Gender as a Principle of the Organization of Clinical Sleep Research

In 2014 the National Institutes of Health (NIH) required researchers to examine sex as a biological variable (SABV). While this approach is necessary to ensure adequate and appropriate female inclusion in research studies, it puts researchers at high risk for attributing their findings to biological sex differences when instead they may be more appropriately attributed to the influence and expectations of gender. In this paper, we specify how gender works as a principle of the social organization of symptoms, experiences, research, and clinical practice using obstructive sleep apnea (OSA) symptomology to illustrate these patterns. We draw from psychologist Sandra Bem’s account differentiating three specific mechanisms of gender: gender polarization, androcentrism, and biological essentialism.

The Society for Women’s Health estimates that only 1 in 10 women with OSA, defined as repeated pauses in breathing (apneas) or reduced airflow with oxygen desaturation (hypopnea) during sleep, are diagnosed with this disorder (Miller, Mong, et al., 2017). As with many health and illness phenomena including cardiovascular disease (Perez, 2019), diagnostic symptoms of OSA have been defined based on the symptoms most common among men: snoring, excessive daytime sleepiness, and waking up during the night snorting or gasping for air (Greenberg, Lakticova, & Scharf, 2017). These symptoms are specific and measurable.

However, when researchers retrospectively investigated sex differences in OSA symptomology, they found that women often report greater rates of depression as well as insomnia at diagnosis (Greenberg-Dotan, Reuveni, Simon-Tuval, Oksenberg, & Tarasiuk, 2007; Shepertycky, Banno, & Kryger, 2005). Together with recent papers examining clusters of OSA symptoms (Keenan et al., 2018; Ye et al., 2014; Zinchuk & Yaggi, 2019) as well as the booklet, Women and Sleep: A Guide for Better Health (Miller, Mong, et al., 2017), these studies suggest that insomnia, fatigue, or depression are symptoms more common in women than in men with OSA. These symptoms overlap with other psychological or physical health problems (Miller, Redline, et al., 2017), making them difficult to use in screening for OSA. Conversely, many researchers have not been able to target specific symptoms of OSA in women, summarizing them in review papers as “vague,” “nonspecific” or “generalized,” none of which can be measured (Franklin, Sahlin, Stenlund, & Lindberg, 2013; Lin, Davidson, & Ancoli-Israel, 2008; Miller, Redline, et al., 2017; Wimms, Woehrle, Ketheeswaran, Ramanan, & Armitstead, 2016; Ye, Pien, & Weaver, 2009). The overlap of symptoms such as depression with other medical disorders, as well as the general uncertainty of the presentation of symptoms in women with OSA, may lead researchers and clinicians to misclassify, dismiss, or miss them altogether, which has likely contributed to the under diagnosis of OSA in women. As a result, researchers are beginning to recognize there is more work to be done in OSA sex differences research (Guadagni & Pun, 2020).

If the symptomology of OSA in women is different than in men, this might suggest a sex-specific biological cause, in which case researchers should explore potential biological mechanisms. However, gendered roles, expectations, and characteristics likely influence perceptions of the causes, features, and consequences of impaired sleep, thus affecting symptom reports. If this is the case, identifying biological mechanisms may be necessary but not sufficient to address OSA in women. To understand why symptoms of OSA are indeterminate in women, we use depression as an exemplar symptom because it is attributed more often to women with OSA than to men to demonstrate how gender polarization, androcentrism, and biological essentialism (Bem, 1993) have organized clinical OSA research.

“Lenses of gender”

Bem’s “Lenses of Gender” provide a means to understand how women are marginalized in clinical sleep research by showing how gender organizes research questions, study designs and methods, and interpretations of results (Bem, 1993). Bem’s lenses – gender polarization, androcentrism, and biological essentialism – provide tools for mapping how research and clinical practice can produce and position women and men as different and unequal, as well as how gender organizes all research. There is a considerable legacy of thinking about gender as a principle of social organization at multiple levels (Messerschmidt, Martin, Messner, & Connell, 2018). Pioneer psychologist of gender Sandra Bem proposes that difference and dominance are produced, perceived, interpreted, and policed through three “lenses of gender” (Bem, 1993). The first lens, androcentrism, is the incorrigible proposition, or unquestioned cultural belief, that men and masculinity define the normal, typical, and default human while women and femininity constitute the abnormal, atypical, or invisible exception and difference. The lens of biological essentialism is the similarly irrefutable proposition that biology – or some natural essence – produces both “opposite” sexes and male dominance. Biological essentialism justifies androcentric perspectives and takes for granted that differences between men and women are binary, natural, evolutionarily adaptive, and rooted in reproductive biology. Both androcentrism and biological essentialism depend on gender polarization, which is the assumption that men and women constitute exhaustive, mutually exclusive categories and beings. Moreover, the notion of “opposite sexes” permeates all aspects of social life; “masculine/feminine binaries are used to categorize phenomena far removed from women’s and men’s bodies” (Carr, Ben Hagai, & Zurbriggen, 2017) so that every decision, preference, interaction, and institution is organized accordingly.

In Figure 1 below, we present the overlapping lenses in a Venn diagram. Each circle represents a specific mechanism (gender polarization, biological essentialism, androcentrism) through which naturalized differences between women and men and a putatively universal and human but actually masculine default organizes research. We make this gendered organization visible by exploring the specific constellation of gender effects on research and practice that produces the marginalization or even invisibility of OSA in women, especially in women with what practitioners consider depressive, vague, nonspecific, or generalized symptomology. We turn next to the gendered organization of sleep research to explain the presumed understanding of depression as a symptom of OSA in women.

Figure 1.

Depressive Symptomology in OSA through Bem’s Lenses of Gender.

Depressive Symptomology in OSA through Bem’s Lenses of Gender

Androcentrism.

Historically, men are the default research subjects in sleep (as in so much other) research, and OSA has been considered a man’s disease. The symptoms that men experience – e.g., waking up gasping for air, snoring, and snorting during sleep, excessive daytime sleepiness - are considered the normal, characteristic diagnostic symptoms of OSA. It remains unclear if women actually express different symptoms of OSA or if they simply perceive, prioritize, and report different symptoms of OSA than men.

Androcentric research and clinical protocols classify depressive symptomology, if present in women, as an atypical symptom of OSA because it is not a primary symptom of OSA in men. Additionally, researchers and clinicians have suggested stronger associations between depressive symptomology and OSA in women than men (Miller, Mong, et al., 2017). There are several alternative explanations that decenter androcentric explanations for depression as a symptom of OSA in women: 1) In most cases, these preconceptions have arisen out of primarily androcentric samples that make it difficult to generalize findings to women. 2) In the most cited studies of sex differences in OSA symptomology (Greenberg-Dotan et al., 2007; Shepertycky et al., 2005), researchers conducting retrospective chart studies found that at OSA diagnosis, women have more depressive symptomology than men. These findings may have resulted from the misinterpretation of depressive symptomology by healthcare providers as a psychological or other psychosomatic disorder (and treated with antidepressants), instead of a symptom of a treatable medical disorder such as OSA (Ejaz et al., 2011). 3) Alternatively, these findings may mirror women’s generally higher rates of depression than men in the general population (Pratt & Brody, 2014) and may not indicate a direct relationship between OSA and depression in women. 4) In the above retrospective studies, participants who were taking antidepressants were considered as depressed. However, patients take antidepressants for reasons other than depression, including insomnia, a sleep disorder. This problem confounds diagnostic criteria because insomnia is a symptom of depression and is especially prevalent in women.

Biological essentialism.

Research rooted in the biomedical model suggests that depression in women is caused by women’s fluctuating reproductive hormones (Albert, 2015). Because sex differences in reported depression rates emerge during adolescence, some theorists have posited that environmental or social-cultural factors unique to adolescent girls produce diagnosable depression (Nolen-Hoeksema, 1994); others integrate both social and hormonal factors (Edwards, Rose, Kaprio, & Dick, 2010). While genetic and gonadal sex, including hormones, are known to produce effects diagnosed as depression in women (Albert, 2015), the role that gender (e.g., divisions of labor, discrimination, socialized emotion) plays in both the causes and the symptoms of depressed affect is often neglected or minimized. OSA’s etiology is assumed to be similarly biological; if women with OSA have depressive symptomology, the biological essentialist assumes a biological mechanism.

As a result, biological essentialist assumptions lead researchers to assume biological origins for the association of depression and OSA in women. For instance, Macey et al. conducted NIH-funded brain imaging studies to examine a possible mechanistic cause for sex differences in mood among people with OSA (Macey, Woo, Kumar, Cross, & Harper, 2010). Research design, data collection, and reported results did not include measures of confounder factors such as men’s and women’s different social roles and the effects they have on mood and sleep were not incorporated. The assumption is that differences in men’s and women’s everyday practices of work, care responsibilities, sleep, and social expectations constitute statistical “noise” irrelevant to the biological mechanisms of poor mood in women with OSA, and therefore can be adequately “controlled” to reveal the natural mechanisms of poor mood in women with OSA. Postulating biological sex differences (Macey et al., 2012) as a cause for women’s greater propensity to depression in the presence of OSA reinforces a biological essentialist framework, minimizes the effects of gendered social roles, and reproduces the gendered organization of research practice, interpretation of results, and recommendations for clinical response.

Gender polarization.

To assume that depression in women with OSA is a result of natural mechanisms and also an atypical symptom of OSA, researchers must first start with the assumption that women and men must always be different. This assumption makes invisible the possibilities 1) that men and women may not be different and do experience the same physical symptoms of OSA and 2) that any differences perceived by researchers are a result of the ways in which gender has organized and evoked interpretations of sex differences.

Overlap of the Lenses

Androcentrism, biological essentialism, and gender polarization are sometimes contradictory and sometimes mutually reinforcing. Their “overlap” in the Venn Diagram further shows how the gendered organization of research and clinical protocols and practice related to OSA and depression further contribute to the uncertainty in women’s ‘real’ OSA symptoms by positioning women and men as naturally different and unequal.

Androcentrism and biological essentialism.

When sleep researchers look to biological explanations for differences in sleep between men and women, they often naturalize androcentrism by considering men as the norm. Until recently, even mouse models have excluded female mice for hormonal reasons (Epstein, 2007; Shansky, 2019). The androcentric perspective assumes that using female specimens – with female-specific hormone cycles – unnecessarily complicates or even compromises the internal validity of a study. However, male mice also produce hormones associated with behaviors that shape research results (Roehr, 2007); androcentrism and biological essentialism together normalize and render as universal the sex-specific hormone pattern of male specimen animals. Depressive symptomology is considered a primary symptom of OSA in women possibly because it also has been associated with the influence of female hormones and characteristics, and because researchers and clinicians have constructed OSA in biologically essentialist, androcentric terms – that is, as a man’s disease. The androcentric perspective considers depressive symptomology in women as hormonally influenced or biologically sex-specific and as a symptom unrelated to the ‘real’ symptoms of snoring and daytime sleepiness.

To the extent that researchers and clinicians base the symptomology for OSA on men’s presentations, phenotypes of OSA privilege men’s physiology. Research suggests there is a phenotype for OSA but it is based on men’s putatively biological traits (Casale et al., 2009). For example, women are more obese than men at the same apnea-hypopnea index (AHI) point and they have a significantly lower mean AHI than men (Wimms et al., 2016). Yet, as the previous sentence illustrates, and in journal articles examining OSA in women, women are viewed as different and usually compared to the norm of men’s presentations. Men have the normal phenotype and researchers compare women to men’s normalized symptoms.

Gender Polarization and Biological Essentialism.

Biological essentialism reinforces gender polarization in the form of sex-specific phenotypes. Women must be studied in terms of their reproductive processes (reproductive years, pregnancy, and menopause or midlife), which are presented as self-evidently and saliently different from men’s reproductive processes. Moreover, the assumption that men and women differ significantly induces credible biological essentialism at the same time that emphasizing biological essentialism naturalizes difference (Lorber, 1993). Reducing sex-specific symptoms, diagnostic criteria, or phenotypes to putatively biological phenomena excludes consideration of the ways assumptions about sexual dimorphism influence identifying and interpreting symptoms and conditions. It may be that men and women with OSA in fact have similar rates of depression, but accountability to “doing” gender in recognizable ways (West & Zimmerman, 1987) organizes the relative importance men and women place on their depressive symptoms, as well as sleep quality or daytime sleepiness (Morris, Rohay, & Chasens, 2018).

Androcentrism and Gender Polarization.

In order to maintain an androcentric perspective, it is essential to maintain the belief that men and women are different, and that male and female are exhaustive, mutually exclusive, heterosexually complementary categories and types of bodies. However, we acknowledge an essential tension within the overlap of androcentrism and gender polarization. Social constructionist theories of gender (Butler, 1990) have provided a possible solution to the androcentrism and biological essentialism that organize research by collapsing sex category differences. However, this approach does not address, and potentially undermines, women’s inclusion and presentation in research. By minimizing concepts of sex difference, researchers risk losing the opportunity to understand health outcomes holistically through the inclusion of an embodied perspective – and more importantly the ability to advocate for the sex category of ‘woman.’ This tension signifies the importance of understanding how gender works in research; understanding gender is the only way to incorporate all perspectives to ensure rigorous and thoughtful health investigations.

Gender Polarization, Biological Essentialism, and Androcentrism.

Men and women have differing DNA, prenatal and circulating hormones and hormone receptors, and genitals, so in sleep research, as in so much other research and clinical practice, researchers expect significant differences in outcomes, process, or conditions between men and women, and attribute observed differences to pre-social biology. When researchers consider male-specific biology to be the norm, androcentrism produces and naturalizes sex differences. This perspective posits that men and women are hormonally dimorphic, and men’s hormones are the norm. Because hormonal influences in women might complicate the conventional symptomology of OSA, it becomes expedient to minimize women’s symptoms because they must be influenced by being female—unlike men’s symptoms, which researchers consider gender-neutral. Gender polarization, biological essentialism, and androcentrism simultaneously influence the experiences and presentations of patients/research subjects and the assumptions, protocols, and practices of researchers and clinicians. As conceptualized, measured, and positioned at the confluence of the three lenses of gender, men and women with OSA have different symptoms because of women’s “atypical” hormones. One consequence is the research practice of studying women with OSA in terms of their reproductive stage of life. As a result, there are multiple and exhaustive sleep studies on women at mid-life, during pregnancy, and post-menopause. Androcentrism contributes to default construction of OSA as a man’s disease, from which a putatively natural pattern of women’s reproductive capabilities and hormones generates different and atypical symptoms of OSA, such as depression. The gendered organization of research and clinical practice reproduces women’s under-representation in research studies of OSA, which is in turn part of the explanation for the uncertainty of women’s OSA symptomology.

These lenses, and the manner in which they overlap, all offer partial explanations for the differences in depressive symptomology found between men and women with OSA. They also shed light on how women have been marginalized in health and sleep research. While it is tempting for OSA researchers to try to find a biological mechanism for sex differences in symptom reports, it is imperative that OSA researchers recognize the complexity not only in biological processes, but in the external influences on behavior that effect symptom reports within in all categories of OSA severity, and most importantly in the androcentric organization of the categories and measures of OSA. Research and clinical practice should recognize the complex ways biological essentialism, gender polarization, and androcentrism jointly produce research and clinical practices as well as shape the causes, presentations, interpretations, and sequela of sleep disorders and symptom expressions.

Conclusion

With the NIH’s call for all researchers to include sex as a biological variable in their studies, there is high risk for bio-essentialist interpretations of research findings. As we have shown, women’ OSA symptomology has been considered atypical compared to men’s symptomology. When their symptoms have been investigated, they has been attributed to natural causes without considering the ways in which the organization of gender affects conceptions of women’s health. Sandra Bem’s lenses of gender constitute a framework for conceptualizing gender not as trait or characteristic of researchers or patients, but as a principle of the social organization of embodied experience, health, symptoms, research, and clinical care. We have used Bem’s lenses to specify some of the ways the gendered organization of research and clinical practice shape the clinical diagnosis of OSA in women, as a model of ways to examine and problematize research questions, design, and interpretations that are central to the gendered organization of biomedical and health research.