How to Choose Between 510(k), De Novo, and PMA for Your Medical Device

Most teams pick their FDA pathway the wrong way round. They decide they want a 510(k) because it is the fastest route, then spend six months trying to bend the device into a predicate that does not really fit. The pathway is not a preference. It falls out of two facts about your device: how much risk it carries, and whether something legally marketed already does substantially the same thing in substantially the same way.

Get those two facts straight early and the route picks itself. Get them wrong and you lose a funding cycle.

What each pathway actually answers

A 510(k), or premarket notification, answers one question: is this device substantially equivalent to a legally marketed predicate? You are not proving the device is safe and effective from first principles. You are proving it is as safe and effective as something FDA already allowed on the market. Most Class II devices go this way.

A De Novo classification request answers a different question: this device is low or moderate risk, but nothing like it exists to compare against — can FDA create a new classification for it, with special controls that make future devices of this type eligible for 510(k)? De Novo is the route for genuine novelty at moderate risk. It is also how many AI-enabled software devices first reached the market, and the granted De Novo then becomes the predicate everyone else uses.

A premarket approval application, or PMA, answers the hardest question: is there valid scientific evidence providing reasonable assurance that this device is safe and effective for its intended use? This is the Class III route — life-sustaining, life-supporting, or substantial-risk devices. It normally requires clinical data of your own and a preapproval inspection of your manufacturing site.

The decision sequence

Work through it in this order. Skipping a step is how teams end up with an unusable submission.

Start with classification. Search the FDA product classification database for the generic device type that matches your intended use, not your technology. A regulation number and device class come back, along with whether the type is 510(k)-exempt, subject to premarket notification, or Class III. This single search resolves the pathway for the large majority of devices.

Then search for a predicate. Use the 510(k) database and read actual clearance summaries, not marketing pages. You are looking for a legally marketed device with the same intended use and either the same technological characteristics or different characteristics that do not raise new questions of safety and effectiveness. If you can only find devices with a similar mechanism but a different clinical claim, you do not have a predicate — you have a starting point for a De Novo conversation. Our companion piece on predicate devices covers the search and its traps in detail.

Next, test the risk controls. If the risks of your device can be managed by general and special controls — labelling, performance standards, post-market surveillance, guidance-based testing — you are in Class II territory and the question is 510(k) versus De Novo. If they cannot, and failure of the device would plausibly cause serious injury or death, you are looking at Class III and a PMA.

Finally, apply the novelty test honestly. Novel intended use, novel patient population, or a novel mechanism with no cleared analogue all push you off the 510(k) route, whatever your engineering team feels about the underlying technology being conventional.

When De Novo is the right answer, not a consolation prize

Teams treat De Novo as the thing that happens when the predicate search fails. That framing costs them. A granted De Novo creates the classification regulation and the special controls for an entirely new device type — and you wrote the evidence that shaped them. Competitors following you have to show substantial equivalence to your device, on the performance benchmarks your submission established.

The trade is time and evidence. De Novo submissions usually need more clinical or clinical-validation data than a 510(k), and the review is longer because FDA is building a classification, not comparing to an existing one. If your device is genuinely first in category and you intend to stay in that category, that work is an asset, not overhead.

When PMA is unavoidable

If your device is already classified as Class III, or FDA determines your De Novo request is not appropriate because the risk cannot be controlled by special controls, PMA is the route. Plan for it properly rather than hoping to argue your way out: an investigational device exemption to run the pivotal study, a full clinical programme, manufacturing readiness for inspection, and post-approval reporting obligations that continue for the life of the product. Our 510(k) versus PMA comparison sets out what that difference costs in money and calendar time.

Software and AI devices need one extra step

If your device is software, and especially if it learns or is updated frequently, the pathway decision is only half the problem. Decide at the same time whether you will file a Predetermined Change Control Plan so that planned model updates do not each trigger a new submission. We covered the mechanics in our piece on PCCPs for AI/ML devices. A PCCP is filed inside the 510(k), De Novo or PMA — so it has to be scoped while the submission strategy is still being written, not afterwards.

Also settle your international sequence now. EU and UK routes classify software differently and will not follow your FDA reasoning; see the EU MDR and UKCA guide before you commit to a filing order.

Three things to do this month

Run the classification search and write the regulation number, class and product code into your regulatory plan. If nobody can name it, you do not have a strategy yet.

Pull three candidate predicates and write a one-page substantial equivalence argument for each. The exercise tells you quickly whether the 510(k) route is real or wishful.

Model the reimbursement consequences alongside the regulatory ones. A cleared device with no payment route is not a commercial product. Our market access guide and MedTech Compass score both tracks together, and DevicePath gives you an open-source starting point for regulatory triage.

Sources

1. FDA — Premarket Notification 510(k): https://www.fda.gov/medical-devices/premarket-submissions-selecting-and-preparing-correct-submission/premarket-notification-510k 2. FDA — De Novo Classification Request: https://www.fda.gov/medical-devices/premarket-submissions-selecting-and-preparing-correct-submission/de-novo-classification-request 3. FDA — Premarket Approval (PMA): https://www.fda.gov/medical-devices/premarket-submissions-selecting-and-preparing-correct-submission/premarket-approval-pma 4. FDA — Product Classification database: https://www.accessdata.fda.gov/scripts/cdrh/cfdocs/cfpcd/classification.cfm 5. FDA — 510(k) Premarket Notification database: https://www.accessdata.fda.gov/scripts/cdrh/cfdocs/cfpmn/pmn.cfm 6. FDA — Investigational Device Exemption (IDE): https://www.fda.gov/medical-devices/how-study-and-market-your-device/device-advice-investigational-device-exemption-ide

This article summarises publicly available FDA guidance and databases. It is not legal or regulatory advice.