SaMD Classification: How to Classify Software as a Medical Device

Software as a Medical Device (SaMD) is the fastest-growing — and most confusing — category in medical device regulation. The rules differ dramatically between markets, the stakes are high (misclassification can mean years of delay), and the landscape is changing rapidly.

If you're building health software, AI diagnostics, clinical decision support, or any digital health tool, this guide will help you understand where you stand.

What Counts as SaMD?

The IMDRF definition: software intended to be used for one or more medical purposes that perform these purposes without being part of a hardware medical device.

SaMD examples:

• AI that reads X-rays and flags potential fractures
• Software that calculates drug dosing based on patient parameters
• An app that monitors cardiac arrhythmia using phone sensors
• Clinical decision support that recommends treatment pathways

NOT SaMD:

• Software that drives a physical medical device (that's "software in a medical device" or SiMD)
• General wellness apps (fitness tracking, meditation, calorie counting)
• Electronic health records (EHR) that just store and display data
• Administrative software (scheduling, billing)

The line between "wellness" and "medical device" is where most disputes happen. If your software makes a claim about diagnosing, treating, or preventing a disease, regulators will likely consider it SaMD.

The IMDRF Framework

The International Medical Device Regulators Forum (IMDRF) created a standardized SaMD classification framework used (with variations) by most major markets. It's based on two factors:

1. Significance of the information provided by SaMD

2. Seriousness of the healthcare situation

The IMDRF Classification Matrix

Most regulatory authorities map these IMDRF levels to their local classification system — but not identically.

FDA Approach to SaMD

FDA has been relatively pragmatic about SaMD:

• Clinical Decision Support (CDS) exemption: Software that meets all four CDS criteria is exempt from device regulation. The key criterion: the software must not be intended to acquire, process, or analyze a medical image, signal, or pattern.
• AI/ML Predetermined Change Control Plan: FDA allows manufacturers to describe future algorithm changes upfront, avoiding repeated 510(k) submissions.
• Most SaMD is Class II. The majority of AI/ML devices cleared by FDA have gone through 510(k) or De Novo.

FDA SaMD classification (typical):

As of mid-2025, FDA has authorized over 1,000 AI/ML-enabled devices (up from ~950 in late 2024), the vast majority as Class II.

EU MDR Approach to SaMD

The EU takes a stricter approach through Rule 11 of MDR Annex VIII:

Software intended to provide information used to take decisions with diagnosis or therapeutic purposes is classified as Class IIa, except if such decisions have an impact that may cause death or irreversible deterioration → Class III.

This means:

• Diagnostic software for serious conditions (cancer, cardiac) → Class III in EU (vs. typically Class II in US)
• General clinical decision support → Class IIa minimum
• Software that monitors physiological parameters → Class IIa–IIb (Rule 10)

The EU MDR SaMD gap:

Key difference: There is effectively no Class I software under EU MDR. Even low-risk SaMD starts at Class IIa, requiring notified body involvement.

Asia-Pacific SaMD Landscape

Singapore (HSA)

• Follows IMDRF framework closely
• SaMD classified as Class B minimum
• HSA issued specific SaMD guidance in 2022
• Reliance pathway available if FDA/CE approved

Japan (PMDA)

• SaMD requires pre-market approval for Class II+
• Japan's DASH (Developing and Advancing SaMD in Healthcare) program offers expedited pathways
• Local clinical validation may be required

South Korea (MFDS)

• Specific SaMD classification rules published 2020
• AI devices require clinical performance validation
• Novel pathway for "innovative medical devices" with expedited review

China (NMPA)

• Published AI/ML-specific guidance in 2023
• Most diagnostic AI → Class II or III
• Requires Chinese clinical validation data
• Local data hosting requirements

Strategic Implications for SaMD Companies

1. Market entry order matters more for SaMD

Because classification varies so widely, a SaMD product classified as Class II in the US might be Class III in the EU. This means:

• US-first is often the fastest path for SaMD
• EU entry may require significantly more clinical evidence
• Singapore can be a bridge to ASEAN with its reliance pathway

2. Intended purpose is your classification lever

Carefully crafting your intended purpose statement is the single most important classification decision. Narrower claims often mean lower classification:

• "Assists clinicians in identifying potential anomalies" (lower class) vs.
• "Diagnoses diabetic retinopathy" (higher class)

Both might describe the same software — the regulatory framing makes the difference.

3. AI/ML adds complexity

Continuously learning algorithms face additional scrutiny:

• FDA's predetermined change control plan helps but has limits
• EU requires re-certification for significant changes
• Most APAC markets don't yet have clear guidance on adaptive algorithms

4. Clinical evidence requirements scale with classification

Classify Your SaMD Instantly

Not sure how your software classifies across global markets? regMD's free classification tool analyzes SaMD alongside hardware devices, giving you instant classification results across 15 jurisdictions.

For the full regulatory strategy — including recommended market entry sequence, evidence requirements, and cost projections — get lifetime access for $99.

This article is for informational purposes only and does not constitute regulatory advice. SaMD regulation is evolving rapidly — always verify current requirements with the relevant authority or a qualified consultant.