This report has been updated with comments from Edwards Lifesciences, Madris Tomes and the FDA.
Abbott (NYSE:ABT) and Edwards Lifesciences (NYSE:EW) may have mislabeled patient deaths as injuries or device malfunctions in hundreds of FDA adverse event reports involving their transcatheter aortic or mitral valve replacement devices, according to a new study.
The study by JAMA Internal Medicine editor Dr. Rita Redberg, former FDA unique device identification (UDI) external program manager Madris Tomes and others examined a total of 10,558 adverse event reports related to Abbott’s MitraClip or Edwards’ Sapien 3 devices from the time the FDA approved them through December 2018. It was published as a research letter in the journal yesterday.
The researchers used an algorithm designed by Tomes to search adverse event reports using terms other than “death” that may also indicate that a death has occurred, including “expired,” “hospice,” “comfort care,” “passed away,” “died,” “autopsy,” “autopsied” and “comfort.”
The researchers found 5,323 injury and malfunction reports for the MitraClip transcatheter mitral valve replacement device (TMVR, approved in 2015) and 295 reports of deaths. Ninety-seven or 1.8% of the injury and malfunction reports also said that the patient had died using the terms “expired,” “hospice” and “passed away.” Thus, 97 of 392 (24.7%) patient deaths were miscategorized as injuries or device malfunction events, the researchers wrote.
The algorithm turned up 9,320 injury and malfunction reports for the Sapien 3 transcatheter aortic valve replacement device (TAVR, approved in 2013) and 1,021 reports of deaths. Of the injury and malfunction reports, 217 (2.3%) also said that the patient had died during or after the implantation of the device. The most commonly used terms to describe a patient death in these reports besides “died” were “expired,” “passed away” and “autopsy,” the researchers found. They concluded that miscategorized reports made up 17.5% of 1,238 total patient deaths.
Anyone may file an adverse event report with the FDA, but medical device manufacturers and hospitals are required to report to the agency when they learn that any of their devices may have caused or contributed to a death or serious injury. Manufacturers must also report to the FDA when they become aware that their device has malfunctioned and would be likely to cause or contribute to a death or serious injury if the malfunction were to recur.
Edwards said that hospitals report data on every patient treated in the U.S. with a Sapien 3 valve directly to the American College of Cardiology/Society of Thoracic Surgeons Registry (ACC/STS) Registry database. The registry makes regular reports to the FDA, but it does not make its information public. Anyone seeking information from it must request written permission from the reporting hospital, according to the registry’s website.
“Outcomes and data from the TVT Registry have been extensively published, and are regularly updated in journals and at medical meetings for reference by physicians and patients,” Edwards said in a statement to MassDevice. “This approach to postmarket surveillance of patient outcomes represents the most complete registry in medical technology.”
The authors of the study in JAMA Internal Medicine note in their research letter that the ACC/TVT registry “submits reports to the FDA in summaries that may omit redacted information and obscure important data.” The authors did not count a report as a death if it wasn’t clear, Tomes said.
Abbott did not immediately provide answers to questions about the study.
Redberg, a cardiologist, told MassDevice that she wanted to research adverse events related to common cardiac devices and found that Tomes’ Device Events database was easier to use than MAUDE.
“I’ve been interested in the limitations and in terms of what is reported, how accurate is it,” Redberg said. “What I took away from (the study) was that there’s room for improvement. It’s totally perplexing to me how a death cannot be categorized as a death. If you’re interested in accurate reporting, if it’s a death, you would want to check off the ‘death’ box.”
FDA analysts review adverse event reports in order depending upon the category assigned by the manufacturer, Tomes told MassDevice in an interview. Reports of deaths are reviewed first, followed by those of injuries and then malfunctions, she said. The number of adverse event reports has risen from about 60,000 per month four years ago when Tomes started Device Events to about 100,000 today, Tomes noted. The FDA has not increased the number of analysts and existing analysts may not be able to review all of the lower-priority reports, she added.
Redberg suggested that the FDA make it easier to report medtech adverse events and that everyone involved in adverse event reporting commit to accuracy because the MAUDE database is the primary source of information for device safety problems.
“As a cardiologist using these devices, I want to know, looking at MAUDE, how many malfunctions and deaths are reported, and I expect the deaths to be in the ‘death’ category,” she said.
Redberg also recommended that the FDA link UDIs to the MAUDE index to give physicians more accurate and timely information. The FDA has implemented the UDI system in some categories of medical devices, but has not completed it or linked it to MAUDE.
The FDA is also not allowed to change an adverse event category on its own but must send the report back to the manufacturer with a request to resubmit it indicating the proper category for the event, according to Tomes.
The FDA said that medical device reports, which are available to the public through MAUDE, are not the only tools the agency uses to monitor device performance, detect device-related safety concerns or signals and contribute to devices’ risk-benefit assessments.
“FDA takes seriously its responsibility to monitor device safety and performance, take action when needed, and help ensure patients and providers have the information they need to make informed healthcare decisions,” an agency spokesperson said in a statement to MassDevice.
“This passive reporting system has limitations, including the potential submission of incomplete, inaccurate, untimely, duplicative, unverified or biased data,” the agency spokesperson added. “As the authors note, this can at times include miscategorization of events, including deaths, by the reporter of the event. All Medical Device Reports are closely monitored by FDA analysts who do not rely solely on the event categorization by the reporters. They also rely on the event narratives in their analyses.”
The FDA said it has been working since 2012 on developing the National Evaluation System for Health Technology (NEST) to “systematically use real-world data in part to rapidly identify and help address safety signals once devices are on the market through active surveillance.” In fiscal year 2019, Congress provided funding to begin to develop active surveillance capabilities for NEST.
In the meantime, the agency said it has “increased focus on manufacturer compliance with reporting requirements and taken action when appropriate.”
Tomes has researched FDA reports for accuracy before, with similar results. A 2016 probe of adverse event reports related to vagus nerve stimulators made by Cyberonics (now LivaNova (NSDQ:LIVN)) led to this report by CNBC, among others. After the CNBC report appeared, Cyberonics began submitting its reports with the “death” box checked when applicable, Tomes said, citing more recent research.
Tomes also applied her algorithm to adverse event reports related to Bayer’s (ETR:BAYN) Essure birth control device in 2016, pointing out links to fetal deaths, miscarriages, stillbirths and ectopic pregnancies. Bayer has denied that Essure has caused harm. The FDA halted sales of the device last year.
“Although the Cyberonics reports appear to have improved, it doesn’t appear that any other manufacturers have discontinued this practice of misreporting,” Tomes said. “If the FDA were to call out this behavior regularly, then companies might improve how they report.”
The FDA has come under more recent scrutiny for failing to take action against manufacturers of devices implicated in injuries or deaths. Part of that scrutiny focused on the agency’s practice of allowing manufacturers to privately report adverse events via its Alternative Summary Reporting (ASR) program rather than publicly through MAUDE. The agency began required manufacturers to file “companion” adverse event reports via MAUDE in 2017, and in June 2019, it made public the ASR reports it received from 1999 through 2019.
“While all of these data were not posted on FDA’s website until June, they have always been available to FDA analysts and have contributed to the agency’s broader ongoing efforts to monitor device performance, detect potential device-related safety concerns or signals and contribute to the benefit-risk assessment of these products,” the FDA spokesperson said. “We can and have used these data to inform our current and past surveillance efforts. Our current policy is that all medical device reports submitted are made available to the public in MAUDE.”
The agency still has work to do, according to Tomes.
“The FDA data would improve if the FDA would more actively enforce their policy on reporting,” she said. “We’re seeing similar mis-reporting from other companies — this isn’t just a one-time thing. It’s happening across the board.”