The Danger Within Us

by Jeanne Lenzer

  The act also includes a provision that allows manufacturers essentially to approve certain of their own devices without FDA input. Here’s how it works: Device manufacturers currently must win FDA approval when they make changes to high-risk implantable devices. Under the act, manufacturers will be able to hire a third party to determine whether their “quality system” is adequate. Once certified, manufacturers will effectively be able to decide whether their own devices are safe and effective.

  The assumption that third-party certification, paid for by industry, will protect the public ignores recent history. Third parties hired to evaluate their paymasters know that return business is dependent on satisfied customers—a dynamic that played a central role in the run-up to the great recession of 2008, when financial analysts, hired by investment bankers, gave high marks to toxic investments. Similarly, evaluators who are financially dependent on the device manufacturers they are evaluating may be tempted to ensure repeat business by not finding too much fault with their paymaster’s products. And if a third party’s evaluations are critical, companies can simply hire a more compliant evaluator. This arrangement assumes a degree of honesty and fair play that has not been the trademark of device makers. For example, consider just a few selected fines levied against Medtronic:

  in 2008, Medtronic was ordered to pay $75 million for false claims regarding the need to hospitalize patients treated with its bone-cement product;

  in 2011, Medtronic was fined $23.5 million for paying kickbacks to doctors to encourage implantation of its pacemakers and defibrillators;

  in 2012, Medtronic agreed to pay $85 million to settle a class-action lawsuit regarding illegal corporate promotion of bone-graft material for off-label uses;

  in 2014, the company was fined $9.9 million for paying kickbacks, again to doctors who implanted the company’s pacemakers and defibrillators (kickbacks included free flights to events, used as “free vacations” by some doctors, along with “gifts of wine and alcohol” and “trips to strip clubs”);

  in 2015, Medtronic was fined $2.8 million for selling its spinal-cord-stimulation device, which “lacks evidence of clinical efficacy,” without FDA approval;

  in another 2015 action, the company was fined $4.4 million for labeling imported devices as “made in the United States” so it could sell them to the US military;

  in 2016, Medtronic admitted that over the course of five years it failed to report more than one thousand adverse events to the FDA related to Infuse, its spine-implant device (the company said the reports were “misfiled” and the failure to report was “unintentional”); no criminal or civil penalties have been assessed to date.229

  Medtronic is not an exception; similar rap sheets exist for many if not most companies. In 2007, Johnson & Johnson was fined $84.7 million by the Justice Department for paying kickbacks to doctors who implanted its artificial hips and knees. In 2011, it was fined $70 million to settle civil and criminal charges that it bribed doctors in Europe and Iraq to sell artificial joints as well as its drugs. Biomet, Inc., was fined $6 million for paying kickbacks to doctors to promote use of its bone-growth stimulators in 2014, and in 2012 the company paid nearly $23 million for giving kickbacks to foreign doctors.

  The list of companies forced to pay civil and criminal fines is extensive—not only in terms of the number of companies involved but also in terms of the number of repeat offenses committed by each company. Concealing harmful side effects, paying kickbacks, and promoting off-label use of devices continue unabated. Fines are simply the cost of doing business and are listed as potential or anticipated costs in each company’s Security and Exchange Commission filings.

  Nor do these actions have adverse consequences for the companies’ CEOs or bottom lines. As of early 2015, Omar Ishrak, CEO of Medtronic, earned $1.5 million in base salary and enjoyed a total annual compensation package of $39.5 million. The average salary and compensation package for the CEOs of the eighteen publicly traded device companies in the US was $1.3 million and $15.4 million respectively as of 2015.

  Civil and criminal fines don’t necessarily slow sales of worrisome devices. Patients and doctors often remain unaware of court rulings and FDA actions, instead relying on professional publications, which can be blissfully upbeat when company-paid researchers publish reports that accentuate the positive while eliminating the negative.

  Medtronic, which moved its headquarters from Minneapolis to Ireland in 2015 in a “tax inversion move,” announced that sales that year of its spinal implant, Infuse, were climbing fast, reaching $740 million globally, beating Wall Street estimates by $20 million. Medtronic, the second-largest device manufacturer behind Johnson & Johnson, saw its fiscal-year 2016 revenues soar to $30.24 billion, up from $16.18 billion in 2012.

  * * *

  The revolving door between private industry and the federal government is one major cause of regulatory capture. Direct pressure on scientists and civil servants from politicians is another: in many cases, a phone call to an FDA manager from an important member of Congress is enough to cause an FDA head to overrule a disapproval by the agency’s medical officers. But sometimes sterner measures are needed. There have been times when Congress has used its power of the purse to thwart regulatory agencies by simply threatening to defund them.

  In the early 1990s, Richard Deyo, a prominent spinal surgery researcher, and his colleagues incurred the wrath of the North American Spine Society (NASS) by publishing several studies suggesting that spinal-fusion surgery is often ineffective and sometimes harmful. Industry backlash was ferocious.230 The NASS, which is heavily funded by the companies that make the plates and screws used in spinal-fusion surgery, attacked Deyo and his sponsor, the federal Agency for Healthcare Policy and Research. The NASS mounted a full-frontal lobbying campaign targeted at key members of Congress, who in 1994 voted to eliminate the agency’s entire budget. Only intensive efforts by the American Medical Association and the American Hospital Association restored 75 percent of the original budget. But the damage was done: the agency was prohibited from issuing comparative effectiveness guidelines and was renamed the Agency for Healthcare Research and Quality (AHRQ)—the word policy notably removed from its name.

  Now the AHRQ outsources research to a mix of academic and contract research organizations. One of the AHRQ research groups, RTI International, based in Raleigh, North Carolina, received approximately $161 million of its $807 million funding in 2014 from “commercial and nongovernmental” sources, according to a spokesperson for the institute. Another, Duke Clinical Research Institute, receives approximately two-thirds of its $331 million annual revenues from industry.231

  Since Congress cut AHRQ’s funding by 25 percent and stripped it of its independent research capabilities, spinal-fusion operations have increased dramatically, says Deyo, and new surgical devices have proliferated. Despite the increase in surgery, there is little evidence of benefit for many patients: three of four randomized controlled trials in Europe found “very little advantage over rigorous rehabilitation for back pain due to worn-out disks,” said Deyo. The long history of industry influence and intimidation, he says, has left the device-approval and surveillance systems in shambles.230

  In the absence of tough, independent regulation, some device problems have become so egregious that manufacturers and the FDA have been forced to issue recalls. There have been recalls for defibrillators that caused continuous shocks, pacemakers that failed to fire, surgical mesh that bred infection, and bone-growth products that have caused paralysis, breathing trouble, and death. Recalls of malfunctioning or poorly designed devices disrupt the lives of hundreds of thousands of people who suddenly have to decide how to handle the danger within them. One study, published in the journal PLoS One, found that in 2004, there were two recalls of stents with design flaws that had led to breakages that could cause perforations, heart attacks, and death, affecting a total of 96,000 people. In 2007, the FDA issued a recall on the Sprint Fidelis defibrillator, affecting 26
8,000 individuals worldwide, 172,000 of them in the US. In 2011, the agency recalled 79,000 more defibrillators because of failures that also led to serious injuries and deaths. A 2010 recall of DePuy metal-on-metal artificial hips affected 90,000 individuals; a 1999 recall of pelvic mesh affected tens of thousands of women. In 2013, 33,000 inferior vena cava filters had to be recalled when it was found that instead of preventing clots from reaching the heart, they actually caused clots.

  Of course no device is perfect. Even the best devices will invariably cause harm and death for some. Lifesaving pacemakers have killed some individuals; artificial hips have poisoned some patients, causing everything from the death of surrounding tissue to heart and thyroid failure. The question is not whether a device ever causes harm but whether the benefits are expected to exceed the harm in a defined population.

  Unfortunately, there is little hope that the FDA will serve as a stalwart defender of public health against potential abuse by device manufacturers as long as it is subject to the long arm of industry. Over the years, the agency’s independence has been grievously compromised, leaving patients like Dennis Fegan virtually unprotected from the harm that oversold and poorly vetted medical devices can do.

  Chapter Eight

  The Power of Illusions

  IN 2008, DENNIS FEGAN came across the online transcripts of the FDA’s Neurological Devices Panel meeting held on Friday, June 27, 1997 to consider approval of the VNS device. Fegan was excited. Perhaps this document would shed some light on the mysterious process by which the company had earned the right to unleash its unproved technology on the American public.

  Levering Keely of the FDA staff called the meeting to order at 10:11 a.m., a late start because of a bomb scare at the building. He said it was important to “push things along,” given that a number of the participants had flown in from other parts of the country and had their return flights booked based on the time the hearing was scheduled to end—4:00 p.m.

  Harold A. Wilkinson, professor of neurosurgery at the University of Massachusetts, served as the chairman. Seeming a bit unfamiliar with his role, he asked FDA staffers a few times how to approach procedural matters. Wilkinson was one of ten voting panelists. The others were all neurosurgeons or neurologists, with one exception: Steven Piantadosi, a professor of oncology and biostatistics at Johns Hopkins, was specially trained in methodology (the ways in which clinical studies are designed and interpreted).

  Other participants included five Cyberonics representatives (four of whom were doctors), a lawyer for Cyberonics, one “consumer representative,” and three FDA staffers. Five members of the public—patients, their family members, and advocates—were also present as scheduled speakers; they would lead off the panel’s discussions.

  The first patient, Tim Fabian from Binghamton, New York, told the panel that before he had his VNS device implanted he had “10 to 15 seizures a day,” but since his device was activated, in January of 1996, he said, “I have had no seizures.”*

  The next speaker, Paulette Machara, chief executive officer of the Epilepsy Foundation, said the VNS would be the “first device treatment option for epilepsy” that might help the “nearly one in three” patients with epilepsy whose seizures aren’t completely controlled by other means. She added her concern that seizures can be “life-threatening” and that medicines for epilepsy can be “harmful to a developing fetus.”

  Machara was followed by Nancy Jean, who told the panel that following implantation with the VNS device, her son, Albert, not only had fewer seizures but also went from being behind in school to winning an academic achievement award and becoming more independent. She concluded, “Albert has been given back his life thanks to the stimulator.”

  Robert Cassidy told the panel that before he was implanted with the VNS device he had two to three seizures a month despite taking three medicines for his seizures. After implantation, he said his seizures dropped to “close to nonexistent,” and he was able to take just one medicine instead of three. He told the panel, “The VNS changed my life.”

  The last member of the public to address the panel was Patricia Kroboth, whose son, George, was just seven months old when he had his first grand mal seizure. Over the years, despite multiple medicines and brain surgery, his seizures persisted, and Patricia was running out of hope, until, she said, “I picked up the newspaper one day and was immediately drawn to a picture of a teenage boy [who] looked exactly like George. The boy had a vagus nerve implant.” Kroboth was overjoyed at the prospect of the VNS device, which meant, she said, “For the first time, there would be no powerful drugs and all of their side effects.…There would be no invasive surgery, with all of the complications that become a parent’s worst nightmare.”

  George underwent implantation in April of 1995 at the age of twenty-three. Kroboth said, “George was not a rapid responder. Slowly, over the coming months, she added, “we started to see a decrease in the numbers of seizures.…In the past two months…we have twice seen a period of 19 days with no seizure activity. The last time we could make that claim, George was 9 years old.” The number of drugs he took was reduced from five to three. Kroboth concluded, “The implant doesn’t need to stop all seizure activity to change someone’s life.”

  The five testimonials from the floor were powerful; even Fegan had to admit that. And they were all positive. Reading the transcript carefully, he tried to understand the gap between the testimonials, the evidence Cyberonics presented that persuaded the panelists to approve the device, and the deaths that continued to mount in the MAUDE database. After the testimonials, the scientific deliberations began. Fegan’s training as a paramedic gave him some familiarity with the terms used by the FDA panelists, but because he had no special training in medical research or how studies were constructed, he found it slow going. He read and reread the statements of the FDA advisers, and when he saw a red flag, a statement of concern, he homed in on the references, looking for meaning and hard numbers.

  It was during this painstaking process that he got a real shock: although some test subjects did have fewer seizures, between 20 and 33 percent of patients had more seizures. He was dumbfounded. It seemed that after implantation with the VNS device, roughly as many patients had more seizures as had fewer seizures.

  How was Cyberonics able to claim that a third of patients benefited from the device if a similar number of patients did worse? Could it be that the entire basis of Cyberonics’ business—the possible reduction in seizures provided by its VNS device—was based on a mere illusion?

  * * *

  One of today’s leading experts on medical illusions is Dr. Jerry Hoffman—the same Jerry Hoffman we met in chapter 2 as a young medical student who aroused the ire of Eugene Braunwald through a bit of irreverent humor. Over the course of his three and a half decades of doctoring, and as a professor of medicine at UCLA, Hoffman has learned why doctors are so often flat-out wrong about diseases and their treatments. As a former director of the UCLA Doctoring Program, he has taught countless medical students and resident physicians, many of whom say he has influenced them more than any other doctor. Doctors from around the world seek out Hoffman for his insight and advice. He has lectured on several continents and held visiting professorships in Paris, Tokyo, London, and Santiago.

  I attended a lecture given by Hoffman in New York City in June of 2011, when he addressed an audience of physicians at Bellevue Hospital. Tall, white-haired, and wearing thick glasses that magnify shiny dark brown eyes, Hoffman looks part elder statesman and part absentminded professor. His clothes are slightly rumpled, and bits of his hair stand out at right angles to his head.

  He begins his talk at Bellevue by focusing not on the body or on any particular treatment but on language. “Words in medicine often have a different meaning than they have in ordinary English usage,” he says. “And how we use language is really important.” As an example, Hoffman draws our attention to the word significant, used in so many studies and marketing campaigns. “In co
mmon English,” Hoffman says, “this word means something entirely different than what it means in statistics—in the larger world out there, if something is described as ‘significant,’ we all understand that it’s important.” If a drug or device is reported to have a “statistically significant” effect, many of us might assume that means it works. In statistics, though, “significant” is, well, less significant. As Hoffman explains, when a study finds a difference that is “statistically significant,” it implies absolutely nothing about whether the difference is important. It simply means that the finding probably wasn’t attributable to the play of chance.

  Hoffman cites a study in which use of a certain drug “significantly” decreased the temperature of patients with fever. “Even if we buy that the observed drop in average temperature, from 101.0 to 100.7, wasn’t due to chance and ‘really’ was due to the drug—is that clinically important? Does it mean patients will feel any better or have a better outcome? What is statistically ‘significant’ may or may not be clinically meaningful.”

  Hoffman then asked his listeners to think a little more deeply about the role of chance, saying, “This, too, is an area where we can easily be misled. We often assign meaning to a chance occurrence in our life when it seems too strange to have happened without a reason. Most of you have surely had an experience where just yesterday you thought about someone you hadn’t seen for a long time…and then today he called! You think that’s too strange to simply be a coincidence—there must be some deeper reason! But extraordinarily rare events actually happen all the time. How many people do you think about, if only for a second, every day…the vast majority of whom don’t call?” For every million “one-in-a-million” events that could happen, the vast majority do not…but one of them will. We just don’t know which one of these rare possibilities that is—and are extremely surprised when it does happen.