DCRI’s Chief Science & Digital Officer Eric Perakslis, PhD unpacks how recently released FDA guidelines will impact data collection and digital health tools immediately and into the future.
The FDA’s recently released guidance on Digital Health Technologies for Remote Data Acquisition in Clinical Investigations represents a significant step forward in the exploration and application of digital health tools. As someone who works with technology innovators, drug and device clinical trial sponsors, and the medical innovation investment community, I’m convinced that this guidance has arrived at an ideal time. Not only will it lay a strong foundation for future policy, but it will also be immediately useful across the current stakeholder ecosystem.
Remote data acquisition refers to data collected outside of healthcare delivery facilities, including at home. Digital health technologies that enable such data acquisition offer both convenience and real-time access to patient data. In addition, because they capture information in “real-world” situations, they also enable observations that would be impossible to capture in traditional clinical settings.
However, successful remote data acquisition depends upon users having access to the internet and being conversant and comfortable with the technologies used. It also often involves methods that are classified by ethicists and privacy advocates as surveillance technologies requiring careful use and strong ethical oversight.
Despite these obvious challenges, this well-constructed guidance for the most part manages to effectively thread the needle. Importantly, it does not seek to solve every riddle, but focuses on remote data acquisition within clinical studies. This is a singularly important denominator, one that spans remote sensors, digital therapeutics, products regulated under the Software as a Medical Device schema, and patient-reported outcome instruments, to name a few.
Below, I list the ways I believe this guidance will be useful, both in the near term and further down the road.
Five Ways the Guidance Will Be Immediately Useful
This new guidance will help demystify what is currently a complex and confusing regulatory landscape through the following attributes:
- It busts myths. I frequently hear debate about whether all devices in clinical trials must be FDA approved for any specific use. Some companies even have policies that prohibit the consideration of devices that aren’t fully approved or cleared for a specific use. By providing essential specifics, the guidance will eliminate confusion and ambiguity and lead to better decision-making among clinical programs.
- It covers specific and essential elements in a detailed fashion. For example, the pros and cons of using patients’ own cell phones for clinical trial data collection are immensely complex. Making decisions based solely upon precedent from a chosen predicate device can be hazardous, because each patient population, indication, and context of use will be different. The FDA’s guidance unpacks these scenarios in detail, allowing methods to be appropriately matched to each trial’s particular needs.
- It spends a lot of time on informed consent. A major hazard within the current digital health ecosystem is the gap between recreational toolsets such as wellness apps, and medical-grade products that include regulated diagnostics and therapies. This gap is frequently exploited, and digital product research can be conducted without the types of protections afforded within regulated clinical trials. Further, big data practices, that are essential infrastructure for remote data utility, may be incompatible with current Informed consent approaches. A solid focus on informed consent is truly essential.
- It creates a space for discussing potential harms. Our collective understanding of the potential harms associated with digital health tools is still nascent. Clinicians receive training on the adverse effects of drugs and medical devices, not on cybersecurity, identity theft, or the psychosocial impacts of social media. It’s heartening to see the beginnings of much-needed work taking place in this domain.
- It will enable innovation. I frequently see sponsors struggling with the risk calculus of including novel digital tools and measures within their high-priority programs. It can be very NIMBY-like: sponsors genuinely want to innovate and drive progress, but the penalties for failure or disruption of any single high-priority program can seem prohibitive. This guidance provides a framework that will encourage smart risk-taking when the potential benefits for patients justify it.
Five Opportunities to Strengthen the Next Steps
In order to make the most of the potential represented in this guidance, I believe that the FDA should:
- Proactively address concerns within the FDA’s Premarket Notification (510k) process. Recent articles have highlighted gaps in the 510k process that could pose hazards to patients. As we explore and attempt to leverage digital health tools for patient benefit, we must not oversimplify or omit known risk factors by trying to shoehorn digital health regulation into existing regulatory precedents. Instead, the opportunities and risks of digital health must be viewed and assessed uniquely.
- Address algorithms more thoroughly. The difference between many digital health tools within single contexts of use is the algorithm, not the hardware – a point the new guidance correctly notes. That said, algorithm ‘quality’ goes beyond GxP software and thus must address resilience, bias, and reproducibility. In fact, any new independent context of use may require reevaluating the underlying algorithms.
- Mandate full Software Bill of Materials for registry clinical studies. The FDA’s (and, more recently, the Biden administration’s) efforts to require a software bill of materials (an ingredient list of all software components used within a system) must be realized. As the recent Log4J cyber vulnerability has once again shown us, systems that operate with unknown components simply cannot be made secure.
- Build and address the corpus of digital harms. Privacy remains an essential element of patient self-determination from the standpoint of medical ethics. However, many of the concepts that underlie these earlier ethical frameworks need to be updated, and enacting explicit protections from the harms that can occur when privacy is lost may assume greater importance. Just as adverse drug events can be studied, tracked, and prevented, so can cyber harms. Cybersecurity already relies on a published taxonomy of harms that occur to victims of cybercrimes; this example should be studied closely and considered. Further, the growing digital divide has demonstrated that underprivileged populations can quickly be left behind. One example of this is digital health toolsets that are only available (or only available and secure) on Apple IOS devices, given that the median income of IOS users is 40% higher than that of Android OS users. These examples represent just a few dimensions of the potential harms of digital health tools and interventions – far more work is needed.
- Accelerate the move from non-binding to binding. I hope that the positive reception of this guidance will encourage rapid formalization and an expeditious move toward a binding guidance. Digital innovation remains stymied by ambiguous regulatory concerns within Big Pharma and other innovation stakeholders. A binding guidance will encourage and enable more prolific innovation.
Where To Next?
When dissimilar exceeds similar.
While the Digital Health Center of Excellence within the FDA’s Center for Devices and Radiological Health (CDRH) is off to a great start, it remains to be seen whether it represents a sufficiently strong approach to supporting innovation and protecting consumers from new and rapidly evolving risks. Just as when the Center for Biologics Evaluation and Research (CBER) was formed because the differences between small-molecule drugs and biologics had to be addressed, we are now approaching a similar tipping point for digital health.
As the opportunities for digital health multiply, so does the complexity. It’s quite likely that we’ll soon see novel digital health solutions that defy association with predicates in ways that require reinvention to understand and regulate. Digital health mixes healthcare with the internet and blurs the lines between medical risk and cyber risk, the clinical environment and the home, engineering and medicine. It also requires a patient trust and participation at a time when basic confidence in medicine and science have been deeply damaged by a relentless pandemic and deep political polarization.
Ultimately, this latest guidance represents a strong step forward. Let’s appreciate the progress being made by our FDA colleagues and stand ready to help, support, and deliver wherever we can.