Matthew Herper, Stat News, January 10, 2023
The Food and Drug Administration’s approach to evaluating some new medical devices might actually increase the risk that those devices will later be recalled over safety concerns.
For some medical devices, the FDA requires data from studies in people to show that the device, including things like implantable heart defibrillators or stents, is safe and effective. That process is roughly equivalent to how regulators review new medicines. But the FDA also approves devices including artificial joints, infusion pumps, and scalpels using a process known as the 510(k) pathway, which simply lets manufacturers show that the new devices are “substantially equivalent” to products already on the market.
Two new papers published Tuesday in JAMA look at a specific wrinkle of the 510(k) problem: What happens if a new medical device is approved based on a device that has been recalled because of a safety problem?
The first paper, by a team that included Kushal Kadakia of Harvard Medical School and Harlan Krumholz of Yale University, looked at cases in which a device underwent what the FDA calls a Class I recall – meaning the device was temporarily removed from the market because of, in the FDA’s words, “a reasonable probability that the use of or exposure to a violative product will cause serious adverse health consequences or death.” Specifically, the researchers looked at cases in which a newer device was cleared using a comparison to the recalled device. They found that devices that were approved based on comparison to a device that had previously undergone a Class I recall were about six times more likely to themselves be recalled than similar devices.
[….]
The agency also argued the number of recalls — “less than 150 over 5 years” — should be “considered in the context” of the more than 120,000 devices that have gone through the pathway.
But Diana Zuckerman, president of the National Center for Health Research, said the data point to “an ongoing problem, and one that is relatively easy to identify.” “That predictive relationship is pretty impressive,” added Zuckerman, who is a critic of approving newer devices based on recalled predicates.
A second group led by Alexander Everhart of the Harvard-MIT Center for Regulatory Science used natural language processing to identify records on 35,176 medical devices, of which 4,007 were recalled. Again the devices were far more likely to be recalled if they were approved based on similar data to a device that had also been recalled. Devices were also more likely to be pulled from the market if they were approved based on comparisons to multiple devices, or to newer devices for which less data were available.
[….]
Kadakia pointed out that there have been efforts to fix the problem. The FDA identified the problem in 201010, and Edward Markey, a former Massachusetts representative and now a senator, tried to pass a 2012 bill to reform the process after the high-profile recalls of transvaginal meshes that harmed thousands of women. But in the end the law was not passed.
To read the entire article, click here.