The New FDA Draft Human Factors Guidance: A Bridge Too Far

More than a decade ago, FDA began systematically to incorporate review of human factors (HF) design validation within 510(k) reviews.  In 2016, the agency issued its definitive HF guidance to guide manufacturers through human factors engineering processes during the development of new medical devices, focusing specifically on the user interface.   All this time, FDA has applied the requirement for HF data to a subset of device types that pose the highest usability‑relate risks, e.g., software controlled devices, some hospital devices, and home use devices.  Accordingly, only a subset of 510(k) submissions have historically included HF validation data.

Now FDA has issued a draft guidance, Content of Human Factors Information in Medical Device Marketing Submissions (Dec. 9, 2022).  This guidance is intended as a complement to the earlier one, focusing on the type of HF data manufacturers should supply in a marketing submission.

In one way, this is good news.  The generation of HF data is a detailed and fairly burdensome process, although certainly short of a full-blown clinical study.  Furthermore, FDA’s HF reviews tend to generate a good percentage of the burdensome additional information (AI) requests in 510(k) reviews.  Many of these AI requests arise because of manufacturer unfamiliarity with the peculiarities in vocabulary and requirements current among FDA’s HF review staff.  And HF‑related AI requests are especially nerve‑racking to manage, because of the ever‑present threat that the study must be repeated to adequately respond.  Industry would certainly welcome guidance intended to smooth the HF aspects of 510(k) reviews.

But there is a very big fly in this ointment.  At least as currently drafted, it appears that the draft HF guidance dramatically expands the number of 510(k) submissions that require HF data.  The threshold is set extremely low:  if there is one “critical task” identified for a new or modified device, HF data must be addressed in some fashion.  (A “critical task” is a “user task which, if performed incorrectly or not performed at all, would or could cause serious harm to the patient or user, where harm is defined to include compromised medical care.”)

To put it another way, according to the draft guidance, the only valid justification for not including human factors data for a new device is that there are no critical tasks.  In the past, FDA made relatively specific determinations as to the HF data requirements (or lack thereof) for various types of device.  This approach is one-size-fits-all.  It will dramatically increases the burden and uncertainty associated with every submission across many device types.

If a single critical task is identified, then human factors validation must be included, regardless of whether such testing was required for the identified predicate.

If the submission is for a modified device which involves a new critical task or which impacts an existing critical task, then a human factors engineering report is required even if usability data were not required for clearance of the original unmodified device/predicate.

The only exception for a modified device is when there were no changes to: the user interface; intended device users; intended device uses; intended use environment(s); training or labeling.  Even in those cases, a high-level summary of the HF evaluation must be included.  FDA could challenge a company’s rationale for concluding that a change did not impact “human factors considerations” and thereby request HF data.  Once again, the uncertainty associated with the submission process will increase.

The draft guidance is at odds with the fundamental purpose of the 510(k) program, to enable an efficient risk-based process for FDA review and clearance of low to moderate risk devices.  A fundamental tenet of the program is that the same testing required for the predicate is generally required for the subject device.  Typically, only changes in technological characteristics from the predicate require use of test methods not required for the predicate.

Consider this scenario:

• A manufacturer files a 510(k) for a Class II device, a surgical mask.
• The subject device is identical to the predicate in all respects.
• Both devices have a single critical task (they must be correctly fitted by the user).
• FDA did not require HF data as a basis for clearance of the predicate.
• The device-specific guidance for surgical masks does not require HF data Surgical Masks – Premarket Notification [510(k)] Submissions | FDA.
• Per this new HF draft guidance FDA would require the highest level, tier 3 HF usability data for 510(k) clearance of the new surgical mask.

As the example shows, this draft guidance departs from the long-standing approach and introduces a new requirement for HF testing, regardless of whether such testing was required for the predicate and regardless of whether alternative methods of addressing differences in technological characteristics are available and less burdensome.  The new policy needs to be justified on a cost/benefit basis, which has not been done, at least as far as can be discerned from the draft guidance.  For example, the agency has not offered how many 510(k) submissions currently require HF data, the adverse effects of not applying an HF requirement more widely, and how many submissions would now present HF validation data after the change goes into effect.

At least to our knowledge, it appears that currently a relatively focused subset of all 510(k)s currently require HF validation.  If so, implementation of the draft guidance will likely cause in dramatic increase in burden on applicants and FDA.  For example, there could be a significant increase in Q‑Subs in order to work out the HF validation protocol or ensure that the agency agrees that HF data is not required.  This issue will increase uncertainty in the 510(k) modification decision making process. For example, if a critical task is introduced or impacted does that trigger the need for a new 510(k)?

The draft guidance can be fixed by simply deferring to existing requirements as to when HF data are required (e.g., as a de novo special control or in a guidance document for a specific device type).  To justify an expansion of the 510(k) burden of this magnitude, FDA needs to provide a clear and detailed accounting of what the problem is, the cost of this change, and why it will improve safety without hindering the development of better devices.

FDA says the primary purpose of this draft guidance is to smooth 510(k) reviews by providing more granular information about how to address the requirement.  It seems like mission creep for FDA to go beyond this primary purpose to impose a dramatic expansion of the requirement.

In short, this draft guidance imposes a major policy change with far-reaching implications.  Simply saying that the approach is “risk-based” is not adequate justification.  The costs and uncertainty of this expansion to HF requirements are significant.  The countervailing need for the change, however, is not well‑documented.  FDA should be careful not to upset the balance inherent in the current 510(k) system, which is used to bring the majority of medical device to market.

Comments on the draft guidance are due on or about March 8, 2023.