On September 24, 2010, Michelle Pfleger, a champion equestrian, former varsity cheerleader, and freshman at Elon University in North Carolina collapsed on her way to class, then died. She was eighteen years old. The autopsy listed the cause of death as pulmonary thromboemboli—a blocked blood vessel in the lung—and stated that she had died of natural causes. Why would a young person in excellent health suddenly die of a blood clot? Pfleger had no risk factors other than the birth control she was on: a pill called Yaz, made by the pharmaceutical manufacturer Bayer.
Yaz was approved by the Food and Drug Administration in 2006. In the FDA’s initial report, released that year, regulators expressed some concern that the drug’s principal ingredient, drospirenone, caused an increased incidence of blood clots. The agency mandated that the company conduct post-market studies, but these were not begun until 2009. Meanwhile, Yaz quickly became one of Bayer’s top sellers, aided by an aggressive advertising campaign, including one commercial that showed balloons marked “irritability,” “headaches,” “bloating,” “fatigue,” and “acne” floating off into the sky and disappearing, as though Yaz could improve moods and cure ailments. In October 2008 the FDA deemed this ad misleading and deceptive and forced Bayer not only to pull it but to run corrective advertising—a rare step. The following year, Pfleger, then a high school senior, was prescribed the drug.
All birth control pills currently on the market carry a small risk of blood clots, but independent observational studies published in The BMJ (formerly the British Medical Journal) in 2009 and 2011 indicated that, as compared with other birth control pills, Yaz more than doubled the chances of a blood clot.1 By 2011 tens of millions of women were taking Yaz. The following year the FDA required Bayer to include a warning on its label stating that drospirenone increased the risk of blood clots relative to birth control pills with the hormone levonorgestrel or some other progestins.
The possibility that Pfleger might have lived if she had been prescribed a different birth control pill haunted her mother, who sued Bayer in 2011. That lawsuit was one of over 11,000 filed over Yaz in state and federal courts. In 2016 Bayer announced that it would pay $2.04 billion to resolve 10,300 pending claims by women who had taken the contraceptive. Such a settlement may seem enormous, but it is not especially large in relation to product sales—in 2012 alone Bayer sold more than $1.3 billion worth of Yaz. By 2015 this figure had dropped to about $918 million.
Of course, for plaintiffs like Pfleger’s parents, money is cold comfort when her doctor might have prescribed a different, safer pill. An ounce of prevention is worth a pound of cure. So often, the American cure is tort: the law of accidents and injuries, which allows people like the Pflegers to seek recourse for all types of harm. The result of using tort law as a remedy for regulatory failure is an alarmingly large number of personal-injury lawsuits related to a single product, ordinarily against a single company. These personal-injury suits, called mass torts, are often consolidated at the federal level before a single judge, which makes patterns easier to observe.
I study mass torts and have noticed that, over the last fifteen years, the number of lawsuits alleging injuries specific to women is remarkably high. These suits include the Yaz lawsuits against Bayer, as well as claims against Bayer that both its Mirena and Essure IUDs perforated the uterus; against GlaxoSmithKline that its anti-nausea drug Zofran caused birth defects; against Merck that its contraceptive Nuvaring caused blood clots, heart attacks, and strokes; against Johnson & Johnson that its talc products caused ovarian cancer; and at least 100,000 lawsuits filed against a number of companies—including Boston Scientific, C.R. Bard, and Ethicon (a subsidiary of Johnson & Johnson)—on behalf of women injured by the implantation of pelvic or transvaginal mesh, a medical product used to treat pelvic organ prolapse, a condition in which the muscles supporting the uterus, bladder, and rectum drop out of position, causing discomfort and sometimes incontinence. There are also mass torts alleging harms that predominantly, but not exclusively, affect women, such as lawsuits against DePuy Synthes (another Johnson & Johnson subsidiary) for its hip implants, which fail more often in women, or lawsuits against Merck alleging that its osteoporosis drug Fosamax causes significant jaw damage, mostly in women. (Not all of these lawsuits have been successful.)
When FDA regulation fails, as it arguably did when Yaz went on the market, tort law fills the breach. My research reveals that 32 percent of the large-scale consolidated mass tort lawsuits pending in the federal courts in 2018 involved products that either exclusively or primarily injured women and menstruating and pregnant persons, and that 27.7 percent of such consolidations involved products that exclusively affected women or menstruating and pregnant persons, because they concerned controlling reproductive health, birth defects, or ovarian cancer. By contrast, only 6.4 percent of the mass torts involved products exclusively affecting men: the erectile dysfunction drug Viagra, the male hair-loss drug Propecia, and Adrogel, a testosterone replacement therapy. The difference seems even starker if individual lawsuits are counted rather than the products that allegedly caused the injury: only 9,969 federal lawsuits in my dataset involved products that exclusively harmed men. Within the same data, 67,085 federal lawsuits were brought because of pelvic mesh alone.
It is true that many mass torts involve products that might adversely affect both sexes, like the blood-thinning medication Pradaxa or the antidepressant Zoloft, but a disproportionate number involve harms exclusive to women. Of the forty-seven mass tort lawsuits pending in the federal courts in 2018, for instance, thirty-one involved products that could affect either sex. Sixteen involved what are clearly sex-specific harms: thirteen for women and three for men. These tort cases, in the aggregate, reflect our medical system’s grisly disregard for women’s health.
In Invisible Women: Data Bias in a World Designed for Men, Caroline Criado Perez, a British activist and journalist, argues that researchers, medical professionals, industrial designers, safety engineers, and others consider the male body as the default in their work: “Failing to include the perspective of women is a huge driver of an unintended male bias that attempts (often in good faith) to pass itself off as ‘gender neutral.’” Women are often either left out of datasets or data are not sex-disaggregated, obscuring how products, safety measures, workplace policies, and the like disproportionately affect women. Because, generally speaking, cisgender women’s bodies are different from men’s—they tend to be shorter, lighter, and have higher body fat and a different shape—a product designed for and tested on the “average man” leaves out approximately half the population.
Criado Perez’s diverse evidence shows that the problems caused by the data gap between genders are pervasive, debilitating, and sometimes fatal in products such as seat belts, airbags, and wearable electronic devices, all of which are designed for the average man. Current seat belt models don’t protect average women as well as they do average men, and no seat belt has ever been designed that safely accommodates pregnant people. This failure of design has dire ramifications: although women are in fewer accidents, they are, according to a Consumer Reports examination published in 2019, 17 percent more likely than men to die in a car crash and 73 percent more likely to be injured in a frontal crash. The industry standard is to use “male” crash test dummies in the driver’s seat, although women make up 50 percent of drivers. Astrid Linder, a director at the Swedish National Road and Transport Research Institute, is working on the first crash test dummy to accurately represent women’s bodies; the US did not start using female dummies in safety tests until 2011, and even those don’t represent the average woman.
Sex-differentiated data were not initially collected in the Covid-19 pandemic. As late as April, the United States was not keeping track of sex differences in infection rates and deaths, though studies in other countries showed that such differences existed. The US finally began tracking comprehensive sex-disaggregated data in May. When those data were analyzed, researchers discovered that men are more likely to contract and die from the disease, though once the data are adjusted for social factors it appears that fatality rates are the same for both sexes.
What about pregnant women? Are they at greater risk, and is there a higher risk of birth defects from the disease? Criado Perez, whose book was first published in 2019, notes that pregnant women’s health outcomes were not tracked in the 2002–2004 SARS outbreak in China, and that, as a result, in the World Health Organization’s words, “it was not possible to fully characterize the course and outcome of SARS during pregnancy.” With Covid-19, studies show an increased risk of hospitalization and need for mechanical ventilation for pregnant women who have been infected.
Even the design of personal protective equipment (PPE), recognized as essential to the safety of medical workers during the pandemic, is gendered; Criado Perez cites studies conducted before the pandemic showing that 57 percent of women in fields ranging from emergency services to construction said that ill-fitting PPE hampers their work: the default design is made to fit the average man. Just 5 percent of women in emergency services, including police officers, reported that their PPE—such as bulletproof vests and jackets—never interfered with their work. Women—who make up approximately 85 percent of registered nurses, 25 percent of emergency room doctors, 38 percent of new emergency medicine residents, and 46 percent of new residents specializing in internal medicine—are at a higher risk of exposure to the coronavirus, or other infectious agents, if their PPE doesn’t fit properly.
Most of the poor policy choices or product design flaws cataloged by Criado Perez are sins of omission. Medical textbooks, for instance, usually depict men, not women, in drawings of human anatomy—even when illustrating parts of the body that everyone has, such as the nervous system. Although Criado Perez’s focus is sex, the point can be made with equal force about race and gender identity. Dermatology textbooks do not typically show examples of brown skin, depicting the typical human as not only male but a man of a specifically Crayola Peach color, leaving doctors ill-prepared to identify cancers on skin of other colors.
Emerging research supports Criado Perez’s gender-gap argument. The FDA database that tracks adverse reactions from pharmaceutical drugs is gender disaggregated, but not the one for medical devices. When the International Consortium of Investigative Journalists (ICIJ) wanted to find out whether there is a greater incidence of serious medical problems—what the health care industry calls “adverse events”—reported for women than men, they had no data to analyze. Collaborating with a team of computer scientists from Stanford University, ICIJ used machine learning to estimate which deaths and injuries reported to the FDA involved female patients and which involved male patients. Of the more than 340,000 incidents in the database whose sex the researchers were able to identify, 67 percent involved women.
This is important to my own findings. Many of the mass tort lawsuits described above involve medical devices that are alleged to injure women, leaving open the possibility that rather than being harmed more often than men by medical devices, women just bring more lawsuits. The only way to answer this question is to look at the data. While the ICIJ study is not definitive, both because the researchers were unable to decode the entire database and because of possible reporting bias among doctors, it suggests that greater numbers of women are injured by medical devices than men. This indicates that the best explanation for the large numbers of mass torts filed by women is probably not that women are more litigious.
If there is a discrepancy, how can we explain it? Are medical products aimed at women more dangerous? Do regulators and drug manufacturers subject products aimed at women to lower-quality testing or more lax criteria than products aimed at men? Or are there simply more products for women? There is some evidence that products for men receive greater scrutiny. A study of the male birth control pill, which appeared to be effective in early tests, was abandoned because the pill caused the male subjects to suffer from depression and mood disorders. Women have been complaining for decades that hormonal birth control has these side effects. Yet the first high-quality study of the emotional effects of birth control was not published until 2016, fifty-six years after the FDA approved the first birth control pill. “Health care professionals should be aware of this relatively hitherto unnoticed adverse effect of hormonal contraception,” the authors of the study wrote. Unnoticed by researchers, perhaps, but not by female patients over the course of half a century.
Questions about disproportionate harm to women from drugs and medical devices remain unanswered. After more than a year of searching, I have found no thorough and reliable statistical study evaluating the propensity of drugs and medical devices to injure patients that disaggregated harms by sex. The best I could find was a study cited by Criado Perez from 2017 that lists adverse drug events that affected only women but does not compare them with men, and the more recent, and incomplete, ICIJ study of adverse effects of medical devices. The gender data gap is a mile wide.
We cannot improve what we don’t measure. To ensure women’s safety, we must first measure it. Yet some medical devices are not even tested before marketing because of the legacy device loophole, a provision in the Medical Device Amendments of 1976 that exempts from testing medical devices sold before 1976. Consider metal-on-metal hip implants. This ostensibly gender-neutral medical device disproportionately injures women, who receive more of them, because it is designed for men. (The implant’s design, which is too shallow for women’s wider hips, has made it more likely that metal particles break off the implant, or that the implant breaks altogether.) Women were at least 29 percent more likely to suffer hip implant failure, a 2013 study published in the prestigious Journal of the American Medical Association found. This type of hip implant was exempt from testing because it was a legacy device, so it was never evaluated the way a new device would have been. A total of 13,700 claims relating to hip implants in women were settled for $3.5 billion between 2013 and 2019. By then the damage had been done.
The underregulation of medical devices affecting women, and the attendant lawsuits, is a yearslong pattern and problem. Between 1970 and 1990, silicone breast implants were implanted in one to two million women, before any systematic safety tests were conducted2; only pending litigation forced medical researchers to begin epidemiological studies. Lawsuits brought in the 1990s alleged that silicone leaking from implants caused autoimmune diseases: Dow Corning had instructed its salespeople to wash the implants before showing them so that tiny leaks, which made the implant bags greasy, were not evident.
Ultimately, studies found no connection between the autoimmune diseases plaintiffs suffered and silicone leaking into the body, and silicone breast implant litigation is now often invoked as an example of useless or even frivolous litigation. But it also demonstrates a severe regulatory failure. The health care establishment was comfortable experimenting on a generation of women, implanting a product that had documented but hidden defects and not studying the health effects of that product for twenty years.
Think, too, of the controversy over prolonged use of pharmaceutical estrogen (called “hormone replacement therapy” for publicity reasons) for women in menopause during that same period of time. Only after many years of being touted as a fountain of youth, hormone replacement was clinically tested on women and, in 1998, acknowledged to increase the risk of heart attacks and strokes, as well as cancer. A well-designed clinical study in the 1950s testing the effect of estrogen on men found it too dangerous to use, but that red flag did not influence its untested use in women two decades later.
The stories of these various lawsuits arising out of untested or undertested products illustrate how the tort system, and the blame it metes out, is only a backstop after regulatory failure has already occurred. The diet drug fen-phen was found, once tested, to cause primary pulmonary hypertension and heart valve defects in some patients and to have only minimal efficacy in promoting weight loss, and pelvic mesh was revealed to be a failure after it had been implanted into tens of thousands of women who now can’t safely have it removed. These debacles ended in lives harmed or destroyed, protracted litigation, and billions of dollars in damages.
Because mass tort lawsuits involve such large sums of money, it is important to consider the profit motive. Perhaps my findings are explained by a selection effect: do lawyers bring more cases involving women than men because such lawsuits are more lucrative? After all, plaintiffs’ lawyers are paid on a contingency basis—their fee can be as high as 40 percent of the recovery—and they can be expected to file suits in which they can make more money. In courtroom dramas on TV, women make more vulnerable victims. In 2019 I confidentially interviewed both plaintiffs’ lawyers and defense attorneys working on high-profile mass tort lawsuits involving products affecting women. All agreed: mass tort cases involving women are in fact not as profitable as those involving men. One reason is that these suits are harder to prove because of data gaps. Another is that the legal system compensates women less than it does men.
In Mass Tort Deals, Elizabeth Chamblee Burch, a law professor at the University of Georgia, scrutinizes mass tort settlements through painstaking review of legal documents buried in court dockets to find out what plaintiffs like Michelle Pfleger’s family actually recover in these suits.3 She finds that while there is initial public disclosure of settlement amounts—like the $2.04 billion settlement in the Yaz cases—there is often little information about how (or even if) the money is paid out. Such cases often settle in large-scale deals that do not award a specific amount to each plaintiff but rather give plaintiffs an opportunity to enter into an alternative dispute resolution system similar to arbitration, in which they will prove their claims. If plaintiffs don’t file claims, or if they lose, the defendant keeps the money. This is consistent with a broader trend in the American justice system away from public courts and toward private dispute resolution. There are fewer and fewer jury trials. Plaintiffs are routinely forced into arbitration through consumer contracts, and mass torts are no different.
The data Burch was able to uncover required extensive detective work, and yet, as she recognizes, it is troublingly incomplete. For example, she found that 20 percent of claimants received compensation in the litigation around Ortho Evra, a birth control pill, and 96.8 percent recovered compensation in the litigation around the anticoagulant Pradaxa, which affected both men and women. But how many received compensation in the Yaz litigation, and how much did they receive? We don’t know. Do women, on average, obtain lower settlement amounts than similarly situated men? We don’t know. Nor do we know why so many claims are rejected in some settlements, and so many paid in others.
Burch’s study illustrates the limits of a researcher’s ability to evaluate whether the tort system is acting as an effective regulatory backstop. Whether it performs this function depends on how much companies pay when they engage in wrongdoing. If companies that sell defective products are not forced to bear the costs of the injuries they cause—if they still profit handsomely even though they pay millions, or even billions, in lawsuits—there is no economic incentive for them to alter their behavior.
What can be done? The problem is pervasive, starting with the scientific and medical community, moving to the private sector, and ultimately into law and regulation. Criado Perez quotes an anonymous comment received by Sabra Klein, a microbiologist at Johns Hopkins who studies gender differences in the development of viral diseases, on a grant application: “I wish you’d stop with all this sex stuff and get back to science.”
These studies raised the alarm, but they did not use the gold standard randomized controlled trial. Lianne Parkin, Katrina Sharples, Rohini K. Hernandez, and Susan S. Jick, “Risk of Venous Thromboembolism in Users of Oral Contraceptives Containing Drospirenone or Levonorgestrrel: Nested Case-Control Study Based on UK General Practice Research Database,” BMJ (April 21, 2011), p. 342; “Hormonal Contraception and Risk of Venous Thromboembolism: National Follow-up Study,” BMJ (August 13, 2009), p. 339. ↩
See Marcia Angell, Science on Trial: The Clash of Medical Evidence and the Law in the Breast Implant Case (Norton, 1997). ↩
I wrote a prepublication comment for Burch’s book and am cowriting an article with her on confidentiality in the courts. ↩