Skip to main content

How Medical Device Companies Test Products Out on Patients—After They're Released

Essure, the latest malfunctioning medical device, shows us exactly what’s wrong with the way medical devices are approved in this country.
Bayer AG, the company that released the birth-control device Essure, conducted only two non-random, non-blinded studies before releasing it on the market. (Photo: H005/Wikimedia Commons)

Bayer AG, the company that released the birth-control device Essure, conducted only two non-random, non-blinded studies before releasing it on the market. (Photo: H005/Wikimedia Commons)

In 2009, 28-year-old Cecilia Bogle went to the doctor to get a new device for permanent birth control called Essure. At the time, the device, a small metal and polyester coil inserted into a woman’s fallopian tubes, was marketed as safer than other options for sterilization because it didn’t require any incisions in the fallopian tubes. Three months after Bogle’s procedure, however, a hysterosalpingogram (an X-ray of the uterus and fallopian tubes) found that her right coil had migrated to her abdominal area. That same week, Bogle learned she was pregnant.

In the next five years, Bogle endured a hysterectomy (a procedure that removed her uterus), a bilateral salpingectomy (a procedure that removed both fallopian tubes), and an estimated three CAT scans, five X-rays, five ultrasounds, and three MRIs as doctors tried in vain to find and remove the Essure device. They were able to remove half of one coil—but told Bogle the rest of the device had fragmented in her body and would be impossible to retrieve. A device that was supposed to improve Bogle’s life instead, she told me, left her with only “ovaries and pieces of metal floating around in my body.”

Bogle was one of several women who testified last month at a Food and Drug Administration panel reviewing the Essure device, an investigation that has raised disturbing questions about the organization’s medical-device approval process. Federal scrutiny has been a long time coming for Essure: Since the device was approved nearly 13 years ago, thousands of women have filed complaints reporting punctured tubes, debilitating muscle fatigue, and severe pain—not to mention the fact that, according to one model based on pre-existing data, nearly 10 percent of patients get pregnant after getting the birth-control device. But Essure is just one example of a problem with malfunctioning medical devices that is far more widespread. Earlier this year, the FDA found that a cardiac device called the Lariat Suture Delivery Device, designed to prevent stroke, caused heart punctures, bleeding, and in some cases, death. Last year, the FDA found that certain devices used to treat uterine fibroids, called laproscopic power morcellators, were actually spreading cancer in women’s bodies. These warnings and recalls highlight a critical failing in the way these medical devices are currently tested before they hit the market—problems that can, and should, be prevented.*

These recalls highlight a critical failing in the way these medical devices are currently tested before they hit the market—problems that can, and should, be prevented.

Standards for clinical trials are shockingly low, according to Diana Zuckerman, president of the National Center for Health Research, a think tank based in Washington, D.C. The "gold standard" for testing the effectiveness of medical treatments are studies that are randomized (meaning that participants were chosen at random) and double-blind (denoting that participants—as well as the researchers—don’t know who is receiving treatment and who is in the control group). Oftentimes, researchers say studies of new therapies should also be compared against a control group of approved therapies, to assess how the new versions measure up. While many drugs approved by the FDA are required to undergo two randomized, double-blind trials, medical devices are more commonly required to undergo one clinical trial—which doesn’t have to be randomized or double-blind.**

Most of these clinical trials aren’t exactly conducted by impartial researchers either. The device manufacturer or a third-party company contracted by the device manufacturer often oversees test trials: One 2011 study found that 80 percent of medical devices recalled for life-threatening or serious hazards from 2005–09 did not even have to go through clinical trials before they hit the market. They were instead ushered through an approval express lane called the 5-10(k) process—which exempts new medical devices from clinical testing if they can prove that they are substantially similar to a device already on the market. “It’s sloppy and disrespectful to the patients generally. It’s not scientifically valid or appropriate,” Zuckerman says. “The standard for high-risk medical devices is lower than the standard for prescription tummy ache medicine.”

In the case of Essure, devastating consequences resulted from one of these typically swift processes. The company that released Essure, Conceptus, a subsidiary of Bayer AG, conducted two non-random, non-blinded studies that didn’t compare the new device to existing permanent birth control methods. Furthermore, these pre-market studies only reported data on Essure’s reliability from 89 percent of study participants. Many women involved in the clinical trials claim their answers were changed. In addition, at the time the device was approved, the company had only followed up with a quarter of those women after two years to see how the device was working. And although the FDA required the company to provide follow-up data on patients after five years, only one of those studies was published (and only seven years after the study was completed).

Asked to comment, Patricia Carney, Bayer AG's medical director for Essure, says that it is difficult to do randomized and blinded trials for permanent birth control methods. "This is not unique to Essure," she says. "It's exactly the same with IUDs and birth control across all companies." Carney says that Conceptus compared their data on Essure with a study of the effectiveness of tubal ligation—an existing permanent birth control method. She says that, although there was an agreement between Conceptus and the FDA to provide follow-up data, there was no requirement to publish that data. "I can’t answer as to why Conceptus chose not to," Carney says. At Bayer, she says, "We certainly like to see things published."

This lack of longitudinal data is troubling to Sanket Dhruva, a cardiologist and a Robert Wood Johnson clinical scholar at Yale University who studies FDA medical device regulation. “We need to follow patients for an adequate amount of time because some of these devices, like Essure, are meant to be put in women’s bodies for a very long time,” he says.  And while it may be difficult to conduct a blind trial for a permanent birth control device, it would be easy to at least do a randomized one, according to Zuckerman. Essure’s failure to do both illustrate the very real and dangerous consequences of the FDA’s low standards for medical device approval.

Essure’s failings illustrate the very real and dangerous consequences of the FDA’s low standards for medical device approval.

The dearth of follow-up is particularly troubling because, once approved to go onto the market, a device can malfunction for years before the FDA realizes it. Currently, the FDA relies on a voluntary self-report complaint system called the Manufacturer and User Facility Device Experience. However, a voluntary self-report system has its blind spots. “Many times when high-risk devices fail, it can result in death,” Zuckerman says. “If you have an elderly patient, there’s often no autopsy. Most of the time, nobody will ever know what caused the death.”

When asked about the clinical trial process for medical devices, FDA spokesperson Eric Pahon writes in an email: “Agency decisions are based on valid scientific evidence and balance the benefits and risks of a given product in the context of the underlying condition and available alternatives.” The agency doesn’t want to withhold “a useful, important therapy from the patients who can benefit from it,” he writes.

While patients may benefit from the latest medical device technology, the evidence conveys that they might be better served if those devices were tested more slowly, and with greater caution first. Although the number of devices recalled between 2002–13 has doubled, the process for device approval has, in the meantime, become quicker and easier. Zuckerman attributes this, in part, to device manufacturer lobby groups such as Advamed, which alone poured $2.4 million into Congress last year. Advamed’s most recent victory is the 21st Century Cures initiative, which passed the House of Representatives this summer and is awaiting Senate approval. Billed as a bi-partisan effort that would “speed the process from scientific discovery to health treatments and cures for people who need them,” the act would effectively allow companies to submit evidence of safety and efficacy based on anecdotal data rather than clinical trials. This means that high-risk devices with the potential to permanently injure or kill patients if they malfunctioned could go to market without going through rigorous scientific testing.

There are efforts to make medical devices safer. In 2012, the FDA passed a resolution to place a barcode on all medical devices—called a Unique Device Identification, or UDI—to track problems with devices before they affect too many patients. Physician groups such as the American College of Cardiology have taken matters into their own hands by creating databases that track the safety and effectiveness of newly approved high-risk devices.

These efforts, however, are slow going. Just this year, the Wall Street Journal reported that the Centers for Medicare and Medicaid Services has refused to put the UDI bar code on medical forms, claiming the change would cost too much. Until the federal government can provide more oversight for medical devices, it falls on doctors to educate the public on risks for new devices—and for the public to listen. “We all understand the challenges of having a family member who is ill,” Dhruva says. “But new devices or new treatments all have risk, and some can make things worse. It’s important for the public and patients to understand that regulation is complex.”

And even if those tracking efforts are finally put in place, it may be too little too late. After all, the best way to prevent patient harm is to test new devices before they hit the market, rather than retract them after they’ve done damage. It appears that with current FDA regulations, many early patients are functioning, perhaps unknowingly, as test subjects for companies. “They’re using patients as guinea pigs,” Zuckerman says. Which, when it comes to devices in our internal organs, isn’t a comforting thought.


*UPDATE — October 9, 2015: This article has been updated to reflect that these figures were based on probability models.

**UPDATE — October 9, 2015: This article has been updated to more accurately reflect the FDA's requirements for medical devices.