Balancing Benefit and Risk in Medical Device RegulationFebruary 20, 2019
Medical devices provide important diagnostic and treatment benefits to patients. Every day, thousands of patients are the beneficiaries of amazing technology that often did not exist even 10 years ago, and certainly not 50 years ago.
At the same time, these medical devices can pose risks to patients that sometimes lead to injuries even if, on net, the benefit of the device exceeds the risk. Additionally, some devices that seemed at first to provide important benefits may not be as effective as thought or may turn out to be unsafe. Unlike drugs, devices are typically improved iteratively, making them safer and/or more effective over time. That still means patients early on may be more prone to suffer adverse events.
The Food and Drug Administration (FDA) is one of the guard rails in place to ensure that devices reach the market with a reasonable assurance of safety and effectiveness. The agency is also entrusted with monitoring post‑market events and manufacturing to ensure that devices that prove to be excessively risky are either made safer or removed from the market. There is no known way to prevent all adverse events.
These thoughts are prompted by an unfortunate op-ed in the Washington Post with this headline: “The FDA is still letting doctors implant untested devices into our bodies.” The headline is literally false, because FDA requires testing on all implanted devices. All implanted devices are subject to bench and animal testing, and most implanted devices are subject to clinical testing as well.
Still, the authors object to the fact that FDA permits marketing of a certain types of implanted devices without clinical testing; they think this rule should be changed to require all implanted devices to be clinically tested. The authors blame “[t]wo key loopholes” they say must be closed to ensure proper clinical testing. One is the 510(k) clearance pathway and the other is the filing of a “supplement” to obtain permission to market modified versions of devices that already have premarket application (PMA) approval. The authors charge that a combination of corruption, regulatory capture and/or obtuseness has prevented FDA from closing these purported loopholes.
The authors are not correct to describe either the 510(k) pathway or PMA supplement pathway as “loopholes.” A “loophole” is “an ambiguity or omission in the text through which the intent of a statute, contract, or obligation may be evaded.” The 510(k) and PMA supplements pathways, to the contrary, are fundamental elements of device regulation. They are centrally enshrined by Congress in the Federal Food, Drug, and Cosmetic Act (FDCA) (see, for example, sections 513 and 515). FDA could not unilaterally close these purported loopholes even if it wanted to do so. An act of Congress would be necessary. The authors do not acknowledge this fact.
Perhaps even more importantly, these so‑called loopholes do not actually exist. The authors seem to think that FDA lacks statutory authority to require clinical data to support either a 510(k) clearance or a PMA supplement approval. In fact, FDA does have such statutory authority, and there are many instances in which it has required clinical data for 510(k) submissions and for PMA supplement filings.
FDA does not require clinical testing for every device type and testing requirements vary widely by device type. But this variation flows from the statutory scheme directing FDA to regulate the heterogenous universe of devices by placing them in Class I, II, or III by level of risk (Class III being the highest level). Even in the realm of implanted devices, there are variations. Some implanted devices need to be clinically tested to provide reasonable assurance of safety and effectiveness. Others do not. FDA has authority to tailor testing requirements to the risk posed by a device type, including implanted devices.
The op‑ed seems to argue that all implanted devices should be subject to the most rigorous level of clinical testing (including for all modifications). It would be not be intelligent, however, to force all implanted devices into a procrustean bed of clinical testing regardless of circumstance. By imposing unnecessary clinical testing requirements in some cases, the authors’ proposal would make some beneficial devices less accessible to patients without good reason. As a result, patients would be harmed. The authors tout the benefits of their proposals but fail to address the possible costs.
It is true, as the op‑ed complains, that few devices are clinically tested in the manner of drug testing. But that is because devices are not drugs. For almost half a century, Congress has repeatedly directed FDA to regulate devices differently than drugs. The authors seem to blame FDA for following the congressional directives. They also seem to imply that FDA should start requiring that implanted devices undergo the same type of clinical testing as drugs. This approach would have to be authorized by Congress. It would also be a terrible idea, because of the profound differences between devices and drugs that have led to different testing requirements in the first place.
Unfortunately, the authors of the op‑ed seem to devote most of their energy to sweeping ad hominem attacks on the honesty and integrity of FDA officials. But FDA officials cannot simply refuse to follow the law or rewrite it as they please. That seems to be what the authors of the op-ed are demanding, because they aim fire at FDA officials and never acknowledge the role of Congress in establishing the current system. Note also that under the current regulatory framework FDA requires many (if not most) implanted devices to be clinically tested. Why would that be so if FDA officials were in the pocket of industry, as the authors allege?
No one contends that the regulatory system for devices operates perfectly. For instance, even FDA would probably agree that surgical mesh intended for transvaginal pelvic organ prolapse repair should have been placed in Class III in 2002 and subjected to intensive clinical study prior to general marketing. But FDA’s errors can also run the other way. If they unduly delay or block beneficial devices, that harms patients as well. These errors are particularly insidious, because the devices kept from the market are invisible. Patients who die from a missed diagnosis or lack of a cure will never know it if the technology that might have saved them is stymied by the regulatory process. Therefore, it is incumbent upon FDA to strike a proper balance that minimizes the risks of new technology but does not unduly delay the benefits. That is easier said than done! It is especially easy to criticize FDA’s decisions with the benefit of 20/20 hindsight.
All would agree that it is wise to continue to scrutinize both the regulatory framework and FDA’s decision making. Both should be the subject of public discussion and proposals for improvement. But such proposals must be rooted in a good understanding of how the system actually operates. Such proposals also need to present a fair account of the benefits and costs of proposed reforms. And, very importantly, mud‑slinging personal attacks have no place in the public discourse about FDA; it is not how we will make progress.