework.