NCBI Bookshelf. A service of the National Library of Medicine, National Institutes of Health.
National Academies of Sciences, Engineering, and Medicine; Health and Medicine Division; Board on Population Health and Public Health Practice; Committee on Community-Based Solutions to Promote Health Equity in the United States; Baciu A, Negussie Y, Geller A, et al., editors. Communities in Action: Pathways to Health Equity. Washington (DC): National Academies Press (US); 2017 Jan 11.
National Academies of Sciences, Engineering, and Medicine; Health and Medicine Division; Board on Population Health and Public Health Practice; Committee on Community-Based Solutions to Promote Health Equity in the United States; Baciu A, Negussie Y, Geller A, et al., editors.
Washington (DC): National Academies Press (US); 2017 Jan 11.Health inequity, categories and examples of which were discussed in the previous chapter, arises from social, economic, environmental, and structural disparities that contribute to intergroup differences in health outcomes both within and between societies. The report identifies two main clusters of root causes of health inequity. The first is the intrapersonal, interpersonal, institutional, and systemic mechanisms that organize the distribution of power and resources differentially across lines of race, gender, class, sexual orientation, gender expression, and other dimensions of individual and group identity (see the following section on such structural inequities for examples). The second, and more fundamental root cause of health inequity, is the unequal allocation of power and resources—including goods, services, and societal attention—which manifest in unequal social, economic, and environmental conditions, also called the social determinants of health. Box 3-1 includes the definitions of structural inequities and the social determinants of health.
The factors that make up the root causes of health inequity are diverse, complex, evolving, and interdependent in nature. It is important to understand the underlying causes and conditions of health inequities to inform equally complex and effective interventions to promote health equity.
The fields of public health and population health science have accumulated a robust body of literature over the past few decades that elucidates how social, political, economic, and environmental conditions and context contribute to health inequities. Furthermore, there is mounting evidence that focusing programs, policies, and investments on addressing these conditions can improve the health of vulnerable populations and reduce health disparities (Bradley et al., 2016; Braveman and Gottlieb, 2014; Thornton et al., 2016; Williams and Mohammed, 2013). This literature is discussed below in the sections on structural inequities and the social determinants of health.
Health inequities are systematic differences in the opportunities groups have to achieve optimal health, leading to unfair and avoidable differences in health outcomes (Braveman, 2006; WHO, 2011). The dimensions of social identity and location that organize or “structure” differential access to opportunities for health include race and ethnicity, gender, employment and socioeconomic status, disability and immigration status, geography, and more. Structural inequities are the personal, interpersonal, institutional, and systemic drivers—such as, racism, sexism, classism, able-ism, xenophobia, and homophobia—that make those identities salient to the fair distribution of health opportunities and outcomes. Policies that foster inequities at all levels (from organization to community to county, state, and nation) are critical drivers of structural inequities. The social, environmental, economic, and cultural determinants of health are the terrain on which structural inequities produce health inequities. These multiple determinants are the conditions in which people live, including access to good food, water, and housing; the quality of schools, workplaces, and neighborhoods; and the composition of social networks and nature of social relations.
So, for example, the effect of interpersonal, institutional, and systemic biases in policies and practices (structural inequities) is the “sorting” of people into resource-rich or resource-poor neighborhoods and K–12 schools (education itself being a key determinant of health (Woolf et al., 2007) largely on the basis of race and socioeconomic status. Because the quality of neighborhoods and schools significantly shapes the life trajectory and the health of the adults and children, race- and class-differentiated access to clean, safe, resource-rich neighborhoods and schools is an important factor in producing health inequity. Such structural inequities give rise to large and preventable differences in health metrics such as life expectancy, with research indicating that one's zip code is more important to health than one's genetic code (RWJF, 2009).
The impact of structural inequities follows individuals “from womb to tomb.” For example, African American women are more likely to give birth to low-birthweight infants, and their newborns experience higher infant death rates that are not associated with any biological differences, even after accounting for socioeconomic factors (Braveman, 2008; Hamilton et al., 2016; Mathews et al., 2015). Although the science is still evolving, it is hypothesized that the chronic stress associated with being treated differently by society is responsible for these persistent differential birth outcomes (Christian, 2012; El-Sayed et al., 2015; Strutz et al., 2014; Witt et al., 2015). In elementary school there are persistent differences across racial and ethnic divisions in rates of discipline and levels of reading attainment, rates that are not associated with any differences in intelligence metrics (Howard, 2010; Losen et al., 2015; Reardon et al., 2012; Skiba et al., 2011; Smith and Harper, 2015). There also are race and class differences in adverse childhood experiences and chronic stress and trauma, which are known to affect learning ability and school performance, as well as structural inequities in environmental exposures, such as lead, which ultimately can lead to differences in intelligence quotient (IQ) (Aizer et al., 2015; Bethell et al., 2014; Jimenez et al., 2016; Levy et al., 2016). One of the strongest predictors of life expectancy is high school graduation, which varies dramatically along class and race and ethnicity divisions, as do the rates of college and vocational school participation—all of which shape employment, income, and individual and intergenerational wealth (Olshansky et al., 2012). Structural inequities affect hiring policies, with both implicit and explicit biases creating differential opportunities along racial, gender, and physical ability divisions. Lending policies continue to create differences in home ownership, small business development, and other asset development (Pager and Shepherd, 2008). Structural inequities create differences in the ability to participate and have a voice in policy and political decision making, and even to participate in the arguably most fundamental aspect of our democracy, voting (Blakely et al., 2001; Carter and Reardon, 2014). And implicit biases create differential health care service offerings and delivery and affect the effectiveness of care provided, including a lack of cultural competence (IOM and NRC, 2003; Sabin et al., 2009).
For many people, the challenges that structural inequities pose limit the scope of opportunities they have for reaching their full health potential. The health of communities is dependent on the determinants of health.
As described above, structural inequities refers to the systematic disadvantage of one social group compared to other groups with whom they coexist that are deeply embedded in the fabric of society. In Figure 3-1, the outermost circle and background indicate the context in which health inequities exist. Structural inequities encompass policy, law, governance, and culture and refer to race, ethnicity, gender or gender identity, class, sexual orientation, and other domains. These inequities produce systematic disadvantages, which lead to inequitable experiences of the social determinants of health (the next circle in the report model, which is discussed in detail later in this chapter) and ultimately shape health outcomes.
Report conceptual model for community solutions to promote health equity. NOTE: Structural inequities are highlighted to convey the focus of this section.
Whether with respect to race, ethnicity, gender, class, or other markers of human difference, the prevailing American narrative often draws a sharp line between the United States' “past” and its “present,” with the 1960s and 1970s marking a crucial before-and-after moment in that narrative. This narrative asserts that until the 1950s, U.S. history was shaped by the impacts of past slavery, Indian removal, lack of rights for women, Jim Crow segregation, periods of nativist restrictions on immigration and waves of mass deportation of Hispanic immigrants, eugenics, the internment of Japanese Americans, the Chinese exclusion policies, the criminalization of “homosexual acts,” and more (Gee and Ford, 2011; Gee et al., 2009). White women and people of color were effectively barred from many occupations and could not vote, serve on juries, or run for office. People with disabilities suffered widespread discrimination, institutionalization, and social exclusion.
Civil rights, women's liberation, gay rights, and disability rights movements and their aftermaths may contribute to a narrative that social, political, and cultural institutions have made progress toward equity, diversity, or inclusion. Highlights of progress include the Civil Rights Act of 1964, the Voting Rights Act of 1965, the Fair Housing Act, Title IX of the Education Amendments of 1972, the Americans with Disabilities Act, the Patient Protection and Affordable Care Act, and, most recently, the Supreme Court case 1 that legalized marriage equality in the United States. With a few notable exceptions—undocumented immigrants and Muslims, for example—these advances in law and policy have been mirrored by the liberalization of attitudes toward previously marginalized identity groups.
Today, polls and surveys indicate that most Americans believe that interpersonal and societal bias on the basis of identity no longer shapes individual or group social outcomes. For example, 6 in 10 respondents to a recent national poll said they thought the country has struck a “reasonable balance” or even gone “too far” in “accepting transgender people” (Polling Report, n.d.). In 2015, 72 percent of respondents, including 81 percent of whites, said they believe that “blacks have as good a chance as white people in your community to get any kind of job for which they are qualified” (Polling Report, n.d.). In another poll, a total of 72 percent agreed that “women and men have equal trouble finding good-paying jobs” (64 percent) or that men have more trouble (8 percent) (Ms. Foundation for Women, 2015). However, when broken down by racial and ethnic categories, the polls tell a different narrative. A recent survey revealed that 70 percent of African Americans, compared with 36 percent of whites, believe that racial discrimination is a major reason that African Americans have a harder time getting ahead than whites (Pew Research Center, 2016). Furthermore, African Americans (66 percent) and Hispanics (64 percent) are more likely than whites (43 percent) to say that racism is a big problem (DiJulio et al., 2015). Here, perceptions among African Americans and whites have not changed substantially; however, Hispanics are much more likely to now say that racism is a big problem (46 percent in 1995 versus 64 percent in 2015) (DiJulio et al., 2015).
Perceptions are confirmed by the persistence of disparities along the lines of socioeconomic position, gender, race, ethnicity, immigration status, geography, and the like has been well documented. Why? For one, historical inequities continue to ramify into the present. To understand how historical patterns continue to affect life chances for certain groups, historians and economists have attempted to calculate the amount of wealth transmitted from one generation to the next (Margo, 1990). They find that the baseline inequities contribute to intergenerational transfers of disadvantage and advantage for African Americans and whites, respectively (Chetty et al., 2014; Darity et al., 2001). The inequities also reproduce the conditions in which disparities develop (Rodriguez et al., 2015).
Though inequities may occur on the basis of socioeconomic status, gender, and other factors, we illustrate these points through the lens of racism, in part because disparities based on race and ethnicity remain the most persistent and difficult to address (Williams and Mohammed, 2009). Racial factors play an important role in structuring socioeconomic disparities (Farmer and Ferraro, 2005); therefore, addressing socioeconomic factors without addressing racism is unlikely to remedy these inequities (Kaufman et al., 1997).
Racism is an umbrella concept that encompasses specific mechanisms that operate at the intrapersonal, interpersonal, institutional, and systemic levels 2 of a socioecological framework (see Figure 3-2). Because it is not possible to enumerate all of the mechanisms here, several are described below to illustrate racism mechanisms at different socioecological levels. Stereotype threat, for example, is an intrapersonal mechanism. It “refers to the risk of confirming negative stereotypes about an individual's racial, ethnic, gender, or cultural group” (Glossary of Education Reform, 2013). Stereotype threat manifests as self-doubt that can lead the individual to perform worse than she or he might otherwise be expected to—in the context of test-taking, for example. Implicit biases—unconscious cognitive biases that shape both attitudes and behaviors—operate interpersonally (discussed in further detail below) (Staats et al., 2016). Racial profiling often operates at the institutional level, as with the well-documented institutionalization of stop-and-frisk practices on Hispanic and African American individuals by the New York City Police Department (Gelman et al., 2007).
Social ecological model with examples of racism constructs. NOTES: The mechanisms by which the social determinants of health operate differ with respect to the level. For the intrapersonal level, these mechanisms are individual knowledge, attitudes/beliefs, (more. )
Finally, systemic mechanisms, which may operate at the community level or higher (e.g., through policy), are those whose effects are interactive, rather than singular, in nature. For example, racial segregation of neighborhoods might well be due in part to personal preferences and behavior of landlords, renters, buyers, and sellers. However, historically, segregation was created by legislation, which was reinforced by the policies and practices of economic institutions and housing agencies (e.g., discriminatory banking practices and redlining), as well as enforced by the judicial system and legitimized by churches and other cultural institutions (Charles, 2003; Gee and Ford, 2011; Williams and Collins, 2001). In other words, segregation was, and remains, an interaction and cumulative “product,” one not easily located in any one actor or institution. Residential segregation remains a root cause of racial disparities in health today (Williams and Collins, 2001).
Racism is not an attribute of minority groups; rather, it is an aspect of the social context and is linked with the differential power relations among racial and ethnic groups (Guess, 2006). Consider the location of environmental hazards in or near minority communities. Placing a hazard in a minority community not only increases the risk of adverse exposures for the residents of that community, it also ensures the reduction of risk for residents of the nonminority community (Cushing et al., 2015; Taylor, 2014). Recognizing this, the two communities could work together toward an alternative that precludes having the hazard in the first place, an alternative that disadvantages neither group.
Most studies of racism are based on African American samples; however, other populations may be at risk for manifestations of racism that differ from the African American experience. Asians, Hispanics, and, more recently, Arabs and Muslims are subject to assumptions that they are not U.S. citizens and, therefore, lack the rights and social entitlements that other U.S. residents claim (Chou and Feagin, 2015; Cobas et al., 2009; Feldman, 2015; Gee et al., 2009; Johnson, 2002; Khan and Ecklund, 2013). The implications of this include threats or actual physical violence against members of these groups. For instance, researchers have found that in the months immediately following September 11, 2001, U.S. women with Arabic surnames who were residing in California experienced increases in both racial microaggressions (i.e., seemingly minor forms of “everyday racism”) and in poor birth outcomes compared to the 6 months preceding 9/11, while women of other U.S. ethnic groups did not (Kulwicki et al., 2008; Lauderdale, 2006). For Native Americans, because tribes are independent nations, the issues of racism need to be considered to intersect with those of sovereignty (Berger, 2009; Massie, 2016; Sundeen, 2016).
The evidence linking racism to health disparities is expanding rapidly. A variety of both general and disease-specific mechanisms have been identified; they link racism to outcomes in mental health, cardiovascular disease, birth defects, and other outcomes (Paradies, 2006a; Pascoe and Smart Richman, 2009; Shavers et al., 2012; Williams and Mohammed, 2009). Which racism mechanisms matter most depends in part on the disease and, to a lesser degree, the population. The vast majority of studies focus on the role of discrimination; that is racially disparate treatment from another individual or, in some cases, from an institution. Among the studies not focused on discrimination, the majority examine segregation. Generally, findings show that members of all groups, including whites, report experiencing racial discrimination, with levels typically, though not always, higher among African Americans and, to a lesser degree, Hispanics than among whites. Gender differences in some perceptions about and responses to racism have also been observed (Otiniano Verissimo et al., 2014). Three major mechanisms by which systemic racism influences health equity—discrimination (including implicit bias), segregation, and historical trauma—are discussed in more detail in the following paragraphs.
The mechanisms by which discrimination operates include overt, intentional treatment as well as inadvertent, subconscious treatment of individuals in ways that systematically differ so that minorities are treated worse than nonminorities. Recent meta-analyses suggest that racial discrimination has deleterious effects on the physical and mental health of individuals (Gee et al., 2009; Paradies, 2006a; Pascoe and Smart Richman, 2009; Priest et al., 2013; Williams and Mohammed, 2009). Significant percentages of members of racial and ethnic minority populations report experiencing discrimination in health care and non-health care settings (Mays et al., 2007). Greater proportions of African Americans than members of other groups report either experiencing discrimination personally or perceiving it as affecting African Americans in general, even if they have not experienced it personally. Hate crimes motivated by race or ethnicity bias disproportionately affect Hispanics and African Americans (UCR, 2015) (see the public safety section in this chapter for more on hate crimes).
Discrimination is generally associated with worse mental health (Berger and Sarnyai, 2015; Gee et al., 2009; Paradies, 2006b; Williams and Mohammed, 2009); greater engagement in risky behaviors (Gee et al., 2009; Paradies, 2006b; Williams and Mohammed, 2009); decreased neurological responses (Harrell et al., 2003; Mays et al., 2007) and other biomarkers signaling the dysregulation of allostatic load; hypertension-related outcomes (Sims et al., 2012), though some evidence suggests racism does not drive these outcomes (Roberts et al., 2008); reduced likelihood of some health protecting behaviors (Pascoe and Smart Richman, 2009); and poorer birth-related outcomes such as preterm delivery (Alhusen et al., 2016). Paradoxically, despite higher levels of exposure to discrimination, the mental health consequences may be less severe among African Americans than they are among members of other groups, especially Asian populations (Gee et al., 2009; Williams and Mohammed, 2009). Researchers have suggested that African Americans draw on reserves of resilience in ways that temper the effects of discrimination on mental health (Brown and Tylka, 2011).
Though people may experience overt forms of racism (e.g., being unfairly fired on the basis of race), the adverse health effects of racism appear to stem primarily from the stress of chronic exposure to seemingly minor forms of “everyday racism” (i.e., racial microaggressions), such as being treated with less respect by others, being stopped by police for no apparent reason, or being monitored by salespeople while shopping (APA, 2016; Sue et al., 2007; Williams et al., 2003). The chronic exposure contributes to stress-related physiological effects. Thus, discrimination appears to exert its greatest effects not because of exposure to a single life traumatic incident but because people must mentally and physically contend with or be prepared to contend with seemingly minor insults and assaults on a near continual basis (APA, 2016). The implications appear to be greatest for stress-related conditions such as those tied to hypertension, mental health outcomes, substance abuse behaviors, and birth-related outcomes (e.g., low birth weight and premature birth) than for other outcomes (Williams and Mohammed, 2009).
Higher socioeconomic status (SES) does not protect racial and ethnic minorities from discriminatory exposures. In fact, it may increase opportunities for exposure to discrimination. The concept of “John Henryism” is used to describe an intensely active way of tackling racial and other life challenges (James, 1994). Though the evidence is mixed, John Henryism may contribute to worse cardiovascular outcomes among African American males who respond to racism by working even harder to disprove racial stereotypes (Flaskerud, 2012; Subramanyam et al., 2013).
Implicit bias John Dovidio defines implicit bias—a mechanism of unconscious discrimination—as a form of racial or other bias that operates beneath the level of consciousness (Dovidio et al., 2002). Research conducted over more than four decades finds that individuals hold racial biases of which they are not aware and, importantly, that discriminatory behaviors can be predicted based on this construct (Staats et al., 2016). The effects are greatest in situations marked by ambiguity, stress, and time constraints (Bertrand et al., 2005; Dovidio and Gaertner, 2000). Implicit bias is not an arbitrary personal preference that individuals hold; for example, “I just happen to prefer pears over apples.” Rather, the nature and direction of individuals' biases are structured by the racial stratification and norms of society. As a result, they are predictable.
Much of the public health literature has focused on the implicit biases of health care providers, who with little time to devote to each patient can provide care that is systematically worse for African American patients than for white patients even though the health care provider never intended to do so (IOM and NRC, 2003; van Ryn and Burke, 2000). The evidence is clear that unconscious racialized perceptions contribute to differences in how various individual actors, including health care providers, perceive others and treat them. Based on psychology lab experiments, functional magnetic resonance imaging (fMRI) pictures of the brain, and other tools, researchers find that white providers hold implicit biases against African Americans and that, to a lesser degree, some minority providers may also hold these biases (Hall et al., 2015). Although not limited to health care professionals, the biases lead providers to link negative characteristics (e.g., bad) and emotions (e.g., fear) with people or images they perceive as being African American (Zestcott et al., 2016). As a result of such implicit biases, physicians treat patients differently depending on the patient's race, ethnicity, gender, or other assumed or actual characteristics (IOM and NRC, 2003; Zestcott et al., 2016).
Given the importance of implicit bias, researchers have considered the role of health care provider–patient racial and ethnic concordance. Even if patients have similar clinical profiles, their care may differ systematically based on their race or ethnicity and that of their health care provider (Betancourt et al., 2014; van Ryn and Fu, 2003; Zestcott et al., 2016). The evidence on whether and how patient–provider concordance contributes to health disparities is mixed (van Ryn and Fu, 2003). Qualitative and quantitative findings suggest that patients do not necessarily prefer providers of the same race or ethnicity; they prefer a provider who treats them with respect (Dale et al., 2010; Ibrahim et al., 2004; Schnittker and Liang, 2006; Volandes et al., 2008). Providers appear to evaluate African American patients more negatively than they do similar white patients; seem to perceive them as more likely to participate in risky health behaviors; and may be less willing to prescribe them pain medications and narcotics medications (van Ryn and Fu, 2003). In a video-based study conducted among primary care providers, the odds ratio of providers referring simulated African American patients to otherwise identical white patients for cardiac catheterization was 0.6 (Schulman et al., 1999). Some evidence suggests minority providers deliver more equitable care to their diverse patients than white providers. For instance, a longitudinal study among African American and white HIV-positive patients enrolled in HIV care found that white doctors took longer to prescribe protease inhibitors (an effective HIV medication) for their African American patients than for their clinically similar white patients. Providers prescribed them on average 162 days earlier for white patients than for comparable African American patients (King et al., 2004). Among African American providers, there was no difference between African American and white patients in how long before providers prescribed the medications.
Racial and ethnic minority providers play an important role in addressing disparities because they help bridge cultural gulfs (Butler et al., 2014; Cooper et al., 2003; Lehman et al., 2012), and greater proportions of them serve minority and socially disadvantaged communities (Cooper and Powe, 2004); however, these providers are underrepresented in the health professions, and they face challenges that may constrain their professional development and the quality of care they are able to provide (Landrine and Corral, 2009). Specifically, they are more likely to serve patients in resource-poorer areas and lack professional privileges associated with academic and other resource-rich institutions. The structural inequities have implications not only for individual clinicians but also for the patients and communities they serve. Pipeline programs that grow the numbers of minority providers may help to address underrepresentation in the health professions. The available data suggest that pipeline participants are more likely to care for poor or underserved patients when they join the workforce (McDougle et al., 2015). Supporting the professional development of and expanding the resources and tools available to providers working in resource-poor communities seems to be one option for improving access to and quality of care; however, the literature does not clearly elucidate the relationship between health care workforce pipeline programs (e.g., to grow the numbers of minority providers) and their impact on the social determinants of health for poor and underserved communities (Brown et al., 2005; Smith et al., 2009). A commitment to equity is not enough to remedy the discriminatory treatment that results from implicit biases because the inadvertent discriminatory behavior co-occurs alongside deeply held personal commitments to equity. Identifying implicit biases and acknowledging them is one of the most effective steps that can be taken to address their effects (Zestcott et al., 2016). Trainings can help health care providers identify their implicit biases. Well-planned allocations of resources, including time, may afford them sufficient opportunity to account for it while serving diverse persons/patients.
Residential segregation—that is, the degree to which groups live separately from one another (Massey and Denton, 1988)—can exacerbate the rates of disease among minorities, and social isolation can reduce the public's sense of urgency about the need to intervene (Acevedo-Garcia, 2000; Wallace and Wallace, 1997). The effects of racial segregation differ from those of socioeconomic segregation. Lower SES whites are more likely to live in areas with a range of SES levels, which affords even the poorest residents of these communities access to shared resources (e.g., parks, schools) that buffer against the effects of poverty (APA Task Force on Socioeconomic Status, 2007; North Carolina Institute of Medicine Task Force on Prevention, 2009). By contrast, racial and ethnic minorities are more likely to live in areas of concentrated poverty (Bishaw, 2011). Indeed, if shared resources are of poor quality, they may compound the low SES challenges an individual faces. Racial segregation contributes to disparities in a variety of ways. It limits the socioeconomic resources available to residents of minority neighborhoods as employers and higher SES individuals leave the neighborhoods; it reduces health care provider density in predominately African American communities, which affects access to health care (Gaskin et al., 2012); it constrains opportunities to engage in recommended health behaviors such as walking; it may be associated with greater density of alcohol outlets, tobacco advertisements, and fast food outlets in African American and other minority neighborhoods (Berke et al., 2010; Hackbarth et al., 1995; Kwate, 2008; LaVeist and Wallace, 2000); it increases the risk for exposure to environmental hazards (Brulle and Pellow, 2006); and it contributes to the mental and physical consequences of prevalent violence, including gun violence and aggressive policing (Landrine and Corral, 2009; Massey and Denton, 1989; Polednak, 1996).
Historical trauma, “a collective complex trauma inflicted on a group of people who share a specific group identity or affiliation” (Evans-Campbell, 2008, p. 320), manifests from the past treatment of certain racial and ethnic groups, especially Native Americans. This is another form of structural (i.e., systemic) racism that continues to shape the opportunities, risks, and health outcomes of these populations today (Gee and Ford, 2011; Gee and Payne-Sturges, 2004; Heart et al., 2011). The past consignment of Native Americans to reservations with limited resources continues to constrain physical and mental health in these communities; however, the methods to support research on this topic have not yet been fully developed (Heart et al., 2011). Additional details on the health of Native Americans are presented in Chapter 2 and Appendix A.
The literature includes a small number of tested interventions. Interventions to address the health consequences of racism need not target racism in order to address the disparities it helps to produce. Furthermore, despite the deeply rooted nature of racism, communities are taking action to address the issue. (See Box 3-2 for a brief example of a community targeting structural racism and Box 3-3 for guidance on how to start a conversation about race.) Policy interventions and multi-sectoral efforts may be necessary to address structural factors such as segregation.
Addressing Structural Racism in Everett, Massachusetts, Through Improving Community–Police Interactions.
