So I had a bunch of things to say this week. I was actually planning a gag where I was going to shut down part of the site for “Internet Maintenance Day“. Then stuff happened that I felt I wanted to talk about more urgently. Than more stuff happened, and I had to bump back the queue again. Specifically, with regards to that last one, John Oliver released a new episode that I have to talk about.
If you don’t care to watch, the central thesis of the episode is, in a nutshell, that our medical device regulation system sucks and needs to be more robust. And he’s not wrong. The FDA is overstretched, underfunded, strung up by political diktats written by lobbyists, and above all, beset by brain drain caused by decades of bad faith and political badmouthing. The pharmaceutical and biotech lobby has an outsized influence on the legislation (as well as executive orders and departmental regulations) that are supposed to govern them.
But, and I’m going to repeat this point, the system isn’t broken. Don’t get me wrong, it’s hardly functional either, but these problems are far more often ones of execution than of structure.
Let’s take the 510(k) exemption that is so maligned in the episode. The way it’s presented makes it seem like such a bad idea, that surely this loophole must be closed. And I’ll agree that the way it’s being exploited is patently unsafe, and needs to be stemmed. But the measure makes sense under far narrower circumstances. To use an example from real life, take insulin pumps. Suppose a pump manufacturing company realizes that it’s replacing a high number of devices because of cracked screens and cases occurring in everyday use. It takes the issue to its engineers, who spend a few days in autocad making a new chassis with reinforced corners and a better screen that’s harder to crack. The guts of the pump, the parts that deliver insulin and treat patients, are unchanged. From a technical perspective, this is the equivalent of switching phone cases.
Now, what kind of vetting process should this device, which is functionally identical to the previous iteration aside from an improved casing, have to go through before the improved model can be shipped out to replace the current flawed devices? Surely it would be enough just to show that the improvements are just cosmetic, perhaps some documentation about the new case and the materials. This is the kind of scenario where a 510(k) style fast track would be good for everyone. It saves time and taxpayer money for regulators, it gets the company’s product out sooner, and consumers get a sturdier, better device sooner. This is why having that path is a good idea.
Not that the FDA is likely to apply section 510(k) in this scenario. Insulin pumps tick all the boxes to make them some of the most regulated devices in existence, even more so than most surgical implants. Any upgrade to insulin pumps, no matter how inconsequential, or how urgently needed by patients, is subject to months of testing, clinical trials, reviews, and paperwork. The FDA can, and regularly does, send applications back for further testing, because they haven’t proven beyond a shadow of a doubt that there is no risk. As a result, improvement to crucial life support devices are artificially slowed by regulations and the market’s reaction to regulations.
Here’s the other thing to remember about medical devices: for as much as we hear about the costs of prematurely releasing devices, there is also a cost to delaying them. And frustratingly, the ones which often have the greatest cost to delaying- devices like insulin pumps, monitors, and other life support -tend to be subject to the greatest scrutiny, and hence the longest delays. For while the FDA examines numbers and research data, real patients continue to suffer and die for want of better care. We might prevent harm by slowing down the rollout of new technologies, but we must acknowledge that we are also consigning people to preventable harm by denying them newer devices. Some argue that this is morally preferable. I staunchly disagree. More than just trying to protect people from themselves, we are denying desperate people the hope of a better life. We are stifling innovation and autonomy for the illusion of security. This isn’t only unhelpful, and counterproductive, but I would argue it’s downright un-American.
Rest assured I’m not about to go and join the ranks of the anarchists in calling for the abolition of regulatory agencies. The FDA is slow, inefficient, and in places corrupt, but this is as much as anything due to cuts in funding, usually made by those who seek to streamline innovation, which have limited its ability to fulfill its mandate as well as ironically made processing applications slower. A lack of respect for the agency, its job, and the rules it follows, have inspired unscrupulous companies to bend rules to their breaking point, and commit gross violations of scientific and ethical standards in pursuit of profit. Because of the aforementioned lack of resources, and a political climate actively hostile to regulatory action, the FDA and the agencies responsible for enforcement have been left largely unable to follow their own rules.
Cutting regulations is not the answer. Improving and reforming the FDA is not a bad idea, but the measures supported by (and implied to be supported by) John Oliver are more likely to delay progress for those who need it than solve the issues at hand. A half-informed politically led moral panic will only lead to bad regulations, which aside from collateral damage, are likely to be gutted at the next changing of the guard, putting us back in the same place. I like to use the phrase “attacking a fly with a sledgehammer”, but I think this is more a case of “attacking a fly with a rapier”, in that it will cause massive collateral damage and probably still miss the fly in the end.
So, how do we do it right? Well, first of all, better funding for the FDA, with an eye towards attracting more, better candidates to work as regulators. If done right, this will make the review process not only more robust, but more efficient, with shorter turnaround time for devices. It might also be a good idea to look into reducing or even abolishing some application fees, especially for those applications which follow high standards for clinical trials, and have the paper trail to prove ethical standards. At present, application fees are kept high as a means to bring in revenue and make up for budget cuts to the agency. Although this arguably does good by putting the cost of regulating on the industry, and hopefully incentivizing quality applications, it constrains the resources available to investigating applications, and gives applying companies undue influence over application studies.
Second, we need to discard this silly notion of a regulatory freeze. Regardless of how one feels about regulations, I would hope that we all agree that they should at least be clear and up to date in order to deal with modern realities. And this means more regulations amending and clarifying old ones, and dealing with new realities as they crop up. There should also be greater emphasis on enforcement, particularly during the early application process. The penalties for submissions intentionally misclassifying devices needs to be high enough to act as a deterrent. Exceptions like section 510(k) need to be kept as exceptions, for special extenuating circumstances, rather than remaining open loopholes. And violating research standards to produce intentionally misleading data needs to be treated far more seriously, perhaps with criminal penalties. This requires not only regulatory and enforcement power, which already exist on the books, but the political will to see abusers held to account.
Third, there needs to be a much greater emphasis on post-market surveillance; that is, continued testing, auditing, and review of products after they reach consumers. This seems obvious, and from conversations with the uninitiated, I suspect it’s where most people believe the FDA spends most of its effort. But the way the regulations are written, and certainly how they’re enforced in practice, post-market surveillance is almost an afterthought. Most of it is handled by the manufacturers themselves, who have an alarming amount of latitude in their reporting. I would submit that it is this, the current lack of post-market surveillance, rather than administrative classifications, that is the gaping hole in our medical regulatory system.
This is also a much harder sell, politically. Industry hates it, because robust surveillance often prevents them from getting away with cutting manufacturing costs after approval, when reducing costs would lead to reduced product quality, and it means they have to keep on extra QA staff for as long as they remain in business. It’s also expensive for industry because of how the current setup puts most of the cost on manufacturers. Plenty of politicians also hate post market surveillance, since it is a role that is ideally redundant when everyone does their jobs. When something goes wrong, we say that it shouldn’t have been sold in the first place, and when nothing goes wrong, why would we pay people to keep running tests?
Incidentally, from what I have been led to understand, this is a major difference between US and EU regulatory processes. Drugs and devices tend to come out in the EU commercially before the US, because the US puts all of its eggs in the basket of premarket approval (and also underfunds that process), while the EU will approve innovations that are “good enough” with the understanding that if problems show up down the line, the system will react and swoop in, and those at fault will be held accountable. As a result, European consumers have safe and legal access to technologies still restricted as experimental in the US, while also enjoying the confidence that abusers will be prosecuted. Most of those new devices are also paid for by the government healthcare system. Just saying.