Last reviewed: September 25, 2026

Human factors testing shows that a device's users can use it safely, and FDA's validation test should in general include at least 15 participants from each distinct user group. Under FDA's final guidance of May 29, 2026, whether your submission includes the results depends on your critical tasks, your user interface and what has changed.

Key takeaways

  • FDA's validation test should in general include at least 15 participants from each distinct user group, usually under simulated use.
  • A critical task is one that, if done wrongly or not at all, would or could cause serious harm, including compromised medical care.
  • FDA's final guidance, issued May 29, 2026, sorts submissions into three human factors categories. Only Category 3 includes validation test results.
  • FDA updated its 2016 human factors guidance on August 3, 2026, to match the new framework.
  • By our count, 76 of the 733 FDA device regulations with special controls name human factors or usability in them.

What is human factors testing?

It is testing with representative users to find use-related risks and reduce them before a device reaches the market [1]. FDA describes two kinds [1]:

Type Purpose When
Formative evaluation Explores the user interface's strengths, weaknesses and unexpected use errors During design, repeated until the design is ready for validation
Human factors validation testing Shows that intended users can use the final device safely and effectively At the end of design, on the final device

Source: FDA, Applying Human Factors and Usability Engineering to Medical Devices, updated August 3, 2026 [1].

Test early. FDA notes that design flaws found in formative evaluation can be fixed more easily and less expensively than flaws found later [1].

FDA notes that validation testing is sometimes called summative testing. It warns that some definitions of summative testing leave out essential parts of validation [1]. Three FDA definitions sit at the center of it [1]:

  • User interface: the means by which users and the device interact, such as packaging, labeling, training materials, controls, displays and alarms. FDA's definition aligns with ANSI/AAMI/IEC 62366-1 [1, 2].
  • Use error: a user action, or lack of action, that differs from what the manufacturer expected. It is not caused solely by device failure, and did or could result in harm.
  • Critical task: a task that would or could cause serious harm to the patient or user if done wrongly or not at all. Harm includes compromised medical care.

Our 510(k) guide shows where human factors fits in a submission.

When does FDA expect human factors test results in a submission?

It depends on your answers to four questions in FDA's final guidance on human factors content, issued May 29, 2026 [2]:

  1. Is it a modification to an existing device? If yes, go to question 2. A new device goes to question 3, unless it shares the user interface of one of your own marketed devices and you reuse that device's human factors work. Then it can follow the modification route.
  2. Does anything change in the user interface, users, uses, use environments, training or labeling? If not, your submission is in Category 1.
  3. Does your use-related risk analysis show critical tasks, or new or affected critical tasks for a modified device? If not, you are in Category 2.
  4. Should validation test data be submitted, given the interface's history of use, its complexity and whether existing risk controls are enough? If yes, you are in Category 3. If not, you are in Category 2.
Category What you submit
1 A conclusion and high-level summary, with a statement justifying that the changes do not affect human factors
2 A rationale explaining why there are no critical tasks, or why validation data is not needed
3 A full human factors and usability engineering (HFE/UE) report, including validation testing

Source: FDA, Content of Human Factors Information in Medical Device Marketing Submissions, final guidance issued May 29, 2026 [2].

FDA's guidance is a recommendation, not a rule. You can use an alternative approach if it satisfies the applicable statutes and regulations [2]. Special controls and the QMSR, by contrast, are regulations, so they are binding [3, 7].

A modified device may need only a rationale. If a change affects a critical task but your risk controls still work, you can consider a rationale instead of new validation data [2].

Keep the work even when you do not submit it. For Class II and III devices, and some Class I devices such as those automated with software, the QMSR requires ISO 13485's design and development clause 7.3, which covers design verification and validation [2, 7]. FDA recommends keeping human factors information regardless of whether you submit it [2].

How do you plan human factors testing, step by step?

Work through these steps in order.

  1. Describe your users, uses, use environments and training. Split users into distinct groups where their characteristics or tasks differ [1].
  2. Analyze use-related risk. List every user task, and identify the critical tasks [2].
  3. Review known use problems with earlier versions of your device and with similar devices, including predicates [1, 2].
  4. Run formative evaluations to improve the design, and repeat them until the design is ready for validation [1].
  5. Work out your submission category using the four questions above [2].
  6. Agree the validation protocol with FDA. FDA encourages you to submit a draft protocol before testing, through a Pre-Submission [1].
  7. Run validation testing on the final device and final labeling, with at least 15 participants per user group [1].
  8. Analyze every use error on a critical task for its root cause, change the design where you can, and test the fix [1].
  9. Write the HFE/UE report at the level your category requires [2].

For step 3, FDA's guidance lists sources for known use problems, including its MAUDE adverse event database, MedSun, recalls and FDA safety communications [1]. Complizen's Superagent has a MAUDE Analyst that checks the adverse event signal on predicate devices. Our guide to FDA's medical device databases shows how to search MAUDE yourself.

How do you run a human factors validation test?

FDA's guidance sets out how the test should be designed [1]:

  • The participants represent the intended, actual users of the device.
  • All critical tasks are performed during the test.
  • The device user interface represents the final design.
  • The test conditions are realistic enough to represent actual conditions of use.

Fifteen is the usual minimum. FDA says the minimum number of participants should in general be 15, and may be higher for some device types [1]. If you have more than one distinct user population, test at least 15 from each [1].

Test the final labeling and realistic training. Labeling used in validation should be the final design [1]. That includes device labels, display information, packaging, instructions for use and quick-start guides. Training should be close to what real users will receive. If most users will get little or no training, test participants should not be trained [1].

Simulated use is the norm. Validation testing is generally done under simulated use [1]. Actual-use testing, or human factors data from a clinical study, is for cases that simulated use cannot properly test [1]. Our clinical trials guide covers clinical studies.

Collect three kinds of data. Observe participants performing all critical tasks. Ask direct questions where understanding cannot be seen in actions, such as knowledge of essential safety information. Then hold a debriefing interview, which supplements observation but cannot replace it [1]. Your protocol should define in advance how use errors are identified and recorded [1].

Protect your participants. FDA's guidance says to consult an institutional review board (IRB) on safeguards for participant safety and privacy, including informed consent forms [1].

Company employees are for formative work only. Employees can take part in formative evaluations, but not in validation testing [1].

What happens if participants make errors on critical tasks?

Errors do not automatically fail the test, but FDA expects you to analyze each one [1]. FDA says it is practically impossible to make any device error-proof, so some residual risk will remain [1].

  • Find the cause. Your analysis should show how each use error happened, including which part of the user interface caused the problem [1].
  • Decide whether to change the design. The analysis should show whether design changes are needed, possible and likely to reduce the risk [1].
  • Prefer design fixes. FDA calls labeling and training important risk controls but not the most preferred, because users must remember them, labeling may not be at hand during use, and training can fade over time [1].
  • Test your fixes. Saying you fixed a problem by changing the instructions for use or other labeling is not acceptable unless new test data shows the change worked [1]. The same applies to adding training [1].
  • Retest only what changed. If your changes affect only some aspects of use, the new testing can focus on those aspects [1].

Serious use errors need a strong case. FDA's guidance says results showing persistent serious use errors are not acceptable unless you analyze them thoroughly [1]. You also need to show that further reduction is not possible or practical, and that the benefits of using the device outweigh the remaining risks [1].

Which devices have human factors in their special controls?

About one in ten device types with special controls, including glucose monitors, insulin dosing controllers and home COVID-19 tests. We ran a count on September 25, 2026 of FDA's device classification regulations in 21 CFR Parts 862 to 892. Of the 733 regulations with special controls, 76 name human factors or usability in them, about 10% [3].

Review area (21 CFR part) Regulations naming human factors or usability
Cardiovascular (870) 14
General and plastic surgery (878) 10
Gastroenterology and urology (876) 8
General hospital (880) 7
Neurology (882) 7
All other areas 30

Source: our analysis of the eCFR text of 21 CFR Parts 862 to 892, current as of September 1, 2026 [3].

Special controls are binding. For an interoperable automated glycemic controller, the special controls require a human factors study. It must show that an intended user can safely use the device [4]. For one microbiology imaging system, they require human factors or usability validation testing with the final version of the labeling [5].

Check your own product code. If your device type's regulation names human factors in its special controls, your submission must meet that control. Our guide to FDA medical device classification shows how to find your regulation. Our De Novo guide shows the ECG app's special controls, which include human factors testing [6].

What goes in a human factors report?

For Category 3, FDA outlines a full HFE/UE report [2]. It covers:

  • a conclusion and high-level summary
  • descriptions of the intended users, uses, use environments and training
  • the device user interface, including its labeling
  • known use problems with earlier models and with similar devices, such as predicates
  • a summary of preliminary analyses and evaluations
  • the use-related risk analysis and critical tasks
  • the details and results of human factors validation testing

FDA gives you the conclusion to aim for. Its outline states that the device has been found adequately designed for the intended users, uses and use environments, such that intended users can use it without serious use errors or problems [2].

What changed in 2026?

Three things, all of which date older guides.

FDA finalized its human factors content guidance. Issued on May 29, 2026, it sets the four-question framework and the three submission categories above [2].

FDA updated its core human factors guidance. On August 3, 2026, FDA revised "Applying Human Factors and Usability Engineering to Medical Devices" to match the new content guidance [1]. It was first issued on February 3, 2016.

Design controls moved into the QMSR. Since February 2, 2026, 21 CFR Part 820 incorporates ISO 13485:2016. The old design controls section, 21 CFR 820.30, no longer exists [2, 7]. Our QMSR guide explains the change.

What should manufacturers outside the US plan for?

FDA's guidance says participants should represent the device's intended, actual users [1]. For a device sold in the US, plan recruitment around the US users who will actually use it.

Validate with your final English labeling. Validation should use the final labeling, and required label statements for the US must be in English [1, 8]. Testing with a translation alone does not test the English labeling you will ship. Our FDA labeling guide covers the English and second-language rules.

Once you are on the market, use errors can be reportable. Under FDA's reporting rule, an event caused or contributed to by user error can be reportable [9]. Our guide to FDA medical device reporting explains when.

Frequently asked questions

What is human factors testing for medical devices?

It is testing with representative users to find and remove use-related risks. Formative evaluations improve the design during development. Human factors validation testing, which some call summative testing, shows that intended users can use the final device safely and effectively under realistic conditions.

How many participants does FDA require for human factors validation testing?

FDA says the minimum should in general be 15 participants, and may be higher for some device types. If your device has more than one distinct user population, such as clinicians and patients, test at least 15 participants from each population.

What is the difference between formative and summative human factors testing?

Formative evaluation explores a user interface's strengths, weaknesses and unexpected use errors while the design can still change. FDA calls the final stage human factors validation testing, and warns that some definitions of summative testing leave out essential parts of it.

What is a critical task in human factors?

A critical task is a user task that, if performed incorrectly or not performed at all, would or could cause serious harm to the patient or user. FDA includes compromised medical care as harm, and its guidance says every critical task should be performed during validation testing.

What happens if participants make use errors in validation testing?

Errors do not automatically fail the test. FDA expects you to find each error's cause, including the part of the user interface involved, and to reduce the risk where you can. FDA's guidance says changing only the labeling, or adding training, is not acceptable without new test data showing the change worked.

When do I need to submit human factors validation data to FDA?

When your submission falls into Category 3 under FDA's final guidance of May 29, 2026. That depends on whether your device has critical tasks, and on its interface history, complexity and existing risk controls. Other submissions need a summary or a rationale instead.

Can company employees take part in human factors testing?

Only in formative evaluation. FDA's guidance allows employees in formative work, though their views may be misleading if they are not representative users. For human factors validation testing, participants should represent the device's intended, actual users. Plan to recruit outside participants who match each user group for the validation test.

Does human factors validation have to use the final device and labeling?

Yes, according to FDA's guidance. The device user interface should represent the final design. The labeling should also be final, including device labels, display information, packaging, instructions for use and quick-start guides. Finalize labeling and training materials before validation testing. Testing an earlier version risks results that do not reflect what users will actually receive.

Should human factors test participants be trained first?

Only as much as real users will be. Training in validation testing should be close to the training actual users receive. If most or all users will get little or no training, FDA says test participants should not be trained. This keeps the test from overstating how well real users will perform.

Is human factors validation testing a clinical study?

Usually not. FDA says validation testing is generally conducted under simulated use. Actual use or clinical study data is for cases where simulated-use methods cannot properly test how users interact with the device. FDA's guidance still advises consulting an IRB on participant safety, privacy and informed consent.

Which standard covers usability engineering for medical devices?

IEC 62366-1, published in the US as ANSI/AAMI/IEC 62366-1. FDA's final human factors content guidance aligns its definition of user interface with that standard. That definition includes labeling and training. So your instructions for use and training materials are tested along with the device.

What changed in FDA's human factors guidance in 2026?

FDA finalized its guidance on human factors content in marketing submissions on May 29, 2026, introducing three submission categories. On August 3, 2026, it updated its 2016 guidance on applying human factors to devices to match. Guides that cite 21 CFR 820.30 are also out of date, since the QMSR took effect on February 2, 2026.

References

  1. FDA. Applying Human Factors and Usability Engineering to Medical Devices. Issued August 3, 2026; originally issued February 3, 2016.
  2. FDA. Content of Human Factors Information in Medical Device Marketing Submissions. Final guidance issued May 29, 2026.
  3. eCFR. 21 CFR Chapter I, Subchapter H, Medical Devices (Parts 862 to 892). Text current as of September 1, 2026.
  4. eCFR. 21 CFR 862.1356, Interoperable automated glycemic controller. Accessed September 25, 2026.
  5. eCFR. 21 CFR 866.2190, Automated image assessment system for microbial colonies on solid culture media. Accessed September 25, 2026.
  6. eCFR. 21 CFR 870.2345, Electrocardiograph software for over-the-counter use. Accessed September 25, 2026.
  7. eCFR. 21 CFR Part 820, Quality Management System Regulation. Accessed September 25, 2026.
  8. eCFR. 21 CFR Part 801, Labeling. Accessed September 25, 2026.
  9. eCFR. 21 CFR Part 803, Medical Device Reporting. Accessed September 25, 2026.