This is all very important context to understand the FDA’s new Data Modernization Action Plan (DMAP), which is a framework for how it will handle, analyze, and respond to data which it receives.
Janet Woodcock (acting Commissioner) and Amy Abernethy (acting Chief Information Officer) wrote, "Even small advances in our ability to gain useful insights from data can represent significant opportunities." I appreciate this perspective, because it suggests a deep truth, which is that there are larger unknown opportunities hidden in plain sight within our industry and its data, and improving how our data are collected and managed can elucidate some of these as-yet unknown mysteries.
Ushering in any new framework within an international regulatory body always sets the industry alight with new activity, and the downstream changes as a result of DMAP will be no different. In addition, here are some other insights likely to occur over the next year as we move from ‘the new normal’ to ‘the pandemic that once was.’
COVID-19 will stay at the top of the pyramid
Therapies, preventives, and testing regimes for COVID-19 and its newer viral variants will continue to take the brunt of FDA’s resources throughout 2021, as the industry is still surging to fill niche and emergent needs within this space. Jeff Shuren, head of FDA’s CDRH division, sees 2021 as a year to ‘reset’ by looking at COVID-19 technologies, but also approaches that are not coronavirus-related.
In my estimation, the idea of tracking epidemiology of viral variants, such as B.1.1.7, P.1, P.2, B.1351 (or 20H/501Y.V2), etc., will be ongoing and will lead to some competitors working in the booster vaccine space for annual or semi-annual variant-boosters.
As Heraclitus of Ephesus said about 2,500 years ago, “Change is the only constant.” And he’s still right.* We’ve seen enforcement actions ebb and flow across the industry over the past many decades. In the last 5 years in medical devices, for example, enforcement actions vis-à-vis FDA Warning Letters dropped by almost 90%. In 2020, FDA put a halt to onsite inspections in a statement dated 18Mar20 titled, “FDA Focuses on Safety of Regulated Products While Scaling Back Domestic Inspections.” It was determined to handle inspections on a case-by-case basis, and while some domestic facility inspections resumed by the end of last summer, there is still a significant backlog to winnow down in terms of site visits and ensuing enforcement actions.
This will have a variable approach applied across pharma and med devices, because for example, IMDRF's Medical Device Single Audit Program (MDSAP) is now in-process, and the FDA is a participating member. This would allow the FDA to satisfy their requirement for a facility audit via another MDSAP-recognized auditing organization having done so as a proxy. The FDA has said it will accept MDSAP audit reports as a "substitute" for routine agency inspections.
Tl;dr: There will be more inspections ramping up throughout 2021 and beyond, wherein the regulatory pendulum is likely to swing to the other direction from where it has been more recently.
Digital Health Center of Excellence
Intended to provide "efficient and least burdensome oversight" while meeting "standards for safe and effective products," the FDA is into Year One of this nascent initiative.
Though there has been a clinical decision support (CDS) software guidance in draft since September 2019, there hasn’t yet been a finalized version. This, along with the FDA’s approaches to Software as a Medical Device (SaMD), are likely to be finalized relatively soon.
Much ado about algorithmic opacity
As you can imagine, this is a difficult set of documents to draft and refine, which may not actually be evident at the outset. These guidances will usher in the future of when more clinical decisions are offloaded to artificial intelligence and machine learning (AI/ML) instead of clinicians, and also how AI/ML algorithms will make decisions as part of medical devices as well. How these are initially written will help define what the future will look like, so they are not guidance documents to take lightly.
There is still the looming reality that machine learning and deep learning (DL) algorithms are supposed to—by their very design and functionality—continually evolve. This very black-box nature of how they do what they do is what seems to, in my estimation, put many of them in violation of 2 specific articles of GDPR, the EU’s data protection regulation. The FDA has so far only approved medical devices that have used locked algorithms that cannot change.
In evaluating AI/ML/DL algorithms in the context of clinical studies, my compliance concerns are what led me, along with some others about 5 years ago, to coin the term ‘algorithmic opacity,’ for it is not easy—and indeed, sometimes not possible—to divine what the algorithm is actually doing to come up with the decisions that it does (much more to come on this aspect). We’re on a collision course with inevitability here in our data-rich healthcare future.
The future of the industry, and ‘futuristics’
Looking ahead, it’s important to note that life will begin to reopen globally, even if some days it doesn’t seem like it. Some regions (e.g., Australia) don’t even really have an endemic COVID-19 situation anymore, which will change once travel re-opens in earnest—and ‘temperature checks’ won’t help and have never helped in this regard. But we’ve all gotten better at adapting to changing environments.
At any rate, those things which could occur—or may not occur—in the industry are ‘futuristics’: All the set of variables which conspire constructively and destructively to eliminate some possible future timelines, and augment others.
As we see how the public behaves on a population scale with vaccination administration and personal freedoms, we’ll garner new opportunities, and others will be closed off from us. Social media in this regard has proved to continue to be our worst enemy. By giving a voice to nonscientific and fraudulent claims by laypeople, conspiratorial threads can be woven and build up momentum to virality. What those who develop these narratives don’t understand is that, on a grand scale, real psychological and physical harms come from these cynical contrarian and divisive viewpoints, and the confusion sown doesn’t actually help anyone—except the ever-increasing ticker of ‘Likes’ from their posts. #Fact.
*End note: Most of the reason that Heraclitus is still right is because of the Second Law of Thermodynamics, the concept of entropy, and nature trying to reach equilibrium. Because entropy tends to increase to a maximum, this brings with it change.
Contributing Editor, International COVID-19 Advisor, Healthcare Futurist
Having spent the past 14 months performing state and national epidemiology of the pandemic, working on vaccine candidates, and developing public health policies, Ben is optimistically looking-ahead to a more freedom-rich future. He has been writing the Clinically Speaking column for seven years; this is about 60% of the orbital period of the planet Jupiter (12 years). In some circles, there is a notion that all the cells in your body are replaced every 7 years. This, however, is scientific rubbish. Cell turnover is continuous and highly variable. Stay safe out there, get vaccinated.