Draft Guidance Announces List of High Priority Devices for Human Factors Review

By Melisa M. Moonan –

FDA has been flooding the zone with new guidance documents. Earlier this week, FDA finalized the human factors guidance, “Applying Human Factors and Usability Engineering to Medical Devices, Guidance for Industry and Food and Drug Administration Staff” (Feb. 3, 2016). At the same time, the agency issued a complementary draft guidance, “List of Highest Priority Devices for Human Factors Review, Draft Guidance for Industry and Food and Drug Administration Staff” (Feb. 3, 2016).  

We will cover the final human factors guidance is an upcoming post. In this post, we focus on the complementary draft guidance, which is intended to help with the question of when human factors data must be included in a premarket submission. Until now, it has always been something of a mystery as to when such data are required and when they are not.

The draft guidance identifies a specific list of devices for which FDA would definitely expect human factors data:

• Ablation generators (associated with ablation systems, e.g., LPB, OAD, OAE, OCM, OCL)
• Anesthesia machines (e.g., BSZ)
• Artificial pancreas systems (e.g., OZO, OZP, OZQ)
• Auto injectors (when CDRH is lead Center; e.g., KZE, KZH, NSC )
• Automated external defibrillators (e.g., MKJ, NSA )
• Duodenoscopes (on the reprocessing; e.g., FDT) with elevator channels
• Gastroenterology-urology endoscopic ultrasound systems (on the reprocessing; e.g.,
• ODG) with elevator channels
• Hemodialysis and peritoneal dialysis systems (e.g., FKP, FKT, FKX, KDI, KPF ODX,ONW)
• Implanted infusion pumps (e.g., LKK, MDY)
• Infusion pumps (e.g., FRN, LZH, MEA, MRZ)
• Insulin delivery systems (e.g., LZG, OPP)
• Negative-pressure wound therapy (e.g., OKO, OMP) intended for use in the home
• Robotic catheter manipulation systems (e.g., DXX)
• Robotic surgery devices (e.g., NAY)
• Ventilators (e.g., CBK, NOU, ONZ)
• Ventricular assist devices (e.g., DSQ, PCK)

FDA indicates that it will expect a human factors report (as described in Appendix A of the Final Human Factors Guidance) for the listed devices unless the submission is for a modification to an approved or cleared device that does not affect usability, i.e., no change to users, use environment, user tasks, or user interface. Alternatively, a submission can provide a detailed rationale for why human factors data are unnecessary based on a risk analysis showing that the severity of potential harm from user error is not serious.

In providing the list, FDA utilized the following standard to identify the device types for which it expects submission of such human factors data: “Device types that have clear potential for serious harm resulting from use error.” FDA states that the draft list was derived by the agency from MDRs and recall information, and the list does reflect recent and historical agency concerns such as AEDs, infusion pumps, and reprocessing methods. However, there is not a detailed analysis as to how FDA applied the standard when generating the list.

Accordingly, we hope the agency will further clarify the terms “clear potential” and “serious harm” in the final guidance, particularly as they relate to established regulatory standards and definitions regarding product risk, such as those for reporting adverse events and malfunctions in the MDR regulation, 21 CFR Part 803, and for reporting recalls in the Corrections and Removals reporting regulation, 21 CFR Part 806.

Of perhaps greater concern, FDA states that additional device types may be identified in various ways, including through guidance, special controls, and classifications, as well as by individual reviewers on a case by case basis. The latter method is troubling. The welcome transparency provided by this draft guidance could be swallowed whole by case by case reviewer decisions to request human factors data for device types that are not on the list.

FDA has attempted to put guardrails around this possibility in two ways. First, the draft guidance puts the onus on manufacturers to proactively submit human factors data when their risk analysis indicates that user error (whether by failing to perform tasks, or by performing them incorrectly) could result in serious harm.

Second, the agency proposes that reviewer requests for such data could be made only if one or more of a group of factors apply. FDA lists certain factors where it states the factor also must be accompanied by a potential for serious harm resulting from use error, and other factors where it appears to presume that the potential for serious harm is inherent.

The factors where a potential for serious harm from user error must be present are:

• The submission is a PMA or a De Novo Petition
• The submission includes a change of intended users, e.g., from professional to lay use
• The device was modified or differs from the predicate in any of the following ways:

• The user interface has been modified (even if simplified)
• User tasks have been added or changed
• The severity of possible harm resulting from use error has increased
• The device will be used in a new use environment (e.g., in a home or vehicle)

The factors for which such potential for serious harm appears to be presumed are:

• The user interface was modified to satisfy a special control or recommendation (in a device-specific guidance document) related to its use. This factor makes sense if such controls and recommendations were promulgated primarily based on a potential for serious harm resulting from use error.

• The device type has been associated with recalls, adverse events, problem reports or complaints for which the cause has been attributed to use error or use error is the only explanation. This factor does not make sense unless the recalls considered are associated with a risk to health and “adverse events” means MDR reported events that have been vetted by the agency and confirmed to represent a potential device use issue. Inclusion of complaints does not make sense, as any complaint associated with use error resulting in serious harm would fall under MDR reported events. It is unclear what is meant by “problem reports,” but those too should meet the criteria for MDR reportability and be vetted by the agency to determine whether they represent a potential device use issue.  

We strongly encourage FDA to provide transparency and opportunity for comment for non-listed device types before permitting reviewer requests for human factors data on a case by case basis. Perhaps the agency could periodically propose any updates to the list under good guidance practices. At a minimum, such requests should require approval by the reviewer’s supervisor. Otherwise, there is a clear potential for surprise requests that could unfairly delay clearances and approvals.

Comments are due within 90 days from February 3rd to Docket No. FDA-2015-D-4599.