25.2.1 Data: What Are We Collecting, What Are We Measuring?

High-quality data that are fit for purpose and meet the criteria of accuracy, validity, completeness, and consistency are a cornerstone of empirical science. Data quality may be affected at the stages of data collection, cleaning, or the numerical transformation that is often used to meet required assumptions such as the normal distribution in statistics. Measurement errors, whether systematic or random, are present in all observational studies. While these errors impact the validity and reliability of data and introduce biases, they are rarely acknowledged or accounted for in the epidemiological literature. Makin and de Xivry address common statistical mistakes [Reference Makin and de Xivry14], and Wagenmakers et al. [Reference Wagenmakers, Sarafoglou, Aarts, Albers, Algermissen and Bahník15] present guidelines on how to report statistical analyses transparently based on four scientific norms of ‘communalism, universalism, disinterestedness and organised skepticism’.

Since we allude above to the notion of variable choice and availability, next, we discuss the importance of clear terminology and data quality in DOHaD research methods. How we define and measure exposures and outcomes impacts inferences, findings, and subsequent interventions and policies. One example of this is the work of researchers who rely on clinically defined groupings based on dichotomisation, such as diabetes diagnosis, or the classification of body mass index (BMI) as obese/overweight/normal-weight/underweight. One obvious risk of using a strict classification of body morphology by BMI alone is undermining the field’s knowledge about fat distribution being a strong determinant for metabolism and cardiovascular health, especially fat within the abdomen (visceral adiposity). Without other markers to corroborate metabolic health risks (blood pressure, cholesterol, visceral fat mass, etc.), some individuals with normal weight, categorised as ‘controls’, may be metabolically unhealthy and ‘at-risk’ of physiological phenotypes. In fact, this group represents 35 per cent of normal-weight individuals [Reference Fan, Qiu, Zhao, Yin, Li and Wang16]. This problem extends to gestational diabetes screening in pregnancy, which is provided to women meeting the BMI > 30 kg/m² criteria in the UK, while those under 30 are assumed to be void of any hyperglycaemia risks during pregnancy. The absence of evidence, however, is not evidence of absence. The consequence of such hidden (latent) subgroups of individuals within a ‘control’ or ‘normal-weight’ category is the introduction of bias into the statistical analyses that epidemiological models rely on for inference, thus leading to inaccurate conclusions.

Similarly, in psychology, the diagnosis of autism as present/absent is common, although autism spectrum disorder (ASD) is typically conceptualised by experts as a continuum (see also Azevedo et al. in this volume). Such a binary diagnosis of ASD ignores potential distinct mechanisms of importance for the DOHaD of autism subtypes, which could possibly relate to the timing of any ‘disruption’ in brain development and could be informative for mechanistic studies and prognosis [Reference Lai, Kassee, Besney, Bonato, Hull and Mandy17].

Overall, what we suggest above is that DOHaD researchers must be conscious of relying on clinical data alone such as those retrieved from EHRs without clarifying sources of biases, the caveats of present/absent dichotomised diagnoses [Reference Altman and Bland18], and the local clinical guidelines from which they derive. We suggest, however, that some classification approaches are available to limit some of these caveats such as collating multiple variables and produce profiles based on similarities of exposure and/or outcomes at one time point (e.g. latent class modelling) or many time points over time to uncover trajectories (latent class growth analysis and piecewise modelling). This could mean, for example, retrieving glucose measures sampled throughout pregnancy and establishing the likely glycaemic status rather than relying only on a single GDM diagnosis. These approaches also help identify profiles of individual responses to interventions and can therefore improve tailored treatment allocation.

Additionally, missing data, either by design (e.g. unmeasured exposures/outcomes) or attrition, negatively impact data quality and challenge causal inference. This can be addressed by powerful analytical tools that recognise data complexity and the impact of missing data [Reference Stavola, De Stavola, Nitsch, dos Santos, McCormack and Hardy19, Reference Johnson20]. Several methods are available to address missing data, including maximum likelihood, multiple imputation, and Bayesian methods. Complete case analysis leads to loss of data and statistical power but is widely used, while other complex but more justifiable methods are not often attempted [Reference Bell, Fiero, Horton and Hsu21]. Assumptions about the properties of missing data, whether these are missing completely at random or not, must be made. With the emergence of Big Data and EHRs that tend to have a high prevalence of missing information, appropriate techniques that deal with missing data need to be carefully applied.

25.3.1 Limitations of Current Big Data Sources and Applications

One first potential caveat of relying on Big Data sources it that DOHaD researchers wanting to use Big Data may be obliged to formulate research questions based on data availability or including data not necessarily designed primarily for DOHaD research. These researchers will have less control over data quality because of the larger distance from data collection, that is the inputting user (clinician for EHR / hospitals, or user of a phone app), and from the decisions made in defining and measuring the variables in these data sets. Overall, sources of error need to be considered when evidence from Big Data is evaluated. Without researchers’ involvement in data collection, it may be impossible to subsequently correct or even identify these errors.

The task of comparing and validating DOHaD models across populations may be further hindered by the heterogeneity in data architectures across national and international sources. Before the term Big Data surfaced, omics-derived data alone (e.g. genomics, transcriptomics, and proteomics) already inferred the outputs of millions of data points [Reference Lapehn and Paquette3]. Formulating a cascading model of these omics layers, which follow biologically downstream from one another, is both necessary and extremely complex. Further, linking biologically derived material to clinical data of different formats, and based on a variety of measurements, including imaging, questionnaires, and diagnoses, requires technologies that facilitate multidimensional integration. Thereafter, powerful methods that support analysis are necessary.

Larger sample sizes improve the power to detect effects, and clearly the whole DOHaD framework requires both large samples and a comprehensive set of exposures and events to be modelled. However, the primary issue is that complex models are more difficult to explain and thus could complicate their practical translation into actionable policies. Users and clinicians equally need to be versed in their use and interpretation.

The use of Big Data also raises issues of data security and representation, particularly data obtained outside conventional academic institutions, in contexts where systems and resources are not fit for this purpose, such as in low- and middle-income countries (LMICs). Users of both healthcare services and digital platforms, such as social media, may represent distinct groups with possibly little overlap in demography, risks, and healthcare needs [Reference Delpierre and Kelly-Irving22]. It is plausible to assume that LMICs are unlikely to have population-wide health records or access to digital data collection and remote health monitoring from which DOHaD modelling could be done. Here, validation and reproducibility of DOHaD models are less feasible and highlight the lack of representation through their exclusion.

25.4.1 Issues of Interpretation and Reporting

Concerns are often raised about the scope, complexity, transparency, reproducibility across different scientific teams and different populations, and the interpretability of prediction models. While AI operates from a ‘black box’ within deep neural networks and unsupervised learning, the biological plausibility and meanings of the output are generated by researchers. Given that so far only a few published prediction models have found utility in clinical practice, the utility of AI when compared to conventional methods remains an open question. It is unclear whether AI can address the questions of causality most pertinent in DOHaD, when DOHaD draws from interpretability and theory and is moving towards the integration of social science and ethnography. (See Richardson, in this volume for a discussion of how DOHaD is characterised by ‘cryptic causality’.)

Guidelines regarding AI are developing rapidly. For example, following a quality assessment of the conduct and reporting of multi-variable prediction models, a 22-item checklist (TRIPOD) was developed [Reference Collins, Reitsma, Altman and Moons34]. The risk of bias tool (PROBAST-AI) for diagnostic and prognostic prediction model studies based on AI has also emerged to ensure that users have key information about the design, conduct, and analysis, alongside a robust standardised tool for bias evaluation that would allow a fair judgement on the utility of these models [Reference Zheng, Xie, Xu, He, Zhang and You31, Reference Sounderajah, Ashrafian, Golub, Shetty, De Fauw and Hooft35, Reference Collins, Dhiman, Andaur Navarro, Ma, Hooft and Reitsma36]. Guidelines should be consulted by authors, reviewers, and editors, to ensure reproducibility, reliability, and validity, and hence safe implementation. Again, even prior to applying AI-derived analytics, the uncertainty of measurements in Big Data and its sources of error must be accounted for and possibly identified systematically, and the data quality validated.

27.1.1 On Terminology

Genetics research and autism studies have a complicated history, so we begin by establishing our choice of terminology and rationale for this. We acknowledge that there are strong and often polarising views about the issues presented in this chapter; however, we hope that we can contribute to meaningful discussion.

The principle ‘nothing about us without us’ communicates that decision-making that impacts a particular group should not take place without the full and direct participation of its members [Reference Charlton1]. Thus, it is crucial that individuals be referred to using their preferred terminology. Person-first language evolved in the 1970s to separate the person from the descriptive trait, for example ‘a person with autism’ rather than ‘an autistic person’, and to give primacy to their identity as a person. Although this is well-intentioned, some disability activists have noted that this forced separation between person and trait reinforces the idea that disability is inherently negative and ignores the integral role that disability plays in shaping a person’s character and experience. As such, there has been a move towards identity-first language.

This terminology is by no means ubiquitous. A person-centred approach to language recommends that on an individual level, words that people use to self-describe should be prioritised.Footnote ¹ For coherence, we have chosen to use ‘autistic person’ here, except where a direct quote incorporates other terminology. This is reflective of the preferred language identified by many autistic individuals and autism self-advocacy organisations [Reference Dwyer2].

27.2.1 Models of Disability

Two dominant models of disability are often contrasted in literature and practice. The medical model ties disability directly to the body, focusing on possible interventions to bring it to a particular type of functioning [Reference Shyman5]. The social model situates disability in the social context and physical environment of the individual and is focused on identifying barriers that prevent full participation in society. The latter model differentiates between impairments – attributes impacting how the body and brain operate – and disabilities – restrictions imposed by societal standards that reflect normative ideas of how bodies should function. A third model is ‘neurodiversity’, a term first coined by autistic sociologist Judy Singer and popularised by Steven Silberman in his book, Neurotribes, in which he defines it as follows:

Like the social model of disability, neurodiversity emphasises the disabling nature of stigmatisation and the prioritisation of brains classified as ‘normal’.

27.2.2 Disability in Research

To de-pathologise disability requires engagement with disability communities and scholars in developing frameworks from research design to knowledge translation. Sometimes referred to as participatory or community-engaged research [Reference Wallerstein, Oetzel, Sanchez-Youngman, Boursaw, Dickson and Kastelic7], evidence indicates this approach contributes to better health and social outcomes. It is critical to respect the contributions of disabled scholars, activists, and organisations and promote collaboration between disabled and non-disabled researchers, and disabled participants and their advocates. Participatory research means an increasing understanding of disabled individuals as co-creators of scientific knowledge, rather than passive subjects.

There is, understandably, hesitation in disability communities regarding participation in medical research. As with many vulnerable communities, the history of unconsented research and other research harms is long and fraught [Reference Bhambra8]. Community members are quick to spot ableist rhetoric and stigmatisation in research documentation and are reluctant to participate if their bodies, lives, and experiences may be used to pursue goals not aligned with the expressed needs of disabled individuals. Thus, participatory research does not begin with inviting disabled individuals as research subjects, but rather, with listening, learning, humility, and trust-building on the part of non-disabled researchers, using the principle of co-design and through participant advisory groups.

27.4.1 The Social Construction of Autism

Criteria for autism in the Diagnostic and Statistical Manual (version 5) include but are not limited to ‘persistent deficits in each of three areas of social communication and interaction plus at least two of four types of restricted, repetitive behaviours’ [31]. This deficit model, at times focused on the external viewer’s perception of autistic experiences, typically shapes research seeking to minimise these behaviours and accompanied distress. By contrast, neurodiversity-focused groups frame autism as a constellation of strengths and challenges across social and sensory spectra and focus on research and resources to support autistic individuals in achieving their best quality of life [32].

Another current view considers autism as a potentially disabling condition that nevertheless may confer various positive traits [Reference Stevenson33]. However, in many cases this view has merely re-circumscribed capitalist values of productivity, for example celebrating those autistic traits, such as hyperfocus, that can be exploited by employers to improve work output. While this view has partly enhanced our understanding of neurodiversity and the need for better neuroergonomics in the workplace, the ultimate focus has not been on promoting quality of life for autistic people. Moreover, other autistic traits that are considered neutral or positive within the autism community, such as stimming to self-soothe and express emotions, are still misunderstood as negatives or viewed with discomfort.

The experiences of autistic individuals in healthcare provide an opportunity for examining pathologisation of their condition via scientific research into DOHaD and epigenetics. These research areas rely heavily on the interpretation of links between social and other environmental factors with biological outcomes. Perhaps the most widely recognised image of the autistic individual is that of a white, masculine-presenting child with an inability to make eye contact, limited or stilted speech, and a fascination with patterns or trains. This perception has recently begun to shift, prompted in part by an increase in later-in-life diagnoses in cisgender women as well as non-binary individuals and transgender men and women, who may present differently from this stereotype.Footnote ²

Autism and diversity of gender identities and experiences overlap substantially, further impacting access to appropriate support and care. Gendered differences in presentation have led autistic girls and women to be underdiagnosed, misdiagnosed, or diagnosed at a later age, sometimes only after their own child’s diagnosis [Reference Rynkiewicz, Janas-Kozik and Slopien34]. This discrepancy has contributed to a lack of understanding of key mental health conditions that co-occur alongside autism, including eating disorders, depression, anxiety, and suicidality. Similarly, disparities in the impact of race and ethnicity on the timing and frequency of diagnosis have led to a paucity of resources and support for racialised/ethnic minority autistic youth, who at the same time experience increased risks and rates of police and other state-sanctioned violence and incarceration [Reference Eilenberg, Paff, Harrison and Long35]. There is an urgent need for an intersectional approach to all disability research, but particularly epigenetic studies examining the social and environmental contexts for the lived experiences of autistic and other disabled individuals.Footnote ³

27.5.1 Biomarker Development for Conditions Classed as Disabilities

We are still far from having reliable predictive or diagnostic genetic or epigenetic biomarkers for conditions such as autism. One major factor that clouds the interpretation of such research is study design. Most classify autism as one entity, whereas it is a highly heterogeneous condition. Furthermore, co-occurring conditions such as ADHD are largely ignored in such studies. Some researchers have turned away from a categorical to a dimensional approach to the origins of autism, using continuously variable dimensions such as anxiety, attention, sensory processing, specific interests, repetition, social interaction, and communication [Reference Ure, Rose, Bernie and Williams36]. We suggest that this method is preferred because it targets traits that can be clinically defined and can identify areas of strength as well as areas in which autistic individuals may require understanding and assistance. This approach also captures intersecting dimensions of co-occurrences, such as ADHD, for example, sustained attention.

We suggest that future studies be based on dimensions of autism with a view to meeting the self-determined needs of autistic individuals and the autism community. A dimensional approach also reflects the spectrum of neurodiversity within and outside the autism community and the reality of the social model of autism rather than the medical model.

As the field of epigenetics moves towards identifying more biomarkers for conditions and associating these with developmental, social, and environmental correlates, the rhetoric surrounding curative approaches to disability could increase. This rhetoric is closely tied to medical and deficit models of disability, with their foundational assumptions that people with disabilities wish to be rid of the disabled parts of themselves. For some, this may be true; the existence of the social and neurodiversity models does not detract from the struggles precipitated by certain features associated with disability, such as chronic pain, anxiety, loss of quality of life, or early mortality. Rather than attempting to categorise disabilities wholesale as ‘bad’ (e.g. where we may aim to repair or alter the body or brain) or ‘good’ (where we may instead target disabling factors in the society or environment), a more useful account would examine components of disability that are unwanted by the individual who experiences them. Again, following the principle of ‘nothing about us without us’, it is important to differentiate between calls for prevention and cure that come from researchers and healthcare providers, and policies based on the lived experiences of disabled individuals and their advocates. In doing so, a stark divide can appear between the expectation of the disabled experience and the reality.

It has long been argued that health economic metrics, such as the quality-adjusted life year (QALY) or disability-adjusted life year (DALY), are not sensitive to the real experiences of disabled people and ignore the significant adaptive ability of individuals [Reference Grosse, Lollar, Campbell and Chamie37]. Inviting more conscious consideration of disabled peoples’ own experiences of their disability can also avoid the tendency to objectify disabled persons’ bodies and view them as separate from the disabled experience. This helps avoid the risk of ignoring meaningful needs assessments conducted by the community. In other words, embracing a neurodiversity model does not mean neglecting to provide support for autistic individuals who consider certain traits to be personally disabling or undesirable. Ethically, it is important to remember another classic phrase in the disability community, coined by autism advocate Dr Stephen Shore: ‘If you’ve met one person with autism, you’ve met one person with autism’ [Reference Flannery and Wisner-Carlson38]. Again, the diversity of manifestations and personal experiences can only be incorporated effectively into epigenetics and genetics research if studies are co-designed and guided by diverse members of the autism community.

27.5.2 Direct-to-Consumer (DTC) Epigenetic Tests

In the past five years, there has been a growing number of companies selling epigenetic tests directly to consumers, that is without the need for a referral from a healthcare practitioner [Reference Dupras, Beauchamp and Joly39]. Despite being unable to define a ‘healthy’ epigenome, DTC companies focus on identifying epigenetic biomarkers in consumers’ blood or saliva samples with the promise of enabling consumers to improve their health outcomes. An ‘altered’ epigenetic status could indicate early-life exposures that increase the likelihood of developing certain conditions, which could be targeted for intervention due to the potential reversibility of epigenetic changes. Identifying environmental risks for the development of a condition also means that prevention strategies could be adopted to reduce the chance of its development, for example, via diet and exercise changes. In the case of autism, there are currently tests being developed to facilitate diagnosis in children as young as 18 months old [40]. Here, the promise is to provide biological data to complement more subjective analyses to expedite autism diagnoses and access to early intervention and resources.

However, DTC epigenetic testing raises various ethico-legal issues, related to the core technical issue surrounding the precision of epigenetic biomarkers for diagnosing complex conditions. Marketing that overestimates the reliability of epigenetic test results could exploit consumer trust in science to sell a product that falls short of its promises. Test results could also affect an individual’s access to insurance policies, particularly life insurance, as the reporting of test results is often a legal obligation of the applicant. With a focus on environmental risks, there is also a tendency to blame individuals for the development of associated conditions. Here we acknowledge the long history of blaming mothers for autism, a fact well demonstrated by the term ‘refrigerator mothers’Footnote ⁴ often used to describe them [Reference Lappé19]. While parents of children with autism seem to support the development of epigenetic testing for improvement of the diagnosis process [Reference Wagner, McCormick, Barns, Carney, Middleton and Hicks41], there is a need for clear regulations of the DTC market to protect consumers, especially vulnerable populations.

27.5.3 Ethical, Legal, and Social Implications of DoHAD Research for People with Disabilities

From an ethical perspective ‘respect for persons’ is one of the fundamental tenets of Western biomedical ethics. Its application in DOHaD research is often more complex than in standard clinical care [Reference Beauchamp and Childress42]. Core ethico-legal issues here include maintaining confidentiality and privacy, and gaining informed consent for medical interventions, which work together to promote autonomy. While protecting sensitive information such as medical diagnoses and treatment decisions may be relatively simple within the practitioner–client communication paradigm, genetic information, for example, might be problematic for the individualistic Western ethico-legal model. (See Karpin in this volume.) From a DOHaD perspective, genetic and epigenetic information might best be conceived of as family information, making privacy concerns more complex. However, the rationale behind protecting this information remains the same: promoting autonomy, including through avoiding potential coercion from those who would misuse sensitive information to discriminate against individuals. The latter is relevant for accessing employment, health and life insurance, and healthcare services. Whether DOHaD and epigenetics research should aim for family or community, rather than individual consent, falls beyond our scope here, but we recognise that when a test impacts more than the individual, there is potential for social harm against others who are impacted by the results. For research on communities with disabilities, especially those with a potential genetic contribution, this suggests the co-design of research studies is important to ensure knowledge about inheritance is not weaponised against the community or used to engage in blaming or labelling of parents or offspring.

Confidentiality is a key pillar in the doctor–patient relationship protected under common law and statutory regulation. For example, privacy laws in Australia governed under the Privacy Act 1988 (Cth) have a broad reach, protecting a range of information, including health information. According to the ‘For your information: Australian Privacy Law and Practice’ ALRC Report No 108) [43], ‘privacy’ covers several aspects, including data protection, such as medical and government records; bodily privacy, such as invasive procedures that may include genetic and epigenetic tests; and communication, such as emails.

The disclosure or privacy of sensitive genetic information in some instances might be problematic. For the disability community, it is possible that epigenetic data collected from one consenting individual or family may have immediate relevance to other community members. For this, the right ‘not to know’ might be as important as the right to know the genetic factors involved in the development of autism. Importantly, once these data exist, they may have wide-ranging impacts on members of the community who did not consent to the research.

27.5.4 Incorporating Perspectives from Multiple Stakeholders

A goal of disability activists has been to reframe conversations about disability, health, and disease away from views that centre concepts of ‘normalcy’ and ‘functionality’ and to instead centre the disabled individual as the core stakeholder in the discussion of their own body and experience. Throughout the twentieth century in particular, the concept of ‘wellness’ came to be equated with ‘virtue’, situating the body as a ‘site for moral action’ [Reference Conrad44] with regard to the pursuit of health. The medical model, in addition to enforcing the idea of a ‘normal’ state of the body to which its owner should aspire [Reference Squier45], increasingly pushed a ‘functionality’ argument that privileged a body’s capacity to contribute to labour [Reference Mitchell DTS46], and disdained disability precisely because of the implication that the disabled individual is of inherently reduced worth under capitalism.

The term ‘stakeholders’ is suggestive of a consumer-driven approach to health and well-being that places disabled individuals immediately at a disadvantage [Reference Migliaccio47]. As a result, the stakeholders most often centred on disability research have been medical practitioners and families of disabled individuals. While both caregivers and practitioners have a significant interest in the disability conversation and valuable experiences to contribute, at times, this has come at the expense of the voices and narratives from disability communities and their advocates. In autism research, this has contributed to frustration and conflict. As this research moves forward, it must include autistic individuals and their caregivers where necessary (as participants, researchers, and scholars) at the centre of the conversation from research development to knowledge translation.

28.2.1 The Colonial Project as an Origin of Ill health and Disease

It is well known that colonisation and colonialism are determinants of Indigenous ill health [Reference Paradies2]. Colonisation by nations including Britain, France, Spain, and others resulted in the mass genocide and/or displacement of Indigenous peoples across the world, and the methodical dismantling of Indigenous lifeways, undermining millennia-long connections to land, culture, and kin. Colonialism, as enacted through historical and ongoing contemporary colonial processes, encompasses a range of risk factors for health and well-being [Reference Mitchell, Arseneau and Thomas3]. This includes interpersonal, institutional, and political processes that, at a minimum, aim to control or dominate a population and, at the extreme end of the spectrum, aim to eliminate or exterminate the population.

In Australia, it is estimated that around 90 per cent of the Aboriginal and Torres Strait Islander population lost their lives between 1788 and 1901 because of widespread colonial violence, introduced diseases, and the theft of land and water resources by British settlers [Reference Williams4, 5]. Those who survived were subjected to systematically racist and discriminatory programmes and policies. Many were dispossessed of their traditional lands and forced to live on Christian-run missions, which often forbade the use of any language other than English and actively suppressed important cultural practices. One particular assimilation policy enacted from 1910 to 1970 in Australia led to 11–24 per cent of all Aboriginal children being stolen from their families by government agents and held in Christian-run institutions for ‘re-education’ and training as domestic servants for the White middle classes [6]. They are known collectively as the Stolen Generations. Colonisation has thus generated significant intergenerational trauma for Indigenous peoples, reinforced by ongoing and cumulative ‘biosocial injury’ inflicted by settler-colonial nation states and individuals [Reference Warin, Kowal and Meloni7].

Settler-colonialism imposes cultural values, religions, laws, and policies that often go against the rights and values of Indigenous peoples. The concept of settler-colonialism as a health risk factor is well understood within Indigenous populations [Reference King, Smith and Gracey8–Reference Czyzewski10]. However, in Australia there is limited ability to examine such links due to an absence of epidemiological data on settler-colonial exposures. This itself is a manifestation of racism, given the lack of research priority and attention to settler-colonialism and its impacts. White racism and discrimination against Indigenous peoples thus perpetuate colonial trauma.

A small number of studies demonstrate poorer outcomes for Aboriginal and Torres Strait Islander peoples who experience individual settler-colonial factors such as discrimination [Reference Priest, Paradies and Gunthorpe11–Reference Thurber, Colonna and Jones15] and the Stolen Generations [16] versus those who do not. However, most, if not all, of these studies are small-scale, use measures that have not been validated, and are not population representative. In addition, there are no quantitative studies that identify factors that can act as a buffer against the negative impacts of settler-colonialism. This is critical in taking the next step past identifying associations and towards identifying practical, strengths-based solutions to guide policy and programme development.

Research conducted by Indigenous peoples, for Indigenous peoples, and with Indigenous peoples is a direct challenge to the ongoing colonisation of our lands, cultures, and communities. As Tuck and Yang [Reference Tuck and Yang17] stated, ‘Decolonization is not a metaphor’. It unsettles the settlers. Indigenous peoples have always conducted research, and we have always used counting as a tool for this research. This fact alone frequently unsettles non-Indigenous researchers in the field of epidemiology who often state that issues of importance to Indigenous peoples, such as culture (the practising of it or revitalisation of it), cannot be quantified. The centring of Indigenous peoples in epidemiology therefore means accounting for settler-colonial-inflicted biosocial injury and centring Indigenous definitions of health and well-being and the determinants thereof within research.

28.2.2 Indigenous Definitions of Health and Well-being

Indigenous identities and concepts of health and well-being are fundamentally connected to the land within an eco-centric relationality or kinship system that defines social and spiritual obligations to family, community, and the land [Reference Dudgeon, Bray, Adams and van de Vijver18, p. 200]. This relationality is reinforced through cultural frameworks where relationships to the past, ancestors, land, and the present are articulated through language, song, dance, storytelling, and maintaining traditional homelands, beliefs, and kinship [Reference Salmon, Doery and Dance19]. As a result, Indigenous definitions of health and well-being are inherently holistic. The Medicine Wheel, or Sacred Hoop, has been used across Turtle Island (North America) to convey Indigenous philosophies of well-being using four quadrants (sometimes represented by the four cardinal directions – North, South, East, and West) to represent the connectedness between the physical, emotional, mental, and spiritual elements of life and the need for balance across these four areas. In Aotearoa (New Zealand), Māori holistic health and well-being also relies on a balance across four fundamentally interconnected elements – wairua (spiritual), whānau (extended family network), hinengaro (the mind), and tinana (physical) [Reference Durie20, p. 1141].

In the Australian context, the most used definition of Aboriginal and Torres Strait Islander health is as follows:

In recent years, several studies have focused on expanding this definition. Garvey and colleagues [Reference Garvey, Anderson and Gall22] outlined their Fabric of Aboriginal and Torres Strait Islander Wellbeing model based on qualitative research with 359 participants, weaving together eight aspects of well-being: culture; community; family; belonging and connection; holistic health; purpose and control; dignity and respect; and basic needs. Butler et al. [Reference Butler, Anderson and Garvey23] conducted a comprehensive national literature review to identify nine interconnected domains of Aboriginal and Torres Strait Islander well-being, including autonomy, empowerment, and recognition; family and community; culture, spirituality, and identity; Country; basic needs; work, roles, and responsibilities; education; physical health; and mental health. Salmon and her colleagues [Reference Salmon, Doery and Dance19] undertook an international literature review to identify and describe six key cultural domains essential for Aboriginal and Torres Strait Islander well-being: connection to Country; Indigenous beliefs and knowledge; Indigenous language; family, kinship, and community; cultural expression and continuity; and self-determination and leadership. Maintaining and reviving connections to land and culture has been found to be protective of Indigenous health and well-being across a range of studies in Australia and internationally [Reference Salmon, Doery and Dance19, Reference Bourke, Wright and Guthrie24].

28.2.3 Indigenous Rights to Data

Indigenous definitions of well-being and its determinants have not been reflected in large-scale epidemiological data collections. For decades, statistics about Aboriginal and Torres Strait Islander communities have been wholly based on the perspective of the White settler-colonial state [Reference Lovett, Prehn and Williamson25]. Hundreds of Aboriginal and Torres Strait Islander cultural groups were homogenised under the umbrella term ‘Indigenous’ for comparison to the (equally homogenised) ‘non-Indigenous’ population. This comparison was based on White settler-colonial definitions of health, social, and economic achievement. Sociologist Maggie Walter (Palawa) argued that the result of this comparison was the production of ‘5D Data’, emphasising the Difference, Disparity, Disadvantage, Dysfunction, and Deprivation experienced within the Aboriginal and Torres Strait Islander population [Reference Walter, Kukutai and Taylor26]. She writes:

When data stressing the ‘Indigenous problem’ present in Australian society are used to inform public health and policy development, the inevitable outcome is a re-colonisation of Aboriginal and Torres Strait Islander communities through programmes and policies designed to correct these perceived deficits. Interventions built on this deficit premise are prone to failure, as has been emphasised time and again by the Australian Government’s own monitoring scheme called Closing the Gap, established in 2008. The ‘Gap’ refers to the statistical gaps highlighted by the above-mentioned data, particularly the difference in life expectancy at birth between the Indigenous and non-Indigenous population, which currently sits at 7.8 years for females and 8.6 years for males [27]. According to Government statistics, over half (53 per cent) of the health ‘gap’ between Aboriginal and Torres Strait Islander Australians and the rest of the Australian population is accounted for by just five social determinants (employment and hours worked, highest non-school qualification, level of schooling completed, housing adequacy, and household income) and six ‘health risk factors’ (binge drinking, high blood pressure, overweight and obesity status, inadequate fruit and vegetable consumption, insufficient physical exercise, and smoking) [28]. The remaining 47 per cent of the health gap is currently unaccounted for.

Research seeking to measure the health of Indigenous peoples but exclude them from the development of such research is, quite simply, bad science. ‘Bad’ in this case may be interpreted in two ways. First, in the production of inaccurate, under-powered, and misleading data that, in Australia and Aotearoa at least, have informed decades of largely ineffective public health policy [Reference Kukutai and Walter29, Reference Walter30]. Indigenous data must have equivalent explanatory power to non-Indigenous data to achieve health equity [Reference Reid, Paine and Curtis31]. Second, it is now widely considered to be unethical to exclude Indigenous peoples from the development of research about their communities. The United Nations Declaration on the Rights of Indigenous Peoples (UNDRIP), which enshrines the right to self-determination for Indigenous peoples [32, Article 3], helped to spur a change in the way Australian human research ethics committees considered applications. Many major research ethics and funding bodies now require applicants to actively partner with Aboriginal and Torres Strait Islander individuals, organisations, or communities involved in the proposed research [33].

To counter the deficit discourse of Indigenous health that has been informed by the 5D data approach, Indigenous Data Sovereignty (IDS) and Indigenous Data Governance (IDG) protocols and principles have been developed for research based on the UNDRIP. For First Nations communities in Canada, this has been solidified through Ownership, Control, Access, and Possession (OCAP®), which became a registered trademark of the First Nations Information Governance Centre (FNIGC) in 2015 [34]. In Aotearoa, the Kaupapa Māori approach has been practised for decades, which upholds Māori self-determination and ways of knowing, doing, and being in research communities. In Australia, possible protocols and principles were discussed at a meeting in 2018 with delegates from the Maiam nayri Wingara Indigenous Data Sovereignty Collective and the Australian Indigenous Governance Institute. They defined Indigenous data as ‘information or knowledge, in any format or medium, which is about and may affect Indigenous peoples both collectively and individually’ [35]. IDS and IDG were defined as follows:

‘Indigenous Data Sovereignty’ refers to the right of Indigenous people to exercise ownership over Indigenous Data. Ownership of data can be expressed through the creation, collection, access, analysis, interpretation, management, dissemination, and reuse of Indigenous Data.

‘Indigenous Data Governance’ refers to the right of Indigenous peoples to autonomously decide what, how and why Indigenous Data are collected, accessed, and used. It ensures that data on or about Indigenous peoples reflects our priorities, values, cultures, worldviews, and diversity.

In line with these definitions, they determined that Indigenous peoples in Australia have the following rights [35]:

• To exercise control of the data ecosystem including creation, development, stewardship, analysis, dissemination, and infrastructure.

• To have data that are contextual and disaggregated (available and accessible at individual, community, and First Nations levels).

• To have data that are relevant and empower sustainable self-determination and effective self-governance.

• To have data structures that are accountable to Indigenous peoples and First Nations.

• To have data that are protective and respect our individual and collective interests.

Supporting these rights generates ‘good data’, which is the antithesis of 5D Data. The concept of good data extends existing global conversations around ethical data and data justice and incorporates IDS and IDG principles to describe a resource that Indigenous communities may use to address their self-determined interests and needs [Reference Lovett, Lee, Kukutai, Daly, Devitt and Mann36].

It is only since 2020 that Aboriginal and Torres Strait Islander representative organisations have been meaningfully engaged by Australian governments in the development of the Closing the Gap agenda and goals. Part of this engagement involved the recognition that Aboriginal and Torres Strait Islander peoples have the right to define our own needs and priorities in the policy arena. This partnership was formalised through the National Agreement on Closing the Gap (the Agreement) that commits to a genuine partnership between all Australian Governments and the Coalition of Aboriginal and Torres Strait Islander Peak Organisations [Reference Song and Chung38]. The goal of this partnership will be to improve the life outcomes of Aboriginal and Torres Strait Islander people, acknowledging that supporting and strengthening Aboriginal and Torres Strait Islander cultures is necessary to achieve this goal [37, p. 4]. Large-scale epidemiological data that reflect Indigenous lifeworlds and reinforce IDS and IDG will be key to driving this national agenda.

28.3.1 Overview of Mayi Kuwayu

Cohort studies, a type of analytical study, have made a considerable contribution to our understanding of human health and have been at the forefront of identifying the influence of social, environmental, and biological processes on health and well-being outcomes. The goal of analytic studies is to identify and evaluate the causes or risk factors of diseases or health-related events [Reference Song and Chung38]. Cohort studies often involve identifying a ‘group’ of people to study and plan the research in advance, collecting data over time. Epidemiology, as the study of the patterns and distribution of disease and arguably health, is often the mechanism or method used in telling the story about the presence or absence of disease and what the likely relationships are between social, environmental, and other exposures under study. Due to their long time frames, expense, and difficulties in proving causation, cohort studies are uncommon when compared to other study designs.

Indigenous cohort studies are not common in settler-colonial states such as Australia, Aotearoa, and Turtle Island (referred to collectively as CANZUS) where Indigenous peoples are often incidentally recruited into studies. This limits the ability to conduct robust analytical studies specific to Indigenous groups, and, critically, the variation in Indigenous populations within CANZUS countries presents numerous challenges. The result is a ‘tyranny of the majority’ and evidence production that is biased towards the majority population while being underpowered for minority groups within those same cohorts. Additionally, these cohorts do not include risk exposures unique to minority groups, nor do they include exposure to unique potentially protective factors.

When the practice of epidemiology is led by Indigenous peoples its potential to contribute towards the achievement of health equity increases exponentially. Mayi Kuwayu is Australia’s largest longitudinal cohort study of Aboriginal and Torres Strait Islander well-being. As of January 2024, over 12,000 individuals have participated in the study that surveys Aboriginal and Torres Strait Islander-identified social and cultural determinants of health and well-being [Reference Lovett, Brinckley and Phillips39]. The Mayi Kuwayu baseline questionnaire was developed in consultation with Aboriginal and Torres Strait Islander community members, organisations, and research experts over the course of four years from 2014 to 2018 and is revised every two to three years. It includes a range of metrics that reflect community-identified determinants of health and well-being, such as connection to Country (e.g. ‘How much of your life have you lived on your tribe’s (mob’s) Country or Island?’) and cultural knowledge and practice (e.g. ‘How much time do you spend learning culture, kinship and respect?’). Mayi Kuwayu data adhere to IDS and IDG principles, and all requests for access to and use of the data must be approved by the Mayi Kuwayu Data Governance Committee, which is a group of Aboriginal and Torres Strait Islander community members and external researchers.

The intent is to conduct world-first analytical work to provide a robust understanding of how settler-colonial risk factors undermine the health and well-being of Aboriginal and Torres Strait Islander peoples, and how culture, a core strength of Aboriginal and Torres Strait Islander communities, can mitigate these adverse effects. This future work aims to understand some of the unexplained 47 per cent health inequity currently experienced by these communities in Australia, while also identifying how the very essence of Indigeneity (cultural maintenance, strengthening, and expression) is fundamental to improving health and well-being and reducing inequities [Reference Dudgeon, Milroy and Walker40, 41]. This work is critical in that settler-colonialism as manifested in historical and contemporary trauma is likely to have a profound impact on adult and intergenerational health.

28.3.2 Mayi Kuwayu Data and Relationship to DOHaD

The Mayi Kuwayu study was developed in response to calls from Aboriginal and Torres Strait Islander peoples to ensure health and well-being concepts were appropriately measured and captured. It is underpinned by a social epidemiological framework, concerned with the influences of social structures, institutions, and relationships on health and well-being [Reference Jones, Thurber and Chapman42, Reference Krieger43]. Therefore, the study is designed to enable the examination of health and well-being, taking into account the varied contexts in which Aboriginal and Torres Strait Islander peoples live, including diversity in exposure to settler-colonial factors, and diversity in opportunities to engage in cultural practice and expression. Further, the study is ideally placed to explore and quantify if, and to what extent, culture buffers the impact of settler-colonial risk factors [Reference Jones, Thurber and Chapman42].

Since its first release in 2018, analyses of the data collected have shown a positive relationship between connection to Country, culture, and health and well-being outcomes in relation to Aboriginal Ranger work in Central Australia [Reference Wright, Yap and Jones44]. Being employed as a Ranger, who uses cultural and environmental knowledge to engage in land management activities on Country, was significantly associated with very high life satisfaction and high family well-being [Reference Wright, Yap and Jones44]. Preliminary analyses of the full Mayi Kuwayu cohort show that key cultural indicators such as spending time on Country, speaking traditional languages, passing on family knowledge and traditions, and feeling in control of one’s life are protective against high psychological distress, diagnoses of anxiety, and low life satisfaction [Reference Lovett45].

Mayi Kuwayu has also identified a range of settler-colonial exposures experienced by Aboriginal and Torres Strait Islander people and developed measures to capture participants’ exposure to these factors from the individual to the systemic level [Reference Lovett, Brinckley and Phillips39, Reference Jones, Thurber and Chapman42]. These exposures have been classified as either Indigenous Historical Trauma (IHT) or Indigenous Contemporary Trauma (ICT). Indigenous Historical Trauma items included in the study are the following:

1. 1. Tribe/mob forcibly relocated to missions or reserves

2. 2. Being unsure of which tribe/mob you belong to

3. 3. Having a parent who was part of the Stolen Generations

4. 4. Having an aunt or uncle who was part of the Stolen Generations

5. 5. Having a grandparent or great-grandparent who was part of the Stolen Generations

ICT items include the following:

1. 1. Feeling disconnected from culture

2. 2. Being dislocated from Country

3. 3. Worrying about being stolen when they were growing up

4. 4. Growing up in foster care

5. 5. Growing up in a children’s home

6. 6. Having children removed in the past 12 months

7. 7. Experiencing interpersonal discrimination

8. 8. Being a part of the Stolen Generations

9. 9. Having a cousin who is part of the Stolen Generations

10. 10. Having child/ren who are part of the Stolen Generations

11. 11. Having grandchild/ren who are part of the Stolen Generations

Over 60 per cent of Mayi Kuwayu participants report at least one exposure to IHT and 85 per cent report at least one exposure to ICT [Reference Lovett, Brinckley and Colonna46]. Exposure to IHT is associated with significant increases in poorer psychological distress and poorer life satisfaction. Stronger links are observed between any experience of ICT and poorer psychological outcomes, poorer general health, and lower life satisfaction. More than half of the Mayi Kuwayu participants have experienced discrimination in some form, and this is significantly associated with a broad range of poor well-being outcomes, ranging from disconnection from culture to high blood pressure and alcohol dependence [Reference Thurber, Colonna and Jones47].

These findings strongly support Hoke and McDade’s [Reference Hoke and McDade1] argument that studies on the DOHaD must consider the broader contexts in which we live as well as the intergenerational and transgenerational events that impact our health and well-being. Centring Indigenous lifeworlds in this research has the extraordinary potential to reveal previously ‘hidden’ factors that either increase the risk for disease or support good health and protect against biosocial harm. With the understanding that such factors and their contribution to health and well-being are heterogeneous across populations, more investment could be made by research institutions and funders in training DOHaD researchers to develop measures for specific salutogenic or risk factors in diverse populations. Intergenerational and transgenerational studies also require secure, long-term funding to provide more robust evidence as it accrues across time.

29.2.1 Environmental Epigenetics

Environmental epigenetics denotes the study of changes in gene expression that occur without changes in DNA sequence [Reference Bollati and Baccarelli10]. Such changes may occur in response to environmental challenges to an organism, including both material (e.g. nutrients and toxicants) and social factors (e.g. discrimination and adversity). They operate through a number of mechanisms. The three most important currently known are methylation, histone- and RNA- modifications. Some of these changes have shown to be mitotically and meiotically heritable, that is they may propagate across generations. Many of the specifics of this field and its implications for DOHaD have been discussed elsewhere in this volume. I therefore keep this point brief.

Epigenetic research challenges the developmental origins of Western biomedical thinking in Mendelian genetics, Weissmann’s germ plasm theory, and – more broadly – the autonomous subject of Western modernity. If indeed heritable changes in germline functionality occur without DNA sequence change, individual human organisms are far more open to their past and present environments than so far appreciated. The DOHaD framework has begun to embrace these findings as they suggest mechanisms for phenomena that have so far only been shown through correlations [Reference Rosenfeld1]. The ‘tracking’ of physiological parameters over time from early life into adulthood might in fact be encoded at least to some degree in epigenetic processes. The details and some implications of this have been debated intensely over the last two decades. I want to focus on three lessons that follow from an epigenetics-inflected DOHaD understanding and that have received somewhat less attention.

First, ‘environments’ are diverse, dynamic, and never innocent. From my own fieldwork in a molecular biology lab working on environmental epigenetics, I have learned that epigenetic response mechanisms are far more subtle than the dominant digital logic of knock-out genetic thinking gives credit for [Reference Niewöhner11]. Oftentimes, the handling of rodents in experimental settings appears to produce stronger epigenetic changes than the actual substance whose epigenetic effects were under investigation. This demonstrates that it might well be very difficult to isolate individual ‘factors’ or ‘causes’ from a complex environment. Instead, material and social factors readily interact in manifold ways with an organismic epigenome (if that term even makes sense), which is in itself a highly dynamic system. Mental models of human–environment relations derived from carcinogenicity and acute toxicity assume a unidirectional and non-reversible dose–response relationship between the environment and human organism (see Rossmann and Samaras in this volume). Epigenetics on the other hand suggests a reversible (e.g. de novo de/methylation) and perhaps bidirectional relationship where dose–response is likely to occur in a non-linear fashion; if indeed it is the right model at all. Lastly, environments are never politically and ethically innocent. The proof of principle experiments around the Dutch Winter Hunger cohort [Reference de Rooij, Painter, Phillips, Osmond, Michels and Bossuyt12], the studies of licking and grooming behaviour under conditions of reduced nesting material and displacement [Reference Weaver, Meaney and Szyf13], or the forensic psychiatric reconstruction of life histories in child abuse and suicide completers [Reference McGowan, Sasaki, Huang, Unterberger, Suderman and Ernst14], all demonstrate that ‘environmental factors’ occur in politically, ethically, and socially charged settings that need to be appreciated in their economic, racial, and gendered complexity (see also the contributions by Meloni et al., Valdez and Lappé, and Cohen in this volume).

Second, while the notion of ‘origins’ in DOHaD was developed against genetic determinisms, it nevertheless still carries unwanted implicit remnants of Mendelian and Weissmannian biological temporality. Origins suggests a starting point that is defined if not genetically then still by some kind of non-human nature. This might not be a blank slate, but it is a starting point that is often reified through methodological designs that rarely reflect the constructed and contingent nature of study subjects – be they ready-to-study mice or Romanian orphans. The dynamic environmental conditions that have evolved historically and cross-generationally – from nutritional environments to political regimes – are seldom explored in social-ecological detail. Epigenetics makes us attentive to these conditions and thus to the historical and social contingency of any starting point. ‘All the way down’ [Reference Haraway15] in the body, we do not encounter some kind of pure biological matter. Rather the body is as much ecosocially entangled at the molecular level as it is at the organismic level. For ‘environments’, the scientific construction work and the ecosocial entanglements are even more obvious. Finding a plausible starting point for one’s research design is a question of cutting the network and making the cut accountable to the field [Reference Strathern16], one’s own discipline, and scientific practice at large. Epigenetics thus helps to challenge the idea of ‘origins’ as a largely unreflected, somehow natural starting point of a developmental process.

Third, for all the attention to environmental factors, the notion of development centres the framework on the human organism. That is perfectly acceptable as it is a framework in human medicine. Epigenetics, however, shows the human organism to be remarkably open to its manifold environments. And vice versa: the human organism contributes to making its own livelihoods and niches. Most ecologists today subscribe to the idea that organisms do not find or adapt to existing niches but that organisms and environments interact in co-producing niches [Reference Odling-Smee, Laland and Feldman17]. In social scientific terms, ‘niching’ as a material and social everyday practice might be the more apt analytic for these forms of world-making [Reference Bister, Klausner and Niewöhner18]. Perhaps, then, one ought to refer to ‘genealogies’ instead of ‘origins’ and ‘development’. This would help to address the multiple histories that run through any origin as well as the necessary contingency of multiple struggles of power/knowledge that mark any genealogy. Genealogies of Health and Disease: GoHaD?

Appreciating the multi-directionality of human–environment relations casts doubt on the developmental thinking in environmental factors, mechanisms, and linear causality for all but the most pervasive and drastic isolated health effects. Most human–environment interaction research, however, remains rooted in a thinking premised on distinct entities. It is either enviro-centric suggesting that the environment as an independent variable causes certain responses in the body or it is organism-centric and thus focused on how human action shapes the environment [Reference Williams, Zalasiewicz, Waters, Edgeworth, Bennett and Barnosky19] or how humans may act as niche modifiers [Reference Gluckman, Low and Hanson20]. These are all entity-based ways of thinking about human–environment interactions. They start from entities with certain characteristics (organisms and niches) that enter into interaction. One might also, however, start from the action and investigate how action produces entities. This results in a process-based approach [Reference Bapteste and Dupré21]. Drawing on the French philosophers Gilles Deleuze and Félix Guattari, British anthropologist Tim Ingold proposes to think of humans in environments not in terms of interacting entities but as ‘lines of flight’[Reference Ingold22]: ‘The line of flight, write Deleuze and Guattari, “is not defined by the points it connects, or by the points that compose it; on the contrary, it passes between points, it comes up the middle”’. What Ingold is essentially challenging his readers to do is think that humans and environments ‘should not be understood as interacting entities, … but as trajectories of movement, responding to each other in counterpoint, alternately as melody and refrain’. The result is a process- or practice-based biology in which organisms and environments are constantly in becoming and in which development occurs rhizomatically rather than along a linear path. The notion of development in DOHaD does not usually take this into account. It rests on the understanding of an entity that is exposed to an environment as a set of factors. I am not suggesting that lines of flight readily translate into biomedical research designs. Yet they present an important conceptual challenge that sensitises researchers to the fact that evolutionary, structural, and systemic thinking never quite captures the situational specificities of human practice and its effects. These require process-based approaches.

29.2.2 Microbiome Research

The human microbiome denotes the aggregate of all microorganisms living on or in human tissue or fluids. Research efforts to better understand the components and dynamics of the human microbiome have rapidly increased over the last decade. The human microbiome comprises around 10–100 trillion symbiotic microbial cells per human individual [Reference Ursell, Metcalf, Parfrey and Knight23]. They match if not outnumber human somatic cells, and their genetic material by far exceeds that of the ‘human proper’. Cells belong to around 500–1000 different species at any given moment within a human [Reference Gilbert, Blaser, Caporaso, Jansson, Lynch and Knight24]. The genetic diversity and hence flexibility of this crowd by far exceed that of human genetic material. Over the last ten years, the field of microbiome research has begun to transition from the description of components to mechanisms and to the tentative development of clinical interventions. It has also invited a rich scene of lay ‘bio hackers’ to self-experiment alongside the emerging science with everything from probiotic foodstuffs to faecal transfer. A scientific understanding of how the microbiome impacts human somatic and mental health and disease onset and progression directly, as well as through complex interactions with the immune, endocrine, and nervous system, is only just emerging. Yet already today it is becoming clear that microbiome research will be another insult to human narcissism and anthropocentrism. After Copernicus, Darwin, and Freud, microbiome research demonstrates that ‘man’ is not even somatically speaking the master in his own house. While the skin as ‘philosophy’s last line of defense’ [Reference Bentley25] remains intact if porous, inside that skin shell emerges a multiplicity of inhabitants and agencies in complex and finely balanced interaction and oftentimes symbiosis.

In the preceding section, I used epigenetics to question whether we should move from origins to genealogies and from development to lines of flight. Microbiome research extends this questioning. Anthropologist Myra Hird discusses how human subjectivity and social form need to be understood as also shaped by bacteria [Reference Hird26]. She demonstrates how deeply biological and economic notions of the self are enshrined in Western modern thought. Westerners think of themselves as individual cognitive agents that act autonomously, often in competition with each other to increase fitness. Society is often understood as synchronised individuals. The vast majority of biological and medical research designs presume the existence of human individuals who act autonomously and are structurally closed to their environments. The subsequent distinction between self and other (along the skin) is foundational to Western self-understanding. Taking bacteria and their actions seriously challenges this understanding. Hird focuses on the understandings of symbiogenesis and a very corporeal interdependence of human and microbial life. The human ‘I’ becomes a multitude or collective. Similar to Ingold, she thus arrives at a world in becoming that is made up of encounters: encounters, for example, between humans and microorganisms that then develop ways of co-existing. Continuous encountering is what makes us what we are. Hence ‘we’ are not only epigenetically open to ‘outside’ environments, but ‘we’ are also open to ‘inside’ environments in the form of microorganismic collectives and their respective habitats.

It is this continuity of encounters that DOHaD also needs to address. Whether framed as contagion, as multi-species thinking, or as making kin, health and disease can rarely be sensibly understood as states of a single organism isolated in interaction from its environment. Dose–response simply does not capture the fact that exposure occurs in dynamic patterns of encounters. Even in the simplest scenario of an isolated single toxicant having an impact on a human organism, it is not only human cells and organs reacting. It is a symbiogenetically evolved social organism that responds. And while this social organism exhibits a distinct meta-stability that most people readily accept as human subjectivity, responses to substances are multiple and differentiated. Substances that occur at levels in the environment that are commonly considered well below toxicity thresholds not only interact to complicate exposure. They interact in differential ways with the human microbiome such that effects may occur that might well then surface as a human health issue. Hence environments are never really environments for only one organism as Uexküll suggested. The ‘environment multiple’ is a direct result of understanding the human body as a multiplicity of encounters.

29.2.3 Planetary Boundaries and Planetarity

So far, I have addressed molecular, cellular, and organismic dynamics. Let me now briefly turn to planetary dynamics and how they might challenge DOHaD. Earth system science is understanding with increasing certainty that the planet’s capacity to sustain life as we know it has boundaries that we are already transgressing through human action [Reference Steffen, Richardson, Rockström, Cornell, Fetzer and Bennett2]. Environmental factors are thus ceasing to be local phenomena and instead need to be contextualised within planetary environmental change and its manifold repercussions for social-ecological systems across the globe. Philosopher Bruno Latour rightly challenges us to develop across all forms of scholarly activity a consciousness for the fragile and restricted conditions of habitability of the earth and life on it [Reference Latour4].

This planetary dimension challenges the DOHaD framework to understand ‘the environment’ and its health-relevant factors as part of earthly subsystems and its associated complexities and non-linear dynamics [Reference Gluckman, Low and Hanson20]. Major efforts have been underway for some years now to better understand and quantify both the global burden of disease [Reference Murray, Abbafati, Abbas, Abbasi, Abbasi-Kangevari and Abd-Allah27] and the bio-geo-physical and increasingly social dynamics of earth’s subsystems as well as their stable state boundaries [Reference Steffen, Richardson, Rockström, Cornell, Fetzer and Bennett2]. Suggestions are being made to integrate earth system science and global health through international consortia and (big) data approaches [Reference Whitmee, Haines, Beyrer, Boltz, Capon and de Souza Dias28]. Such approaches are commonly shaped by systems thinking and various forms of computational modelling that foster data-driven integration. The rise of earth system modelling from the late 1980s onwards has undoubtedly been an amazing process of knowledge production culminating in the 6th Assessment Report of the IPCC on the ‘physical science basis’ of global climate change. Never has a comparable global evidence machine been built of such scale and with such rigour. This evidence machine runs on a positivist epistemology that addresses ‘the planet’ through data aggregation and integration as part of system dynamics modelling. Its dominant if not its only mode of speaking about large phenomena (aka ‘the planet’) is through scaling up by means of data aggregation and integration. Everything else is anecdotal or an opinion, that is not considered evidence. This approach aligns well with the data-driven calculations of global burdens of disease and exposomes.

Yet feminist literary theorist Gayatri Spivak [Reference Spivak29] and others use the notion of ‘planetarity’ [Reference Spivak30] to point to the data-driven construction of the planet asking where this leaves significant differences in ways of thinking and being on this planet – differences to which the social sciences and humanities attend. An altogether different perspective on this new planetary dimension is thus possible and important. The social sciences and humanities have long developed conceptual and empirical alternatives to address large-scale phenomena such as globalisation. Thinking in scapes and flows [Reference Appadurai31], global assemblages [Reference Ong and Collier32], global entanglement [Reference Conrad, Randeria, Conrad, Randeria and Römhild33], and post- and decolonial critique [Reference Escobar and Mignolo34] approaches ‘the global’ very differently. These approaches start from significant social differences and ask how they spread, reach, infect, travel, transform, and resonate. These approaches either focus on the forces that make phenomena graspable as ‘global’ or they move around inside the phenomena understood as global, showing their heterogeneity and multiplicity. Globality is about differences and what these differences (can) mean for respective others. Planetarity, in contrast, is often about trying to produce the one true representation of a whole. Of course, globality has been concerned with primarily social dynamics, while planetarity is rooted in biogeophysical phenomena. In the Anthropocene, however, this neat division of labour is being challenged as social inquiry becomes interested in material dynamics and ontological questions, while modellers of physical systems are incorporating not only economic exchanges into their models but increasingly also social dynamics.

Latour’s ‘Terrestrial’ [Reference Latour4] might be a useful point of contact between these very different ways of addressing phenomena that span the world in various ways. It is these reconfigurations of socio-material relations and the ensuing debate of how to address them that form the context within which the DOHaD framework can make an important contribution. How do we conceptualise ‘environmental factors’ when they cease to be local phenomena; when we try to understand them within a terrestrial context? One response would be to scale ‘environmental factors’ up, for example, to produce global estimates of ‘novel entities’, that is ‘new substances, new forms of existing substances and modified forms of life’ [Reference Persson, Carney Almroth, Collins, Cornell, de Wit and Diamond8], and assess their impact on earth’s subsystems. Microplastics, lead, or persistent organic pollutants serve as examples. This might be related to the global burdens of disease in a data-driven integrative approach. In a very different approach, one might contextualise ‘environmental factors’ in highly political patterns of exposure related to colonial and racialised histories of exploitation and production, imperial debris, and the ruins of capitalism [Reference Nading35].

The Anthropocene and its demand to think in planetary terms, then, challenge biomedically and epidemiologically established notions of environment, environmental factors, exposure, dose–response, temporality, and scale. These notions are also embedded within the DOHaD framework. However, the framework does not inherently offer a singular and straight path to address these challenges. It is not a foregone conclusion that these challenges will be solved with data collected and analysed in the empirico-analytical frameworks of twentieth-century biomedicine. Rather, the DOHaD framework might afford a reflexive moment, a moment of producing ‘theory out of science’ [Reference Roosth, Schrader and Jentsch36], which might enable a diverse set of approaches to address anthropogenic challenges and to address more-than-human liveability on a fragile planet. The DOHaD framework might offer a space within which data-driven approaches might complement other approaches that open up environments to a politics of exposure and habitability and that understand the human body as historically shaped, socialised, and habituated in patterns of practice [Reference Roepstorff, Niewöhner and Beck37]. Realising this potential and situating DOHaD in this sense, however, requires a diversity of approaches.

29.3.1 Three Modes of Interdisciplinarity

Moving DOHaD into the Anthropocene cannot be achieved with a single logic or methodological approach. It would be a futile task to try to develop an overarching framework that can fully integrate the existing conceptual and methodological diversity, the desire to reduce and explain with the need to contextualise and interpret, and the strengths of rigorous data-based analysis with the strengths of critical inquiry. The present volume features this diversity, and it is clear that this does not cohere in an overarching heuristic or integrative framework – nor should it. Instead, it might be useful to distinguish between different forms of interdisciplinary engagement and forms of collaboration to give orientation and take away some of the pressure towards integration.

Andrew Barry and Georgina Born [Reference Barry and Born9], drawing on their investigation of using ethnography within various other epistemic cultures, usefully outline three modes of interdisciplinarity: subordination–service, integration–synthesis, and agonistic–antagonistic. In the subordination–service mode, the research question and design come from one lead discipline, while the other discipline delivers additional data to extend or deepen the analysis. This is a fundamentally asymmetrical approach shaped by one discipline. In the integration–synthesis mode, two different disciplines readily find ways of addressing problems that are of interest to both. Biochemistry is the typical example within the natural sciences. It is a little less obvious across the natural/social science divide. This approach is symmetrical with both disciplines staying within their comfort zone but addressing a new topic in a shared way [Reference Otto, Donges, Cremades, Bhowmik, Hewitt and Lucht3]. The agonistic–antagonistic mode is perhaps the most demanding. It arises when two disciplines differ in their understanding of the research object. Often, this disagreement is ontological in nature. For example, human differences might be considered a material and bodily phenomenon by biologists while social scientists would insist on it being primarily a social phenomenon constructed through the social interaction of subjects positioned in social space. In an agonistic–antagonistic mode, these differences between disciplinary perspectives are not levelled. Instead, they need to be worked with to turn them into something generative from which both disciplines can learn without necessarily agreeing to the respective other perspective. This approach is symmetrical but not as comfortable as in the integration mode. It is about letting oneself be irritated by other ways of thinking and designing research and by sustaining significant differences to learn and develop one’s own perspective.

In opening up DOHaD further to social science thinking, this form of agonistic interdisciplinarity will play a key role. It is important to note that co-laboration [Reference Niewöhner, Jouhki and Steel38] is possible: co-laboration is temporary joint knowledge production between two disciplines without necessarily having a shared goal. Two scholars from different disciplines might work together without necessarily getting a shared result. Rather they might take different results and insights away from the co-laboration and integrate those into their respective disciplines. To make this rather abstract typology of possible interdisciplinary research more tangible, I want to briefly sketch examples of research questions and designs for each of these three modes of interdisciplinarity that might help the current DOHaD framework embrace the challenges of the Anthropocene.

29.3.2 Subordination–Service

Environmental epigenetics offers an obvious entry point for subordination–service interdisciplinarity. Currently, research within the DOHaD framework tends to operationalise environmental factors as independent variables, for example social disadvantage, zip code, and nutritional status (cf. Liz Roberts in this volume). For many social situations and structural inequalities [Reference Charlesworth, Gilfillan and Wilkonson39], such operationalisations are not only too crude. They also miss the entire dimension of subjective and collective experience that social science would consider fundamental to the emergence of ‘the social’ and thus to pathways from ‘objective’ disadvantage to actually ‘living inferiority’ [Reference Charlesworth, Gilfillan and Wilkonson39] and associated individual bodily health and disease. Hence social science could contribute its understanding of social dynamics to research on social drivers within the DOHaD framework. At a time where many countries – and in particular the major metropoles – are undergoing fundamental transformations trying to meet climate targets, this kind of social science service work could also contribute to making sure that ambitious climate targets are not reached at the expense of increasing inequality and thus worsening individual and public health outcomes.

Subordination–service approaches might also work the other way around. The long-term social inquiry into living with chronic respiratory diseases, for example, might benefit from global to local climate projections, air quality modelling, and associated health data. The Anthropocene foregrounds the dynamics of nature–culture relations, making such social inquiries into ‘natural’ phenomena such as disease aetiologies even more relevant [Reference Timmermans and Haas40]. Planetary health is currently conceptualised largely in biomedical and public health terms with the social sciences adding knowledge about social dynamics. This could be – perhaps ought to be – turned around placing well-being and environmental justice at the centre and putting the medical disciplines in the service role.

29.3.3 Integration–Synthesis

Recent work at the interface of sociology and biogeochemistry presents an outstanding example of biosocial interdisciplinarity in the integration–synthesis mode [Reference Neubauer and Landecker6]. Hannah Landecker, a sociologist and historian of science, and Cajetan Neubauer, a biochemist, in rather serendipitous fashion, began to work on the planetary availability and bodily effects of methionine together. Methionine is an essential amino acid, that is it cannot be synthesised by the human body and needs to be taken up through food. Working together on the methionine metabolism both globally and within the body, the two perspectives together were able to ‘establish the scale and historical trajectory of the methionine industry and provide a preliminary model for tracing this amino acid through the food supply into the human body’ [Reference Neubauer and Landecker6]. The study shows how planetary-scale anthropogenic activity changes ‘environments’ and consequently also human metabolisms with so far largely unknown consequences. Human biology and ecology, that is both environment and body, are understood as anthropogenic. The DOHaD framework thus requires biosocial, or rather ecosocial, collaboration as it is the social science perspective that can reveal how environmental factors come to be what they are through anthropogenic activity, specifically through analyses of dominant political economies and ecologies.

Hannah Landecker reflects on the back of the methionine study: ‘I have always felt that my contribution has been to enable the asking of experimental questions, the parameterisation of models, and the forming of hypotheses that would not otherwise have been possible … What Harry Collins termed “interactional expertise” has also been important in helping teams of different kinds of knowledge practitioners recognise ways in which they don’t understand one another, or facilitating the synthesis of different modes of proof or reasoning….’ [Reference Landecker41] Landecker reports her interdisciplinary research as an example of integration–synthesis cooperation, that is both disciplines, biochemistry and history, contribute from within their comfort zone to arrive at a new question and analysis. This is cleverly done, and it is this approach that enables Landecker to conduct a historical and social study of chemicals of metabolic significance – as a social scientist, albeit one with deep knowledge of natural science and the dynamics of the particular field in question.

29.3.4 Agonism–Antagonism

Oftentimes, however, social science perspectives do not readily complement or integrate with existing natural science or medical knowledge. Take approaches to social dynamics as an example. Biomedical or public health operationalisations of social dynamics often do not resonate with state-of-the-art social scientific knowledge and critique. The reason for such dissonance often lies in profound differences in the understanding of the research object. Much medical expertise works with a concept of ‘the social’ that is based on individuals interacting within a value system or culture. The notion of the individual tends to be under-socialised, that is based on ideas of individual decision-making that might be found in behavioural economics. Whereas the notion of the ‘value system’ tends to be over-socialised, that is assuming a firm grip of an abstract but homogeneous set of values on the framing of individual behaviour. Social dynamics from most social science perspectives sit in between these two perspectives and often work with notions of agency, subjectivity, and practice that combine structural (‘value system’) and individual (‘decision-making’) elements in highly dynamic and reflexive ways.

The agonistic–antagonistic approach starts from such dissonances and tries to make them generative. Agonism here refers to the work of political scientist Chantal Mouffe, who argued for agonism as a democratic form that does not solely rely on consensus but rather is able to work with differences as a potentially positive democratic force. Agonism within interdisciplinary research then means working with and through differences rather than searching for an integrative framework. This takes some work between the right partners. Working with differences requires explicating methodological, epistemological, and ontological assumptions. Differences need to come out to enable research that searches for common ground while critically reflecting on one’s own disciplinary perspective.

In this regard, the DOHaD framework needs to address the question of how much it can allow environments and bodies to be ‘situated’ in a social science sense [Reference Niewöhner and Lock42]. As it stands, DOHaD often rests on a universal body that responds to environmental factors. While this rests on a perfectly plausible set of assumptions, it does appear to underestimate what anthropogenic biology means. If all-pervasive anthropogenic activity has begun to change environments and bodies in significant ways, the ‘universal’ body is perhaps not the most prudent assumption anymore. Of course, at some general level, most bodies share some basic structures and processes. Yet these structures and processes are shaped by widely different trajectories through individual life histories and start to differentiate in significant ways. Most study designs assume a universal body and investigate the response to exposure. Yet ‘exposure’ to chemicals, social disadvantage, infrastructural violence, and to colonial legacies is systematically patterned and has been so for many centuries. The ‘inward laboratory’ of the body adjusts and contributes to these patterns and to living in, with, and through these patterns in significant ways. Rather than starting from the universal body at a national or even global scale, is it not high time that we should also try to scale approaches a little differently? Sure, trying to assess the global burden of disease, the global presence of chemicals and novel entities, and planetary boundaries remains important. Systems thinking and data-based integration remain important, and attempts at measuring the exposome and conducting exposome-wide association studies will certainly be productive in many ways even if they are bound to never reach their objective [Reference Vermeulen, Schymanski, Barabási and Miller43]. Yet other exposure patterns also matter. We might start with landscapes and how they have been shaped by political histories and economies, social dynamics and forms of dwelling, as well as biogeophysical contexts of climate, topography, and ecology. Urban and rural landscapes [Reference Tsing, Mathews and Bubandt5] can be investigated in great ecosocial detail to better understand how they afford particular exposure patterns and how these contribute to shaping deeply situated bodies [Reference Pentecost and Cousins44]. Inside such landscapes, situated bodies are thoroughly historicised, socialised, and politicised and constantly in becoming. They are not simply local as they relate to many transnational exchanges and flows of people, goods, and information as well as planetary environmental change. Such ecosocial analyses of the situated genealogies of health and disease require the agonistic and antagonistic struggle between critical social science perspectives and rather more solution-oriented medical perspectives.