Regulating For Patient Safety And Success With Combination Products

By Tom von Gunden, Chief Editor, Drug Delivery Leader

The two days I spent at the recent AFDO/RAPS Combination Products Summit, the annual meeting for those in biopharma responsible for the regulatory aspects of combination product development, were particularly reassuring. Why do I choose reassuring as the “put it in a word” takeaway from my experience at the conference? Well, I suppose it stems from my own combination of empathy projected onto a patient landscape I editorially monitor, and b) potential self-interest should I one day join that combo product-receiving population or have loved ones who do. With those professional and personal contexts in mind, I was pleased to discover so many presentations, commentaries, and conversations focused on patient-centric considerations for combination products. That set of interrelated topics included human factors and usability, risk and safety, and post-market reporting and changes.

Perhaps the most bracing (in a good way) table-setter was a data-validated reminder of just how voluminous is the acceleration of advanced, biologics-based therapies being put into conveniently administered delivery devices, often for in-home self-administration. That reminder occurred early in the event when, in his “From the Desk of the Office of Combination Products” remarks, FDA OCP Director James Bertram highlighted a slide showing that the highest number of submissions, by type of combination product, to FDA over the past year was for prefilled biologics syringes. That statistic suggests significant potential benefit in targeting unmet needs with increased efficacy and added convenience. It also underscores the need for diligent oversight of device usability and safety.

Crosstalk Between Drugs, Devices, And The People Who Combine Them

A key challenge in combination product development – particularly when patient safety enters the thinking (as it should) — continues to be bridging role-based functional silos commonly characterized as the “drug side” or “pharma side” versus the “device side” or “delivery side.” We have addressed this challenge in other editorial here at Drug Delivery Leader, including insights on organizational alignment offered by our senior industry and technical advisor, Fran DeGrazio.

In the opening Solutions Exchange at AFDO/RAPS, co-moderator Khaudeja Bano, VP of Combination Product Quality at Genentech/Roche, echoed the observation of an audience member (who had asserted, “nomenclature matters, words matter”) about ways in which even the terminology used in different departments can make cross-functional communication about patient safety difficult. Bano noted challenges stemming from “the interpretation of definitions, especially between the pharma perspective and device perspective” even around seemingly “plain English words” such as correction and recall.

A point of group consensus emerging from Solutions Exchange table conversations was that differences in how functional areas view and talk about patient risk and safety often become most noticeable when documenting and responding to post-market data. It can be challenging to determine and, therefore, to address which pieces of feedback point to the drug as a potential problem, which feedback points to the device, which points to both, and, in some cases, which ultimately points to neither. Biopharma companies’ systems and procedures for handling patient and HCP complaints often illuminate uncertainty about cause and remedy (if remedy is even needed). And that uncertainty often reflects gaps in visibility, despite multi-department efforts at collecting and analyzing clinical and patient data on usability, efficacy, and safety.

As one audience member put it, “Complaints coming from the device side are often siloed [from those coming in from the drug side], and the integration of the two can be quite difficult for companies, especially in navigating how to set their post-market reporting systems.” Clearly, there is opportunity for expanding while tightening approaches to collecting and reviewing field data. If done well, doing so would foster more effective cross-functional conversations about what to make of what is coming in.

Post-Market Reporting: Responding To Inputs Responsibly Yet Reasonably

The use of post-market data was a focal point of a session later that first day at the AFDO/RAPS summit. In setting the context, moderator Jiayang Shen, Executive Director of Medical Device and Combination Product Development at Merck, referred to the FDA having begun daily publication of adverse event data through the FAERS [FDA Adverse Event Reporting System] public dashboard. After acknowledging the agency’s “commitment to data transparency and real-time protection of the public health,” Shen queried the panel about potential implications for industry from this broadening of visibility.

While joining Shen in championing transparency, the panel of seasoned industry insiders, as expected, offered cautionary reminders that raw data, presented without the full situational context, exposes it to limited and/or limiting interpretations lacking what could otherwise be instructive nuance. Panelist Susan Neadle, President at Combination Products Consulting Services and long-time biopharma industry vet, expressed concern that knee-jerk, “the sky is falling” reactions, or worse – opportunistic rushes to litigation – could result. “I worry that, when there is free access for everybody, real-time … they’re not necessarily applying the educated context of the medical implications of the specific data,” Neadle said.

Post-Market Panel At AFDO/RAPS CP Summit

Panelist Khaudeja Bano, returning from her earlier moderator duties, similarly weighed in with concerns about the use of the FAERS database. For her, not only might unfiltered adverse event reports trigger overreactions, but those overreactions could threaten to wall off some patients from receiving the real benefits of products that may have become unduly demonized by the reporting. Bano painted the potential for reductive and, therefore, concerning interpretive scenarios in this way: “If you can imagine [that] this data set includes every piece of serious adverse event [reporting] within adverse event aggregate data, it mixes things up for a lay person who is going out to review this data. Especially with the innovative therapies that are out there, my concern is it is going to negatively, potentially influence the consumers [away] from certain therapies.”

While reiterating industry’s regulatory obligation to safety reporting, especially around serious adverse events, Bano proposed that risk analysis shouldn’t be conducted without similar attention to benefit analysis. “What's completely missing from that dynamic is the benefit-risk,” Bano explained. “When you can't share the benefit aspect of post-market surveillance data, which we struggle with even within our organizations: We are so focused on the risk aspect, we sometimes completely forget that what it means to extending a life or saving a life is very different when put together in context of a benefit-risk [analysis].”

Is The Devil In The (Device) Details?

As mentioned above, while combination product post-market data may indeed clearly point to concerns related to the drug or biologic constituent part, sometimes the investigative focus should be on the device. But exactly when may not always be clear. What is clear, nevertheless, is the need for biopharma organizations, particularly those moving from traditional drug product development into combination product development, to conduct what I’ll refer to, in this context, as anticipatory learning. Especially given that combination products are often on the forefront of innovation for targeting unmet needs via previously uncharted administrative routes, anticipatory learning would include making educated guesses about potential device issues, especially in novel combinations. One way of approaching this is by extrapolating possibility or likelihood even from imprecise or incomplete data gleaned from scenarios only peripherally applicable to the envisioned use case at hand.

Susan Neadle described something like that when she advised, “As you look into [adverse event] data, you can also look at other products that are out there —  not even in the same therapy area — but with the same type of device constituent part. You won’t be able to get to the granularity that you would with the MAUDE [Manufacturer and User Facility Device Experience] database. But you can learn things about, oh, these are potential issues I might run into for my type of product.”

Bano echoed the anticipatory nature of combination product development, emphasizing the need to think carefully about what may be unique about the intended use scenarios. However, she warned against allowing apprehensive acts of envisioning lead to undue hesitation in moving forward the dial of innovation. Otherwise, as both she and Neadle had previously cautioned regarding responses to isolated instances of concerning post-market data, worries in a “the sky may fall” vein could inordinately influence pre-market ideation away from ground-breaking advances. Bano described a measured approach: “Bring together a diverse cross-functional team that can imagine the potential use, abuse, off-label use, and intentional abuse of those products,” she advised, then went on to offer this reaction-tempering suggestion: “Keep it reasonable, likely. When you imagine all the possibilities that can happen, you have to stay pragmatic. You cannot design for the one-in-a-million.”

Human Factors At The Intersection Of Patients And Delivery Platforms

A key trend in drug delivery, generally, and, therefore, a rising phenomenon in combination product development is the lure of repeatability offered by platform approaches. Whether it’s putting new formulations in the same device technologies or creating bespoke variations that expand a combination product portfolio, questions arise about the degree to which pre- and post-market studies of patient safety and device usability need to be repeated.

In a Day 2 session on human factors at the AFDO/RAPS summit, moderator Rumi Young, Director of Regulatory Policy at Novo Nordisk, engaged the panel in considerations of when and how often to conduct and document studies or analyses in accordance with FDA draft guidances on human factors [HF] studies and/or on URRA [Use-Related Risk Analysis].

To illustrate a tailored approach based on determining the applicability of existing data, Senior Usability Engineer Matt Marber, also from Novo Nordisk, pointed to the FlexTouch® pen injector. Since its initial approval in 2010, the device has been employed for various drugs and indications. Those differences are clearly noticeable, if for no other reason, than the different colors and sizes of the devices in the market. Yet, at a basic design and use level, it’s essentially the same device, used the same way – i.e., there’s more of the more in a “more or less the same” characterization. As Marber described, in all permutations, the user steps remain a fairly simple, straightforward set of four: “Put the needle on [the pen injector], dial your dose, insert the needle into your selected site, and push the dose button.”

Human Factors Panel At AFDO/RAPS CP Summit

Marber’s explaining and reiterating the standard steps for device usage usefully framed the question of how much variation in subsequent treatment scenarios employing a platform technology would require additional HF studies and to what extent. As Marber acknowledged about a product that has been on the market for 15 years, “That is hundreds, probably thousands of usability test participants, representing patients, caregivers, and healthcare professionals. That is a significant investment in time, money, and what we have learned.”

Marber’s point in noting the volume of existing usability data was not to suggest that additional studies shouldn’t be conducted, even though doing so might seem to be an act of data collection overkill. “The whole point of usability studies is, you want to learn something: whether your risk control measures are effective, whether or not a specific user group has difficulties using a product,” Marber explained. “But after you’ve seen 30 participants in your first study, you’re not really learning that much. When you multiply this [number of studies by the number of participants], that’s a lot of nothing new.” Despite the temptation to rest-on-one’s experiential laurels, Marber described pushing past potential complacency when leveraging platform technology. “That [‘nothing new’] raises flags for us,” he offered. “Maybe we can do something a little bit different when we have the same product, maybe a new variant of it. How can we be more strategic when it comes to the human factors practices with the same device?”

The strategic approach Marber went on to illustrate employs a flow chart, based on FDA guidance documents (e.g., URRA), which maps differences, including those potentially worthy of additional study. “When we conduct these analyses for each new variant, we have a team member or multiple team members go through each of 10 assessment areas and identify if there are any gaps that exist between the proposed product and the reference product,” he explained. “From there, you can follow a comparative analysis guidance document that's referenced within a threshold analysis guidance document to classify those differences as no differences, minor, or other.” With this approach in place, decision-informing steps include identifying and assessing such key considerations as risk, gaps, and critical user tasks, all toward determining if additional full (less likely) or supplemental (more likely) HF validation studies are warranted.

Risk Analysis: Leveraging Prior Learning To Forestall Study Sprawl

Panelist Tina Rees, Director of Human Factors at Regeneron, furthered the consideration of how earlier studies can inform decisions about later combination product development in terms of patient usability and safety. Rees focused on the application of URRA guidance and methodology in assessing patient risk. In setting the context, Rees pointed to varying indications and patient populations for which the monoclonal antibody Dupixent® has been developed and approved for delivery — initially, as a prefilled syringe and safety system and, subsequently, as an autoinjector. Across the various conditions Dupixent targets, common human factors assessed during initial and follow-on development include vision, hearing, physical strength, cognition, injection experience, and comorbidities.

Echoing Marber’s comments about product variations, Rees acknowledged that, as the therapeutic indications and use cases for the Dupixent expanded, questions arose about whether and to what extent additional user studies would be needed. The answers came in zeroing in on unique differences in subsequently targeted populations/indications in terms of the ability of each next group of users to properly and effectively operate the device. “We were seeing in the URRA that we have the same critical tasks identified. The device itself is exactly the same,” Rees explained, then added, “ But the users are different.”

To illustrate when user population differences necessitate additional studies and when they don’t, Rees traced the development trajectory of Dupixent indications from atopic dermatitis [AD] to eosinophilic esophagitis [EoE] to chronic obstructive pulmonary disease [COPD]. When moving from AD to EoE, for example, the company determined that one of the (in Rees’ term for it) primary performance-shaping characteristic for AD patients — limited hand strength or manual dexterity — was not likely to be of significant severity for EoE patients. “What we determined is that [for] patients who have EoE, the use of the device by them would be equivalent or easier compared to the already validated atopic dermatitis data,” Rees explained. “So, we documented this justification in our URRA to show that there were no additional or novel use-related risks and submitted that to the FDA. And the agency agreed that the data was sufficient and could be leveraged.”

Rees went on to share the contrasting example of moving from AD to COPD. Because of a preponderance of occurrences in geriatric populations, COPD patients may similarly experience physical strength and manual dexterity limitations. Yet a unique difference for them, compared to, say, the AD population, comes in potential challenges with cognition. For that reason, the existing AD study data didn’t apply for that particular risk assessment item. “If we had just had a physical strength issue, we could have done a bench study to show that their force capabilities were still acceptable,” Rees explained. “However, because it was a cognition issue, we determined that we needed to run a supplemental [HF validation] study to make sure that those patients were able to interact with the product similar to AD patients. And the results were [that] FDA did approve our additional indication based, in part, on that supplemental HF data that we submitted.”

Safe Innovation Relies On Incremental Insights, Intentionally Acquired

As I collected my thoughts on the collected commentary from the AFDO/RAPS Combination Products Summit, I arrived at an insight about how product development and delivery innovators across the biopharma ecosystem work to advance treatments that improve health outcomes while doing so safely. The insight is insight – their commitment to it, and how they strategically set out to glean insights from product use and then diligently map approaches to building on the learnings from each.

And, as they gather, review, and make decisions based on data inputs, especially those pertaining to risk, safety, and usability in a regulatory context, they model an instructive set of “what to do, what not to do’ response behaviors. It’s a practice that I and no doubt others in the “general public” would do well to apply in the face of doubt-driving anecdotes and assertions intended to challenge medical merit. In characterizing the industry approach, I contend that it is a matter of favoring one “pair of P’s” over a less helpful yet often competing pair. The AFDO/RAPS Combination Products Summit showcased this tendency as follows:

As they work responsibly in a highly regulated environment and in service of the public good, these industry innovators self-regulate when confronted by data inputs some might consider worrisome. Rather than indulge in the temptation to panic and project (in a catastrophizing way), they embrace a different “P set.”  They pause and ponder. And by doing so, they demonstrate a patient-centric intentionality perhaps accurately captured in a third “pair of P’s”: protect and provide.