Diversity and Inclusion Training in Health Care: What the Evidence Suggests

On September 22, 2020, then-President Donald Trump signed Executive Order (EO) 13950, titled “Combating Race and Sex Stereotyping” (1). EO 13950 barred federal employers, contractors, and granting agencies from engaging in diversity and inclusion (DI) training on the basis of nine “divisive concepts” that allegedly caused “race or sex stereotyping” or “race or sex scapegoating” (1). Although the order did not mention terms such as critical race theory, systemic racism, or unconscious bias, the Office of Management and Budget asked federal agencies to identify all DI training programs and budget submissions for fiscal year 2020 with these phrases for possible violations (2). Hundreds of organizations inside and outside the health care sector denounced EO 13950 for impeding efforts to fight discrimination.

The punishments for violating EO 13950 included lawsuits and debarment from federal awards, contracts, and grants (1). The order would have reversed two generations of research and practice on DI within the mental health sector. The federal government is the largest funder of mental health services in the United States, dispersing more than $12 billion for mental health–specific programs in fiscal year 2020 across the Substance Abuse and Mental Health Services Administration, National Institutes of Health, Administration for Families and Children, and Departments of Justice and Housing (3). The total amount is much higher because this figure omits federal funding for the Centers for Medicare and Medicaid Services, the U.S. Department of Veterans Affairs, and other agencies that provide mental health services within their general portfolios. Any mental health care entity that works with any federal agency would have been subject to significant consequences under EO 13950.

EO 13950 was another reprehensible example of discriminatory presidential actions from the Trump administration. Other EOs prevented Iranians, Iraqis, Libyans, Somalians, Sudanese, Syrians, and Yemenis from traveling to the United States, commonly known as the “Muslim Ban”; forbade high-skilled foreign workers on H-1B visas, who were mostly from India, from working on federal contracts; froze green cards for immigrants through December 2020 to prevent them from becoming permanent residents; and revoked requirements for companies seeking federal contracts to prove that they do not discriminate on the basis of sexual identity or orientation.

On December 22, 2020, the U.S. District Court for the Northern District of California issued a preliminary nationwide injunction to prevent the Office of Federal Contract Compliance Programs from enforcing EO 13950, even while former President Trump was still in power. Because the injunction was preliminary, President Joseph R. Biden repealed it permanently on his first day in office by signing EO 13985, which is devoted to promoting racial equity as part of an all-of-government program across federal agencies. Nonetheless, former President Trump’s order revealed how a government that prioritizes political ideology over scientific research can undermine progress in DI. In this column, I expose the racism behind EO 13950; examine DI training sessions in their current form; and suggest evidence-based solutions, such as embracing antiracism and structural competence, for organizations that are committed to dismantling racism.

The Racism of EO 13950

EO 13950 perpetuated White supremacy, denied White privilege, and sanitized the roots of American society through a deliberate misreading of scientific evidence. It defined “race or sex stereotyping” as “ascribing character traits, values, moral and ethical codes, privileges, status, or beliefs to a race or sex, or to an individual because of his or her race or sex” (1). It defined “race or sex scapegoating” as “assigning fault, blame, or bias to a race or sex, or to members of a race or sex because of their race or sex” (1). The order claimed that some DI training “perpetuates racial stereotypes and division” (1). EO 13950 permitted federal agencies and contractors to support DI training that did not view individuals predominantly on the basis of their race. In addition, so-called “divisive concepts” could be taught “as part of a larger course of academic instruction” and “in an objective manner and without endorsement” (1).

These arguments have no scientific basis. First, EO 13950 treated stereotyping and scapegoating in parallel even though the two terms have different evidence bases. Research on people of color being stereotyped in mental health services has existed since at least the 1970s (4). No research agenda exists for what the Trump administration labeled scapegoating. Second, EO 13950 equated asking White people to acknowledge their privilege with stereotyping and scapegoating. However, best practices for cultivating antiracism throughout mental health services have demanded that all individuals locate their historical positions in systems of oppression and privilege and that White people reflect on their greater access to resources at the expense of people of color (4). Because of the deliberate disinformation in EO 13950, it is necessary to clarify the evidence base for DI training and future opportunities.

Predominant Approaches to DI Training

Comparing EO 13950’s conception of DI training against scholarly evidence proves that EO 13950 was based on political ideology, not science. In the most comprehensive review of its kind, a meta-analysis of DI training across a variety of settings included 260 studies from 1973 to 2013 that examined pre- and posttraining outcomes with 29,407 participants (5). The authors defined DI training as a “set of instructional programs aimed at facilitating positive intergroup interactions, reducing prejudice and discrimination, and enhancing the skills, knowledge, and motivation of participants to interact with diverse others” (5). DI training sessions were classified as targeting awareness skills, behavioral skills, or both. Awareness skills were defined as “getting participants to be more aware of their own and other cultural assumptions, values, and biases.” Behavioral skills were defined as “monitoring one’s own actions and appropriate responses to specific differences, such as identifying and overcoming interracial communication barriers” (5).

More hours of DI training predicted better skill development, but training sessions that were designed to improve awareness skills showed a smaller effect than forms of training that were designed to improve awareness and behavioral skills (Hedges’ g=0.31 vs. 0.46, p<0.05). The authors noted that adding behavioral skills helped participants practice changes within their actual routines, such as how to communicate with people from other backgrounds (5). The length of DI training was more important than other factors of training in explaining variations in skill development; it was less important whether the training focused on a specific ethnic, gender, racial, or sexual group or discussed how to avoid discrimination based on any type of demographic difference (5).

In contrast, today’s standard approach to DI training in most health organizations is to briefly discuss race (but not racism) in single, standalone sessions. These sessions were moved online even before the COVID-19 pandemic without ongoing opportunities for in-person behavioral practice and knowledge reinforcement (6). The situation is worse in mental health settings. Evidence for DI training in mental health organizations mostly comes from educational institutions, such as medical schools and residency programs (7). A consistent finding shows that DI training, often framed as a “cultural competence” or “cultural competency” class, triggers intense emotions; moreover, it is dismissed as “diversity talk” or “political correctness” because many participants do not see themselves as working in institutions that are shaped by histories of slavery or racial tensions (7).

This type of DI training and the negative responses that it generates persist despite incontrovertible evidence that racism is a social determinant of health that exacerbates health disparities for people of color (8). Hence, most DI training sessions in mental health organizations do not offer basic, requisite knowledge about racism as a social determinant of health or ongoing opportunities for participants to consider how their actions may promote or mitigate racism. In today’s era of evidence-based medicine, the single, standalone DI session can no longer be justified when the data suggest otherwise.

Evidence-Based DI Training Embraces Antiracism and Structural Competence

The current approach of DI training in health organizations is to concentrate on the individual trainee’s responsibility to counteract racism on the basis of their perceptions of others (6). Although this strategy may be effective in remediating individual racism, it does not tackle institutional racism, which has been defined as “discriminatory policies and practices carried out within and between individual state or non-state institutions on the basis of racialized group membership” (6). Mental health organizations have perpetuated discriminatory policies and practices, such as excluding Black students and patients from predominantly White institutions for centuries, staffing outpatient clinics for uninsured people of color with medical students and residents who have less experience than fully trained physicians, and allocating less funding for clinics that serve mostly people of color (8). DI training for individuals that was integrated with other institutional initiatives to combat racism improved participants’ knowledge, attitudes, and skills more than single, standalone sessions (5) because more integrated training illustrates an institutional commitment to health equity and leads to better outcomes (6).

For this reason, mental health organizations must embrace antioppression and antiracism training sessions. Driven by social justice and activism, both frameworks recognize and redress power imbalances that constrain personal, social, and economic opportunities (4). Antiracism and antioppression frameworks share many features; however, the former identifies racism as the main contributor of power imbalances, whereas the latter does not predefine the source of oppression and acknowledges that ethnicity, gender, race, and sex can create intersectional identities that lead to multiple types of oppression (4). Organizations can enact antiracism and antioppression practices by sharing service utilization data with consumer groups to lobby for more funding and quality improvement, educating all people about an institution’s history of race relations, and building alliances with local communities (4).

Institutional antiracism initiatives raise the broader question of how institutions fit within society. Therefore, leaders must work toward structural competence, which seeks to understand the totality of ways that societies foster discrimination through interconnected systems and institutions that reinforce discriminatory beliefs, values, and distribution of resources (6). One tool that organizations can use for institution-wide DI initiatives is the U.S. Department of Health and Human Services’ checklist of National Standards for Culturally and Linguistically Appropriate Services in Health and Health Care (9).

Beyond just DI training for individuals, other initiatives include incorporating DI in mission statements; recruiting a workforce from the local communities served by the organization; developing mentorship programs to promote meritorious but overlooked employees of color; collecting racial, ethnic, and linguistic data on all patients to improve service provision; assessing language assistance needs; and utilizing certified interpreters with patients (10). An integrated approach to DI training for individuals complements institutional activities to dismantle racism and the inequitable distribution of resources (6, 8, 9).

Critics may question the rationale for funding more DI initiatives, but the business case has already been well established. Companies with diverse management teams have 19% more revenue than companies with poorer diversity scores because people with diverse lived experiences add new ideas and find opportunities for innovation (11). Most millennials and Generation Zers will not join or visit organizations where DI is ignored (11). Furthermore, people of color and women, as historically disadvantaged groups, are likely to mentor others and demand pay parity, which improves retention and reduces attrition for all employees (11). Mental health organizations must catch up and expand this evidence base. President Biden’s EO 13985 requires federal agencies to engage organizations in underserved communities to pilot demonstration projects that measure equity; moreover, administrators within mental health service organizations can use this opportunity to conduct research to improve DI training.

DI training has been criticized for several reasons: the gains are not immediate, not every intersectional identity can be encompassed, all institutions have unique local histories, and their remediation practices may not be generalizable (4). Still, DI training sessions produce skills that alleviate health, economic, and social disparities among people of color. With a robust evidence base across multiple lines of scholarship, DI initiatives must be implemented so that mental health services are responsive, equitable, and effective. A commitment to greater DI training would undermine the spirit of EO 13950.