Call to Action on the Categorization of Sex, Gender, Race, and Ethnicity in Surgical Research

Words matter. This statement has taken on heightened meaning as we collectively struggle to achieve equity for groups who experience systemic barriers to equity. Language is powerful and reflects, even shapes, the prevailing culture of the times.

As researchers, we know words matter for operationalizing variables for research. Disparities in risk and outcomes exist for those with surgical diseases. Because words and definitions influence research and society as a whole, the Eastern Association for the Surgery of Trauma made the commitment to look critically at how studies use language to define sex, gender, race, and ethnicity. With this in mind, a group of surgical researchers, a medical anthropologist, and a social epidemiologist sought to make recommendations to promote inclusive language in the dissemination of surgical research.

SEX AND GENDER

The US population has a slightly higher percentage of female than male people, however, fewer than one-third of phase I US Food and Drug Administration trial participants are female.¹ This is despite the fact that nearly 20 years ago, it was determined that the majority of drug safety issues had greater health risks for the female population. In addition, there continue to be differences in outcomes when comparing male and female patients, including cancer and cardiovascular outcomes.^2,3

The exact method of reporting sex and gender (self-reported vs chromosomal data) and use of the correct term (sex vs gender) is lacking in many surgical studies. Sex is a biologic component defined by chromosomes, sex organs, and endogenous hormones. Gender is a social construct, including behaviors occurring within a historical and cultural context.⁴ Both can independently affect disease presentation, treatment response to medications,⁵ and outcomes. However, there is a paucity of high-quality data stratified by sex and gender.

Definitions of sex and gender are vague or omitted in surgical databases, such as the American College of Surgeons NSQIP and TQIP. They either provide no definition of sex or simply define it as, “The patient's sex,” adding that “patients who have undergone a surgical and/or hormonal sex reassignment should be coded using the current assignment.” In addition, to our knowledge, no national surgical database distinguishes sex from gender.

Mansukhani and colleagues⁶ uncovered considerable sex bias in human surgical research, with fewer than one-third of studies within prominent surgical journals performing data analysis according to sex. Few studies included women in equal proportion to men, even with diseases that affect both at a similar rate.⁶ Also, the Sex and Gender Equity in Research guidelines recommend researchers convey within their methods how sex was determined (self-reported vs via genetic testing). Also, Sex and Gender Equity in Research recommends all trials analyze and report sex-stratified outcomes data, including tables.⁷ This enables future aggregation of data into meta-analyses. Therefore, surgical trials should provide, at minimum, supplementary tables with sex-stratified outcomes.

Data on the independent effects of gender in surgical research are sparse, as there is no agreed-on, validated tool to assess gender.⁸ At a minimum, the use of an “other” or “nonbinary” category might facilitate progress toward capturing patients who do not identify as a woman or a man. In addition, gender can be the same (cis) or different (trans) from an individual's sex. However, despite being a vulnerable population, few research studies report whether subjects are transgender.

RACE AND RACISM

Surgery research often uses the language of “biologic essentialism,” suggesting differences in outcomes by race might be due to fundamental biologic differences between racial groups based on biology and ancestry. However, race is a social construct, with the majority of the variation of the 0.1% of human genes that differ between individuals present within and not between racial groups.⁹

Social scientist Dr David Williams defines racism as “an organized social system in which the dominant racial group, based on an ideology of inferiority, categorizes and ranks people into social groups called ‘races’ and uses its power to devalue, disempower, and differentially allocate valued societal resources and opportunities to groups defined as inferior.”¹⁰ Structural racism refers to the totality of ways in which society fosters racial discrimination through mutually reinforcing inequitable systems in, for example, housing, education, employment, and healthcare. This inequitable system, in turn, reinforces discriminatory beliefs and distribution of resources, which affect the risk of adverse health outcomes. In fact, when class-related variables, such as quality of education, wealth, and racism, are accounted for, the effect of race on health outcomes is reduced greatly.¹¹

Bailey and colleagues¹² noted the small body of research on racism and health has focused predominantly on interpersonal racial and ethnic discrimination, with less emphasis on structural racism. This is supported by a systematic review that demonstrated a lack of intentionality when conducting research on the cause of healthcare disparities, namely institutional racism.¹³ An analysis of the reporting of race and ethnicity in the top-10 academic journals found only 4% of studies provided a formal definition of race and ethnicity.¹⁴ Racism, structural or institutional, was not mentioned. In addition, many journals do not have reporting guidelines on race.

In 1999, the Institute of Medicine noted that researchers should move beyond race in favor of ethnicity, as race is hopelessly imprecise as a population taxonomy to describe complex genetic, behavioral, environmental, and socioeconomic states of populations.¹⁵ Higher accuracy can be obtained by examining other dimensions, such as behavioral, ethnic, and socioeconomic markers to determine true associations rather than falsely reinforcing race as a causal agent.

ETHNICITY

In 1993, the NIH Revitalization Act mandated the inclusion of racial and ethnic minorities in federally funded research.^16,17 Despite this, there remains a failure to create ethnically diverse cohorts in research today. In addition, health disparities continue to exist between different ethnic populations, as health outcomes are directly related to socioeconomic and environmental disadvantages and social determinants of health, all of which are intertwined with ethnicity.

Adding to this complexity is the interchangeable use of race and ethnicity, although their definitions differ. Self-identified race is often dependent on physical attributes that, although heritable, correlate poorly with genetic similarity or ancestry; and ethnicity is defined as large groups of people classified by common racial, national, tribal, religious, linguistic, cultural origin, and background. Therefore, there is a need to differentiate ethnicity as a data variable to address ethnic disparities.

Surgical research frequently relies on the ethnicity reporting categories based on NIH classifications. Across the field of surgery, ethnic minorities present with higher incidence of surgical disease and worse postoperative outcomes.¹⁸ Approximately 87 different ethnic groups exist in the US, and the NIH separates ethnicity into only 2 categories: Hispanic/Latino and not-Hispanic/Latino.¹⁹ The use of this standard within surgical databases results in incomplete and likely incorrect conclusions.

CONCLUSIONS

We propose the following call to action related to categorizing sex, gender, race, and ethnicity. First, researchers should follow best practice as it relates to collecting, documenting, analyzing, reporting, and discussing sex, gender, race, and ethnicity. This includes how the categories were determined (eg self-reported) and providing a pathophysiologic basis for use of these variables before evaluation of a potentially causal relationship. In addition, control for other social determinants of health should occur before ascribing an effect to one of these variables. In addition, there is a much needed expansion beyond the binary categorization of gender to include at least a nonbinary gender variable and a more inclusive than Hispanic/Latino and not-Hispanic/Latino ethnicity category. Collaboration with historians, sociologists, and/or other experts, as is done with biostatisticians, might prove helpful in these area. Second, surgical journals must ensure published articles include appropriate terminology and data for these topics. Some journals have already begun to take a lead by incorporating within their author instructions to report the sex distribution of study participants or explain why the other sex is not reported. Additional efforts are needed in relation to gender, race, and ethnicity. Third, surgical professional societies and funding agencies should require distinct evaluation of sex, gender, race, and ethnicity in studies they support. Fourth, national surgical databases should include sex, gender, race, and ethnicity, and clearly define how these variables are collected and reported, with additional consideration for adding other social determinants of health to foster higher quality outcomes research. Fifth, the Enhancing the Quality and Transparency of Health Research Network should address the standardized approach to reporting and publishing sex, gender, race, and ethnicity in their series for reporting tools. Sixth, researchers must address intersectionality, where overlapping identities (sex, gender, race, and ethnicity) contribute to inequity (Table 1). Because words matter, we believe that with implementation of these measures, surgical research will be improved, especially for marginalized patients.

Table 1: A Call to Action Related to Categorizing Sex, Gender, Race, and Ethnicity

Author Contributions

Study conception and design: Nahmias, Zakrison, Haut, Gurney, Joseph, Hendershot, Ghneim, Stey, Hoofnagle, Bailey, Rattan, Richardson, Santos, Zarzaur

Acquisition of data: Nahmias, Zakrison, Haut, Gurney, Joseph, Hendershot, Ghneim, Stey, Hoofnagle, Bailey, Rattan, Richardson, Santos, Zarzaur

Analysis and interpretation of data: Nahmias, Zakrison, Haut, Gurney, Joseph, Hendershot, Ghneim, Stey, Hoofnagle, Bailey, Rattan, Richardson, Santos, Zarzaur

Drafting of manuscript: Nahmias, Zakrison, Haut, Gurney, Joseph, Hendershot, Ghneim, Stey, Hoofnagle, Bailey, Rattan, Santos, Zarzaur

Critical revision: Nahmias, Zakrison, Haut, Gurney, Joseph, Hendershot, Ghneim, Stey, Hoofnagle, Bailey, Rattan, Richardson, Santos, Zarzaur

ACKNOWLEDGMENT

The authors thank Christine Eme and Rachel Dixon for their support and administrative leadership.