The contribution of ethnography to epigenomics research: toward a new bio-ethnography for addressing health disparities

An introduction to ethnography

The term ethnography refers to the systematic study of people and cultures in which investigators attempt to elicit from respondents (known as informants) their understanding about any given topic in question. It is a method in which the point of view of the subjects under study is regarded as central to the activity of the investigators, and its findings are typically informed by a mixture of interviews, participant observation and general immersion within the communities under study. Such findings are qualitative, and have been characterized as providing ‘thick description’, a term first created by Ryle [1], but then used extensively by Geertz [2]. At its best, ethnography is inherently participatory and entails actively working with a community to define the scope of the research and the methods that would be most appropriate to understanding the experiences that are important to them [3]. Its goal should be to build trust with and empower the community being studied in order to create research output that the community itself can also use to better their lives [4].

Ethnography – as a methodology and technique for designing research – is particularly well suited for challenging, and even changing, prevalent societal ideas and behaviors because it allows researchers to uncover extremely local and diverse human thoughts, practices and behaviors that have not yet been identified or understood by others [5,6]. As we shall discuss in this article, uncovering these significant experiences through ethnography will be crucial to a new research paradigm that combines these ethnographic insights with biological and molecular research in order to better understand human biological diversity and the production of health disparities.

When conducting ethnography, the most important task is to situate people and their experiences in time and space, thus bringing to the fore the coalescence of history, society, culture, politics, disease, and bodily experience. What ethnography reveals are local moral worlds', which are 'the commitments of social participants in a local world about what is at stake in everyday experience' [7]. Through working closely with informants, ethnographic investigators are able to isolate and understand what is at stake in their everyday experiences. The qualitative depth and personal engagement of ethnography excel in tracing all of the different aspects of life, society, history and health that permeate these experiences. For these reasons, ethnographic methodologies are perhaps most useful when trying to identify and understand the local experiences of those within communities that remain under-represented in society, especially where these are significantly different from the investigator's own experience.

Until relatively recently, however, the majority of anthropologists who conduct ethnographic research have confined their research to investigating historical, socioeconomic, political and cultural factors alone. The physical body has been discussed in relationship to these factors, but much more rarely have ethnographic researchers concerned themselves with putting their ethnographic analyses into dialogue with biological and molecular analyses of the human body. Although recent ethnographic research has engaged conceptually with the body and physiology, seen most prominently in research on embodiment [8], human biology itself has often been normatively set to one side by ethnographers on the assumption that this is the domain of biologists and natural science and beyond the purview of social scientists and ethnographers [9].

When trying to understand how the experiences revealed through ethnography influence and shape human health, however, more is needed than ethnography alone. Ethnography cannot help to reveal and analyze biological and physiological changes that occur as a response to one’s local world or environment (especially when these changes are ones that a person may or may not feel but cannot yet adequately verbalize). Likewise, what a medical or biological analysis cannot do is reveal all of the different significant experiences and environments that may have come together to create the observed physiological change. We argue that, especially when the goal is to better understand health inequality and disparities among under-represented communities, what is needed is a more complete research methodology that can combine both of these approaches. What is needed is a new ‘bio-ethnography’ [10] that can simultaneously use ethnography to uncover the most salient lived experiences and environmental conditions experienced by communities, while also using biological, molecular and epidemiological research methods to investigate how these local experiences and environments lead to changes in biological function, molecular activity, and, ultimately, patterns of health and disease. We see bio-ethnography as a new, transdisciplinary research process that incorporates all of these methodological approaches into a single project.

As we shall discuss throughout the rest of this article, environmental epigenomics is one such field that is well suited to be combined with ethnographic methods in the new bio-ethnography we are calling for. Together, ethnography and epigenomics will be critical to elucidating the biological pathways through which environmental disadvantage influences risk of disease, and in doing so will be crucial to better understanding and addressing health disparities and inequality.

Connecting environmental epigenomics & anthropology

Epigenetics has expanded into an enormous field of inquiry that includes stem cell biology, cancer biology, investigations into genome instability, DNA repair and molecular epidemiology. The sub-field of environmental epigenetics is specifically concerned with how physical environments (i.e., toxins, pollution, and green space), social environments (i.e., adversity, stress, trauma, malnutrition and poverty), and other external measures directly influence a number of epigenetic mechanisms, including DNA methylation, histone modifications and noncoding RNAs [11–16], which in turn dysregulate gene expression and can directly shape disease susceptibility [17–19]. The lessons learned from the last few decades of environmental epigenomics have been particularly powerful in understanding how our outside environments lead to epigenetic regulation of a number of biological processes.

The approach most common in environmental epigenomics research is to take a single factor from the environment (e.g., air pollution) and analyze the contribution of this factor to dysregulation of a known epigenetic mechanism. However, the impact of the environment on health is rarely this straightforward. As has been shown repeatedly throughout the public health literature, the impact of ‘environment’ on health is a complex, ecological process that involves an interplay of factors from multiple material, biological, social, and cultural dimensions [20] that come to bear on the health of individuals through several levels of influence (individual, interpersonal, institutional, community and structural/policy) [21]. This broad, ecological approach to conceptualizing, researching and intervening upon ‘environment’ has been increasingly adopted in public health over the past several decades [22,23], and has been strongly recommended by the US Institute of Medicine [24] and the WHO [25].

Research in environmental epigenomics has frequently focused on commonly studied individual-level factors such as smoking, toxin exposure, and maternal care [26], however recent research has also begun to show that structural environmental factors, such as neighborhood living conditions and childhood socioeconomic deprivation, are also associated with epigenome-wide changes in DNA methylation [27–30]. It has been shown that research and intervention aimed precisely at these structural levels of the environment may achieve the greatest success in alleviating health disparities [31,32], although relatively fewer public health interventions have been developed to date that focus on community- and policy-level structural factors [21]. Epigenomics research investigating how these different environmental factors may combine to influence disease and produce health disparities is greatly needed.

In order to transform research in the field of environmental epigenomics into a coherent body of literature that will have the maximal impact on public health and addressing health disparities, we propose that the field is in need of a common conceptual framework that ties together each of the individual environmental factors investigated between different studies into a common understanding of the ‘total’ environment. A recently published systematic review of the environmental epigenomics literature demonstrates this need well [26]. In this review, the authors found that the ‘environment’ studied between environmental epigenetics studies is usually understood and operationalized in different ways, with different studies isolating and analyzing factors within different dimensions and levels of environmental influence, and usually without reference to a common understanding of what constitutes ‘environment’.

We propose here that a multi-level, ecological model that understands the ‘environment’ as encompassing factors from multiple dimensions (material, biological, social and cultural) and levels of influence (individual, interpersonal, institutional, community, and structural/policy) will be a fruitful common point of reference for an environmental epigenomics focused on health disparities. What is needed now are tools that can extend measurement of the environment within epigenomics research to allow it to see the greater ecological connections between these environmental influences, and which will make it more attuned to engaging with the environments and local worlds of under-represented communities. To be clear, we are not arguing against the approach of studying single factors in the environment in relationship to epigenetic regulation. Rather, we are arguing that these approaches are in need of companion analyses and a conceptual organization that allows for greater contextualization of results to understand how environmental factors analyzed in isolation relate to a broader set of environmental factors and, ultimately, where to prioritize subsequent investigations.

To take one example, recent research has shown repeatedly that maternal stress, anxiety, and depression have lasting effects on infant development, often culminating in psychopathology later in life [33,34]. A growing body of epigenetics literature demonstrates that one mechanism through which maternal distress may impact child behavioral and health outcomes pre and postnatally is through altered DNA methylation in the placenta and cord blood [35,36]. This can lead to the development of positive interventions and providing guidance to mothers during postnatal development. However, resituating these exposures within an ecological model of the environment may help us to realize and identify that the ever-increasing poverty and often violent living conditions that mothers experience, especially in disadvantaged communities, are often at the root of the anxiety, distress, poor diet, smoking, and excess alcohol consumption measured as maternal exposures in these studies. Conducting ethnographic work with some of the mothers in a study such as this would be a particularly effective and expedient way to identify some of these other inter-related factors and structural issues that may be deeply connected to these more proximal measures of distress. A logical next step that would have a maximal positive impact on public health would then be to use these ethnographic insights to develop or identify survey measures to capture these other structural factors and then analyze these variables in a larger epigenomics research project. This future epigenomics research could assess whether maternal distress is merely a mediator between poverty or violence and placental DNA methylation, or whether it is associated with dysregulation of separate epigenetic pathways.

Uncovering local & situated biologies through ethnography

The term ‘local biologies’ was coined by Lock to account for the manner in which both individual biology, and that of groups of people whose daily lives are intertwined, reflect the environments and cultural practices common to them [37,38]. There are many recent ethnographic research examples that have explicitly incorporated this concept into their analyses [39–41]. The term local biologies was first developed as a result of research into menopausal experience in Japan [42]. This seminal work made use of both qualitative and quantitative findings, based on a sample of over 1200 female informants distributed equally among the USA, Canada and Japan, and demonstrated that significant differences existed in biological symptom reporting in connection with the menopausal transition among these women as a result of their different histories, societies, cultures, diets, and environments [37,43–45]. One particularly salient discovery was that menopausal women in Japan report few experiences of hot flashes, whereas women in Canada and the USA report significantly higher levels. Follow-up in-depth ethnographic research conducted in these communities strongly suggested that lifelong dietary practices likely explain much of the difference in biological symptoms among postmenopausal women in Japan and North America [46].

It is important to emphasize that a focus on local biologies is not meant to reinforce any ideas about inherent biological differences between different groups or communities. Instead, it shifts the focus to how environmental variables – diet, toxic exposures, psychosocial stressors, access to healthcare, climate change and so on – give rise to biological differences in a dynamic process over time and in specific locations. It does not refer to the idea that members of a specific community (e.g., American–Indians or African–Americans) share inherent biological traits that differ from other communities. Instead, it highlights how local biological differences are diversely distributed in a heterogeneous fashion even within these groups as a result of different local ecologies. It may therefore be useful to think of these as ‘situated biologies,’ which are not inherent to any given locality or people [47].

We believe that using ethnographic and epigenomic methods together is a particularly powerful way that researchers can investigate and understand the mechanisms through which local biologies emerge. What follows below are illustrative, ethnographically informed accounts that provide concrete examples of the ways in which local environments and histories can give rise to local biologies.

Intergenerational transmission of toxin exposure

Ethnographic work conducted by Gammeltoft has documented the devastating effects that the chemical defoliant Agent Orange continues to have on reproduction in Vietnam more than 40 years after the Vietnam War [48]. During the war, the US military conducted an aerial defoliation program as part of a forced urbanization strategy to force peasants to leave the countryside and move to the cities dominated by US forces, spraying nearly 20 million gallons of chemical herbicides and defoliants and destroying all plant material in 2 days.

Agent Orange contains the highly toxic chemical dioxin, which is known to have long-lasting effects on the environment and human biology. Today, at least 3 million people in Vietnam suffer from serious health problems due to defoliant exposure and the rate of severe congenital abnormalities is higher in exposed versus unexposed individuals [48]. The media has even reported cases of third-generation Agent Orange victims, in which individuals exposed during the war have produced apparently healthy children whose grandchildren are born severely disabled. Some families, reluctant about abortion and longing for a healthy child, are raising three or four children with birth defects and deformities. The most common of these is hydrocephalus (an accumulation of cerebrospinal fluid in the brain), which can cause brain damage. A range of severe illnesses are also associated with dioxin exposure, including deadly cancers, Parkinson’s disease and spina bifida, in addition to those associated specifically with pregnancy [48].

Gammeltoft’s interviews with affected families demonstrate that many people refuse to acknowledge that Agent Orange is the cause of an anomalous fetus detected by ultrasound or the birth of a horribly deformed child [48]. They are too worried that the stigma attached to families afflicted by Agent Orange all but ensures that finding marriage partners, even for those family members who are healthy, would be virtually impossible. It is more socially acceptable to claim publicly that the anomaly resulted from a cold that the mother had or the heavy work that she did while pregnant.

Through this powerful example, we can already see the complicated interplay and feedback between the intergenerational transmission of toxin exposure; stigma; social, historical, and political factors; and the establishment and perpetuation of biological differences and abnormalities across generations. This interplay of factors would likely be missed in a standard epidemiological study of birth outcomes in Vietnam, where many participants would be likely to falsely reply ‘no’ to a measure asking whether they or their family members have been exposed to dioxin because of the stigma associated with such a label.

Historical trauma among Indigenous & First Nations Canadians

The case of Indigenous and First Nations Canadians further highlights the ways in which ethnographic inquiry can uniquely excavate the viewpoints of individuals and communities in order to contextualize the relationship between their local histories and local biologies (Figure 1). As of the 2016 census, Canada is home to over 1.6 million individuals who endorsed the category ‘Aboriginal’. Well over half of these individuals live in communities that continue to endure the devastating legacy of settler colonialism, including entrenched poverty, toxic environments and discrimination. Substance dependence, depression and violence are widespread [49]. Suicide is estimated in some Inuit communities to be six-times the rate in other parts of Canada [50]. These high rates of pathology are intertwined with a complex experience of historical trauma and exposure to environmental toxins.

Although levels are on the decline recently, since roughly the 1960s the Arctic has seen the accumulation of persistent organic pollutants (POPs) such as dioxin and polychlorinated biphenyls [51]. Over decades, these POPs have permeated the Arctic food web and accumulated in the body fat of Arctic marine mammals, which are the main traditional food source of indigenous Arctic peoples [52]. Consequently, individuals in Arctic Inuit communities have been shown to have greatly elevated levels of blood POPs [51]. As one politically active Inuit grandmother put it, “When women have to think twice about breast feeding their babies, surely that must be a wake-up call to the world” [53].

Beyond these more recent assaults, Indigenous leaders are also acutely aware of the effect that colonization of their lands, beginning 5 centuries ago, has had on the health of their communities. The concept of ‘historical trauma’ has been adopted to call attention to the collective and intergenerational psychosocial effects that this past colonial subjugation carried forward to the present day [54]. Among the early travesties was the introduction of infectious diseases like tuberculosis, measles and smallpox between 1700 and 1900. Exposure to measles among Indigenous peoples proved to be particularly deadly. By 1915, the population of the Haida Nation, to give one example, had declined to just 588 people as a result of smallpox and other diseases brought by Europeans [55]. Given that Indigenous communities’ subsistence was based on hunting, the ability of those who survived to procure sufficient food for the larger community was effectively destroyed and led to starvation among those who survived [56].

Further added to these insults was the establishment of residential schools throughout Canada. These schools, over 130 in all, were found throughout the entirety of Canada. School administrators, members of the Royal Canadian Mounted Police and agents from the Bureau of Indian Affairs gathered and sent young children great distances from their homes in order to be housed in institutions where they were not permitted to speak their own languages or participate in anything regarded as cultural practice [57]. Medical and nutritional experimentation was also carried out on some of the First Nations students at residential schools. There was neither informed consent nor parental notification for these experiments, and many children suffered from sickness, malnutrition and death [58,59]. Tuberculosis was also rampant, with few attempts made to curb it [60]. In one residential school, the death rate of children was 75% during the first 16 years of the school’s operation [54,57].

As part of the Truth and Reconciliation Commission (TRC) of Canada that met from 2008 until 2015, it was also discovered that repeated sexual abuse took place in these schools, with one investigating Supreme Court justice characterizing a particular school as practicing ‘institutionalized pedophilia’ [57]. Overall, the TRC final report found that the residential school system amounted to ‘cultural genocide’ [61]. These residential schools, the last of which was only closed as recently as 1996, are regarded among First Nations and Inuit communities as the primary source of their current suffering as a community.

Today, nearly half of First Nations people live on reserves [62]. Shockingly, 53% of First Nations children living on reserves live in poverty (41% for those living off reserve) [63]. Indigenous peoples make up less than 5% of the Canadian population, but Indigenous peoples represent 26.8% of federal prison inmates (41.4% of all federally incarcerated women) [64]. Furthermore, since 2010, the Indigenous portion of the incarcerated Canadian population has grown by more than 50% overall and 74% for in-custody Indigenous women [64]. Suicide rates, substance abuse and the disappearance and death of young Indigenous women continue to be extraordinarily high [65]. These children have usually been removed from their homes due to what is described by social workers as neglect, while the impoverished environments in which their families live remain unaddressed. In 2015, the Canadian government launched a ‘National Inquiry into Missing and Murdered Indigenous Women and Girls’, which published its final report in 2019. The report found that, ‘colonial violence, as well as racism, sexism, homophobia and transphobia against Indigenous women, girls and Two Spirit, lesbian, gay, bisexual, transgender, queer, questioning, intersex and asexual (2SLGBTQQIA) people, has become embedded in everyday life…this amounts to genocide’ [66].

As this example demonstrates, better understanding the health of Indigenous and First Nations Canadians in order to identify potential remedies for this kind of structural disadvantage should be a top priority. However, research attempting to engage with these communities to understand the complex interplay of historical trauma, societal disadvantage and discrimination, environmental toxin exposure and mental illness will need to deeply engage with these communities on a qualitative level in order to understand how to formulate research questions and survey measures for larger epidemiological and epigenomics work that accurately reflect what is at stake in their daily lives.

Toward a new bio-ethnography: engaging with the experiences of under-represented communities for a new environmental epigenetics to address health disparities

The stories and experiences of disenfranchised individuals and communities remain invisible to us unless we look. It is the power of ethnography to reclaim the stories and experiences of communities such as these and place them within the contexts that community members themselves claim to be most salient to their lives and health. It is the power of biological and molecular research to then show us how these powerful experiences impact different biological pathways and contribute to the development of disease. Crucially, ethnographic findings are limited in their impact and usefulness for developing public health interventions to empower and revitalize communities without being able to characterize and quantify deeper biological changes occurring in response to these environments. As Gisli Palsson remind us, “Humans may usefully be regarded as fluid beings, with flexible, porous boundaries; they are necessarily embedded in relations, neither purely biological nor purely social, which may be called ‘biosocial’; and their essence is best rendered as something constantly in the making” [67].

The term bio-ethnography has been discussed previously by the anthropologists Roberts and Sanz [10]. As they describe it, bio-ethnography is “a research platform that combines data derived from biological and ethnographic methods to arrive at a better understanding of the larger histories and life circumstances that shape health, disease and inequality.” As such, it allows for an understanding of “how phenomena (i.e., disease or biological change), including geopolitical processes and participant’s life conditions, are created through intra-active looping” [10].

Anthropologists have recently begun engaging with and conducting research on epigenetics, particularly with regards to the pathways through which sociocultural experiences such as poverty, trauma, psychosocial stress and social inequality influence health [68–75]. And a small subset of recent anthropological genetics work has also successfully made use of ethnographic interviewing [76] to develop culturally relevant measures of stress to be assessed in relationship to DNA methylation of stress-related genes [69,70]. The type of bio-ethnography that we are calling for is somewhat different from this, however, and refers instead to the incorporation of ethnographic methods in large-scale, population-based molecular epidemiology and epigenomics research that often takes place in large cohort and consortia study populations. It is this latter approach that has not been accomplished to date.

In this section, we outline our preliminary suggestions for what the methodologies of such a bio-ethnography might look like. Some of the approaches that we outline below may be more relevant to one study design or set of hypotheses, whereas another may be more appropriate in another setting. We are not proposing a blanket approach. Rather, we are merely suggesting a range of possible transdisciplinary considerations and research strategies that could be used individually or in combination for disparities-focused epigenomics research that helps link identified epigenetic changes to broader, multifactorial influences experienced across different communities. These methods will need to be tested and refined in actual practice, and future researchers will no doubt devise new, creative ways to combine environmental epigenetics and ethnography that have escaped us here.

The imperative for transdisciplinary research

In contrast to multidisciplinary or interdisciplinary work, this new bio-ethnography calls for transdisciplinary research. Multidisciplinary work can be characterized as a process where researchers work in parallel, but separately, from their specific disciplines to address a common problem, whereas interdisciplinary work refers to when researchers work together but still from the perspectives of each of their own distinct disciplines. Transdisciplinary research, on the other hand, requires that ‘researchers work jointly using a shared conceptual framework that draws together disciplinary-specific theories, concepts, and approaches to address a common problem’ [77]. If we think of a ‘continuum of collaboration’ along which these research approaches fall, transdisciplinary research requires the highest degree of synthesis [78]. It requires that methods be fully synthesized across disciplines to develop new frameworks, concepts, and methods. As Roberts and Sanz describe it, bio-ethnography allows us to “ask new kinds of research questions that could not be asked with ethnographic or biological data alone” [10]. This means that transdisciplinary bio-ethnography research also demands real work on the part of all members of the research team to learn and integrate new methods and concepts from other disciplines.

Heinzmann et al. [77] have written a particularly useful summary of how a transdisciplinary research approach is optimal for engaging with American–Indians to understand social determinants of health in those communities, noting that it should be characterized by a participatory approach; focusing on real-world, socially relevant issues; and transcending and integrating different disciplinary paradigms.

The question of generalizability

A common critique of ethnography as a method has been that it cannot provide results that are generalizable to larger populations, as good science always strives to do. We argue, however, that what ethnography lacks in generalizability to larger populations it more than makes up for with the depth of new knowledge it generates about the specific, local community being studied.

The sampling frame for the ethnography component in bio-ethnographic work needs to be a random subsample from the larger cohort or population being interrogated in the epidemiological or epigenomics work. It also needs to encompass a clearly defined study population that matches that of the larger epigenomics study. If multiple sub-populations are present in the larger study (i.e., different self-identified racial/ethnic or socioeconomic groups or participants of different geographical ancestry), then a random subsample will need to be taken from each. If this is done, then the smaller sample sizes that are practically feasible for ethnographic work are more than ample for reaching thematic saturation [79,80] in qualitative analysis (i.e., the point after which further ethnographic data collection generates no new information). Once thematic saturation is reached, these themes and findings can then also be discussed with community members to verify whether they reflect their experiences. If these steps are followed, the results will be generalizable to the specific, local community being studied.

These results, however, will not be generalizable outside of the community studied to more general population groups, nor will they be generalizable to members of the same community outside of the specific locale or environment in which the research took place. Again, we argue that this greater specificity of ethnography is a unique strength that will help to disrupt preconceptions about homogeneity within populations. It will also prove indispensable in identifying and understanding the most salient issues affecting the health of under-represented and disenfranchised communities, as experienced and articulated by community members themselves.

Bio-ethnographic methodologies

Below we describe the different contributions that ethnographic methods can make to environmental epigenomics studies throughout the research process to improve environmental epigenomics research (Figure 2). At the very least, ethnographic interviewing can be utilized in all of the methods discussed here. However, we recommend deeper ethnographic work to uncover less obvious experiences and factors that investigators may be able to observe but informants may not be able to articulate themselves. In epigenomics work where investigators are using previously collected data from cohort studies or consortia, it is often possible to get permission to recruit cohort participants for follow-up qualitative work. This means of course that ethnographic activity will not be able to inform survey development efforts for existing cohort data, as these will have already collected by cohorts, but it can still be used to better contextualize and interpret results and even inform the development of new survey items to be administered in those cohort or consortia populations.

Generating new knowledge to create de novo survey measures that capture exposures & experiences unique to the study community

Ethnography is perhaps most powerful as a methodology to ascertain whether one is asking the right questions. Especially when working with under-represented communities, researchers will not always be able to anticipate the most important experiences and environmental influences that have shaped the lives and biologies of their study participants. Nonetheless, it will be crucial to identify and measure these experiences and analyze them in relationship to biological variables if the aim is to better understand the production of health disparities. When designing their research project or the development of a new cohort study, investigators can propose an initial ethnographic phase with a random subsample of the enrolled participants. Even if the investigators already have hypotheses about environmental variables that may influence the biological variable or health condition they are interested in, this ethnographic phase can still be proposed as initial empirical work to potentially verify or augment the scope of environmental exposure variables to be tested in their epidemiological research. This ethnographic phase would then be carried out after participant enrollment, but before any baseline survey data, environmental data or biospecimens are collected.

While it is clearly not feasible to conduct ethnographic interviews with each individual from a large epigenomics study, especially when sample sizes need to be on the order of thousands or tens of thousands, conducting this smaller scale ethnographic work at the outset of the study in a random subsample of the larger study population(s) is an optimal method for generating new knowledge and discovering which environmental influences are most salient across the various sub-populations to be included in the study. In studies involving different sub-populations according to: socioeconomic status, race/ethnicity, genetic ancestry or some other category; ethnography can be used to identify the most salient experiences, perceptions and themes in each community, which will help to define the scope of exposures to be included in the study.

In this initial ethnographic phase of the research project, investigators can ask participants to describe the experiences, both past and present, that have had the greatest impact on their lives and on their community. This early-stage ethnography may very well uncover new, salient experiences and exposures that the researchers will not have thought to measure. If validated survey measures of these experiences already exist, these can be added to the study’s main questionnaire. If no survey measures of these experiences exist, then de novo survey items could be developed to capture these important but unmeasured aspects of the psychosocial and physical environments of the intended study population. These measures developed in the ethnography component do not limit or define the scope of what needs to be investigated in the epidemiological portion of the research, but merely augment and add on to the other exposure variables that investigators have identified a priori as of interest to analyze. These new survey items can then be administered to the entire study population alongside other study questionnaires, and then quantified and analyzed in relationship to biological variables such as DNA methylation.

Contextualizing knowledge to assess the adequacy & relevance of existing questionnaire measures to the particular community being studied

During the same initial ethnographic phase described above, ethnographic work can also act as a method of cognitive testing. For example, a molecular epidemiology study team may have identified existing survey measures that they would like to use in their study questionnaires and subsequently analyze in relationship to epigenetic factors, but these measures may not have been psychometrically validated in their specific study population. In this case, an initial round of ethnographic interviewing could be strategically deployed to understand how the meanings of those survey questions differ across important sub-populations included in the study, and whether these measures are capturing what the researchers intend to capture among diverse populations. Ethnographic methods may also uncover specific language that different communities within the study population may use to describe a particular phenomenon that differs significantly from the wording of a validated measure, leading investigators to either tailor these measures or use new measures that better reflect the language and understanding of the specific population.

This process of assessing the validity of existing survey measures is especially important in disparities-focused studies, as validated survey scales are often developed in white or majority populations and fail to capture important exposures among minority populations. For example, recent anthropological research has shown that existing survey scales do not capture the complexity of stress and inequality experienced by minority populations in the USA such as Hispanic/Latina women [81].

Another example comes from qualitative research and focus groups conducted by two of the authors (MA Argentieri and A Shields) on religion and spirituality among American–Indians in the Dakotas. Religious service attendance is often used as the default measure of religiosity in epidemiological research because it has been robustly associated with mortality in many cohort studies [82–84]. Our work has demonstrated, however, that because of the poverty on American–Indian reservations, many families differ in which church services they attend, if at all, depending on the social services or material support particular churches might be providing at any given time. In this context, traditional measures of church attendance would completely fail to capture the intended construct of belonging to a religious community. Rather, we learned that participation in traditional sweat lodges or other traditional American–Indian rituals better captured the essence of belonging to a spiritual community for many American–Indians we interviewed. Based on prompting from community spiritual leaders and community members, we included sweat lodges as one of the response options in assessing communal religious activities in our larger epidemiological research.

Contextualizing & interpreting epigenomics study results

Ethnography may also be usefully carried out after the study team has completed their main epidemiological and epigenomics analyses to aid in interpreting study results. This may be particularly useful where unexpected results appear for a defined sub-population of the study sample. Ethnographic interviews with a random subset of study participants post-study may reveal hitherto unidentified sources of mismeasurement, where respondents believed a question was asking about content different from that which the investigators intended or where perhaps respondents were offended by a question and thus did not answer it at all or selected ‘not applicable’.

To return to the example of Lock's research on menopause in Japan, many researchers and Japanese doctors at the time were attempting to attribute the low numbers of hot flashes reported in Japan to a perceived shyness among Japanese women, who they understood as likely to ‘play down’ their discussion of bodily symptoms [42]. Lock’s ethnographic work showed, however, that women were open to discussing their bodies but were simply not experiencing the same biological symptoms as North American women. Conducting ethnography in this way therefore also allows scientists to challenge their social and cultural assumptions about the populations that they study. These assumptions are often what lead researchers to ask the wrong questions, develop incorrect interpretations, or ignore significant aspects of life when working with under-represented communities.

Probing deeper into significant variables to inform future research

Ethnography can also be used to probe deeper on a specific environmental variable found to account for a significant degree of variation in the biological variable of interest (e.g., DNA methylation or risk of disease). Through ethnographic interviews with a random subsample of the study population after biological analyses are complete, researchers can dive deep into understanding the larger historical, cultural, economic and social context in which a significant exposure used in the epigenomics study becomes important to participants’ lives and health. In this way, ethnography not only facilitates greater contextualization and interpretation of results, but also allows researchers to identify the structural issues at play in a given issue, thereby opening up expanded horizons for imagining health interventions that might engage a particular social/environmental influence in a meaningful way across diverse populations or local communities. These insights gleaned from ethnographic methods can also meaningfully inform the design of future epigenomics research aimed at analyzing the effect of different understandings of the same phenomena in relation to changes in DNA methylation.

It is equally important to note that results from environmental epigenomics work can also inform future ethnographic and qualitative work with the study population. Such biological analyses can reveal a number of different biological and physiological patterns and changes among the study population that participants were unaware of, and thus unable to verbalize to the ethnographic investigators. The results from these analyses can give researchers a more complete picture of which types of social and environmental experiences contribute to biological variation and health differences among their participants, and thus which new topics and interview questions might be important to include in future ethnographic work with the study community.

Transdisciplinary funding opportunities as a mechanism for change

We anticipate that many epidemiologists and epigeneticists will at first be reluctant to invest time and resources in developing a bio-ethnographic study design. Here, we think that funding organizations have a large role to play, as a particular barrier for many researchers will likely be a lack of relevant funding to pursue this new research agenda. While transdisciplinary research programs and Requests for Applications (RFAs) such as the NIH Transdisciplinary Tobacco Use Research Centers (TTURCs) [85], the National Institute on Minority Health and Health Disparities’ (NIMHD) Transdisciplinary Collaborative Centers for Health Disparities Research Program (TCC) and the Canadian Institutes of Health Research's (CIHR) Strategic Training Initiatives in Health Research [86] have now become more common, scientific research teams have nonetheless been slow to incorporate the epistemological and methodological perspectives of the social sciences. Those working in anthropology, disparities research and ethics have occasionally been incorporated into genomics research teams, but few large-scale genomics research efforts include anthropologists or have incorporated the disciplinary insights and methods of anthropology, which relies on ethnography as a core method for producing new knowledge. We believe the availability of an even greater number of RFAs from major funding institutions that require transdisciplinary approaches bridging the physical and social sciences will be a strategic way forward to ensure that these future collaborations take hold.

Conclusion

A focus on ethnography is not merely a focus on social or cultural variables. There is a danger in ‘conflating structural violence with cultural difference’ [87], where focusing too heavily on cultural difference as the driver of health disparities will lead investigators to mistaken assumptions and blind them to the basic historical, economic and structural factors that contribute to the health and illness of a community [88,89]. The type of rich ethnographic inquiry that we have argued for instead helps researchers contextualize informants’ lives in relation to these broader trajectories. It is a deeply participatory method that calls for empathy and engagement with communities about what’s at stake in their lives. In doing so, it engenders trust with communities, unveils novel exposures to investigate in epigenomics work, and helps to enrich, contextualize and deepen those measures of the environment already in use. These features of ethnography become all the more paramount when undertaking research with under-represented communities, whose stories and experiences are often far less visible to and appreciated by those from outside their community.

Reciprocally, these ethnographic methods and insights will be significantly strengthened and augmented by the biological and molecular insights that can be understood from larger-scale, computational epigenomics work using population-based data. Environmental epigenomics analyses, for instance, can elucidate which experiences, identified through qualitative work with members of the study community, seem to have the greatest biological impact and thus might be worth exploring further in future ethnographic investigation. The bio-ethnography we are calling for will help anthropologists to better understand human biological diversity through the methods learned from epigenomics. This new transdisciplinary method will therefore not only be mutually beneficial for all disciplines involved, but will also facilitate more meaningful engagement with under-represented populations in a way that respects their expertise as key informants in their communities and thereby builds authentic and trusting relationships between researchers and members of the community under study. Such relationships are essential to the success of any future efforts aimed at improving the health of underserved communities.

Ultimately, while we focus on particular ways that ethnography and epigenomics can be employed together in constructing a new bio-ethnographic approach to disparities-focused research, such an approach is not limited to these two disciplines. Rather, we argue that the intellectual and methodological resources of multiple different disciplines are needed to optimally study the complexity we seek to understand. Collaborative models of research that take seriously the concepts, methods and insights gleaned from multiple social and biological sciences to develop a new transdisciplinary research paradigm will, we believe, have the best chance of generating new insights and identifying new leverage points to address health disparities and inequality.

Future perspective

We have outlined practical recommendations for integrating ethnographic methods into large-scale, epigenomics research in order to understand the impact of environment on health in under-represented communities. These recommendations will need to be tested and refined in cohort- and consortia-based research projects in order to optimize them for research environments and study populations in which large-scale epigenomics research projects are commonly conducted. Future RFAs and funding opportunities that not only sponsor collaboration between (epi)geneticists and social scientists, but require these collaborations for disparities-focused research as a condition of funding, will prove most effective in encouraging in-depth transdisciplinary collaboration. It is also well documented that minority populations are not well represented in (epi)genomics research [18,90], and thus it will be even more important that program announcements require substantive engagement with minority communities as study populations in order to better understand how their local environments influences local biological changes and disease.

Ultimately, this new bio-ethnographic approach will be a crucial step in better measuring and understanding how specific adverse social, economic, historical and physical environments influence dynamic biological and genetic responses within disadvantaged populations, and how these translate into racial/ethnic and socioeconomic disparities in the burden of disease.