Invisible Women: Exposing Data Bias in a World Designed for Men
Page 20
Michelle’s parents rushed her to the emergency room. A doctor there asked her (in front of her parents) if she could be pregnant. No, she couldn’t be, Michelle explained, because she hadn’t had sex, and in any case, the pain was in her intestines. ‘They wheeled me into an exam room and without any explanation, placed my feet into stirrups. The next thing I knew, a large, cold metal speculum was crammed in my vagina. It hurt so badly I sat up and screamed and the nurse had to push me back down and hold me there while the doctor confirmed that indeed, I was not pregnant.’ She was discharged with ‘nothing more than some overpriced aspirin and the advice to rest for a day’.
Over the next decade Michelle sought help from two more doctors and two (male) gastroenterologists, both of whom told her that her problems were in her head and that she needed to be less anxious and stressed. At the age of twenty-six Michelle was referred to a female GP who scheduled her for a colonoscopy: it revealed that the entire left side of her colon was diseased. She was diagnosed with both irritable bowel syndrome and ulcerative colitis. ‘Funnily enough’, Michelle says, ‘my colon is not in my head.’ As a result of the extended delay in receiving a diagnosis and treatment she has been left with an increased risk of colon cancer.
It’s hard to read an account like this and not feel angry with the doctors who let Michelle down so badly. But the truth is that these are not isolated rogue doctors, bad apples who should be struck off. They are the products of a medical system which, from root to tip, is systematically discriminating against women, leaving them chronically misunderstood, mistreated and misdiagnosed.
It begins with how doctors are trained. Historically it’s been assumed that there wasn’t anything fundamentally different between male and female bodies other than size and reproductive function, and so for years medical education has been focused on a male ‘norm’, with everything that falls outside that designated ‘atypical’ or even ‘abnormal’.1 References to the ‘typical 70 kg man’2 abound, as if he covers both sexes (as one doctor pointed out to me, he doesn’t even represent men very well). When women are mentioned, they are presented as if they are a variation on standard humanity. Students learn about physiology, and female physiology. Anatomy, and female anatomy. ‘The male body’, concluded social psychologist Carol Tavris in her 1992 book The Mismeasure of Woman, ‘is anatomy itself.’3
This male-default bias goes back at least to the ancient Greeks, who kicked off the trend of seeing the female body as a ‘mutilated male’ body (thanks, Aristotle). The female was the male ‘turned outside in’. Ovaries were female testicles (they were not given their own name until the seventeenth century) and the uterus was the female scrotum. The reason they were inside the body rather than dropped out (as in typical humans) is because of a female deficiency in ‘vital heat’. The male body was an ideal women failed to live up to.
Modern doctors of course no longer refer to women as mutilated males, but the representation of the male body as the human body persists. A 2008 analysis of a range of textbooks recommended by twenty of the ‘most prestigious universities in Europe, the United States and Canada’ revealed that across 16,329 images, male bodies were used three times as often as female bodies to illustrate ‘neutral body parts’.4 A 2008 study of textbooks recommended by Dutch medical schools found that sex-specific information was absent even in sections on topics where sex differences have long been established (such as depression and the effects of alcohol on the body), and results from clinical trials were presented as valid for men and women even when women were excluded from the study.5 The few sex differences that did get a mention were ‘hardly accessible via index or layout’, and in any case tended to be vague one-liners such as ‘women, who more often have atypical chest discomfort’. (As we’ll see, only one in eight women who have a heart attack report the classic male symptom of chest pain, so in fact this description is arguably not only vague, but inaccurate.6)
In 2017 I decided to see if much had changed, and set off to a large bookshop in central London with a particularly impressive medical section. Things had not changed. The covers of books entitled ‘Human Anatomy’ were still adorned with be-muscled men. Drawings of features common to both sexes continued to routinely include pointless penises. I found posters entitled ‘Ear, Nose & Throat’, ‘The Nervous System’, ‘The Muscular System’, and ‘The Vascular System and Viscera’, all of which featured a large-scale drawing of a man. The vascular-system poster did, however, include a small ‘female pelvis’ off to one side, and me and my female pelvis were grateful for small mercies.
The gender data gaps found in medical textbooks are also present in your typical medical-school curriculum. A 2005 Dutch study found that sex- and gender-related issues were ‘not systematically addressed in curriculum development’.7 A 2006 review of ‘Curr-MIT’, the US online database for med-school courses, found that only nine out of the ninety-five schools that entered data into the system offered a course that could be described as a ‘women’s health course’.8 Only two of these courses (obstetrics and gynaecology classes taught in the second or third academic years) were mandatory. Even conditions that are known to cause the greatest morbidity and mortality in women failed to incorporate sex-specific information. Ten years later, another review found that the integration of sex- and gender-based medicine in US med schools remained ‘minimal’ and ‘haphazard’, with gaps particularly identified in the approach to the treatment of disease and use of drugs.9
These gaps matter because contrary to what we’ve assumed for millennia, sex differences can be substantial. Researchers have found sex differences in every tissue and organ system in the human body,10 as well as in the ‘prevalence, course and severity’ of the majority of common human diseases.11 There are sex differences in the fundamental mechanical workings of the heart.12 There are sex differences in lung capacity,13 even when these values are normalised to height (perhaps related is the fact that among men and women who smoke the same number of cigarettes, women are 20-70% more likely to develop lung cancer14).
Autoimmune diseases affect about 8% of the population,15 but women are three times more likely to develop one, making up about 80% of those affected.16 We don’t fully know why, but researchers think it might be down to women being the child-bearing sex: the theory is that females ‘evolved a particularly fast and strong immune response to protect developing fetuses and newborn babies’,17 meaning that sometimes it overreacts and attacks the body.18 The immune system is also thought to be behind sex-specific responses to vaccines: women develop higher antibody responses and have more frequent and severe adverse reactions to vaccines,19 and a 2014 paper proposed developing male and female versions of influenza vaccines.20
Sex differences appear even in our cells: in blood-serum biomarkers for autism;21 in proteins;22 in immune cells used to convey pain signals;23 in how cells die following a stroke.24 A recent study also found a significant sex difference in the ‘expression of a gene found to be important for drug metabolism’.25 Sex differences in the presentation and outcome of Parkinson’s disease, stroke and brain ischaemia (insufficient blood flow to the brain) have also been tracked all the way to our cells,26 and there is growing evidence of a sex difference in the ageing of the blood vessels, ‘with inevitable implications for health problems, examination and treatment’.27 In a 2013 Nature article, Dr Elizabeth Pollitzer points to research showing that male and female mice cells have been found to respond differently to stress; that male and female human cells ‘exhibit wildly different concentrations of many metabolites’; and to ‘mounting evidence’ that ‘cells differ according to sex irrespective of their history of exposure to sex hormones’.28
There are still vast medical gender data gaps to be filled in, but the past twenty years have demonstrably proven that women are not just smaller men: male and female bodies differ down to a cellular level. So why aren’t we teaching this?
The inclusion of sex-specific information in textbooks is dependent on the availa
bility of sex-specific data, but because women have largely been excluded from medical research this data is severely lacking. Even the very basics of sex determination have a sex data gap: since the landmark 1990 paper that identified the Y chromosome as ‘the’ sex-determining region, the female sex has – the irony – been seen as the default. But in this case, the default didn’t mean we focused on the female. Rather, research instead focused on testes development as the supposedly ‘active’ process, while female sexual development was seen as a passive process – until 2010, when we finally started researching the active process of ovarian determination.29
Most early research into cardiovascular disease was conducted on men, and women continue to be under-represented, making up only 25% of participants across thirty-one landmark trials for congestive heart failure between 1987 and 2012.30 Women represent 55% of HIV-positive adults in the developing world,31 and in parts of Africa and the Caribbean women aged five to twenty-four are up to six times more likely to be HIV-positive than young men of the same age.32 We also know that women experience different clinical symptoms and complications due to HIV, and yet a 2016 review of the inclusion of women in US HIV research found that women made up only 19.2% of participants in antiretroviral studies, 38.1% in vaccination studies and 11.1% in studies to find a cure.33
Because of their routine exclusion from clinical trials we lack solid data on how to treat pregnant women for pretty much anything. We may not know how a disease will take hold or what the likely outcome may be, although the WHO warns that many diseases can have ‘particularly serious consequences for pregnant women, or can harm the foetus’.34 Some strains of influenza virus (including the 2009 H1N1 swine flu virus) have ‘particularly severe symptoms during pregnancy’. There is also evidence that SARS can be more severe during pregnancy. It is of course understandable that a pregnant woman may be reluctant to take part in medical research, but this doesn’t mean that we have to just throw our hands up in the air and accept that we know nothing: we should be routinely and systematically tracking, recording and collating pregnant-women’s health outcomes. But we aren’t – not even during pandemics: during the 2002-4 SARS outbreak in China, pregnant-women’s health outcomes were not systemically tracked and ‘consequently’, the WHO points out, ‘it was not possible to fully characterize the course and outcome of SARS during pregnancy’.35 Another gender data gap that could have been so easily avoided, and information that will be lacking for when the next pandemic hits.
Like the failure to include women in anatomy textbooks, the failure to include women in medical trials is a historical problem that has its roots in seeing the male body as the default human body, but this traditional bias was radically enhanced in the 1970s, to the great detriment of women’s health, following one of the biggest medical scandals of the twentieth century.36
In 1960 doctors began prescribing thalidomide to pregnant women who suffered from morning sickness. The drug, which had been available as a mild over-the-counter sedative in many countries since the late 1950s, was considered safe because its developers ‘could not find a dose high enough to kill a rat’.37 But while it didn’t kill rats, it did affect foetal development (something that in fact the manufacturers knew as early as 1959).38 Before the drug was taken off the market in 1962, over 10,000 children had been born around the world with thalidomide-related disabilities.39 In the wake of the scandal, the US Food and Drug Administration (FDA) issued guidelines in 1977 excluding women of childbearing potential from drug trials. This exclusion went unquestioned.40 The acceptance of the male norm went unquestioned.
The male norm continues to go unquestioned by many today, with some researchers continuing to insist, in the face of all the evidence, that biological sex doesn’t matter. One public-health researcher revealed that she had received the following feedback on two different grant applications: ‘I wish you’d stop with all this sex stuff and get back to science’, and ‘I’ve been in this field for 20 years and this [biological difference] doesn’t matter’.41 It isn’t just anonymous notes, either. A 2014 op-ed published in the journal Scientific American complained that including both sexes in experiments was a waste of resources;42 in 2015 an op-ed in the official scientific journal of the US National Academy of Sciences insisted that ‘focusing on preclinical sex differences will not address women’s and men’s health disparities’.43
Alongside insisting that sex differences don’t matter, some researchers advocate against the inclusion of women in research on the basis that while biological sex may matter, the lack of comparable data arising from the historical data gap makes including women inadvisable (talk about adding insult to injury).44 Female bodies (both the human and animal variety) are, it is argued, too complex, too variable,45 too costly to be tested on. Integrating sex and gender into research is seen as ‘burdensome’.46 It is seen as possible for there to be ‘too much gender’,47 and for its exclusion to be acceptable on the basis of ‘simplification’48 – in which case it’s worth noting that recent studies on mice have actually shown greater variability in males on a number of markers.49 So who’s too complicated now?
Beyond the argument that women’s bodies, with their fluctuating, ‘atypical’ hormones, are simply inconvenient research vessels, researchers also defend their failure to include women in trials by claiming that women are harder to recruit. And it is certainly true that, due to women’s care-giving responsibilities they have less leisure time and may find it harder to make, for example, clinic appointments during the school run. However, this is an argument for adapting trial schedules to women, rather than simply excluding them, and in any case, it is possible to find women if you really want to. While reviews of FDA-mandated medical product trials found that women made up only 18% of participants in trials for endovascular occlusion devices (used if your foetal blood vessel hasn’t closed of its own accord)50 and 32% of participants in studies on coronary stents (which, incidentally, are another device where women have worse outcomes than men),51 women represented 90% and 92% of participants in facial wrinkle correction trials and dental device trials, respectively.
A more novel approach to addressing the problem of female under-representation in medical research is simply to claim that there is no problem, and women are represented just fine, thank you very much. In February 2018 a paper was published in the British Journal of Pharmacology entitled ‘Gender differences in clinical registration trials: is there a real problem?’52 Following ‘cross sectional, structured research into publicly available registration dossiers of Food and Drug Administration (FDA)-approved drugs that are prescribed frequently’, the all-male-authored paper concluded that, no, the problem was not ‘real’.
Leaving aside any philosophical debate over what an unreal problem might be, the authors’ conclusions are baffling. For a start, data was available for only 28% of the drug trials, so we have no way of knowing how representative the sample is. In the data researchers were able to access, the number of female participants in over a quarter of trials did not match the proportion of women in the US affected by the disease the drug was supposed to treat. Furthermore, the study did not address trials for generic drugs, which represent 80% of prescriptions in the United States.53 The FDA describes a generic drug as ‘a medication created to be the same as an already marketed brand-name drug’ and they are sold after the patent for the original branded drug runs out. Drugs trials for generic drugs are much less rigorous than original trials, having only to demonstrate equal bioavailability, and they are conducted ‘almost exclusively’ in young adult males.54 This matters because even with the same active ingredient, different inactive ingredients and different fabrication technology can affect a drug’s potency.55 And sure enough, in 2002 the FDA’s Center for Drug Evaluation and Research showed ‘statistically significant differences between men and women in bioequivalence for most generic drugs compared with reference drugs’.56
Despite all this, the authors claimed that there was no evidence of a
ny systematic under-representation of women in clinical trials because in phase two and three trials women were included at 48% and 49%, respectively. But the study authors themselves report that in phase one trials women represented only 22% of participants. And, contrary to what their conclusion might imply, the under-representation of women in phase one trials does matter. According to the FDA, the second most common adverse drug reaction in women is that the drug simply doesn’t work, even though it clearly works in men. So with that substantial sex difference in mind: how many drugs that would work for women are we ruling out at phase one trials just because they don’t work in men?
Digging deeper into the numbers, another issue the authors completely failed to address is whether or not the drugs were tested in women at different stages in their menstrual cycles. The likelihood is that they weren’t, because most drugs aren’t. When women are included in trials at all, they tend to be tested in the early follicular phase of their menstrual cycle, when hormone levels are at their lowest – i.e. when they are superficially most like men. The idea is to ‘minimise the possible impacts oestradiol and progesterone may have on the study outcomes’.57 But real life isn’t a study and in real life those pesky hormones will be having an impact on outcomes. So far, menstrual-cycle impacts have been found for antipsychotics, antihistamines and antibiotic treatments as well as heart medication.58 Some antidepressants have been found to affect women differently at different times of their cycle, meaning that dosage may be too high at some points and too low at others.59 Women are also more likely to experience drug-induced heart-rhythm abnormalities60 and the risk is highest during the first half of a woman’s cycle.61 This can, of course, be fatal.