Human Factors: A Key Factor In Regulatory Approvals

By Doug Mead, CP Pathways LLC

In a previous article about bolstering regulatory intelligence, “GenAI: The Muscle Behind Strong Regulatory Intelligence For Combination Products,” I offered the foundational recommendation of employing GenAI search tools to conduct precedent research on agency decisions and communications. As I explained there, pointing GenAI queries at regulatory databases (FDA foremost among them) is a crucial, new best practice in the pursuit of regulatory approvals for combination products and other drug delivery devices.

In this article, I offer a case in point: The role of Human Factors (HF) in device design, development, and use. One of the most challenging areas of product development for an injectable biopharmaceutical is ensuring that the patient interface, i.e., the device packaging and IFU, is ideally suited for the intended patients and that the evidence for that is accepted by the FDA for product approval. This expectation is one of the highest regulatory risks for approval and one for which precedent research can yield insights that complement a regulatory strategy and related submission content.

The Context For Conducting Precedent Research On Human Factors (HF)

In preparing key  documents for FDA submission review of a BLA or NDA, the  Human Factors Validation Study (HFVS) report and a current Use Related Risk Analysis (URRA) will be critical documents. If your company has already had the expected prior engagement with FDA’s Division of Medication Error and Prevention Analysis (DMEPA) and received its input on the HFVS protocol, current Use Related Risk Analysis (URRA), and the proposed IFU, this history should be considered and reflected in the final submission documents.

When deficiencies in the study or the study results are identified by DMEPA and are a concern of the reviewing CDER Division, the outcome can sometimes be a Complete Response Letter (CRL) denying BLA or NDA approval. In a CRL, the FDA may recommend  changes to the patient interface and request a new HFVS.

In the aforementioned article on regulatory intelligence and GenAI search tools, I discussed the importance of learning about FDA policies and review precedents disclosed in drug approval review memos provided in the drugs@fda database. In relation to human factors considerations, these communications often include DMEPA reviews that disclose detailed study findings from prior applicants and the FDA's concerns raised during its review. For companies developing combination products similar to those already approved, such memos can provide critical insights into FDA expectations for the patient interface and IFU regarding device suitability, usability, and safety.

What Do GenAI Searches Tell Us About CRL Risk From HF Study Deficiencies?

If you search for examples in the drugs@FDA and CRL databases from the last five years where a human factors issue was a proximate cause of a Complete Response Letter, you will likely find that such occurrences are much less common than in prior periods. However, that reduction alone does not diminish the risk that DMEPA would play a heavy hand in the review or carefully scrutinize study findings.

The reduced incidences may, instead, be attributable to sponsors and human factors research groups becoming much more diligent in planning the studies, conducting formative studies, acting on the results, and justifying root causes and mitigations for the success misses in the validation study for both observed and knowledge tasks.

A ChatGPT5 search found the following instances for relevant CRLs in the last five years:

• BESREMI™ (BLA 761166) received a complete response on March 12, 2021, for a prefilled syringe (PFS), due to Hunan Factors and facility inspections. The PFS required dose adjustments using the graduated syringe.
• Risvan™ (NDA 214835) received a complete response on September 24, 2021, that cited human factors deficiencies. The reconstitution system included a diluent PFS and a lyophilized drug PFS that required careful intermixing.
• Other notable CRL hits included Sofdra™ (NDA 217347), on September 22, 2023, for deficient patient instructions, which required an additional human factors study after IFU revisions, the MydCombi™ ophthalmic spray/dispenser (NDA 215352) on October 22, 2021, for Human Factors Study deficiencies and the lack of a URRA, and  Ycanth™  topical solution ( NDA 212905 ) with a pre-filled topical applicator and enclosed ampule, on July 13, 2020, (beyond the five-year search parameter) that cited study use errors and palmar pinch strength limitations.

Beyond CRLs, DMEPA review memos provide comments on all use errors, use difficulties, and close calls; they review root causes and mitigations in effect; and they provide commentary about them. Companies submitting a combination product with a Human Factors study would be well advised to review several similar products to learn how DMEPA addressed study deficiencies and their recommendations for IFU revisions. Even though CRLs do not appear to be that common, all of my combination product clients continue to treat this submission element as a high regulatory risk.

There is one way to avoid this risk: obtain a waiver for a Human Factors Validation study, which is the next topic about which GenAI searches of precedents can be highly informative.

What Do GenAI Searches Tell Us About Human Factors Study Waivers?

One of the industry's hottest topics in Human Factors regulations is whether you can leverage prior HF studies on the same or similar delivery device for a similar patient population. Might FDA waive an HF study requirement for the new product, and, if so, in what kinds of scenarios?

In my prior work with Rhizome AI, a commercial GenAI search tool, we refined the tool to extract content from the FDA review database and provide a comprehensive list of instances in which these waivers were granted. Typically, the applicant provided a URRA and a comparative analysis justifying a waiver prior to filing the BLA or NDA  — and obtained a favorable determination. This history is usually reported in the DMEPA review memo, along with suggestions for IFU revisions to include “patient-friendly” instructions.

In one search I conducted using Rhizome AI, I found multiple examples of waivers of Human Factors studies by the FDA, including the following:

• PREVDUO (Glycopyrrolate; Neostigmine Methylsulfate)
• Midazolam Hydrochloride Autoinjector
• Zymfentra  Radicava ORS (Edaravone)
• Izervay (Avacincaptad Pegol Sodium)
• Simlandi (Adalimumab-RYVK)
• Pavblu (Aflibercept-ayyh)
• Enflonsia (Clesrovimab-cfor)
• IMULDOSA (Ustekinumab-SRLF)
• Zavegepant Nasal Spray (ZAVZPRET)

The search tool helped me understand the FDA’s rationale for the waiver, and, in my consulting role, I could convey these precedents and my recommendations to clients considering similar opportunities. As subcutaneous injectables become more common and as companies develop platform autoinjectors for multiple drugs, it’s worth considering a URRA approach to obtain a waiver.

What Does GenAI Tell Us About Use Errors?

For products without an HFS waiver possibility, I use a GenAI search tool to research the types of use errors that the FDA emphasized in their memos for any new or novel combination products recently approved. The results show that FDA often proposes IFU changes or, if the issues are significant enough to warrant a Complete Response Letter (CRL), the reasons for a CRL.

I review these FDA-proposed edits to an IFU and determine whether they are actionable for IFUs for similar products. When I review these memos, I also see instances where DMEPA states that a particular use error seen in the sponsor’s study has been reported in other studies of similar products and is adequately mitigated.

What Does GenAI Tell Us About Delivery Device Critical Tasks?

In reviewing clients’ URRAs and HF study protocols, I have seen multiple instances in which critical tasks are inconsistently identified across applicant submissions. Companies can devote significant time trying to determine whether a use task is critical to meeting FDA expectations. While I’m not sure whether FDA has ever comprehensively designated critical tasks, the agency does seem to scrutinize these categories in applicant document reviews.

I performed a Rhizome AI search to see if there has been any consistent interpretation of these for observable HF study tasks and in IFU language. The query response provided the following examples, and it referenced the product review memo in which each FDA assessment was documented. The list below summarizes the key findings in a query response I received:  

Critical Handling Tasks

Several tasks related to the preparation and handling of prefilled syringes (PFS) or autoinjectors were identified as non-critical. These tasks generally involve steps that, while important for proper use, do not pose significant risks to safety or efficacy if errors occur. Examples include:

• Inspecting the device or packaging: Tasks such as checking the expiration date, inspecting for damage, or verifying drug condition were deemed non-critical
• Allowing the product to warm to room temperature: For example, allowing a PFS to warm for 15-30 minutes was identified as non-critical
• Storing the product: Tasks like storing pens in the refrigerator or ensuring proper storage conditions were considered non-critical
• Opening packaging: Tasks such as breaking tamper-evident seals, opening cartons, or removing devices from trays were identified as non-critical
• Injection site preparation: Tasks such as selecting an injection site, cleaning the site, or pinching the skin were identified as non-critical

For this response, the search tool found specific FDA opinions in review memos for HULIO ADALIMUMAB-FKJP, Skytrofa, Bonsity Teriparatide, Entyvio (vedolizumab), and Zepbound (tirzepatide) delivery devices. Those opinions consistently showed that certain tasks identified would be designated as non-critical.

Despite those findings, I recommend that most tasks performed by a user should be listed as “critical” in a URRA. DMEPA focuses on use errors of critical tasks, defined as those that can lead to any harm or to medication errors even if the harm from a medication error is not serious. It can be hard for companies to parse out risk scenarios and seriousness for any task, let alone the potential for over- or under-dosing.

When errors are observed in any study for any task, the FDA may comment on the need for additional mitigation in its review for approval, despite the task's criticality. Regarding the above-listed examples, which may be DMEPA reviewer-specific, user-specific, or product-specific:  As the saying goes, past precedents provide regulatory insights but are not always indicative of current FDA expectations.

What Does GenAI Tell Us About DMEPA Concerns?

There are other illuminating precedents related to Human Factors that can be found by searching the FDA database. One query from which I found the results to be particularly revealing was: Provide a list of FDA drug approvals where DMEPA recommended against approval due to human factors study results.

This search query surfaced explicit examples in which DMEPA raised serious questions about the HFVS (Human Factors Validation Study) results. However, these did not always rise to the level of a CRL, as some commentary reflected formative study reviews, instances where the issues were resolved in the final BLA or NDA review or where DMEPA wanted to emphasize a particular concern:  

1. Teriparatide (Teriparatide Injection, 250 mcg/mL)
   • Company: Apotex
   • Application Type: ANDA 211097
   • Year of Decision: 2024
   • DMEPA Recommendation: DMEPA indicated that the formative human factors study conducted did not adequately demonstrate that the proposed device does not pose significant risks to patients switching between the reference listed drug and the proposed product. They recommended conducting additional human factors studies with the proposed device and the reference drug.
2. Mounjaro (Tirzepatide)
   • Company: Eli Lilly and Co
   • Application Type: NDA 215866
   • Year of Decision: 2024
   • DMEPA Recommendation: The human factors study results raised concerns, although specific recommendations against approval were not detailed. However, the review indicated the need for further evaluation of the human factors study.
3. Lyumjev (Insulin Lispro-Aabc)
   • Company: Eli Lilly and Co
   • Application Type: BLA 761109
   • Year of Decision: 2021
   • DMEPA Recommendation: The human factors study results indicated potential usability issues, although specific recommendations against approval were not explicitly stated.
4. SYMJEPI (Epinephrine)
   • Company: Adamis Pharmaceuticals Corp
   • Year of Decision: 2017
   • DMEPA Recommendation: DMEPA concluded that the human factors study did not demonstrate that the Symjepi prefilled syringe could be used safely and effectively, noting critical failures in product use and methodological flaws.
5. ADMELOG (Insulin Lispro)
   • Company: Sanofi-Aventis US
   • Year of Decision: 2020
   • DMEPA Recommendation: DMEPA found that the study participants were not representative of the intended user population, raising concerns about usability and leading to recommendations for further evaluation.
6. BESREMI (Ropetinterferon Alfa-2b-NJFT)
   • Company: PharmaEssentia Corp
   • Year of Decision: 2021
   • DMEPA Recommendation: DMEPA found the human factors validation study protocol unacceptable, indicating significant reservations about the usability of the product.

Of course, each search result requires careful regulatory analysis of the source document to fully understand the issues the FDA raised. GenAI search results that provide a link to the FDA source document are critical. The explanations listed above need to be researched to understand their implications for new products in development. For example, in the case of the BESREMI™ review, while DEMPA recommended against approval, it was not the predominant reason for its CRL.

Expanding Your GenAI Regulatory Search Targets

The power of GenAI searches can go well beyond general questions described above. In my practice, I use the searches to find the latest precedents on many other questions, including:

• What companies have included training arms in their Human Factors Validation Studies, and what did the FDA say about the results from these study groups?
• When has the FDA cited the potential for study bias when a moderator script implied that an IFU was “available” for the participant?
• When has DMEPA accepted patient participant surrogates or proxies when a rare disease or patient population is extremely difficult to recruit, and how has this been justified by the sponsor?

Because Human Factors studies and URRA development are significant efforts — and risks — for drug development programs, it's worth spending the time to check FDA precedents for the likely scenarios specific to the company’s product in development. Regulatory staff should use these GenAI search tools regularly as a new resource for managing regulatory risk.

Doug Mead is Principal Consultant and President of CP Pathways LLC, a combination product consultancy. He helps combination product companies with their regulatory strategies and works closely with delivery device teams and regulatory staff to prepare submissions in line with the latest regulatory expectations.