When her tests came back normal, Carolyn Thomas was mortified. Even from her desk in the palliative care department of British Columbia’s Royal Jubilee Hospital, Thomas knew how busy the ER could get. She had just wasted her coworker’s precious time only to be diagnosed with acid reflux.
Less than an hour earlier, on her morning walk, the 58-year-old had been stopped in her tracks by a wave of nausea and a crushing pain in her chest. Suddenly breathless and sweaty, Thomas had slumped against a nearby tree. When the pain had spread to her left arm, it confirmed her worst fear: She thought she was having a heart attack.
Yet by the time she got to the ER, her textbook symptoms had all but passed, and the doctor said, according to the standard tests, her heart rhythm looked normal. He sent her home with instructions to talk to her general practitioner about antacids.
It was May, 2008. After two weeks of ever-more-debilitating episodes, Thomas returned to the ER, desperately hoping they might have a more potent medication to treat her crippling acid reflux. But this time, they called in a cardiologist. A few minutes into the examination, the cardiologist insisted that she be rushed to the operating room immediately.
“Are you saying, I’m going to have a heart attack?” Thomas asked.
“I’m saying, you’re having a heart attack,” the cardiologist replied.
Tip of the Iceberg
Misdiagnoses are not rare, particularly for patients like Thomas. Just last month, the latest of many studies showed that doctors overlook cardiovascular disease in women more often than in men, even though the disease is more common in women.
“People still think of it as a man’s disease,” says lead author Dr. Raphaëlle Delpech, a general practitioner at Paris-Sud University. “But women still get the disease more than men, disease rates are dropping slower for women than for men, and doctors more often fail to diagnose the disease in women than in men,” she says. The American Heart Association emphasized these points earlier this year in a scientific statement.
Cardiac disease is just the tip of the iceberg. As a rule, women suffer from higher rates of many chronic diseases, experience more health complications, and suffer more adverse drug reactions than men do. Between 1997 and 2000, eight of the ten drugs withdrawn from the market had more detrimental side effects in women than men.
In 2012, researchers discovered that women metabolize zolpidem, the active ingredient in Ambien and other sleep medications, more slowly than men. This meant that for 20 years, women, on the advice of their doctors, had been overdosing on a drug known to significantly impair driving. The drug took a particularly noticeable toll in 2010, when women experienced more than two-thirds of all 64,000 zolpidem-related emergencies. In January of 2013, the FDA warned that the recommended dose for women should be cut in half, a change now specified in the drug directions.
“To my knowledge, that’s the first sex-specific dosage labeling for a drug,” says Dr. Janine Clayton, director of the National Institute of Health’s Office of Research on Women’s Health. “We need to do better.”
Clayton first became aware of these inequities in 2001 when she discovered yet another shocking statistic: worldwide, nearly twice as many women suffer from visual impairment and blindness as their male counterparts. This left Clayton, an ophthalmologist at the time, reeling. “Surely this can’t be the case in the U.S.,” she had said to herself. But, she found out, it was.
Clayton was determined to find the cause for these stark differences. One explanation, she suspects, is rooted in a centuries-old bias.
Underrepresented in Research
Across biological and medical research, scientists more often study males than females. Historically, the male-dominated scientific community claimed that female hormonal fluctuations introduced complicating variables to their experiments. Others thought it a moral imperative to protect women, especially pregnant ones, from the risks of research.
Regardless of their motivations, scientists systematically sidestepped sex differences by studying men and then applying the findings to women. Today, we’re discovering that this has left yawning gaps in our biomedical knowledge.
This is not to say that male and female biomedicine are always different. Analyses of 163 FDA approved drugs between 1995 and 2005 uncovered only 11 with largely differing effects on those with two X chromosomes. And some drugs originally thought to act differently in the sexes have been found indistinguishable upon further investigation. But where there are true differences, biomedical science is often unaware or ill-equipped to deal with them.
The bias went largely unaddressed until 1993, when, for the first time, the National Institutes of Health (NIH) required that women and minorities be included as study subjects whenever appropriate. (Some studies are exempt for obvious reasons, such as those studying testicular cancer or melanoma rates in people of northern European descent.)
Since then, things have progressed. A 2015 study from the Government Office of Accountability reported 57% of NIH-funded clinical trial subjects since 2005 were women—a drastic improvement.
However, this progress has limits. The NIH only funded about 6% of total clinical trials in the U.S. in 2014. Since the GAO statistic was aggregated across all types of biomedical science, specific research fields—like cancer research or neuroscience—may be inequitable for all we know. That’s to say nothing of the studies performed before 2005, many of which are still referenced by researchers and clinicians. And for some fields, like sports medicine and cardiovascular disease, we know for sure that the bias persists.
Notably missing from the GAO’s statistic are trials conducted by pharmaceutical companies, a secretive body of research that’s much harder to appraise.
A Big Unknown
In 2014, the pharmaceutical industry conducted about 35% of all clinical trials in the U.S. (The percentage is even higher in Eastern Europe and Central and South America.) During the last 25 years, the private sector is estimated to have conducted anywhere from two-thirds to nearly all of U.S. drug development.
The Food and Drug Administration (FDA) reviews all publically and privately developed drugs before they are released onto the market, a process which, according to the FDA website, ensures that drugs “work correctly and that their health benefits outweigh their known risks.” However, just because the benefits trump the risks for the entire population does not guarantee that a drug is effective or safe for subgroups like women.
For years, organizations like the Society for Women’s Health Research and the National Women’s Health Network pressed the government to address this issue. In 2014, the FDA responded with a 27-point action plan and a new tool called Drug Snapshots to make clinical trial data more accessible. “The goal of the Snapshots was to increase transparency by putting sex, race, and age-related information in an easy-to-read format,” says Dr. John Whyte, who oversees their creation.
Whyte was glad that advocates pushed for this information, but he is concerned that it could be misinterpreted. He predicts advocates may balk at any trial that uses less women than men, even when enough women participated to know how the drug affects them. “These numbers are not a question of social justice. They are a question of science,” Whyte says.
Still, the 72 Snapshots released to date reveal a handful of drugs for which the effects on women are inconclusive. For three drugs, there were not enough women in trials to determine if they are as effective in women as in men. For another three, it’s unclear whether they could cause more side effects in women. Five are definitely known to have worse side effects in women. (Even less is known about such effects in people of color, particularly African Americans.)¹
And for privately developed drugs released before May 8, 2014, when the first Snapshot was published, patients and physicians are left guessing.
Even for new drugs, critical information is omitted from the Snapshots, including results from experiments conducted before the final round of clinical trials (Phase III trials). This conceals another minefield of sex bias—animal testing.
Excluded from Experiments
In 2011, neuroscientist Dr. Annaliese Beery published the first literature survey to document the inadequate use of female rats, mice, monkeys, and other animals in research. Beery examined a total of 800 studies published in 2009 in the top journals across ten areas of biology. The degree of bias varied, but it was most pronounced in her own field of neuroscience: Of the studies that reported animals’ sex, 88% of them included males while only 38% included females. Pharmacology was similarly lopsided, with at least 63% of the studies using only male animals.
“If you only studied male animals, you might develop drugs and give them to females, not even knowing that the underlying biology is fundamentally different,” Beery says. “For many phenomena, we never stopped to ask, ‘Is there a sex difference?’”
Sometimes there is. A 2015 research paper in Nature Neuroscience showed that certain classes of pain medication worked in male mice, but not in females. Ironically, women suffer from more chronic pain issues than men.
Beery says the bias may be even more extreme than the data show. Scientists often fail to specify the sex of the animal subjects, cells, and tissues used in their research, which makes it impossible to know just how pervasive the bias is.
Failure to label an animal’s sex is not just a social problem but a scientific one, because it makes the experiment impossible for others to replicate. The NIH’s Janine Clayton suspects that the lack of labeling and the sex bias issue in general may be contributing to the reproducibility crisis plaguing many areas of science. “A key aspect of scientific rigor is transparency and a lack of bias. Nobody can recreate your experiment if you don’t tell them the sex of your animals,” she says.
Even when researchers specify the sex, they often fail to show the results for males and females separately, preventing other researchers from using the data to answer sex-specific questions.
Hurdles to Progress
In 2014, Clayton championed a new policy requiring researchers to consider the sex of their study subjects when applying for NIH funding. As of January 2016, the applications are evaluated, in part, based on that consideration. “This has been the most concrete advance at the policy level,” Beery says, herself in the process of applying for a grant.
Not everyone thinks this is a good idea. “Studying sex differences is too important to just tack on to every other research project,” says Douglas Fields, a neuroscientist in Bethesda, Maryland.
Fields thinks that requiring all researchers to simultaneously study sex differences and their original scientific question will be quite costly, requiring more time and more animals. And it might not even work. “People spend their whole careers studying sex differences. It’s an insult to them to pretend that everyone can do what they do. This just promotes bad science and allows researchers to check off the ‘sex differences box’ without having done the research justice,” he says.
Ani Manichaikul agrees with Fields’ first point. “If there is a difference, you’ll only find it by analyzing the two sexes separately. It’s basically asking researchers to double their sample size,” says Manichaikul, an assistant professor of biostatistics at the University of Virginia.
But she still supports the NIH policy. “It seems kind of a shame if people go off in a really developed research direction and find interesting results without realizing until later that it’s only relevant for half the population,” Manichaikul says.
While this policy may reduce such problems in the public sector, no similar policy yet addresses private companies. Privately developed drugs released before May 18, 2014, remain a black box. Those released after that date are slightly less opaque, thanks to Snapshots. But even when a Snapshot suggests a bias, little can be done. Since private pharmaceuticals provide their own funding, Clayton’s new grant policy holds no sway over how they conduct research.
A Perfect Storm
Back in the hospital, Carolyn Thomas woke from sedation hours later to learn that a major artery feeding her heart muscle had been 95% blocked. The misdiagnosis had nearly killed her. The cardiologist had implanted a stent just in time. Though the heart attack would force her to retire early, Thomas had survived.
Retirement gave her time to read up on misdiagnosis. She was surprised by what she found. “Many physicians still think of heart disease as a man’s problem,” Thomas says. “Even the kind of heart attack I had … physicians still call it the ‘widow maker’ even when it happens in women, because when a man gets this heart attack, it makes a widow of his wife.”
Dr. Sharonne N. Hayes, cardiologist and founder of the Women’s Heart Clinic at Mayo Clinic, agrees. “Physicians who see a younger woman come in with classic symptoms sometimes don’t even run the tests. It’s not even on their radar.”
Heart disease has become a flash point for the sex bias issue, partly because bias affects women’s cardiovascular health in everything from drug development to clinical treatment. Cardiovascular disease is more often overlooked in women. Many cardiovascular drugs have different and more adverse effects in women. Even non-cardiovascular drugs contribute to the problem; four of the ten drugs pulled from the market in 2001 caused arrhythmia in females.
Researchers are only just understanding how heart disease presents in women, which could further explain why it’s underdiagnosed. All these problems are exacerbated by the continued dearth of women in heart-health-related clinical trials.
This also causes treatment to lag behind. In 2009, one research team in Michigan found that women were not adequately represented in the clinical trials for implanted defibrillators, which require a risky invasive surgery. By combining the data from several studies they discovered that the devices provided no survival difference for women at all.
However, the study led by Dr. Delpech, the general practitioner at Paris-Sud University, might reveal a weapon for combatting bias—female doctors. While all doctors are more prone to misdiagnose women with cardiovascular disease than men, female doctors make fewer mistakes than males do. (Her current unpublished research suggests this is true for more than just cardiovascular disease.)
“Women were more systematic in their assessment of patients than men,” Dr. Delpech says. “They were more likely to follow protocols for each patient. They also took longer.” This means that physicians could potentially use standard protocols to counteract their own biases.
She suspects that adherence to protocols is internalized early. “At school, teachers tell girls to work hard and leave fooling around to the boys,” she says. Boys, in contrast, are not as often expected to follow rules. She doubts these socially constructed stereotypes are limited to France where the study was conducted.
Regardless of its origins, today’s physicians need a check for their bias. Dr. Sharonne Hayes believes that survivors like Thomas can help play this critical role.
In 2002, Hayes became medical director of the Science and Leadership Symposium for Women With Heart Disease, a program that trains women survivors to share their experiences at home and among medical professionals. The program is held at Mayo Clinic in partnership with the National Coalition for Women With Heart Disease. “Over 60% of the women in this program have been directly told that they have saved a life,” Hayes says.
The program even introduced Hayes herself to a whole new line of research. Hayes had been taught that a certain form of heart attack—spontaneous coronary artery dissection or SCAD—was too rare to study properly. In 2009, a patient at the symposium asked her a fateful question. “What if I told you that there was an online community of 70 women who experienced this heart attack who’d be willing to help?” Hayes has since developed the largest ongoing SCAD study that now includes over 600 patients.
Carolyn Thomas was also inspired by this community. Since her heart attack, she has given hundreds of talks to medical professionals, women’s clubs, and more. What was originally just an online speaking calendar became Heart Sisters, her blog which currently has received over 8 million views from 190 countries.
Thomas thinks that the American Heart Association’s scientific statement indicates significant progress. “Finally! This is the first time in its 92-year history that the AHA has said anything about heart attacks in women. It’s a small step, but it’s a start,” Thomas says.
Beery, the neuroscientist, thinks the animal research community is also becoming more aware of their bias. In a 2014 meta-analysis, her colleagues debunked the myth that female mice show more variability than male mice, making them perfect candidates for almost all types of animal research. This finding along with the NIH’s new initiative to consider the sex of animals and tissues, may bring equity to the earliest stages of biomedical research.
Thomas, Dr. Clayton, Dr. Hayes, and Dr. Delpech all agree that the medical and research professors must join forces to successfully address sex bias. It will require structural and cultural changes at every stages of research and treatment. “It will take a village,” Dr. Hayes says.
- There were not enough women in trials for Cholbam (cholic acid), Odomzo (sonidegib), or Stiolio Respimat (tiotrpium bromide and olodaterol) to determine if they are as effective in women as in men. It also remains unclear whether Praxbind (idarucizumab), Strensiq (asfotase alfa), or Vistogard (uridine triacetate) could cause more side effects in women. Several drugs are known to have worse side effects in women, including Kengreal (cangrelor), Lenvima (lenvatinib), Zepatier (elbasvir and grazoprevir), Zontivity (vorapaxar), and Zurampic (lesinurad). ↩︎