A recent surgeon general’s report and various studies document racial and ethnic disparities in mental health care, including gaps in access, questionable diagnostic practices, and limited provision of optimum treatments. Bias is a little studied but viable explanation for these disparities.
It is important to isolate bias from other barriers to high-quality mental health care and to understand bias at several levels (practitioner, practice network or program, and community). More research is needed that directly evaluates the contribution of particular forms of bias to disparities in the area of mental health care.
RACIAL AND ETHNIC DISPARITIES are as widespread in the diagnosis and treatment of mental illness as they are in other areas of health. In 2001, then–Surgeon General David Satcher issued the report Race, Culture, and Ethnicity and Mental Health,1 in which he convincingly documented disparities in access and treatment that leave too many minority individuals untreated or improperly treated.
One possible reason for disparities is that practitioners and mental health program administrators make unwarranted judgments about people on the basis of race or ethnicity. Their inappropriate expectations lead to inappropriate decisions and actions. In a strict sense, it is these unwarranted views, reactions to a person “on the basis of perceived membership in a single human category, ignoring other category memberships and other personal attributes,”2 that constitute bias. Biased views can be held knowingly or unknowingly and can result in action or a failure to act.
Taking account of racial and ethnic differences does not in itself constitute bias. Indeed, some critics argue that responding to racial and ethnic differences is essential, that mental health interventions must be varied to allow for differences in race, culture, and ethnicity.3 They claim that appropriate treatment necessitates awareness of critical differences between minority individuals and others in beliefs and sensitivities related to mental health, in expression of symptoms, and in treatment preferences. From this perspective, to ignore racial and ethnic differences reflects a kind of bias.
There may be greater reason for concern about bias in mental health than in other areas of health. Some continue to doubt the very existence of mental illness, believing that difficulties labeled as such, however troublesome, are no more than universal problems in everyday living.
Consensus has increased about appropriate methods of diagnosis and treatment, but a large role remains for discretion. There is great variation in practice norms, and the advent of well-founded protocols4 is recent. These protocols are far from achieving full acceptance.
Decisionmakers other than mental health professionals, including business owners, neighbors, and the public at large, as well as police and courts, play an important role in assessing mental illness and in deciding whether troublesome behavior warrants treatment or punishment.5 Mentally ill persons can be detained by the police and required to undergo treatment against their will, a practice with few counterparts elsewhere in health. Institutional and community decisionmakers also enjoy considerable discretion, and there is great opportunity for bias to intrude.
It is useful as a starting point to consider disparities, examining the research literature for clues about bias. What is the evidence on disparities in mental health? What does it tell us about bias?
As noted by Surgeon General Satcher, epidemiological research consistently reveals that African, Asian, Native, and Latino Americans needing outpatient care are unlikely to receive it.1 Disparities persist after differences in socioeconomic status, region of residence, and other sociodemographic factors are controlled. They have been shown to occur among Mexican Americans, despite lower levels of need,1 as well as among children, adolescents, and the elderly.6
Some regional studies point to a lessening of differences between racial/ethnic groups in regard to treatment rates.7 Specialized programs, including those operated by the Department of Veterans Affairs, have reported encouraging results.8 At the same time, recently published national data suggest that, in the nation as a whole, access disparities persist.9
When sought, assistance for mental health problems is especially likely to come from providers in the general medical sector.10,11 For example, one study showed that, among individuals treated by the Indian Health Service, mental health and social problems were associated with one third of requests for services and that “[m]ental health was identified as the top health problem by 10 of 12 IHS areas and the Urban Indian Health Programs in [fiscal year] 2001.”12
There are disparities as well among members of minority groups who do seek mental health specialty treatment. African Americans, Latinos, Asian Americans, and Native Americans have been shown to be more likely than Whites to leave treatment prematurely.13 The “dropout problem” includes large numbers of individuals who attend only one treatment session and are unlikely to have received any benefit.
African American populations have received the greatest attention from researchers, and African American–White disparities have been revealed. A persistent finding has been that, along with Native Americans, African Americans are greatly overrepresented in inpatient settings.14 African Americans are overrepresented too in psychiatric emergency rooms. Dramatic changes in the mental health care system, including the advent of managed care, have had little impact on the overrepresentation of African Americans and Native Americans in emergency care settings.
Along with problems involving access, researchers have paid increasing attention in recent years to the quality of mental health care provided to members of minority groups. Young et al.15 reported that African Americans and Latinos were less likely than Whites to receive guideline-adherent treatment when suffering from anxiety disorders and depression. Similarly, Wang et al.16 found that African Americans were relatively unlikely to receive guideline-based care.
Gaps in quality extend to elderly and seriously mentally ill African Americans. Schneider et al.,17 investigating enrollees in Medicare health plans, discovered that African Americans were less likely than Whites to obtain treatment meeting a Health Plan Employer Data and Information Set guideline calling for a follow-up visit within 30 days after psychiatric hospitalization. Wang et al.18 found that African Americans were overrepresented among persons suffering from serious mental illness who failed to receive “minimally adequate” treatment.
Investigators have studied African Americans10 and Latinos19 visiting primary care physicians with mental health–related complaints. African Americans and Latinos have proved to be less likely than Whites to receive a prescription for psychotropic medication. Among elderly community residents as well, African Americans have been found to be relatively unlikely to receive antidepressant medications.20
Studying Medicaid recipients, Melfi et al.21 found that African Americans were less likely than Whites to receive an antidepressant medication and less likely, if they were in fact prescribed such a medication, to receive selective serotonin reuptake inhibitors. In another study of Medicaid recipients, Kuno and Rothbard22 reported that African Americans were less likely than Whites to receive newer atypical antipsychotic medications that have fewer side effects and more likely to receive injectable antipsychotics. Other researchers as well have demonstrated greater receipt of injectable antipsychotic medications among African Americans.23
In addition, when they are prescribed psychotropic medications, minority individuals sometimes receive suspiciously high doses. Segal et al.24 and Chung et al.25 reported that African Americans seen in psychiatric emergency services and inpatient settings were prescribed higher doses than others of antipsychotic medications. Other researchers have reported similar results.26,27
Although disparities in access, continuity, and quality are well established, it is hazardous to infer bias solely from the presence of disparities. To avoid misidentifying the problem and misdirecting attempts at finding solutions, we must consider explanations other than bias.
Socioeconomic differences are typically proposed at the outset. Often, researchers control for socioeconomic differences; after doing so, they typically continue to find evidence of disparities.1 Critics next turn to insurance coverage as a possible explanation. Minority individuals lack private health insurance in disproportionate numbers, a gap that is not eliminated by coverage obtained from public sources.1
Health insurance coverage facilitates treatment seeking, and widening the scope of coverage would benefit members of minority groups disproportionately. Investigators have found, however, that despite having lower incomes, minority individuals exhibit a less active response than Whites to reductions in the cost of mental health treatment.28
Thus, it appears that even after financial barriers have been removed, other factors continue to prevent minority individuals from seeking treatment. Bias is an important hypothesis, but it must be considered in a context of alternatives. Chief among them are lack of familiarity with mental illness–related concepts, preferences for interpreting mental health problems in spiritual or other culturally sanctioned terms, stigma, and coping habits that stress self-reliance and family reliance. Among Asian Americans, for example, researchers have identified cultural barriers to treatment seeking such as coping strategies that favor willpower and avoidance of morbid thoughts and kinship and family orientations that allow for recognizing mental illness only if it disrupts social harmony.29
Evidence is sparse, however, and suggests a complicated picture. Diala et al.,30 analyzing the National Comorbidity Survey, found that African Americans had more favorable attitudes than Whites toward mental health services before using them but less favorable attitudes after using them. Another study31 indicated that African American women were more likely than White women to affirm religious or supernatural causes of mental illness. Latinas were less likely than members of other groups to indicate stigma and less likely to subscribe to medical causation. On the other hand, an analysis of the General Social Survey revealed few differences between African Americans, Latinos, and Whites in terms of beliefs about causes of and appropriate treatment for mental illness.32
Establishing disparities by eliminating alternative competing alternative reflects an indirect approach based on circumstantial evidence. Direct evidence establishing bias, however limited, is important to consider. More than a decade ago, Lopez33 carefully weighed studies evaluating whether mental health decisionmaking, as conducted by practicing clinicians, was tainted by bias. He began by enlarging the definition of bias, noting 2 kinds of distortion: overpathologizing bias and minimization bias. The former occurs when unfamiliar behavior of minority individuals is interpreted as a manifestation of mental illness. The latter occurs when practitioners ignore genuine manifestations of mental illness. Lopez noted that, perhaps operating from a sense of misguided cultural sensitivity, practitioners could dismiss real mental illness, attributing genuine symptoms to variations in cultural beliefs and practices.
Lopez’s review was restricted to African Americans and Latinos, groups on whom data were available, and in fact few of the available studies addressed Latinos. Lopez found mixed evidence for most kinds of bias but stronger evidence of bias when clinicians diagnosed mental illness among African Americans.
The evidence of bias in diagnosis underscored a robust phenomenon in African American mental health. For more than 2 decades, researchers have documented that African Americans have higher than expected rates of diagnosed schizophrenia and lower rates of diagnosed affective disorders.34 These differences have aroused suspicion that clinicians indeed are biased in the course of routine practice. However, recent research suggests a complex picture, and differences between African Americans and Whites in how they present symptoms of mental illness to clinicians play a crucial role.35
There are indeed reasons to believe that clinicians misinterpret problems of minority individuals in making diagnoses and in formulating overall assessments of mental health problems. Translation, which is often necessary, leaves room for confusion. In some instances, important mental health–related concepts lack true equivalents in languages other than English, opening the way to misunderstanding of complaints.12 When faced with standardized assessment procedures, for example, some Asian Americans approach the very task of responding with tendencies different from those assumed by developers of the procedures.16
Bias might intrude in the formation of clinical relationships. The therapeutic alliance36 through which practitioner and client engage each other can be adversely affected by bias. The therapeutic alliance is compromised not only by outright rejection but also by lack of commitment to overcoming estrangement. The result can be alienation and lack of trust,37–39 compounded by cultural misunderstanding.40,41
Bias occurs in the beliefs and actions of individual clinicians, and it is at this level that it has received the greatest amount of attention. Bias also occurs when unfounded assumptions become normative beliefs shared by members of practitioner networks or treatment organizations. Bias occurs too when authorities and community members become particularly intolerant of minority individuals with mental illnesses and differentially enforce conformity norms of acceptable behavior.
Practice styles are local norms governing diagnosis and intervention. They reflect shared understandings of how clinical decisions should be made over the wide range of discretionary action open to clinicians. Practice styles come about because, in the course of informal interactions, people develop common understandings about how uncertainty should be handled.
Organizational culture also refers to shared, often unspoken understandings about procedures and goals. Researchers have measured dimensions of interorganizational differences in culture: quality emphasis, performance goals, coordination of care, communication, and conflict resolution.42
Shared understandings might include biases about the mental health status of or treatment expectations for ethnic minority clients. That clients from certain backgrounds are unreceptive to treatment, hostile, naive, superstitious, or otherwise unpromising might represent a prevailing view in a practitioner network or organization.
Shared understandings also can express themselves in neglect. Among minority communities as well as individuals, engagement can require overcoming reluctance and mistrust. Positive steps toward community engagement reflect necessary norms of commitment.43–46
Behaviors defining mental illness violate societal expectations of acceptable behavior; mental illness is a kind of deviance. Mental illness can elicit forces of social control.
Police and courts, as well as employers, merchants, neighbors, and family and friends, determine whether boundaries of acceptable behavior have been transgressed. When inconvenienced or threatened, community agents decide whether to respond and whether an appropriate response is personal, legal, or medical. Bias can be found in differential degrees of tolerance.
Researchers have documented notable differences between African Americans and Whites in rates of involuntary civil commitment.47 These differences are associated with differences in how mentally ill individuals are presented to the emergency room; African Americans are more likely to be brought in by police.48 African Americans and Latinos are overrepresented in jails and prisons, institutions with substantial representations of individuals who are mentally ill. The question of why differences occur in rates of civil commitment and in rates of incarceration associated with mental illness remains to be answered, and the role of bias in decisionmaking is yet to be determined.
In related research on the “visibility hypothesis,” investigators have found evidence that mentally ill individuals are more likely to be challenged when, as members of minority groups, they are visibly different from other community residents.49 They stand out as more worthy of attention than others—more visible—and deviant behavior is recognized more readily.
Community tolerance varies not only from community to community but also with the passage of time. Coerced treatment has been shown to increase with economic decline,50 the reason being that economic contraction produces greater insecurity, greater frustration, and less tolerance. Tolerance of members of racial and ethnic minorities especially appears to decline. Forces of social control appear aimed more at minority individuals, especially male African Americans, who exhibit symptoms of mental illness (R. F. Catalano, L. R. Snowden, and M. Shumway, unpublished data, 2002).
We know that there are disparities in access, treatment, and quality of mental health care, but we do not yet know the extent to which these disparities are attributable to bias. We know even less about bias operating at levels beyond that of the individual practitioner: in the practice network, treatment organization, and community.
Nevertheless, whether intentional or inadvertent, whether by active decisionmaking or by default, it is reasonable to believe that bias partially explains disparities. Social scientists have established that bias need not be blatant but rather can be “automatic, cool, indirect, ambiguous, ambivalent.”2 The ambiguity surrounding mental illness and appropriate treatment invites bias, including bias of a well-intentioned kind (i.e., minimization bias).33
Missing is knowledge of where, when, how, and to what extent bias operates in mental health decisionmaking and treatment. In addition, the contribution of bias relative to that of other factors has not yet been assessed. Determining the role of bias in mental health assessments is important in establishing a comprehensive explanation of disparities and, ultimately, efforts to effectively address them.