What a PMA actually is — when FDA approves a device

A PMA is not a bigger 510(k), and the difference is not effort — it is the question being answered. Under section 515 of the FD&C Act and 21 CFR Part 814, premarket approval asks whether your device is safe and effective, demonstrated independently, on its own evidence. A 510(k) argues a comparison to a predicate; a PMA has no predicate to stand behind. This is the pathway for Class III devices — the ones that sustain life or present unreasonable risk of illness or injury — and it is the one place in device regulation where “FDA-approved” is the correct phrase.

Why the framing matters

Teams raised on 510(k)s write their first PMA like a very long comparison — and comparisons are precisely what the file cannot lean on. There is no predicate whose testing absorbs your risk questions, no substantial-equivalence argument to bound the evidence. Every question the device raises, the application must answer with data. That inverts the economics of the programme: in a 510(k) the expensive decision was choosing the predicate; in a PMA it is designing the clinical study, because the study you run is the approval you can get. Indications you did not study are indications you will not be approved for.

The framing also corrects the vocabulary. A cleared device may not be marketed as “approved” — clearance is a finding of equivalence, and the agency polices the distinction. A PMA device has earned the word: FDA has independently concluded, on valid scientific evidence, that reasonable assurance of safety and effectiveness exists. The same discipline applies on the drug side, where approval is an affirmative action and never a lapsed clock — across every pathway, what the agency actually did is the only thing you may claim it did.

The terms, precisely

Class III device

A device for which general and special controls are insufficient to provide reasonable assurance of safety and effectiveness — typically one that supports or sustains human life, is of substantial importance in preventing impairment of health, or presents a potential unreasonable risk. Class III is the default destination of a not-substantially-equivalent finding.

Reasonable assurance of safety and effectiveness

The approval standard. Not certainty, and not superiority — a demonstration, on valid scientific evidence, that the probable benefits to health outweigh the probable risks, and that the device will perform as labeled in a significant portion of the target population.

Valid scientific evidence

Defined in 21 CFR 860.7: well-controlled investigations, partially controlled studies, objective trials without matched controls, well-documented case histories, and significant human experience. Isolated case reports and unsubstantiated opinions are expressly excluded. For a Class III device, the centre of the file is almost always a clinical investigation under an IDE.

PMA supplement

The instrument, under 21 CFR 814.39, by which an approved PMA is changed. Modifications affecting safety or effectiveness — design, labeling, indications, manufacturing — require approval before implementation, through panel-track, 180-day, or real-time supplements, or a 30-day notice for certain manufacturing changes.

What the file actually carries

The manufacturing leg deserves the emphasis it rarely gets. A PMA is the only device pathway where approval routinely waits on a facility inspection — the agency verifies that the quality system can actually produce the device the clinical data described. Programmes that treat the submission as a writing exercise discover, at month eight, that the gating item is not a statistical question but a corrective action in the cleanroom.

The 180 days, honestly

Section 515 gives FDA 180 days to act, and the MDUFA performance goals commit the agency to decisions within 180 FDA days for a PMA that does not go to an advisory panel, and longer for one that does. Neither number is your calendar. The file first survives a 45-day filing review. A major deficiency letter stops the review until you answer completely. A novel device may be taken to a panel of outside experts in public session. The inspection has to be scheduled, passed, and closed. And the actions available at the end are graded — an approval order, an approvable letter naming the conditions that remain, a not-approvable letter, or denial.

So the honest programme plan runs backwards from the evidence, not forwards from the submission: the IDE study is the critical path, the manufacturing readiness runs in parallel rather than afterwards, and the submission date is an output of both. Teams that ask “how fast can FDA review a PMA” are usually asking the wrong question a year too late.

Where PMAs go wrong

Writing it as a scaled-up 510(k)

The habits of equivalence — leaning on other devices' history, bounding the evidence by comparison — have no purchase in a PMA. The standard is independent demonstration on your own data. Material that argues similarity instead of safety and effectiveness spends pages answering a question nobody asked.

Studying a narrower population than the label you want

The approval is bounded by the evidence. An indication broader than the studied population is not a stretch goal — it is an unapprovable claim, and the panel will say so in public. Decide the label first, then design the study that supports it.

Treating manufacturing as a post-approval problem

The pre-approval inspection verifies the methods and facilities described in the application. A quality system that cannot pass inspection holds the approval hostage regardless of how clean the clinical data are — and afterwards, changes to what was approved run through PMA supplements, not engineering discretion.

Frequently asked questions

Does every PMA need a clinical trial?

Plan on it. The reasonable-assurance standard is met with valid scientific evidence under 21 CFR 860.7, and for a Class III device that evidence is, as a rule, clinical data from a study conducted under an IDE. The occasional exception — extensive prior clinical experience, a well-understood technology — is argued to the agency, not assumed.

How long does PMA review actually take?

The statute sets a 180-day review period, and FDA's MDUFA performance goals run to 180 FDA days for a PMA without an advisory panel and longer with one. In practice the calendar includes a 45-day filing review, possible panel review, a pre-approval inspection, and major-deficiency letters that stop the clock — the submission is the short leg of a programme measured in years.

What is a PMA supplement?

The approval covers the device, labeling, and manufacturing described in the application. Changes affecting safety or effectiveness need FDA approval before implementation under 21 CFR 814.39, through a graded set of supplements — panel-track, 180-day, real-time, and 30-day notices for certain manufacturing changes.