To repair a weak or failing heart, a patient may undergo a delicate procedure to help restore normal pumping action. Patients may receive an implantable cardioverter defibrillator or perhaps a ventricular assist device to mimic the role of a healthy heart muscle.
The hope is that implanting such a device will improve, even extend, a patient’s life. Although any surgical procedure comes with risks, most people likely assume that the device itself is safe.
But in the past few years, a string of high-profile recalls of medical devices has led experts to question just how safe high-risk medical devices really are.
“We are clearly witnessing an increase in medical device recalls. The data comes from the FDA’s own reports,” said George Ball, PhD, assistant professor of operations and decision technologies at the Kelley School of Business, Indiana University. Dr. Ball is a former medical device industry manager who now researches causes of medical device recalls.
In 2014, the FDA’s Center for Devices and Radiological Health (CDRH) published an in-depth analysis of medical device recalls over the past 10 years. What the FDA discovered suggests that there may be cause for concern: Between 2003 and 2012, the annual number of medical device recalls had increased by 97%. Recalls almost doubled over the 10-year study period, increasing from 604 to 1,190, while the total number of products recalled in that time increased from 1,044 to 2,475.
Breaking down the numbers, the FDA found that, although the lowest risk recalls, designated as class III, had declined during that period—from 137 in 2003 to 90 in 2012—the most severe class I recalls, which could cause serious harm or death, rose from seven to 57 in that time. Still, class I recalls only represented a small fraction of all recalls in any given year: 1.1% in 2003 and 4.8% in 2012. The majority of recalls were designated as moderately severe, class II, which means the device may result in temporary or medically reversible adverse health consequences. During the study period, class II recalls increased from 460 in 2003 (76% of all recalls) to 1,043 in 2012 (87.5% of all recalls).
The FDA concluded that the observed increase in medical device recalls is likely due to the fact that firms have become more vigilant in reporting adverse events. “Based on the data we have reviewed and our experience, we believe the increase is the result of increased awareness prompted by targeted interactions with industry and individual manufacturers,” the agency wrote.
But greater attentiveness and reporting is likely not the whole story. Research published this year indicates that, between 2004 and 2014, manufacturers delayed reporting more than 160,000 adverse drug events to the FDA, which is about 10% of all adverse event reports and includes more than 40,000 deaths (JAMA Intern Med 2015; published online July 27).
So what might account for this rise in recalls? Although the evidence is limited, some experts have voiced concern that the lack of premarket clinical testing may increase our uncertainty about device safety, which may in turn lead to recalls after devices reach the market.
“Although many devices are well designed, if a device has a safety issue, the more you study it in a premarket environment, the more likely you will be to uncover the issue,” said Joseph Ross, MD, MHS, associate professor of medicine at Yale School of Medicine, in New Haven, Conn., and assistant professor of public health at the Institute for Social and Policy Studies. “It’s a matter of probability.”
But the FDA, which is responsible for medical device approvals and clearances, can only do so much to ensure that medical devices are ready for prime time. Specifically, the agency can only ask manufacturers to take the “least burdensome” route—invest the minimal effort—to determine the safety and effectiveness of devices.
“The FDA is bound by the limits of its authority,” said Dr. Ross. “Its hands are tied in terms of the level of testing it can require.”
As a result, most medical devices do not undergo any safety testing before reaching the market (Milbank Q 2014;92:114-150). The majority of medical devices are cleared through the 510(k) pathway, which typically requires companies to show a device is “substantially equivalent” to an already approved device. Thus, 510(k)-cleared devices often bypass clinical testing, even when the predicate device was recalled or never assessed for safety and effectiveness,explained Rita Redberg, MD, MSc, a cardiologist in the Department of Medicine at the University of California, San Francisco.
Overall, less than 2% of approved medical devices go through the most rigorous FDA regulatory pathway, premarket approval (PMA), which requires companies to perform clinical tests. But even PMA-approved medical devices may not be examined rigorously enough. A report published in the Journal of the American Medical Association last August revealed that the evidence for evaluating the safety and effectiveness of high-risk medical devices is often insufficient (JAMA2015;314:604-612). Dr. Ross and his colleagues from Yale School of Medicine examined clinical studies of 28 high-risk devices that received PMA in 2010 and 2011. In that time, the authors found that 82 clinical studies were performed on the 28 devices prior to approval, but only 30 of these (approximately one study per device) were considered pivotal studies, designed to evaluate safety and effectiveness. Furthermore, almost half of the pivotal studies did not compare outcomes between the new device and other similar devices on the market.
Dr. Ross’ team also reported that after devices were approved, the FDA required manufacturers to conduct 33 postmarket studies (about one additional study per device), yet only six of those studies were completed in the three to five years after approval.
“Postmarket studies aren’t being completed as quickly as we’d expect, but it’s unclear whether they’re not completed within five years because there are delays in the FDA ordering the postmarket study, they are longer follow-up studies or there are delays in the design, initiation, recruitment and thus completion of studies whose results should have been finished up,” Dr. Ross said.
As a physician, Dr. Ross does not think the level of evidence is robust enough, but “this is probably the level of evidence that we as a society are satisfied with. I don’t think most people want longer or more studies.” The result is a “have your cake and eat it too problem,” Dr. Ross said. “We want things available as soon as possible, but we need to balance this speedier approval with reliable evidence.”
Inconsistencies in the quality of clinical trials and lack of testing altogether may increase the likelihood that unsafe or subpar devices are approved. In a 2011 study, Diana Zuckerman, MD, and her colleagues evaluated recalls of high-risk medical devices between 2005 and 2009 (Arch Intern Med 2011;171:1006-1011). In that period, the FDA recalled 113 high-risk devices, 80 (71%) of which had been cleared through 510(k) and 21 (19%) of which had been approved via PMA. An additional eight (7%) were exempt from FDA review. According to the authors, 13 of the 80 510(k) recalled devices should have gone through PMA, given the severe risk they posed to patients if a malfunction occurred.
Overall, Dr. Zuckerman and her colleagues found that most medical devices recalled for life-threatening or serious hazards were cleared through 510(k), and concluded that the 510(k) pathway poses greater dangers to patients because it is less likely to uncover design or manufacturing flaws.
But the rate of PMA recalls may actually be higher overall when considering the total number of devices approved each year, explained Ariel Dora Stern, PhD, assistant professor at Harvard Business School, in Boston. Dr. Zuckerman’s analysis revealed that about four times more 510(k)-cleared devices are recalled each year than PMA-approved devices. But according to the FDA’s recall database, in any given year, about 100 times more devices reach the market through 510(k) than PMA. In 2014, for instance, more than 3,100 devices were cleared through 510(k), while only 25 went through the original PMA pathway—a 124-fold difference.
“We approve orders of magnitude more products through 510(k) each year, so proportionally fewer 510(k) devices are getting recalled compared to PMA devices,” Dr. Stern said. “Also, PMA-approved devices tend to be higher risk, so in theory, they are more likely to kill you if they malfunction.”
Although the quality and extent of clinical testing may be insufficient, Dr. Ball does not think it is a major driver of recalls. “Products that don’t require clinical testing are generally those that are substantially equivalent to currently sold products and probably wouldn’t benefit from additional clinical testing,” Dr. Ball said. “In our research, we have seen that it’s not really a lack of understanding of how the product interacts with the customer, something that clinical testing would reveal, but more about process management, [whether] design, supplier development or manufacturing processes. Problems in these areas would not, in my opinion, be impacted by additional clinical trials.”
Efforts to better understand why recalls happen are currently underway. Dr. Ball and Rachna Shah, PhD, associate professor in the Supply Chain and Operations Department, Carlson School of Management at the University of Minnesota, Twin Cities, have been investigating medical device recall causes, including FDA inspection results, as predictors of future medical device recalls with Carlson School colleague Enno Siemsen, PhD.
Amid the growing concern about device recalls, it’s important to take a step back and try to understand the risk in context. “More often, we observe when something goes wrong, but we are less aware of the years saved by devices that do what they’re supposed to,” Dr. Stern said. “There are many safe products that are helping people, even saving lives. But currently, we don’t have [a] good way to measure the benefit of billions of dollars [worth] of medical devices we use each year.”
Leave a Reply
You must be logged in to post a comment.