FDA revised "Applying Human Factors and Usability Engineering to Medical Devices" on August 3, 2026 — the first revision since February 2016. The documentation provisions moved substantially. The participant provisions didn't move at all. Here's the diff, and what it means for planning a study.
The revision was issued under Level 2 guidance procedures (21 CFR 10.115(g)(4)), the mechanism FDA uses for changes that don't represent significant new policy.
The guidance history table describes the work as updating the definitions and Section 9 (Documentation) for consistency with the "Content of Human Factors Information in Medical Device Marketing Submissions" final guidance, and deleting Appendix A.
That companion guidance is the driver. It went final on May 29, 2026 and applies to submissions received on or after August 1, 2026. It establishes risk-based HF Submission Categories that determine what human factors information belongs in a given 510(k), De Novo, PMA, or HDE.
The 2016 guidance did two jobs: describe the development process and specify what to file. Appendix A held an eight-section HF report outline, and Section 9 pointed to it. Both are gone. Section 9 now defers to the companion guidance. What remains is a process document.
Section 8.1.1, which governs test participants in validation testing, is unchanged from 2016. So is the sample size appendix, including the Faulkner (2003) data behind the number 15.
The provisions that determine recruitment scope all survived verbatim: 15 participants minimum, 15 per distinct user population, US residency, no sponsor employees, and representation of functional limitations where the device treats a condition that produces them.
We cover what each of these means in practice in How Many Participants Do You Need for a Usability Study? Nothing in that article needs revising.
Documentation expectations moved. Participant composition didn't. That has a practical consequence for how programs sequence their work.
A protocol naming three distinct user populations has committed to 45 completed sessions. No submission strategy reduces that number.
The harder problem is quota stacking. Credential requirements, procedure volume thresholds, and availability windows are each reasonable on their own. Combined, they can define a population smaller than the number of completes the protocol requires. That constraint exists whether or not anyone checks for it before the protocol locks.
The check is a funnel: qualified pool size, reachable fraction, response rate, qualification rate, show rate. It takes a few days and it belongs before the protocol is final, not after.
Section 6.4.3 permits company employees to serve as participants in formative evaluation. Section 8.1.1 prohibits them in validation. Under budget pressure, that reads like an opportunity to staff formative work internally.
The same section explains the risk. Employee participants may not represent intended users, may already be familiar with the device, and may be hesitant to give honest feedback. FDA's framing is that their performance and opinions could be misleading or incomplete.
The guidance is blunter elsewhere: if no formative evaluation is conducted and design flaws surface in validation, that validation test has become a formative evaluation.
Sections 8.1.3, 8.1.4, and 8.1.7 close the exits. If validation produces use errors on critical tasks, you can't mitigate in a submission by revising the instructions for use, adding training, or deferring to a future version — not without new data showing the fix worked.
The guidance history table understates the diff. Also changed:
None of this changes recruitment scope. It does mean the change log isn't a substitute for the document.
If you're scoping a validation study, the feasibility question comes before the protocol question. How many distinct populations, how large is the qualified pool for each, and does the geography support it.
Usability House provides the facility, recruitment, and study logistics behind submission-grade human factors programs. If your team needs labs and participants, we can quote within 24 hours.
We recruit clinical and patient populations against your screener and run studies in purpose-built labs in Minneapolis.
Get in Touch