For digital therapeutics companies, the FDA has often been viewed as the gatekeeper to getting products to the patients. However, today the conversation around efficacy and research is branching beyond just what ticks the FDA boxes, into what kinds of studies are needed to meet the expectations of stakeholders and end users. This morning at DTx East, a group of panelist discussed the current regulatory landscape.
When it comes to digital therapeutics, there are two buckets of regulation: those requiring a prescription, which need 510(k) or de novo clearance, and those that do not, which can fit into the enforcement discretion designation. The latter has less regulatory oversight, but this has caused some confusion around what kind of research is expected from the agency – and from industry.
“For enforcement discretion, the idea is that this is a body of lower risk products for which FDA does not have the resources to focus their compliance efforts. So, they have decided that the guidance that they have given is that you are under FDA jurisdiction and you have to behave accordingly, but we are not going to focus our compliance efforts on making sure you are doing that,” Dr. Marisa Cruz, chief medical officer for Limbix, said on the panel.
"So those kinds of requirements that you would be subject to, like registration and enlisting, like the quality management system, like a certain level of clinical evidence of generation to support a finding of relative safety and efficacy, is not necessarily something the FDA is going to look at unless there is a cause for concern.”
Although companies seeking the discretion designation may not have to demonstrate the same research as those undergoing the 510(k) process, Cruz noted that the agency does prioritize safety and efficacy in all products.
Acacia Parks, chief science officer of Happify Health, noted that, in her experience, working on the 510(k) process the FDA has been keen on using shams apps in research and efficacy studies.
“[The FDA] have voiced somewhat of a sliding scale of evidence generation based on risk,” Cruz said. “By the very nature of enforcement discretion being a low-risk noninvasive device, I think that they are willing to allow for comparative groups that may not be up to the same rigor as something that might be a prescription therapeutic.”
While shams may not be required for every product or designation, in general, the panelists agreed that sham studies were helpful to prove efficacy, not just to the FDA, but also to payers, providers and other stakeholders.
“I think [FDA] are just really concerned about the placebo effect and just the impact of having something that you pay attention to that is an engaging app, and we make these apps with all these engaging elements on purpose to draw people in, and is that driving your effect?” Parks asked.
Dr. Ravi Nunna, medial advisor for CureApp, noted that whether a company is planning to do a 510(k) clearance, de novo, or enforcement discretion, the more thorough the research, the better the company’s prospects are of finding users.
His company got its start in Japan, and worked with the country’s version of the FDA, called the PMDA.
“With discussions with some of the regulatory individuals in Japan, we realized, being the first DTx company, we would essentially need a very high level of evidence. So we endeavored to make a highly, highly rigorous study,” Nunna said.
During the regulatory process, the company conducted research that included 600 participants, two groups, and a sham application. The research was published in Nature Digital Medicine Journal.
“I think what we have found is, the stronger the clinical evidence there is, there is really so much synergy beyond ... from a regulatory perspective,” Nunna said. “We have so many more physicians interested in our product, we have so many more payers interested in our product, we get contacted all the time, because we have [a] legitimate academic publication and people can see we have a legitimate trial. It’s really, really helped our product expand tenfold. Whereas, if we did a smaller trial and went for whatever the minimal necessary component was, I don’t think we would be laying the foundation for a springboard effect.”