The Regulatory Pathway Your CTO Mapped Out Is Based on the Wrong Device Classification

Your CTO stayed up until 2 a.m. reading the FDA product code database and came back convinced your device is a Class II, 510(k)-exempt. The team celebrated. Someone updated the investor deck. A timeline got built around it.

Photo by Markus Winkler on Pexels.

That classification is probably wrong.

Device classification mistakes are one of the most expensive errors a medtech startup can make, not because the FDA penalizes you for trying, but because your entire commercialization plan calculates from that one assumption. Manufacturing timelines, capital requirements, reimbursement strategy, clinical study design: everything downstream depends on getting classification right. Get it wrong at month four and you may not discover the problem until month twenty-two, when a Q-sub response lands in your inbox and the reviewer politely explains that your predicate doesn't hold.

Here's what actually happens inside most early-stage medtech teams.

A technical founder (or, generously, a consultant hired on the cheap) searches the 510(k) database for something that looks like their device. They find a cleared device with a similar description, note the product code, assume equivalence, and move on. The problem is that device classification is not a keyword search. Two devices with nearly identical physical descriptions can land in completely different regulatory buckets depending on intended use, patient population, and the specific claims you need to make in your labeling.

Software is where this goes most catastrophically sideways. A decision-support tool that surfaces data for a clinician to act on looks very different to FDA than a tool that autonomously drives a clinical decision. The first might clear as a Class II device under a Software as a Medical Device framework. The second could require a De Novo request or a full PMA. Founders consistently underestimate how much their marketing copy, sales pitch, and product roadmap will shape which of those buckets FDA puts them in.

Consider the sequence of how this plays out:

graph TD
    A[Intended Use Defined] --> B{Classification Analysis}
    B --> C[Class I: Exempt]
    B --> D[Class II: 510k or De Novo]
    B --> E[Class III: PMA Required]
    D --> F[Predicate Search]
    F --> G{Substantial Equivalence?}
    G --> H[Cleared]
    G --> I[De Novo or PMA]

Most teams treat the "Classification Analysis" step as a 30-minute exercise. Regulatory consultants who specialize in this spend days on a single device, reading the relevant guidance, checking enforcement discretion policies, reviewing recent clearance letters for comparable predicates, and often calling FDA directly through the Q-sub process before committing to a pathway.

A few specific traps worth knowing:

The combination product problem. If your device includes a drug, biologic, or software component, FDA may route it to a different center entirely. A wound care product with an antimicrobial agent might be a CDRH device or a CDER drug product depending on the primary mode of action. Getting the assignment wrong means you've been building a regulatory package for the wrong reviewer.

The predicate graveyard. FDA occasionally removes 510(k) clearances from the database when post-market problems emerge. If your predicate was recalled or had its clearance rescinded and you didn't check, you're building on sand. The agency won't tell you this upfront.

The indication creep problem. Your device starts as a screening tool. Then a sales conversation happens, someone adds a phrase to the website about diagnosis, and suddenly your labeling claims have drifted into a different classification category without anyone noticing. This is how companies get untitled letters.

The global market divergence. EU MDR and FDA classifications don't map one-to-one. A Class IIa device under MDR might be a Class III under FDA's scheme. Founders who plan US and EU simultaneously using a single regulatory strategy frequently discover the mismatch after they've already committed resources to clinical evidence generation that satisfies one market but not the other.

The fix is not complicated, but it requires spending money early that feels premature.

Hire a regulatory affairs professional with specific device-type experience (not just general medtech experience) before you've committed to a pathway, not after. Ask them to produce a written classification rationale, not a verbal opinion. If there's genuine ambiguity, file a Q-sub to FDA requesting a classification determination. That meeting costs you a few months and a filing fee. Discovering the problem in the middle of a Series A due diligence process costs you the round.

Your CTO is brilliant at building the device. Classification is a different discipline entirely, and treating it as a research problem solvable with a database and good intentions is where smart technical teams consistently come undone.