SaMD is software intended for a medical purpose that performs that purpose on its own, without being part of a hardware medical device. The term comes from the International Medical Device Regulators Forum (IMDRF) and is used by regulators worldwide.
Free sandbox · No credit card · 21 CFR Part 11 aligned
Software as a Medical Device
noun · clinical research
Software intended to be used for one or more medical purposes that performs those purposes without being part of a hardware medical device (IMDRF).
Related terms
In short
Explained
IMDRF introduced the term SaMD in 2013 to replace older labels such as stand-alone software, so regulators could share one vocabulary. According to the IMDRF key definitions document, the concept is independent of technology and platform: a mobile app, a cloud service and a desktop programme can all be SaMD if they meet the two-part test. Many national regulators have adopted or mirrored the IMDRF wording, though each has its own rules on classification and approval. Check the regulator for your market for the current text.
Intended use is the deciding factor. The same algorithm might be a wellness tool in one claim and SaMD in another. If a developer says the software helps diagnose, monitor, predict or treat a disease, it is moving toward SaMD. Common examples include software that analyses medical images, apps that calculate insulin doses, or digital therapeutics that deliver a clinical intervention. General wellness apps, electronic health record storage and software that only transmits data are usually treated differently, but the boundary depends on the claims and the jurisdiction.
SiMD, software in a medical device, is software that is part of a hardware device or needed to drive it, such as the code running a pacemaker. Regulators assess it as part of that device. SaMD stands alone, so it is assessed and updated on its own schedule, which is why change management and version control matter so much for it.
IMDRF also published a risk categorisation framework for SaMD that considers the significance of the information the software provides and the seriousness of the health situation. Regulators then map those considerations to their own device classes, for example in the EU, the US and the UK. The class decides how much evidence and what kind of review is needed before the product reaches users.
Depending on class and claim, a SaMD manufacturer may need a prospective clinical investigation, a retrospective validation on existing data, or a literature-based evaluation. When a study is run, the same rules on consent, data integrity and audit trails apply as in drug trials. Our guide to EDC for medical device clinical trials covers study set-up, and EDC for AI medical device clinical validation covers algorithm validation studies.
| Medical purpose? | Part of a hardware device? | Likely category | |
|---|---|---|---|
| App that analyses retinal images to flag disease | Yes | No | SaMD |
| Firmware controlling an insulin pump | Yes | Yes | SiMD |
| General step counter | No | No | Not a medical device |
| Digital therapy app for insomnia | Yes | No | SaMD |
In practice
The steps differ from a drug trial mainly in the endpoints and the regulatory pathway.
Write the exact claim the evidence must support. Changing it later usually means new evidence.
For diagnostic software this is often sensitivity and specificity against a clinician-adjudicated reference. For therapeutic software, it is a clinical outcome, often patient-reported.
Record which software version each participant or case used, because a change in code can change performance.
Device studies follow device rules, which differ from drug rules by region. Confirm the pathway early.
Use a system with an audit trail and role-based access, then analyse against the prespecified plan.
Build forms for the reference standard, the software output and participant-reported outcomes, then test them free. Every feature is open in the sandbox.
In Capture
Capture is an electronic data capture and ePRO platform that studies of digital health products can use to collect clinical data: eCRFs for reference-standard results and software outputs, ePRO diaries for participant-reported outcomes, eConsent, and a field-level audit trail, with 21 CFR Part 11-aligned controls. The platform is not described here as a medical device itself; it is the system in which a study team records the study data.
Weight changed 68.0 to 68.5 kg
Site coordinator · Reason: Transcription error
Query Q-0014 answered on Weight
Site coordinator · Reason: Source checked
Vital signs form locked
Data manager · Reason: Visit cleaned
Software as a Medical Device. It is software intended for one or more medical purposes that performs them without being part of a hardware medical device.
SaMD stands alone. SiMD, software in a medical device, is embedded in or drives a hardware device, such as firmware in a pump. Regulators assess SiMD as part of the device.
No. Intended use decides. General wellness or fitness apps with no medical claim are usually not medical devices, while software that diagnoses, monitors or treats a condition may be.
Often they need clinical evidence, but not always a prospective trial. Depending on the class and claim, a regulator may accept a retrospective validation or a literature-based evaluation. Check the requirements for the market.
The International Medical Device Regulators Forum (IMDRF), a group of regulators, published the definition and a risk categorisation framework. National authorities then apply their own rules.
Yes. The definition does not depend on the platform, so mobile apps, cloud services and desktop programmes can all be SaMD if they meet the intended-use test.
Often yes, when it delivers a clinical intervention for a medical purpose and runs independently of a hardware device. Regulatory status still depends on the claims and the regulator.
Keep exploring
EDC for medical device trials
Set up a device study.
EDC for AI device validation
Validation studies for algorithms.
EDC for digital health trials
Collect data for digital products.
FDA digital health technologies guidance
Technologies used in trials.
Post-market clinical follow-up
Evidence after launch.
Good Clinical Practice (GCP)
The wider quality standard.
eCRFs, ePRO and an audit trail in the free sandbox. No credit card.