FDA 510(k) Substantial Equivalence: How FDA Evaluates Your Predicate and What Breaks It

What Substantial Equivalence Actually Means -- and Why Most Startups Get It Wrong

Substantial equivalence is the legal and scientific backbone of every 510(k) submission. Yet it remains one of the most misunderstood concepts in medical device regulatory strategy. Founders often assume that finding a device that looks like theirs on the FDA 510(k) database is enough. It is not -- and that misunderstanding costs companies months of delay, additional testing costs, and sometimes a Not Substantially Equivalent (NSE) determination that forces a De Novo or PMA pathway.

This post breaks down exactly how FDA evaluates your predicate comparison, where submissions fall apart, and what you can do to build a defensible substantial equivalence argument from the start.

The Statutory Framework: What FDA Is Actually Asking

Under Section 513(i) of the Federal Food, Drug, and Cosmetic Act, FDA must determine whether a new device is substantially equivalent to a legally marketed predicate. The agency evaluates this through a structured comparison defined in 21 CFR 807.87(f) and clarified extensively in FDA's guidance document Recommended Approaches to Integration of and General Principles of Software Validation -- but more directly in The 510(k) Program: Evaluating Substantial Equivalence in Premarket Notifications [510(k)], finalized in 2014.

FDA's analysis follows a sequential logic:

• Does the new device have the same intended use as the predicate?
• Does the new device have the same technological characteristics as the predicate?
• If technological characteristics differ, do those differences raise new questions of safety and effectiveness?
• Do the performance data demonstrate that the new device is at least as safe and effective as the predicate?

Each step in this chain must hold. A break at any point -- even after you have demonstrated equivalent performance -- can result in an NSE determination or an interactive review request that delays your clearance by 90 days or more.

Predicate Selection: The Decision That Shapes Everything Downstream

Your predicate choice is not administrative -- it is a strategic and scientific decision that defines the entire evidentiary standard your submission must meet. A poorly chosen predicate forces you to explain away differences you cannot easily bridge, or worse, anchors you to performance standards that require expensive clinical or bench testing you had not planned for.

Strong predicates share your device's intended use at the level of specificity FDA will expect. If your predicate clears a broad indication and your device targets a narrow sub-population, FDA may challenge whether the intended use is truly the same. Conversely, if you select a predicate with a narrow cleared indication, you may inadvertently constrain your own labeling.

Under 21 CFR 807.87(f), you are required to identify the predicate and explain the comparison. FDA expects you to address design, materials, energy source, software, and any other technological characteristics that differ. The 2014 substantial equivalence guidance makes clear that you may use multiple predicates -- a split predicate approach -- to address different aspects of your device, but you cannot mix and match intended use from one predicate with technological characteristics from another to create an indication that neither predicate actually supports. FDA has explicitly warned against this practice in subsequent Q-Submission feedback and in the 2023 content of review guidance updates.

What Breaks a Substantial Equivalence Argument

Based on common patterns across FDA's Additional Information (AI) requests and NSE letters, here are the failure points that derail 510(k) submissions most often:

• Intended use drift: Labeling or indications-for-use language that extends beyond what the predicate was cleared for -- even subtly -- triggers FDA scrutiny. Words like 'real-time,' 'continuous,' 'autonomous,' or specific patient populations can shift the intended use analysis entirely.
• Unacknowledged technological differences: Failing to identify a material difference -- such as a new software algorithm, a novel material, or a different energy delivery mechanism -- signals to FDA that the comparison was not rigorous. Reviewers will find it and it will damage your credibility.
• Performance data gaps: When technological differences exist, 21 CFR 807.87(f) requires data demonstrating equivalent safety and effectiveness. Submitting bench test results without linking them explicitly to the predicate's cleared performance specifications leaves FDA without the comparison it needs to make a determination.
• Predicate obsolescence or modification: Using a predicate that has since been recalled, had its clearance rescinded, or been substantially modified through a subsequent 510(k) without acknowledging that history creates a fragile foundation FDA will question.
• Software without sufficient documentation: For devices containing software, FDA expects alignment with the Deciding When to Submit a 510(k) for a Change to an Existing Device guidance and the Software as a Medical Device (SaMD): Clinical Evaluation framework. Missing a Software Level of Concern determination or omitting the required documentation per the 2023 Software Pre-Specs guidance is an automatic AI request.

Building a Submission That Holds Up to Review

The strongest 510(k) submissions treat the substantial equivalence argument as a structured scientific narrative, not a checklist exercise. Every claimed similarity should be supported by comparative data or specifications. Every acknowledged difference should be accompanied by a clear articulation of why that difference does not raise new safety or effectiveness questions -- or by test data showing equivalent or superior performance.

Your predicate comparison table is not a formality. It is the analytical spine of your submission. FDA reviewers use it to anchor their entire review. If your table is vague, inconsistent with your device description, or silent on known differences, it signals a submission that was not prepared by someone who understands how clearance decisions are actually made.

Final Thought: Regulatory Strategy Starts at Predicate Selection

By the time many device companies engage a regulatory consultant, they have already locked in a predicate choice, drafted indications-for-use language, and begun testing -- all without validating that their substantial equivalence argument is defensible. Correcting that late in the process is expensive and sometimes impossible without restarting the submission.

The right time to stress-test your predicate strategy is before your design freeze, not after your submission receives an AI request.

At ADB Consulting and CRO Inc., Andre Butler and the team work directly with device founders and regulatory professionals to evaluate predicate strategy, identify submission risks, and build 510(k)s that are designed to clear -- not just to be submitted. If your team is preparing a 510(k) or reassessing a stalled submission, book a free discovery call at adbccro.com and get a straight assessment of where your strategy stands.