64PART 1 The Profession of Medicine

In addition, in the setting of even a minimal language barrier,

provider–patient communication without an interpreter is recognized

as a major challenge to effective health care delivery. These communication barriers for patients with limited English proficiency lead

to frequent misunderstanding of diagnosis, treatment, and follow-up

plans; inappropriate use of medications; lack of informed consent for

surgical procedures; high rates of adverse events with more serious

clinical consequences; and a lower-quality health care experience than

is provided to patients who speak fluent English. Physicians who have

access to trained interpreters report a significantly higher quality of

patient–physician communication than physicians who use other

methods. Communication issues related to discordant language disproportionately affect minorities and likely contribute to racial/ethnic

disparities in health care.

CLINICAL DECISION-MAKING Theory and research suggest that variations in clinical decision-making may contribute to racial and ethnic

disparities in health care. Two factors are central to this process: clinical

uncertainty and stereotyping.

First, a doctor’s decision-making process is nested in clinical uncertainty. Doctors depend on inferences about severity based on what

they understand about illness and the information obtained from the

patient. A doctor caring for a patient whose symptoms he or she has

difficulty understanding and whose “signals”—the set of clues and

indications that physicians rely on to make clinical decisions—are hard

to read may make a decision different from the one that would be made

for another patient who presents with exactly the same clinical condition. Given that the expression of symptoms may differ among cultural

and racial groups, doctors—the overwhelming majority of whom are

white—may understand symptoms best when expressed by patients of

their own racial/ethnic groups. The consequence is that white patients

may be treated differently from minority patients. Differences in clinical decisions can arise from this mechanism even when the doctor has

the same regard for each patient (i.e., is not prejudiced).

Second, the literature on social cognitive theory highlights how natural tendencies to stereotype may influence clinical decision-making.

Stereotyping can be defined as the way in which people use social categories (e.g., race, gender, age) in acquiring, processing, and recalling

information about others. Faced with enormous information loads and

the need to make many decisions, people often subconsciously simplify the decision-making process and lessen cognitive effort by using

“categories” or “stereotypes” that bundle information into groups or

types that can be processed more quickly. Although functional, stereotyping can be systematically biased, as people are automatically

classified into social categories based on dimensions such as race,

gender, and age. Many people may not be aware of their attitudes, may

not consciously endorse specific stereotypes, and paradoxically may

consider themselves egalitarian and not prejudiced.

Stereotypes may be strongly influenced by the messages presented

consciously and unconsciously in society. For instance, if the media

and our social/professional contacts tend to present images of minorities as being less educated, more violent, and nonadherent to health

care recommendations, these impressions may generate stereotypes

that unnaturally and unjustly impact clinical decision-making. As signs

of racism, classism, gender bias, and ageism are experienced (consciously or unconsciously) in our society, stereotypes may be created

that impact the way doctors manage patients from these groups. On

the basis of training or practice location, doctors may develop certain

perceptions about race/ethnicity, culture, and class that may evolve

into stereotypes. For example, many medical students and residents

are trained—and minorities cared for—in academic health centers or

public hospitals located in socioeconomically disadvantaged areas. As

a result, doctors may begin to equate certain races and ethnicities with

specific health beliefs and behaviors (e.g., “these patients” engage in

risky behaviors, “those patients” tend to be noncompliant) that are

more associated with the social environment (e.g., poverty) than with

a patient’s racial/ethnic background or cultural traditions. This “conditioning” phenomenon may also be operative if doctors are faced

with certain racial/ethnic patient groups who frequently do not choose

aggressive forms of diagnostic or therapeutic intervention. The result

over time may be that doctors begin to believe that “these patients” do

not like invasive procedures; thus, they may not offer these procedures

as options. A wide range of studies have documented the potential for

provider biases to contribute to racial/ethnic disparities in health care.

For example, one study measured physicians’ unconscious (or implicit)

biases and showed that these were related to differences in decisions to

provide thrombolysis for a hypothetical black or white patient with a

myocardial infarction.

It is important to differentiate stereotyping from prejudice and

discrimination. Prejudice is a conscious prejudgment of individuals

that may lead to disparate treatment, and discrimination is conscious

and intentional disparate treatment. All individuals stereotype subconsciously, yet, if left unquestioned, these subconscious assumptions may

lead to lower-quality care for certain groups because of differences in

clinical decision-making or differences in communication and patientcenteredness. For example, one study tested physicians’ unconscious

racial/ethnic biases and showed that patients perceived more biased

physicians as being less patient-centered in their communication.

What is particularly salient is that stereotypes tend to be activated

most in environments where the individual is stressed, multitasking,

and under time pressure—the hallmarks of the clinical encounter. In

fact, in a survey of close to 16,000 physicians, 42% admitted that bias—

including by race and ethnicity—impacted their clinical decisionmaking. Interestingly, emergency medicine physicians, who work in

environments of stress, time pressure, risk, and where they are multitasking, topped the list by discipline at 62%.

Patient-Level Factors Lack of trust has become a major concern

for many health care institutions today. For example, an IOM report,

To Err Is Human: Building a Safer Health System, documented alarming

rates of medical errors that made patients feel vulnerable and less trustful of the U.S. health care system. The increased media and academic

attention to problems related to quality of care (and of disparities themselves) has clearly diminished trust in doctors and nurses.

Trust is a crucial element in the therapeutic alliance between patient

and health care provider. It facilitates open communication and is

directly correlated with adherence to the physician’s recommendations

and the patient’s satisfaction. In other words, patients who mistrust

their health care providers are less satisfied with the care they receive,

and mistrust of the health care system greatly affects patients’ use of services. Mistrust can also result in inconsistent care, “doctor-shopping,”

0

20

40

Total White African

American

Hispanic Asian

American

19%

16%

23%

33%

27%

Percent of adults with one or more communication problems*

Base: Adults with health care visit in past two years

*Problems include understanding doctor, feeling doctor listened,

had questions but did not ask.

FIGURE 10-9 Communication difficulties with physicians, by race/ethnicity. The

reference population consisted of 6722 Americans ≥18 years of age who had made

a medical visit in the previous 2 years and were asked whether they had had trouble

understanding their doctors, whether they felt that the doctors had not listened, and

whether they had had medical questions they were afraid to ask. (Reproduced with

permission from the Commonwealth Fund Health Care Quality Survey, 2001.)

Racial and Ethnic Disparities in Health Care

65CHAPTER 10

self-medication, and an increased demand by patients for referrals and

diagnostic tests.

On the basis of historic factors such as discrimination, segregation,

and medical experimentation, blacks may be especially mistrustful of

providers. The exploitation of blacks by the U.S. Public Health Service

during the Tuskegee syphilis study from 1932 to 1972 left a legacy

of mistrust that persists even today among this population. Other

populations, including Native Americans/Alaskan Natives, Hispanics/

Latinos, and Asian Americans, also harbor significant mistrust of the

health care system. A national survey conducted by the Kaiser Family

Foundation found that there is significant mistrust for the health care

system among minority populations. Of the 3884 individuals surveyed,

36% of Hispanics and 35% of blacks (compared to 15% of whites) felt

they were treated unfairly in the health care system in the past based on

their race and ethnicity. Perhaps even more alarming—65% of blacks

and 58% of Hispanics (compared to 22% of whites) were afraid of being

treated unfairly in the future based on their race/ethnicity (Fig. 10-10).

This mistrust may contribute to wariness in accepting or following recommendations, undergoing invasive procedures, or participating in clinical research, and these choices, in turn, may lead to

misunderstanding and the perpetuation of stereotypes among health

professionals.

■ KEY RECOMMENDATIONS TO ADDRESS RACIAL/

ETHNIC DISPARITIES IN HEALTH CARE

Unequal Treatment provides recommendations to address the root

causes of racial/ethnic disparities organized as health system interventions, provider interventions, patient interventions, and general

recommendations.

Health System Interventions • COLLECTING, REPORTING, AND

TRACKING OF DATA ON HEALTH CARE ACCESS AND USE, BY PATIENTS’

RACE/ETHNICITY Unequal Treatment found that the appropriate

systems to track and monitor racial and ethnic disparities in health

care are lacking and that less is known about the disparities affecting

minority groups other than African Americans (Hispanics, Asian

Americans, Pacific Islanders, Native Americans, and Alaskan Natives).

For instance, only in the mid-1980s did the Medicare database begin

to collect data on patient groups outside the standard categories of

“white,” “black,” and “other.” Federal, private, and state-supported

data-collection efforts are scattered and unsystematic, and many

health care systems and hospitals still do not collect data on the race,

ethnicity, or primary language of enrollees or patients. A survey by the

Institute for Diversity in Health Management and the Health Research

and Educational Trust in 2015 found that 98% of 1083 U.S. hospitals

collected information on race, 95% collected data on ethnicity, and

94% collected data on primary language. However, only 45% collected

data on race, 40% collected data on ethnicity, and 38% collected data

on primary language to benchmark gaps in care. A survey by America’s

Health Insurance Plans Foundation in 2008 and 2010 showed that the

proportion of enrollees in plans that collected race/ethnicity data of

some type increased from 75 to 79%; however, the total percentage of

plan enrollees whose race/ethnicity and language are recorded is still

much lower than these figures.

COLLECTING, REPORTING, AND TRACKING OF SDOH DATA In 2014,

the IOM Committee on Recommended Social and Behavioral Domains

and Measures for Electronic Health Records recommended the routine

collection, in the electronic health record, of a parsimonious panel of

clinically significant SDOH measures that may be obtained by selfreport in advance of or during the health care encounter and, when

used together, provide a psychosocial vital sign. The IOM-recommended

questionnaire includes 25 items addressing the following domains: race

and ethnicity, education, financial resource strain, stress, depression,

physical activity, tobacco use, alcohol use, social connection or isolation, intimate partner violence, residential address, and geocoded census tract median income. Implementation studies have demonstrated

that collection of these data takes about 5 minutes, and both patients

and providers saw this data collection as appropriate and important.

Given that data access and monitoring is an essential component to disparities elimination, we highlight several important sources of up-todate racial/ethnic disparities monitoring initiatives that are available to

the general public and are updated regularly. We highlight only three

examples of national data sources.

• Since 2003, the Agency for Healthcare Research and Quality has

led the yearly compilation of The National Healthcare Quality and

Disparities Report, which reports trends for measures related to

access to health care, affordable care, care coordination, healthy living, patient safety, and the quality of care across acute and chronic

disease management by race/ethnicity, income, and other SDOH

(https://www.ahrq.gov/research/findings/nhqrdr/index.html).

• Since 2011, the Geospatial Research, Analysis, and Services Program

(GRASP) created and maintains the Centers for Disease Control

and Prevention Social Vulnerability Index. This database maps, for

all U.S. Census tracts, 15 social factors (grouped in four SDOH categories: socioeconomic status, housing composition and disability,

minority status and language, and housing and transportation) and

is updated every 2 years (https://www.atsdr.cdc.gov/placeandhealth/

svi/index.html).

• Launched in 2018, the Health Opportunity and Equity (HOPE) Initiative benchmarks and tracks 27 indicators by race, ethnicity, and

socioeconomic status. The indicators measure social and economic

factors, community and safety, physical environment, access to

health care, and health outcomes for the United States (https://www

.nationalcollaborative.org/our-programs/hope-initiative-project/).

INCREASE INSURANCE COVERAGE AND ACCESS Lack of access to

high-quality health care is an important driver of racial/ethnic disparities.

Signed into law in 2010, the Affordable Care Act (ACA) fundamentally

transformed health insurance by decreasing the uninsured population

from 16.3% in 2010 (~49.9 million) to 8.8.% in 2016 (~28.1 million). This

represents the largest expansion of health insurance since the creation

of Medicare and Medicaid in 1965. Prior to the ACA, non-Hispanic

blacks were 70% and Hispanics nearly three times more likely to be

uninsured than non-Hispanic whites. Of note, Medicaid expansion

accounted for an estimated 60% of the ACA’s effect through a combination of expanded eligibility and increased enrollment of previously

eligible but unenrolled people. This is important given the higher

number of racial/ethnic minorities who obtain insurance through

Medicaid. Many studies have demonstrated that increased insurance

coverage has also translated to greater improvement for blacks and

Hispanics in access to care, more access to a usual source of care, and

improved health outcomes.

ENCOURAGEMENT OF THE USE OF EVIDENCE-BASED GUIDELINES AND

QUALITY IMPROVEMENT Unequal Treatment highlights the subjectivity of clinical decision-making as a potential cause of racial and

0 20 40 60 80

Future unfair

Tx based on

race/ethnicity

Past unfair

Tx based on

race/ethnicity

Whites

Blacks

Latinos

15

35

36

22

65

58

Percent

FIGURE 10-10 Patient perspectives regarding unfair treatment (Tx) based on race/

ethnicity. The reference population consisted of 3884 individuals surveyed about

how fairly they had been treated in the health care system in the past and how fairly

they felt they would be treated in the future on the basis of their race/ethnicity. (From

Race, Ethnicity & Medical Care: A Survey of Public Perceptions and Experiences.

Kaiser Family Foundation, 2005.)

66PART 1 The Profession of Medicine

ethnic disparities in health care by describing how clinicians—despite

the existence of well-delineated practice guidelines—may offer (consciously or unconsciously) different diagnostic and therapeutic options

to different patients on the basis of their race or ethnicity. Therefore,

the widespread adoption and implementation of evidence-based guidelines is a key recommendation in eliminating disparities. For instance,

evidence-based guidelines are now available for the management of

diabetes, HIV/AIDS, cardiovascular diseases, cancer screening and

management, and asthma—all areas where significant disparities exist.

As part of ongoing quality-improvement efforts, particular attention

should be paid to the implementation of evidence-based guidelines for

all patients, regardless of their race and ethnicity.

SUPPORT FOR THE USE OF LANGUAGE INTERPRETATION SERVICES IN

THE CLINICAL SETTING As described previously, a lack of efficient

and effective interpreter services in a health care system can lead to

patient dissatisfaction, to poor comprehension and adherence, and

thus to ineffective/lower-quality care for patients with limited English

proficiency. Unequal Treatment’s recommendation to support the use

of interpretation services has clear implications for delivery of quality

health care by improving doctors’ ability to communicate effectively

with these patients.

INCREASES IN THE PROPORTION OF UNDERREPRESENTED MINORITIES

IN THE HEALTH CARE WORKFORCE Data for 2018 from the Association of American Medical Colleges indicate that of active physicians,

56.2% identified as white, 5.8% identified as Hispanic, 5.0% identified

as black or African American, and 0.3% identified as Native American

or Alaskan Natives. Furthermore, U.S. national data show that only

3.6% of full-time faculty are black or African American, and 5.5% are

Hispanic, Latino, or of Spanish origin (alone or in combination with

another race/ethnicity), compared to 63.9% who identified as white.

Longitudinal data demonstrate that minority faculty are more likely to

be at or below the rank of assistant professor, while whites composed

the highest proportion of full professors. Similarly, several studies have

found that both Hispanic and black faculty were promoted at lower

rates than their white counterparts. Despite representing ~30% of the

U.S. population (a number projected to almost double by 2050), minority students are still underrepresented in medical schools. In 2018,

matriculates to U.S. medical schools were 6.2% Latino, 7.1% African

American, 0.1% Native Hawaiian or Other Pacific Islander, and 0.2%

Native American or Alaskan Native. These percentages have decreased

or remained nearly the same since 2007. It will be difficult to develop a

diverse physician workforce that can meet the needs of an increasingly

diverse population without dramatic changes in the racial and ethnic

composition of medical student bodies. Long-term investment in pipeline programs and the nearly universal adoption of holistic admissions

(a process by which schools consider each applicant individually to

determine how they might contribute to the learning environment

and the workforce instead of relying just on test scores and grades)

have produced modest results. Institutional change in medical schools,

focused on creating nurturing, inclusive, and equity-focused environments that dismantle the structural racism that has created the opportunity gap faced by many minority students, is needed to address this

important workforce challenge.

Provider Interventions • INTEGRATION OF CROSS-CULTURAL EDUCATION INTO THE TRAINING OF ALL HEALTH CARE PROFESSIONALS

The goal of cross-cultural education is to improve providers’ ability

to understand, communicate with, and care for patients from diverse

backgrounds. Such education focuses on enhancing awareness of

sociocultural influences on health beliefs and behaviors and on building skills to facilitate understanding and management of these factors

in the medical encounter. Cross-cultural education includes curricula

on health care disparities, use of interpreters, and effective communication and negotiation across cultures. These curricula can be incorporated into health professions training in medical schools, residency

programs, nursing schools, and other health professions programs, and

can be offered as a component of continuing education. Despite the

importance of this area of education and the attention it has attracted

from medical education accreditation bodies, a national survey of

senior resident physicians by Weissman and colleagues found that up

to 28% felt unprepared to deal with cross-cultural issues, including

caring for patients who have religious beliefs that may affect treatment,

patients who use complementary medicine, patients who have health

beliefs at odds with Western medicine, patients who mistrust the health

care system, and new immigrants. In a study at one medical school,

70% of fourth-year students felt inadequately prepared to care for

patients with limited English proficiency. Efforts to incorporate crosscultural education into medical education will contribute to improving communication and to providing a better quality of care for all

patients.

INCORPORATION OF TEACHING ON THE IMPACT OF RACE, ETHNICITY,

AND CULTURE ON CLINICAL DECISION-MAKING Unequal Treatment

and more recent studies found that stereotyping by health care providers can lead to disparate treatment based on a patient’s race or ethnicity.

The Liaison Committee on Medical Education, which accredits medical schools, issued a directive that medical education should include

instruction on how a patient’s race, ethnicity, and culture might unconsciously impact communication and clinical decision-making.

Patient Interventions Difficulty navigating the health care system and obtaining access to care can be a hindrance to all populations,

particularly to minorities. Similarly, lack of empowerment or involvement in the medical encounter by minorities can be a barrier to care.

Patients need to be educated on how to navigate the health care system

and how best to access care. Interventions should be used to increase

patients’ participation in treatment decisions.

General Recommendations • INCREASE AWARENESS OF RACIAL/

ETHNIC DISPARITIES IN HEALTH CARE Efforts to raise awareness of

racial/ethnic health care disparities have done little for the general

public but have been fairly successful among physicians, according to

a Kaiser Family Foundation report. In 2006, nearly 6 in 10 people surveyed believed that blacks received the same quality of care as whites,

and 5 in 10 believed that Latinos received the same quality of care as

whites. These estimates are similar to findings in a 1999 survey. Despite

this lack of awareness, most people believed that all Americans deserve

quality care, regardless of their background. In contrast, the level of

awareness among physicians has risen sharply. In 2002, the majority

(69%) of physicians said that the health care system “rarely or never”

treated people unfairly on the basis of their racial/ethnic background.

In 2005, less than one-quarter (24%) of physicians disagreed with

the statement that “minority patients generally receive lower-quality

care than white patients.” More recently, a survey by WebMD showed

that 42% of 16,000 physicians admitted that their own personal biases

impact their clinical decision-making, including on characteristics

such as race and ethnicity. Increasing awareness of racial and ethnic

health disparities, and their root causes, among health care professionals and the public is an important first step in addressing these disparities. The ultimate goals are to generate discourse and to mobilize action

to address disparities at multiple levels, including health policymakers,

health systems, and the community.

CONDUCT FURTHER RESEARCH TO IDENTIFY SOURCES OF DISPARITIES

AND PROMISING INTERVENTIONS While the literature that formed

the basis for the findings reported and recommendations made in

Unequal Treatment provided significant evidence for racial and ethnic

disparities, additional research is needed in several areas. First, most of

the literature on disparities focuses on black-versus-white differences;

much less is known about the experiences of other minority groups.

Improving the ability to collect racial and ethnic patient data should

facilitate this process. However, in instances where the necessary systems are not yet in place, racial and ethnic patient data may be collected

prospectively in the setting of clinical or health services research to

more fully elucidate disparities for other populations. Second, much

of the literature on disparities to date has focused on defining areas in

which these disparities exist, but less has been done to identify the multiple factors that contribute to the disparities or to test interventions to

address these factors. There is clearly a need for research that identifies

promising practices and solutions to disparities.