How Early Talks With The FDA Can Pave Your DTx Approval Pathway

A conversation with Acacia Parks, PhD, MBA, CEO, Liquid Amber

Though the FDA doesn’t have its own definition of DTx, the Digital Therapeutics Alliance says they’re "evidence-based therapeutic interventions delivered through digital tools that can be used independently or integrated with other therapies to prevent, manage, or treat a medical condition." Rather than regulating DTx based on their name, the FDA regulates them by function and risk profile. 

Critical to any level of success in bringing a DTx to market are frequent, forthcoming, and early conversations with the FDA. Consultant Acacia Parks discusses when and how DTx developers should incorporate them into their overall development and clinical trial processes.

Clinical Leader: Understanding they might be treading an unknown path toward approval, how and when should medical device and DTx companies begin conversations with the FDA? Are presubmissions part of that process?

Acacia Parks: Presubmissions are an important part of the process for receiving regulatory approval, particularly when doing something innovative where there aren’t a lot of precedents to build on. While there are also other less formal mechanisms for engaging with the Agency, such as interacting at conferences, submitting emails to the Digital Health Center of Excellence’s inbox, and participating in the Total Product Life Cycle Advisory Program (TAP) program. The presubmission is the primary way to receive extensive written feedback in an official capacity.

The sooner a company engages with the FDA, the better — especially before anything is finalized, because the FDA may give feedback that changes what the sponsor wants to do going forward. Presubmissions can help de-risk at each stage of the product life cycle.

For example, a sponsor may want to decide whether to build a product for a prescription use case depending on what they learn in a presubmission about the evidence that would be required to get such a product cleared. A sponsor may adjust the design of their pivotal trial depending on what they learn in a presubmission about the expectations for sample composition, endpoints, or statistical approach. It would be a mistake to ask the FDA for input after the product is “done” or after the research is already underway, as the purpose of feedback is for it to be implemented.

What information is contained in a presubmission?

Broadly, a presubmission is an opportunity to ask the FDA a limited number of questions about a product. Ideally, the presubmission document itself contains all the information the FDA needs to provide answers to a sponsor’s questions. In general, the FDA prefers to have as much detail and background information as possible, — so it can be a lengthy document, anywhere from 20 pages to over 100, depending on the nature of the questions being asked and the background needed to answer them.

For example, if the sponsor hopes to get the FDA’s feedback on a full study protocol (often ~80 pages) then the package will be especially large. Other presubmissions may focus on higher-level strategic questions and may require less documentation. That said, it is important not to flood the FDA with unnecessary information; the package should be carefully constructed to meet their needs.

Key areas of the presubmission could include the reason for the presubmission (what information the sponsor is looking to gain, about which product); the proposed indications for use (IFU); background on the product including product strategy, device description, details about software development, maintenance planning, and cybersecurity; regulatory information including the history of the product, the benefit/risk evaluation and expected pathway and special controls; and clinical performance testing plans, including study protocols or statistical plans.

The document, which takes about three months for a sponsor to assemble, is submitted to the FDA, which responds about 70 days later with written feedback. The sponsor may then opt to meet with FDA to discuss the feedback. After the meeting, the sponsor submits draft minutes for the FDA’s approval. All in all, it’s about a six-month process.

As it relates to trial design, what are the can’t-miss items to cover or questions to ask in an FDA presubmission?

The contents of a presubmission vary dramatically depending on the sponsor’s objectives for a given presubmission. One presubmission may focus on higher-level product strategy and product features, while another might zero in on clinical performance testing requirements. The questions asked will be different for these different types of meetings, and in general, no single meeting can cover every type of topic and objective. However, here are some general do’s and don’ts:

Do:

• Tell the FDA what you propose and ask if they agree with your approach.
• Give the FDA enough information to comment on the topics you ask about.
• Be honest; the panel can help you best if they know what you are trying to achieve and how you are trying to achieve it.
• Use the meeting to explore why the FDA gives the answers it gives so that you can understand your options for addressing their concerns.

Do not:

• Ask the FDA to tell you what to do. This is not their role, but it is also not to your advantage to be told, anyway!
• Ask the FDA to comment on topics that you have not given them details about.
• Wait to talk to the FDA until you feel that everything is perfect.
• Press a topic that FDA has said it cannot comment on.
• Be afraid to propose what you think is reasonable! Sometimes, the FDA is figuring out what it thinks in the process of answering your presubmission, so use the opportunity to help mold their thinking.

What types of questions will the FDA respond — or not respond — to?

If you ask the FDA what pathway your product belongs on (510k vs. De Novo, for example) they may decline to answer and instead refer you to the 513(g) mechanism, which is designed exactly for that type of question. While they can explore certain questions related to predicates in presubmission meetings, which can indirectly give you an idea of whether your product is eligible for 510k, they cannot rule formally on a regulatory pathway in a presub.

If you ask the FDA what to do — for example, what type of study to run or what endpoints to use — they are likely to bounce that question back to you as the sponsor. Even if they are willing to answer, which they sometimes are, this is not the type of question I recommend asking, because it pushes you into doing what they recommend. Instead, propose what you think is reasonable, and help them educate you on what their concerns are around study design,  endpoints, etc. which lets you then consider your options.

If a company is designing a trial and seeking approval for a companion diagnostic, what are the unique regulatory considerations the company should be aware of?

There are not a lot of precedents for FDA-cleared digital therapeutics that provide diagnoses. Any time there is a use case that is new or uncommon, it is in a company’s interest to talk to the agency about it.

Depending on what the sponsor wants to claim about the product, it may or may not need clearance. The word “diagnostic” holds significance and typically necessitates an FDA clearance to use, but there are phrases like “non-device clinical decision support” that apply to products that do not need to go through clearance. However, the product characteristic differences between a diagnostic product and a clinical decision support product can be negligible; it is largely the intended use and marketing claims that determine regulatory status.

When it comes to writing a presubmission, companies can choose to write their own or outsource the task to a consultant. What are the pros and cons of each?

While it is rare for presubmissions to go badly, it’s somewhat common for them to fall short of the sponsor’s goals in one way or another. Maybe the FDA panel feels they cannot answer a key question, either due to a lack of background information or due to the phrasing of the question. Maybe the sponsor gets an answer that makes them wish they had not asked the question in the first place. Maybe the sponsor is left with a request for more information and submission of a new presubmission if they want to get their questions answered. Maybe the sponsor does not understand how to utilize the feedback they received.

To avoid outcomes such as these, sponsors should work alongside someone who writes presubmissions often and is experienced with the ins and outs of FDA communication when submitting its presubmissions. While there is a cost to employing or contracting someone with this type of expertise, the cost of walking away from a time-consuming and labor-intensive process without the desired insights is arguably higher.

About The Expert:

Acacia Parks is a digital health executive whose experience revolves around scientific, product, and regulatory strategy for digital therapeutics/software as medical device (SAMD). As chief science officer of Twill (fka Happify) for 10 years, she built evidence generation strategies for direct-to-consumer, enterprise, health care, pharma, and FDA audiences. She then spent a year as chief behavioral health officer at Found, a telemedicine weight loss company, building their behavior change program, both digital product and coaching offerings. Now, she runs Liquid Amber, a consulting firm that delivers strategic advising to DTx companies and investors with a focus on products with high scientific and regulatory complexity, including novel mechanisms of action and delivery methods and combination drug/DTx products, across a variety of therapeutic areas (including cancer, obesity, treatment-resistant depression, Alzheimer's, social anxiety, ADHD, and chronic disease). She serves as an advisor to Lindus Health.