As FDA has noted, artificial intelligence (AI) and machine learning (ML) technologies have the potential to transform healthcare. These technologies could be used at the agency, by other agencies involved in healthcare regulation and finance, by private participants in the healthcare delivery system, and by medical device manufacturers. They can also be embedded in conventional medical devices or, indeed, serve as standalone medical devices. For several years now, FDA has been exploring how its existing medical device framework applies, or should be adapted to apply to, software “as” a medical device as well as software “in” a medical device. In this post, I review four recent law review articles exploring legal and policy issues presented by the emergence of artificial intelligence and machine learning in medical devices and the healthcare setting more generally.Continue reading “Artificial Intelligence & Machine Learning: Recent Scholarship”
The 2020 annual conference of the Petrie-Flom Center for Health Law Policy, Biotechnology, and Bioethics (at Harvard Law School) will focus on the future of medical device regulation. If you’re interested in participating, abstracts are due on October 14, final papers are due March 27, and the conference is May 8. (Here’s a link to the call for papers.) I’m working on an empirical project relating to device premarket approval and patent term restoration, but sadly the dataset won’t be ready in time. Meanwhile, here’s a run-down of some interesting legal scholarship on medical device regulation in the last year and a half.Continue reading “SSRN Reading List . . . or Device Regulation: What Role for Tort Law?”
Continuing with FDA-related scholarship posted on SSRN in February and March: Professor Noah explores the possibility of “reverse switches” from over-the-counter (OTC) status back to prescription status (or perhaps, he suggests “behind the counter” status) in response to new safety information. Professor Paradise explains the provisions of the 21st Century Cures Act that contemplate greater patient involvement in medical product development, and Professors Evans and Ossorio evaluate the 21st Century Cures Act exclusion of clinical decision software from the agency’s medical device authority.
Very few folks posted papers in December and January relating to FDA law, but hopefully the February law journal submission cycle will yield a rich crop. Here are two essays and one article of potential interest – one each on medical devices, biological products, and food. (I am excluding my own paper.)
Sarah Duranske, Reforming Regenerative Medicine Regulation
In this article forthcoming in the Georgia State Law Review, Duranske (currently a fellow at Stanford) (* edited to correct the spelling of her last name!) evaluates proposals for regulation of regenerative medicine. She has several interesting ideas tucked in here, any of which could have been the basis for an article in its own right. Section I contains a nice overview of the current regulatory paradigm for therapies that fall within the umbrella of “regenerative medicine” — including the recently enacted accelerated approval pathway for “regenerative medicine advanced therapies.” Section II responds to deregulation arguments, arguing that FDA regulation is necessary to protect patients and to ensure the development of meaningful data. Much of this retreads familiar ground, but the discussion of “Baptists and bootleggers” alliances with respect to regenerative medicine is very interesting. This phenomenon is pervasive in food and drug law and merits more discussion in scholarship. Section III is where the article gets interesting. Here, she considers proposals for “adaptive licensing” of regenerative products. The basic idea is that FDA would approve a product on the basis of less evidence, but would restrict access while the sponsor gathered more evidence from clinical use. She characterizes adaptive licensing as a type of adaptive management — a particular type of process that an agency might use to produce a regulatory outcome. She then assesses the suitability of regenerative medicine for adaptive licensing by running it through the various rationales in the administrative law literature for adaptive management at agencies. This leads her to the conclusion that the benefits of adaptive licensing do not outweigh its risks. Section IV contains a brief discussion of her proposals — for instance, shifting some regenerative therapies to the more loose regulatory paradigm governing human tissue and cell products. But I wanted to read much more about her ideas.
Jane R. Bambauer, Dr. Robot
In this essay published in the UC Davis Law Review, Professor Bambauer considers whether health and medical artificial intelligence (AI) should be regulated more like physicians or medical devices. When the application is a “knowledge” app rather than a “measurement” app, she argues, physicians are the better analogy. Some of the duties of a physician (such as the duty of competence and the duty of confidentiality) translate well, but she is more guarded about other rules (rules of informed consent, for instance, and the duty to disclose conflicts of interest).
Laurie Beyranevand, Regulating Inherently Subjective Food Labeling Claims
In this essay published in Environmental Law, Professor Beyranevand essentially argues that FDA should ban claims like “natural” and “healthy” in food labeling. The statute imposes clear rules governing specific types of claims, such as “health claims” and “nutrient content claims.” She is focused on claims that are not covered by these specific claims-authorizing provisions of the statute. And she argues that FDA should subject these claims to a standard of “significant scientific agreement” — that is, permitting them only if there is significant scientific agreement. The FDCA already uses this standard for health claims, which generally characterize a specific relationship between a food product and a health condition or disease. She also believes it would be virtually impossible to support a claim like “natural” and “healthy” under this standard. So, in essence, she is arguing for a ban. Finally, although the federal courts have concluded that the First Amendment requires FDA to consider disclaimers for health claims that lack significant scientific agreement, she contends that the First Amendment is no impediment to the proposal. I think the idea is that if a claim is inherently subjective, then a disclaimer isn’t going to clear things up (there’s x amount of data, but not y). It’s just going to confuse consumers more.
Here are three more new and noteworthy articles from October and November, including Professor Robin Feldman’s new empirical study of the pharmaceutical industry.
Cross-posted on Stanford’s Law and the Biosciences Blog
As I have previously written about here, in January FDA published a controversial revision to its regulations defining “intended use,” and then, in the wake of procedural and substantive objections to the revised definition, the agency delayed the effective date of the new rule until March 2018. These revisions are important because the “intended use” of a product is crucial for determining whether the product is a drug or device subject to FDA jurisdiction at all, and if so, whether the drug or device is in compliance with various FDA requirements. Accordingly, there is significant interest in the kinds of evidence that FDA considers relevant to determining a product’s intended use. The January revision to FDA’s regulations explained that that FDA would use a “totality of the evidence” approach to determining intended use, which would permit the agency to look to “any relevant source of evidence,” including, perhaps most controversially, a manufacturer’s knowledge about consumers’ and patients’ actual uses of the product. The procedural “logical outgrowth” arguments against this standard do not persuade me for the reasons I explained here. Likewise, I am not sure the substantive arguments against the revised regulations convince me.
On October 26, Senator Cruz introduced the “Reciprocity Ensures Streamlined Use of Lifesaving Treatments Act of 2017” (S. 2022), which is interesting from an FDA law perspective as well as an administrative law perspective. We have seen this proposal before — in 2015 (S. 2388, introduced by Senator Cruz) and in 2016 (H.R. 6241, introduced by Congressman DeSantis). Rachel Sachs wrote about it from a policy perspective in December 2015, and Zach Brennan offered more details in his own piece the same month. I am going to dig into the details a bit more than they did and explain why I call it the “Send All the FDA Employees Home Act of 2017.”
Data exclusivity for drugs and biological products gets all the attention. (In fact, recently I read a law review article asserting that medical devices are not entitled to any sort of exclusivity period after approval. But this is wrong!) It is apparently not as well known, but sponsors of premarket approval applications (PMAs) enjoy six years of data exclusivity. Folks interested in FDA and innovation policy should know about the device scheme because it has a unique history (with a novel and clever – though unworkable – approach in place for seven years) and because at a high level it is still analogous to drug and biologic exclusivity even though the regulatory paradigms are different.
Cross-Posted on Notice & Comment
In January FDA published a controversial revision to its regulations defining a product’s “intended use” that, among other things, has raised an interesting logical outgrowth question. “Intended use” is an important concept in FDA law because a product’s intended use—judged by the “objective intent of the persons legally responsible for the [product’s] labeling”—can be crucial to determining whether a product is a drug or device subject to FDA oversight at all, and whether an FDA-authorized drug or device is in compliance with FDA requirements. (Readers can find more about “intended use” generally, and the background behind the current controversy, here). Because “intended use” is so important in the FDA world, it should come as no surprise that stakeholders that disagree with the revised definition in the January final rule—which has yet to go into effect—have lodged both procedural and substantive arguments against the revision (see, e.g., here and here).
Cross-posted on Stanford’s Law and Biosciences Blog
A few news stories over the past week or so—one in the Wall Street Journal about “neurotech,” one in Geek Gadget about “neuroscience wearables,” one in the Washington Post about baby monitors for measuring an infant’s vital signs, and one in Gizmodo about “vaginal wellness products” marketed on Etsy—reminded me how much I enjoy questions of intended use. As I wrote last week, intended use is a critical concept in FDA law, in part because a product’s intended use is crucial to determining whether it meets the law’s definition of drug or device within the FDA’s jurisdiction. And, for whatever reason, I have an unabashed and—as far as I can tell—limitless love for thinking through questions about whether, and how, products fall with the definition of a drug or device.
As for the reported neurotech, neuro-wearable, baby monitor, and vaginal wellness products, it seems to me that many of these products may fall within the Federal Food, Drug, and Cosmetic Act’s (FDCA) definitions of drugs or devices. Why is that?