The bare bones of race

Research priorities and considerations for nutrition research: methods of sex and gender analysis for biomedical and nutrition research

For some 20 years, science funding bodies have been asking for the integration of sex- and gender-related factors into the content of research and innovation. The rationale for those requirements has been the accumulated evidence that sex and gender are important determinants of health and disease. The European Commission (EC) has been the first, since 2002, to seriously ask for the integration of sex and gender into research and innovation in the context of their multi-annual framework programmes. When introduced, this condition was not immediately applauded by the research community, who perhaps lacked training in methods for the integration of sex- and gender-related factors. The EC Expert Group on Gendered Innovations sought to fill this gap. This review describes the work of this international collaborative project which has resulted in the development of general and field-specific methods for sex and gender analysis and 38 case studies for various research domains (science, health and medicine, environment, engineering) to illustrate how, by applying methods of sex and gender analysis, new knowledge could be created. Since 2010, science funding bodies in Canada, the USA and several EU member states have followed the example of the EC issuing similar conditions. Although the effects of nutritional patterns on a range of (physiological and health) outcomes may differ for men and women, sex and gender analyses are rarely conducted in nutrition research. In this review, we provide examples of how gender is connected to dietary intake, and how advancing gender analysis may inform gender-sensitive policies and dietary recommendations.

Using generative AI to investigate medical imagery models and datasets

BACKGROUND

AI models have shown promise in performing many medical imaging tasks. However, our ability to explain what signals these models have learned is severely lacking. Explanations are needed in order to increase the trust of doctors in AI-based models, especially in domains where AI prediction capabilities surpass those of humans. Moreover, such explanations could enable novel scientific discovery by uncovering signals in the data that aren't yet known to experts.

METHODS

In this paper, we present a workflow for generating hypotheses to understand which visual signals in images are correlated with a classification model's predictions for a given task. This approach leverages an automatic visual explanation algorithm followed by interdisciplinary expert review. We propose the following 4 steps: (i) Train a classifier to perform a given task to assess whether the imagery indeed contains signals relevant to the task; (ii) Train a StyleGAN-based image generator with an architecture that enables guidance by the classifier ("StylEx"); (iii) Automatically detect, extract, and visualize the top visual attributes that the classifier is sensitive towards. For visualization, we independently modify each of these attributes to generate counterfactual visualizations for a set of images (i.e., what the image would look like with the attribute increased or decreased); (iv) Formulate hypotheses for the underlying mechanisms, to stimulate future research. Specifically, present the discovered attributes and corresponding counterfactual visualizations to an interdisciplinary panel of experts so that hypotheses can account for social and structural determinants of health (e.g., whether the attributes correspond to known patho-physiological or socio-cultural phenomena, or could be novel discoveries).

FINDINGS

To demonstrate the broad applicability of our approach, we present results on eight prediction tasks across three medical imaging modalities-retinal fundus photographs, external eye photographs, and chest radiographs. We showcase examples where many of the automatically-learned attributes clearly capture clinically known features (e.g., types of cataract, enlarged heart), and demonstrate automatically-learned confounders that arise from factors beyond physiological mechanisms (e.g., chest X-ray underexposure is correlated with the classifier predicting abnormality, and eye makeup is correlated with the classifier predicting low hemoglobin levels). We further show that our method reveals a number of physiologically plausible, previously-unknown attributes based on the literature (e.g., differences in the fundus associated with self-reported sex, which were previously unknown).

INTERPRETATION

Our approach enables hypotheses generation via attribute visualizations and has the potential to enable researchers to better understand, improve their assessment, and extract new knowledge from AI-based models, as well as debug and design better datasets. Though not designed to infer causality, importantly, we highlight that attributes generated by our framework can capture phenomena beyond physiology or pathophysiology, reflecting the real world nature of healthcare delivery and socio-cultural factors, and hence interdisciplinary perspectives are critical in these investigations. Finally, we will release code to help researchers train their own StylEx models and analyze their predictive tasks of interest, and use the methodology presented in this paper for responsible interpretation of the revealed attributes.

FUNDING

Google.

(Predictable) performance bias in unsupervised anomaly detection

BACKGROUND

With the ever-increasing amount of medical imaging data, the demand for algorithms to assist clinicians has amplified. Unsupervised anomaly detection (UAD) models promise to aid in the crucial first step of disease detection. While previous studies have thoroughly explored fairness in supervised models in healthcare, for UAD, this has so far been unexplored.

METHODS

In this study, we evaluated how dataset composition regarding subgroups manifests in disparate performance of UAD models along multiple protected variables on three large-scale publicly available chest X-ray datasets. Our experiments were validated using two state-of-the-art UAD models for medical images. Finally, we introduced subgroup-AUROC (sAUROC), which aids in quantifying fairness in machine learning.

FINDINGS

Our experiments revealed empirical "fairness laws" (similar to "scaling laws" for Transformers) for training-dataset composition: Linear relationships between anomaly detection performance within a subpopulation and its representation in the training data. Our study further revealed performance disparities, even in the case of balanced training data, and compound effects that exacerbate the drop in performance for subjects associated with multiple adversely affected groups.

INTERPRETATION

Our study quantified the disparate performance of UAD models against certain demographic subgroups. Importantly, we showed that this unfairness cannot be mitigated by balanced representation alone. Instead, the representation of some subgroups seems harder to learn by UAD models than that of others. The empirical "fairness laws" discovered in our study make disparate performance in UAD models easier to estimate and aid in determining the most desirable dataset composition.

FUNDING

European Research Council Deep4MI.

Health Disparities, Clinical Trials, and the Digital Divide

In the past few years, there have been rapid advances in technology and the use of digital tools in health care and clinical research. Although these innovations have immense potential to improve health care delivery and outcomes, there are genuine concerns related to inadvertent widening of the digital gap consequentially exacerbating health disparities. As such, it is important that we critically evaluate the impact of expansive digital transformation in medicine and clinical research on health equity. For digital solutions to truly improve the landscape of health care and clinical trial participation for all persons in an equitable way, targeted interventions to address historic injustices, structural racism, and social and digital determinants of health are essential. The urgent need to focus on interventions to promote health equity was made abundantly clear with the coronavirus disease 2019 pandemic, which magnified long-standing social and racial health disparities. Novel digital technologies present a unique opportunity to embed equity ideals into the ecosystem of health care and clinical research. In this review, we examine racial and ethnic diversity in clinical trials, historic instances of unethical research practices in biomedical research and its impact on clinical trial participation, and the digital divide in health care and clinical research, and we propose suggestions to achieve digital health equity in clinical trials. We also highlight key digital health opportunities in cardiovascular medicine and dermatology as exemplars, and we offer future directions for development and adoption of patient-centric interventions aimed at narrowing the digital divide and mitigating health inequities.

The Myth of Innate Racial Differences Between White and Black People's Bodies: Lessons From the 1793 Yellow Fever Epidemic in Philadelphia, Pennsylvania

In this commentary, I take up the question of why beliefs in fundamental, innate racial differences between Black and White people's bodies persist in medical discourse, despite evidence to the contrary.I locate the origin of some of these beliefs in the infamous yellow fever epidemic that struck Philadelphia, Pennsylvania, in 1793. During that early public health crisis, White physicians and lay people erroneously thought that Black people were immune to yellow fever because of their race. I then highlight the efforts of Philadelphia's Black leaders during the epidemic-namely Absalom Jones and Richard Allen-to challenge the belief in fundamental and innate differences between Blacks and Whites.I conclude by asking us to consider how the false belief that there is something peculiar about Black people's bodies has become a feature, not an aberration, in the production of medical knowledge. Indeed, I point out how medical experimentation in the 20th century and in the marketing of new drugs in the 21st century have been buttressed by this persistent yet incorrect assumption that innate racial differences exist.

Reconsidering osteoarthritis as a skeletal indicator of age at death

OBJECTIVES

Osteoarthritis (OA) generally increases with age, but it is a complex, multifactorial disease. This study investigated whether obesity, physical activity, and antemortem trauma preclude the use of OA for skeletal age estimation.

MATERIALS AND METHODS

The temporomandibular joint (TMJ) and all appendicular joints were scored for skeletal indicators of OA in 408 modern European-Americans (Bass Collection, TN). Binomial generalized linear models (GLMs) assessed the contributions of self-reported demographic data to OA, including: age; body mass index (BMI); and metabolic values for physical activities. Repeated resampling tested whether observed mean OA scores for joints with trauma consistently exceeded mean scores for unaffected joints. Single-variable GLM probit models were generated for OA presence/absence data in relevant joints.

RESULTS

Age was the only statistically significant predictor of OA in most multivariable GLMs. Occupation and age were both significant predictors of male hand OA; BMI was the only significant predictor of female ankle OA. Trauma significantly affected OA in most joints. Age cut-offs calculated from the single-variable probit models (representing ages of transition to "OA present") ranged from 29.7 to 77.3 years (90%) and 32.7 to 96.6 years (95%), but were problematic for the male TMJ.

DISCUSSION

Ankle OA should not be used to age females; TMJ OA should not be used to age males. For other joints, using OA to inform age estimates appears valid (in absence of antemortem trauma). While skeletal evidence of OA is not a primary age indicator, its presence can refine age ranges and provide essential age data in fragmentary cases.

A TALE OF TWO MOLECULAR CALIFORNIAS

Molecular identification technologies are often framed in terms of their societal benefits. Forensic uses of DNA databases benefit society through the efficient identification of criminal suspects, while consumer DNA services empower individuals by identifying ethnic, health-related, and potentially sexual, molecular genealogies. Two examples of these technologies are California's criminological database CAL-DNA and the revitalized project to find a 'gay gene'. Both examples show how molecular identification technologies are also entangled with histories of coercion and stigmatization. The search for a 'gay gene' is premised on the historical stigmatization of homosexuality as deviant as well as contemporary concerns with resisting the idea that it is a lifestyle choice. The CAL-DNA database demonstrates that stigmatization still underpins contemporary identification technologies. This 'race-neutral' database puts racial minorities at increased risk of getting caught up in the criminal system precisely because of a racist history of identifying men of color as potential criminals. While the increasing criminological and consumer applications of molecular identification technologies are spearheaded in California, their uses emerge in a futurist culture that decontextualizes them from historic and contemporary coercion. The molecular identities these technologies create tell a tale of two Californias; one of empowerment and another of surveillance and stigma.

A reconsideration of the role of self-identified races in epidemiology and biomedical research

A considerable number of studies in epidemiology and biomedicine investigate the etiology of complex diseases by considering (self-identified) race as a relevant variable and focusing on the differences in risk among racial groups in the United States; they extensively draw on a genetic hypothesis--viz. the hypothesis that differences in the risk of complex diseases among racial groups are largely due to genetic differences covarying with genetic ancestry--that appears highly problematic in the light of both current biological evidence and the theory of human genome evolution. Is this reason for dismissing self-identified races? No. An alternative promising use of self-identified races exists, and ironically is suggested by those studies that investigate the etiology of complex diseases without focusing on racial differences. These studies provide a large amount of empirical evidence supporting the primacy of the contribution of non-genetic as opposed to genetic factors to the risk of complex diseases. We show that differences in race--or, better, in racial self-identification--may be critically used as proxies for differences in risk-related exposomes and epigenomes in the context of the United States. Self-identified race is what we need to capture the complexity of the effects of present and past racism on people's health and investigate risk-related external and internal exposures, gene-environment interactions, and epigenetic events. In fact patterns of racial self-identifications on one side, and patterns of risk-related exposomes and epigenomes on the other side, constantly coevolve and tend to match each other. However, there is no guarantee that using self-identified races in epidemiology and biomedical research will be beneficial all things considered: special attention must be paid at balancing positive and negative consequences.

Re-considering the turn to biology in feminist theory

This article argues that feminist theorists should conceive of the life sciences not only as a factual resource but also as a figural resource. It proposes that in shifting our conceptual orientation to biological science from fact to figure, feminists will be able to give theoretical life to scientific findings about the ways in which social environments and material habitats are processes integral to our development, growth, and social and political well-being. The figuration of ourselves as specifically biocultural creatures will enable feminists to gain a stronger theoretical and analytic purchase on the ongoing mutual shaping of the biological and social dimensions of existence. This, in turn, will enable feminists to creatively re-imagine the fields and fora of feminist politics.

Coronary artery disease and the contours of pharmaceuticalization

Coronary artery disease (CAD) has dominated mortality for most of the past century, not just in Europe and North America but worldwide. Treatments for CAD, both pharmaceutical and surgical, have become leading sectors of the healthcare economy. This paper focuses on the therapeutic landscape for CAD in the United States. We hope to add texture to the broader conversation of pharmaceuticalization explored in this issue by situating pharmaceutical therapies as just one element in the broader therapeutic terrain, alongside cardiac surgery and interventional cardiology. Patients with CAD must navigate a therapeutic landscape with three intersecting paths: lifestyle change, pharmaceuticals, and surgery. While pharmaceuticals are often seen as a quick fix, a way of avoiding more difficult lifestyle changes, it is surgery and angioplasty that promise patients the quickest fix of all. There also is another option, often overlooked by analysts but popular among physicians and patients: inaction. The U.S. context is often critiqued as a site of excessive treatment with respect to both drugs and procedures, and yet there is deep stratification within it--over-treatment in many populations and under-treatment in others. People who experience the serious risks of CAD do so in a racialized terrain of durable preoccupations with difference and unequal access to care. While the pharmaceuticalization literature disproportionately attends to lifestyle drugs, which some observers consider to be medically inappropriate or unnecessary, CAD does remain the leading cause of death. Thus, the stakes are high. Examination of the pharmaceuticalization of CAD in light of surgical treatments and racial disparities offers a window into the pervasiveness and persuasiveness of pharmaceuticals in an increasingly consumer-driven medicine, as well as the limits of their appeal and their reach.

Genetic factors influencing bone mineral content in a black South African population

Bone mass differs according to ethnic classification, with individuals of African ancestry attaining the highest measurements across numerous skeletal sites. Elevated bone mass is even maintained in those individuals exposed to adverse environmental factors, suggesting a prominent genetic effect that may have clinical or therapeutic value. Using a candidate gene approach, we investigated associations of six candidate genes (ESR1, TNFRSF11A, TNFRSF11B, TNFSF11, SOST and SPP1) with bone mass at the hip and lumbar spine amongst pre-pubertal black South African children (mean age 10.6 years) who formed part of the longitudinal Birth to Twenty cohort. 151 black children were genotyped at 366 polymorphic loci, including 112 previously associated and 254 tagging single nucleotide polymorphisms (SNPs). Linear regression was used to highlight significant associations whilst adjusting for height, weight, sex and bone area. Twenty-seven markers (8 previously associated and 19 tag SNPs; P < 0.05) were found to be associated with either femoral neck (18) or lumbar spine (9) bone mineral content. These signals were derived from three genes, namely ESR1 (17), TNFRSF11B (9) and SPP1 (1). One marker (rs2485209) maintained its association with the femoral neck after correction for multiple testing (P = 0.038). When compared to results amongst Caucasian adults, we detected differences with respect to associated skeletal sites. Allele frequencies and linkage disequilibrium patterns were also significantly different between populations. Hence, our results support the existence of a strong genetic effect acting at the femoral neck in black South African children, whilst simultaneously highlighting possible causes that account for inter-ethnic bone mass diversity.

Political challenges to biomedical universalism: kidney failure among Egypt's poor

Why do patients in need of kidney transplants in Egypt decline offers of kidney donation from their family members out of reluctance to cause them harm? Is it not universally the case that a living donor could live in complete health with a single remaining kidney? To address this conundrum, I discuss a case study from Egypt, in which patients reveal social, political, and environmental stresses on organ function that challenge the presumed universal efficacy and safety of kidney transplantation. I demonstrate that the biomedical position on the tolerable risks posed to the living donor is conditional and premised on particular social and historical contingencies that can be misaligned when applied in other contexts. Drawing on the work of Margaret Lock, I illustrate how analytical contributions of medical anthropologists can shed light on a political and public health impasse about how to legally regulate organ transplantation in Egypt.

Resisting the obfuscation of ‘race’: The importance of acknowledging the sociopolitical in normative data research

The concept of ‘race’ is, rightly, associated with a myriad of discriminatory practices, including its ontological justification through the misuse of science, including that of Psychology. Seeking to distance the discipline and profession from such a problematic history, most psychologists have abandoned the notion of ‘race’ or, as in social psychology, have turned to the study of racism. Within the context of contemporary South Africa, now almost two decades beyond the legislated presence of apartheid, the abandonment of the term ‘race’ has been especially motivated. This article asks the apparently controversial question of whether such an urgent distancing is justified. This is especially salient given that it has resulted in some researchers turning to alternate terms such as ‘culture’ and ‘ethnicity’ as though the simple exchange of these seemingly apolitical terms for ‘race’ will allow the business of traditional research to continue as usual. However, we cannot escape the effects of the past so easily; these terms are themselves not innocent but are in a complex relationship with each other and society. Furthermore, ‘race’ is more than a word, and its consequences lie well entrenched via past and present ideologically based social and material engineering. In this article, I argue that we need to return, as an ethically and conceptually sound act, to the concept of ‘race’ in normative data research. This requires that we abandon naïve notions of scientific practice and acknowledge both the sociopolitical context in which we conduct scientific research and the historically embedded and contingent nature of the concepts we utilise in our work and the data we produce.

Goldberg’s brain and the sex/gender distinction

The sex/gender binary has proven to be a profoundly useful conceptual distinction in the furthering of the feminist project. It has also been a controversial opposition that has given rise to an ongoing and productive debate. In this article we utilise neuroscience, specifically a text by the neuropsychologist Elkhonon Goldberg, to trouble this binary in the hope of furthering the critical project. We argue that a cautious negotiation with the biological may be theoretically and politically productive. By taking seriously Goldberg’s notions of functional-morphological and corporeal-environmental intimacy in reading his claim of distinct gender-based cognitive styles it is possible to glimpse the variation of sex itself. This, we argue, demonstrates both the limits of binaries and celebrations of difference and reveals the complexity which we have to negotiate in the search for emancipatory change.

Sexing the baby: Part 2--Applying dynamic systems theory to the emergences of sex-related differences in infants and toddlers

During the first three years of life, children acquire knowledge about their own gender and the gendered nature of their environment. At the same time, sex-related behavioral differences emerge. How are we to understand the processes by which bodily differentiation, behavioral differentiation and gendered knowledge intertwine to produce male and female, masculine and feminine? In this article, we describe four central developmental systems concepts applied by psychologists to the study of early human development and develop them in enough depth to show how they play out, and what sort of knowledge-gathering strategies they require. The general theoretical approach to understanding the emergence of bodily/behavioral difference has broad applicability for the health sciences and for the study of gender disparities. Using dynamic systems theory will deepen and extend the reach of theories of embodiment current in the health sciences literature.

Sexing the baby: Part 1--What do we really know about sex differentiation in the first three years of life?

The most common paradigm used to analyze health differences between men and women, is to divide the body from the social environment. In such a model, the bodily contribution and the social contribution add up to 100%. A few health science researchers offer more sophisticated approaches. None, however, offer an intensive study of the first several years of life in order to offer a model which integrates biology and culture in a fashion that demonstrates the productive processes by which gender itself emerges. In this article, we identify the earliest known sex-related biological and behavioral differences in young infants, toddlers and their parents and indicate how these might relate to health and disease. We frame these differences using unifying concepts from the study of neuroplasticity and dynamic systems theory.

A critical review of racial/ethnic variables in osteoporosis and bone density research

Racial and ethnic variables are common in research on variation in bone density. This literature review describes some of the common flaws associated with the use of these variables and provides some suggestions for how bone density research may be able to better document and address skeletal health disparities.

INTRODUCTION

Racial/ethnic differences in bone density have been commonly documented in the research literature. While effective identification of the specific factors underlying these trends might go a long way in informing treatment and screening for osteoporosis, this would require careful consideration of exactly what these variables are capturing. However, the basis and implications of what racial/ethnic variables represent have not carefully been examined in bone density research.

METHODS

For this paper, we systematically reviewed 55 articles that included bone density and race/ethnicity as key variables. Our analysis reveals that racial/ethnic terminology in these articles is highly variable, and discussion of how race/ethnicity is determined is often vague and idiosyncratic. Racial/ethnic variables are being used for a wide range of analytical purposes in statistical tests, which may not be appropriate for such a complex and poorly defined variable.

RESULTS

Many articles attribute racial/ethnic differences in bone mass/bone density to genetic causes, although few studies actually examine genetic data.

CONCLUSION

This analysis indicates that more rigorous examination of what race/ethnicity actually captures, more careful definitions of group labels and the procedures for assigning them, and attention to the limitations of how such variables can reliably be used in data analyses is needed to help address the problems and issues outlined in this review.

The postconventional body: retheorising women's health

We propose that women's health-both theory and practice-is a powerful arena in which to re-align and change the modernist theoretical underpinnings of current biomedical paradigms, which limit our understanding both of concepts of health and illness and of the impact of health care technologies on the body. We highlight the necessity of a move to a more dynamic paradigm for health and illness in the clinic, as well as a theoretical fluidity that allows for the real messiness of lived bodies. We argue that postmodernist thought, within wider feminist theory, is one of many perspectives that can contribute to contemporary biomedicine by providing theoretical underpinnings to develop 1) an understanding of bodies in context, 2) an epistemology of ignorance, and 3) an openness to the risk of the unknown. While these all entail a commitment to self-reflection and a willingness to be unsettled, which may not seem practical in the context of medical practice, we argue that self-reflection and unsettledness will provide pathways for grappling with chronic conditions and global bodies. Overall, we suggest that women's health practice can serve as a site in which both sides of the humanistic/scientific divide can engage with a human self in all its corporeal variety, contingency, and instability. More specifically, by providing a space within the clinic to examine underlying ontological, epistemological, and ethical assumptions, women's health can continue to contribute to new forms of biomedical practice.

Race, genetics, and disease: questions of evidence, matters of consequence

This special issue of Studies of Science highlights ongoing debates concerning race, genomics, and disease. Some of the papers examine the production of disease etiology research, pharmaceutical drug response, or DNA genealogy tests, while others analyze institutional consequences and challenges arising from contemporary biomedicine, such as medical education and recruiting subjects for clinical research. In this introduction, we outline major issues that provide background and foreground for the specific studies that follow, and end with a brief description of the papers. First, we briefly outline the debates around contemporary genetics research on race, ancestry, population, and disease. Second, we describe genomics and disease research projects on the genetics of populations that provide the ground on which the past debates have played, as well as introduce very recent projects that may change the tenor of future debates. We discuss why some scientists argue that their research does not biologize race, while others argue that their findings do demonstrate racial differences. Finally, we relate these complex genomic sciences and their biopolitical debates to relevant STS themes.