How to Find an FDA 510(k) Predicate Device: A Practical Workflow

Use the free tool

FDA Predicate & Pathway Agent

Cross-search FDA classifications and premarket databases for product-code, pathway, and predicate directions

Open tool

Start with a device description, then build a comparison

To find an FDA 510(k) predicate device, describe your device's intended use and technology, identify relevant product codes, search cleared devices, and compare the strongest candidates against the underlying FDA records. A useful search ends with a documented shortlist and a list of evidence gaps.

The Pure Global FDA Predicate & Pathway Agent helps with the research stage. Enter a device description to search across FDA classification, 510(k), De Novo, PMA, and GUDID data. Its report presents a preliminary pathway assessment, possible 510(k) predicate candidates, and retrieved source records. You can use it without an account.

This guide walks through the tool and gives you a comparison worksheet to take into a regulatory planning meeting. The example is hypothetical; it does not identify an actual cleared device or establish a submission strategy.

What makes a useful predicate candidate?

A predicate is a legally marketed device used to support a substantial equivalence comparison. FDA's framework considers the same intended use and either the same technology or technological differences that do not raise different safety and effectiveness questions, supported by appropriate evidence. A similar appearance or shared product code cannot establish that conclusion by itself. See FDA's explanation of substantial equivalence.

For an initial shortlist, ask whether you can locate the candidate's regulatory identity, understand its cleared use, and explain the differences from your product. A popular competitor with an incomplete public record may be a useful research lead, but you still need evidence for the comparison.

Keep three working labels in your notes: candidate to investigate, candidate with documented comparison, and excluded with reason. These are suggested research labels, not FDA statuses or automated tool scores. They prevent a promising search hit from quietly becoming an assumed predicate in the next presentation.

Step 1: Prepare the description you will paste

Open the FDA research tool and find the Device description field. Before running it, collect the facts that distinguish your product from neighboring device types.

Your input should cover:

• The clinical purpose and the proposed intended use, using your team's current wording.
• The patient population, intended user, and use environment.
• What the device measures, delivers, treats, or helps a user decide.
• Its operating principle and important software functions.
• Patient contact, reuse, and sterility where relevant.
• Known comparator names, with a clear note that they are unverified leads.
• Unresolved claims or design decisions that should remain questions.

For a fictional thermometer project, a useful starting input would be:

We are researching a handheld, non-contact infrared thermometer intended to measure human body temperature at the forehead in adults. The proposed users are healthcare professionals in clinical settings. It provides a single temperature reading on the device display. It has no phone application and does not make an autonomous diagnosis. Pediatric use and home use are outside the current proposed scope. Please investigate relevant FDA product codes, potential 510(k) candidates, and uncertainties that need source verification. Do not assume that an industrial surface thermometer is an appropriate comparator.

Replace every example fact with your actual product information. If pediatric use is unresolved, state that it is unresolved instead of excluding it merely to obtain a cleaner result. A search based on an easier, narrower product would answer the wrong business question.

If your source material is spread across brochures, specifications, and emails, first organize it with the Regulatory Product Master. Keep the description version you actually used for the search.

Step 2: Run the agent and read each evidence branch

Select Run free FDA research. When the report appears, read the preliminary summary and its explanation before reviewing Possible 510(k) predicate candidates. A confidence label describes the report's assessment; it is not a probability of FDA clearance.

The report brings together several types of records. Use them for different research questions:

Do not give every record the same evidentiary role. For example, a matching brand or identifier can help you locate the right product, while the clearance record and supporting documents are where you investigate the candidate's cleared scope. Copy the source link alongside the fact it supports.

If the result appears to describe a different product, correct the input and rerun it before developing a comparison. Check for an omitted intended user, a vague word such as “monitor,” or a missing distinction between measuring a value and interpreting it.

Step 3: Refine the search with product codes and names

Use the candidate product code as another search route, not as the final selection criterion. The FDA product code lookup and our product code lookup guide explain how to inspect the category before carrying it into a clearance search.

FDA recommends product code searches as an effective way to find similar device types. Its search guidance also notes that the 510(k) database retains the original applicant and trade name, so later ownership changes or rebranding can obscure a familiar product. Try historical applicant names and spacing or hyphen variations when a name search fails. See FDA's predicate search instructions.

Keep a short search log in your own worksheet: date, description version, code or name searched, source, and result. Record unsuccessful searches too. If a colleague repeats the exercise next month, the log will show whether new results reflect a different query, a revised product scope, or newly available records.

Step 4: Open the FDA record for every serious candidate

Use the source link on the candidate card. Record the exact K number, applicant, device name, and decision information. Then inspect the available summary and indications for use. Keep the document link and page or section reference for each important comparison fact.

Read the scope closely. A family name on a marketing page may cover several models, but the record you opened may concern only one configuration. Likewise, a current brochure may describe features introduced after the clearance you found. Flag those mismatches instead of combining information from different versions into one imaginary comparator.

When a summary is unavailable or silent about a characteristic, write not established from the retrieved document. That is more informative than filling the cell from memory. Give the unresolved item an owner and a next step, such as locating the relevant labeling or obtaining technical documentation through an appropriate channel.

Step 5: Build a predicate comparison worksheet

FDA describes a comparison table together with a discussion of similarities and differences as part of the substantial equivalence section. The table below is an initial research worksheet, not a complete submission template. Tailor it to the device and applicable guidance. See FDA's description of 510(k) content.

For the fictional thermometer, “both measure temperature” would be an incomplete comparison. The team would still need to investigate the measurement site, intended population, use environment, operating principle, and supporting performance information. Do not replace that investigation with a made-up similarity percentage.

For each difference, add a plain-language explanation of why it matters and what evidence could resolve it. Separate facts already demonstrated in your records from tests that are merely planned. This makes the worksheet useful for budgeting research and testing work without pretending that the budget itself proves equivalence.

Step 6: Separate the proposed predicate from other useful references

FDA distinguishes a primary predicate from reference devices used to support scientific methods or reference values, and states that split predicates are inconsistent with the 510(k) standard. Identify the primary predicate most similar in indications and technology, rather than assembling a favorable intended use from one device and favorable technology from another. See FDA's guidance on using predicates.

In your working file, give each additional device an explicit purpose. “Useful source for a test method” is a different rationale from “proposed primary predicate.” If the role is unclear, leave it as an open question for the submission team.

Add a separate review of public post-market information where relevant. Our MAUDE search workflow explains how to investigate reported events while keeping the limits of those reports visible. Keep those findings separate from your similarity assessment.

Troubleshooting: irrelevant hits, missing records, and no candidates

A zero-result search means the query did not establish a candidate. It does not, by itself, prove that no predicate exists or that a particular alternative pathway is required.

Frequently asked questions

Is the same FDA product code enough to choose a predicate?

No. Use the code to organize the search, then review the intended use, technology, and supporting evidence under FDA's substantial equivalence framework. Keep your reason for retaining or excluding each candidate in the comparison file.

Can AI choose the final predicate for my 510(k)?

The agent helps retrieve and organize evidence and proposes candidates. Your submission team must verify the records, evaluate differences, and develop the supported argument. An AI-generated report is preliminary research and does not establish FDA acceptance.

What should I bring to the next regulatory planning meeting?

Bring the searched device description, a source-linked shortlist, the comparison worksheet, exclusion reasons, and unresolved questions with owners. That package gives the team something concrete to review and turns a predicate search into a plan for the next evidence-gathering steps.

Start with the FDA Predicate & Pathway Agent, then use the worksheet above to document what the retrieved records actually support.