番茄社区app

In this exclusive video, cardiologist Harlan Krumholz, MD, SM, director of the Center for Outcomes Research and Evaluation at Yale University and Yale New Haven Hospital in Connecticut, discusses research on recalled medical devices, including a study by his group that found that devices subject to Class I recalls were often used as predicates for subsequent devices authorized by the FDA.

Krumholz is also the Harold H. Hines, Jr. Professor of Medicine and a professor in the Institution for Social and Policy Studies at Yale School of Medicine.

The following is a transcript of his remarks:

I'm here to talk about two articles that were published in the January 10th issue of JAMA: One of them came from , one of them from , on regulatory science and devices.

Now that may seem like an esoteric topic, and one that may not interest you immediately if you're a clinician and taking care of patients. But we need to know that medical devices have really revolutionized the way in which the medical practice has occurred over the last 20, 30, 40, 50 years. Almost every single procedure that's being done now is using a medical device; sensors are using medical devices. Actually, the FDA even considers tongue depressors as a medical device. But here the focus is on medical devices that are very important and can be lifesaving.

Now most medical devices are authorized under FDA's 510(k) pathway. This requires manufacturers to demonstrate that the new device that they want to sell is substantially equivalent to a previously authorized device. That is, something else that was on the market. The idea here is if manufacturers make incremental changes in a device, they shouldn't have to start over. There are lots of tweaks that may occur in the course of developing devices, and instead of having them start over and do a whole bunch of clinical studies -- if it's substantially equivalent and it's being used for the same purpose -- then FDA can authorize it and allow it to be sold. So many devices go through this pathway.

These two studies take a look at particularly the issues regarding recalls and recalled devices. Now recall has various classes, and Class I is the most severe of the recall levels. That's [used] for withdrawing a device that could be life-threatening or cause really substantial harm.

In the , we did a cross-sectional study of all the 510(k) devices that had been subject to this most serious of recalls, this Class I recall, from January 2017 through December 2021. What we discovered was that many of these devices that were subject to a Class I recall had been authorized based on their substantial equivalence to a prior device that was also subject to a Class I recall. In addition, we found that these devices that were subject to a Class I recall were sometimes used by subsequent devices as the basis for their authorization.

So, it's sort of a flaw or a problem that was seen in the device to which it's said to be substantially equivalent was propagated down the line through generations. If a certain device is authorized based on a device that had a Class I recall, its risk of having a Class I recall was much higher -- six times as high -- than average.

Now, this isn't an FDA issue. This is, in fact, in the law. The law says that if the FDA withdraws a device because of a Class I recall, they cannot use that device as a predicate for a subsequent device. But if it's voluntarily withdrawn by the manufacturer, it can be. So, this represents a loophole in the law.

The , similarly, began to look cross-sectionally at devices that were approved by the Food and Drug Administration [FDA] and identified that of all the ones that were approved, a substantial proportion -- again, in the grand scheme of things, most devices don't have Class I recalls. When they went and looked back and asked 'What predicted the Class I recall?' they found that, in fact, having a prior device upon which you were based for the authorization to be sold, if that had a Class I recall, your risk of having a Class I recall also was a predictor; much higher.

Now, you may ask, what kind of devices are we talking about? Well, for example, CareFusion's Alaris Infusion Pump, this device was used to control the delivery of IV fluids and medications for patients in the hospital. This device is subject to multiple Class I recalls with hundreds of thousands of units affected.

The device cites three predicates, three devices upon which it said it was substantially equivalent and then therefore didn't need to do studies in patients because it was already similar to other devices. And in fact, two devices also had been subject to recall. So the FDA was put in a position where it needed to use those as the predicate because it was being sold and it couldn't preclude it. It's a loophole in the law that needs to be addressed.

One of the things that we're recommending as a result of our study, and now even reinforced by the other study, is that Congress needs to provide the FDA with the ability to reject devices subject to Class I recalls from being used as the basis to authorize future devices. That is, if you've had a Class I recall, then you shouldn't be used as a predicate, as a basis for future devices to be used.

We're also adding that we think there ought to be greater transparency about the lineage of these devices. Look, as clinicians, we're often using devices. One of the things we're advocating for is that we need, as clinicians, to have more knowledge, deeper knowledge about these devices and the evidence upon which they're based. We're used to that with drugs. In fact, we all demand to know exactly what each drug is based on and why it's beneficial, what the quality of the studies were.

For devices, that can often be hard to figure out, especially if they're through this 510(k) pathway, exactly what was the basis because they're said to be authorized based on similarities to another device. Then you have to go all the way back to figure out where the evidence was that supported the safety and effectiveness of that original device, if it even existed.

I think that as devices become even more consequential in the future, and as data science makes them more complicated in addition to their consequence, we're going to be in a position where it's going to be more and more important for us to understand the evidence that stands behind each of these devices.

Comment