Digital Health and Software as a Medical Device (SaMD)
HSA's digital health programme
HSA has a dedicated Digital Health programme to address the growing number of software-based medical devices, mobile health applications, and AI-assisted diagnostic tools entering the Singapore market. The digital health section of HSA's website provides current guidance and updates:
hsa.gov.sg/medical-devices/digital-health
What is SaMD?
Software as a Medical Device (SaMD) is software that is intended to be used for medical purposes and meets the definition of a medical device under the Health Products Act — without being part of a hardware medical device.
SaMD includes:
- AI-powered diagnostic imaging analysis tools;
- Clinical decision support software that recommends specific diagnoses or treatments;
- Remote patient monitoring apps that analyse physiological data;
- Apps that calculate drug dosing based on a patient's clinical parameters and disease state;
- Software that classifies or predicts patient risk for a specific condition.
SaMD does not include:
- Software that merely stores, archives, or displays data without analysing it;
- Administrative or workflow software (scheduling, billing, EHR);
- General wellness apps with no medical intended purpose;
- Software that controls or drives a hardware medical device (this is part of the hardware device, not standalone SaMD).
HSA's digital health guidance follows the IMDRF SaMD framework (including the IMDRF SaMD N41 document on nomenclature and the N12 framework for risk-based categorisation).
Classification of SaMD
SaMD is classified using the same Class A–D framework as other medical devices. Classification depends on:
- The significance of the information provided by the software (informing, driving, or treating a condition);
- The severity of the condition the software is intended to address (critical, serious, or non-serious);
- Whether incorrect output could cause direct patient harm or harm through a clinical action based on the output.
As a general guide:
- SaMD providing information for non-serious conditions — Class A or B;
- SaMD driving clinical decisions for serious conditions — Class C;
- SaMD driving clinical decisions in critical/life-threatening situations — Class D.
Use the HSA risk classification tool for a device-specific assessment.
Immediate registration route for standalone SaMD
A dedicated Immediate registration route is available for standalone medical mobile applications (Class B and Class C SaMD), provided:
- The SaMD has approval from ≥1 overseas reference regulatory agency; and
- No known global safety issues; and
- No prior rejection or withdrawal by HSA or any overseas reference agency.
This route provides a significantly faster path to market for Class B and C SaMD products that have already been cleared or approved in a reference jurisdiction (e.g. FDA De Novo, CE marking under MDR).
AI and machine learning in medical devices
Artificial intelligence and machine learning (AI/ML) software used for medical purposes is regulated as SaMD where it meets the medical device definition. HSA follows IMDRF guidance on AI/ML-based SaMD, including considerations for:
- The transparency and explainability of AI outputs;
- Continuous learning algorithms and how software updates are managed;
- The role of the clinician vs the algorithm in clinical decision-making.
AI/ML medical device products with novel architectures or intended uses that do not fit existing classification guidance are encouraged to engage HSA's Innovation Office early: hsa.gov.sg/medical-devices/io
Cybersecurity for medical devices
HSA has published guidance on cybersecurity requirements for medical devices, particularly software-based and network-connected devices. Manufacturers and registrants should ensure their technical documentation addresses:
- Cybersecurity risk management throughout the device lifecycle;
- Software bill of materials (SBOM);
- Post-market cybersecurity monitoring and patching.
HSA follows IMDRF's Principles and Practices for Medical Device Cybersecurity guidance.