In 2001 the landmark Surgeon General's report Mental Health: Culture, Race, and Ethnicity examined disparities in mental health care and treatment in racial and ethnic minority groups (1). The report highlighted three main issues regarding these disparities: persons from racial and ethnic minority groups have less access to health care than do non-Hispanic whites; persons from minority groups are less likely than persons from nonminority groups to receive treatment for mental illnesses; and when minority populations do receive treatment, the care is more likely to be of poor quality. In addition, racial and ethnic disparities in mental health expenditures appear to be widening over time (2).
Recent studies using representative, national samples have suggested that compared with non-Hispanic white persons, certain minority populations have a higher prevalence of chronicity and disability associated with psychiatric disorders (3). For example, among 6,082 participants surveyed in the National Survey of American Life, people of African-American and Caribbean race-ethnicity were more likely (57% and 56%, respectively) than people of non-Hispanic white ethnicity (39%) to have chronic major depressive disorder (4). In addition, Caribbean blacks and African Americans reported greater severity and disability associated with their depression compared with non-Hispanic whites (4). Inasmuch as treatment can foster remission and reduce disability, these findings highlight the particular importance of timely and appropriate treatment in these specific populations.
Despite the greater persistence and severity of mental illnesses among many minority populations, access to mental health treatment often lags behind that of non-Hispanic white populations (4,5). Analysis of U.S. national probability sample surveys shows that people of African-American and Hispanic or Latino race-ethnicity receive approximately half the amount of outpatient mental health treatment as non-Hispanic whites (6). In addition, when they are treated, persons from racial-ethnic minority groups are more likely to receive mental health treatment from primary care physicians than from psychiatrists or other specialty mental health professionals (7,8,9). These differences persist after analyses have controlled for socioeconomic status and insurance coverage, suggesting a need to better understand cultural and attitudinal factors driving these disparities in care.
Some limited previous research has investigated the impact of stigma on differences in access to mental health treatment. Stigma can be thought of generally as psychological and social attributes that lead to prejudice, devaluation, and overt discrimination (10). In a review of the evidence, Schraufnagel and colleagues (11) found that persons from minority populations were less likely to utilize mental health care because of stigmatizing beliefs about mental illness that originated from families, social networks, and religious communities. Stigmatizing perceptions about mental illness are often major impediments to accessing mental health services for all populations, particularly for minority populations (12). Nadeem and colleagues (12) found that among 1,577 U.S.-born women, African-American women were more likely than their non-Hispanic white counterparts to report perceived stigma about depression and that African-American and Hispanic-American women expressed less desire for treatment for mental illnesses than non-Hispanic white women. Persons with mental illnesses who are from racial-ethnic minority groups often experience "double stigma," or discrimination associated with living with a mental illness in conjunction with racial-ethnic discrimination, which causes some persons either to avoid seeking treatment or to not adequately participate in treatment (13).
Despite these findings, conflicting data exist regarding the impact of stigma on minority populations and their accessing of services. More recent examinations of attitudes of racial-ethnic minority populations and help-seeking behaviors, based on data from the National Comorbidity Survey (NCS) and the National Comorbidity Survey Replication (NCS-R), have shown that African Americans report a greater likelihood than non-Hispanic whites of seeking mental health services from professionals and are less likely than non-Hispanic whites to report embarrassment for seeking treatment (14,15). These newer data suggest that when compared with non-Hispanic whites, some minority populations actually report more favorable attitudes toward mental health treatment. The contradictory results of the available research reflect the complexity of this issue and the fact that there are no simple interpretations.
The goal of this study was to examine differences in the attitudes of people of non-Hispanic white, African-American, and Hispanic or Latino race-ethnicity toward seeking mental health treatment in the NCS-R, a nationally representative, U.S. sample. The investigation focused on attitudes related to willingness to seek help, comfort level in talking to a mental health professional, and perceived embarrassment about seeking help. It was hypothesized that African Americans and Hispanics or Latinos would be less willing to seek help, would be less comfortable talking to a professional, and would have greater perceived embarrassment about seeking help than non-Hispanic whites. This investigation sought to clarify the conflicting information in previous studies by applying a straightforward analytic approach to an existing large, representative data set.
The sample included individuals who participated in the NCS-R, a nationally representative, psychiatric epidemiological, cross-sectional survey of household populations of the 48 contiguous states (16). The NCS-R interview was administered between February 2001 and December 2003 to 9,282 English-speaking adults aged 18 years and older. The response rate was 71% (17).
The NCS-R interviews were conducted in two parts, with part 1 being a core diagnostic assessment module administered to all participants and part 2 including more specific diagnostic assessments and questions about attitudes and beliefs, among other detailed queries. Part 2 was administered to 5,692 of the 9,282 respondents. Respondents with clinically significant psychopathology were oversampled in part 2 (17).
In this analysis, three key dependent variables for evaluating attitudes toward mental health treatment seeking from the NCS-R were assessed. Each was measured with a closed-ended question that provided four response options. Question 1 was "People differ a lot in their feelings about professional help for emotional problems. If you had a serious emotional problem, would you definitely go for professional help, probably go, probably not go, or definitely not go for professional help?" Question 2 was "How comfortable would you feel talking about personal problems with a professional—very comfortable, somewhat, not very, or not at all comfortable?" And question 3 was "How embarrassed would you be if your friends knew you were getting professional help for an emotional problem—very embarrassed, somewhat, not very, or not at all embarrassed?" The latter item was used as a proxy for stigma. There are many ways to operationalize the complex stigma construct; for this analysis, stigma associated with seeking mental health treatment was analyzed by the level of embarrassment perceived by respondents.
For this study, the four responses to each question were dichotomized into two response categories, either positive or negative. That is, for question 1, positive responses were "definitely go" and "probably go," whereas negative responses were "definitely not go" and "probably not go;" for question 2, positive responses were "very comfortable" and "somewhat comfortable," and negative responses were "not very comfortable" and "not at all comfortable;" and for question 3, positive responses were "very embarrassed" and "somewhat embarrassed," and negative responses were "not very embarrassed" and "not at all embarrassed."
Categories for race-ethnicity in the NCS-R included the following: Vietnamese, Filipino, Chinese, all other Asian, Cuban, Puerto Rican, Mexican, all other Hispanic, Afro-Caribbean, African American, non-Latino white, and "all other." For this study, Vietnamese, Filipino, Chinese, all other Asian, Cuban, Puerto Rican, Afro-Caribbean, and "all other" were excluded because of small sample sizes (<284 in each of those race-ethnicity groups). Mexican and all other Hispanic categories were combined into a general Hispanic category. As a result, three categories of race-ethnicity were used in the analysis: non-Latino white, African American, and Hispanic, herein referred to as non-Hispanic white, African American, and Hispanic or Latino.
Additional sociodemographic variables included age, gender, region of the country (Northeast, Midwest, South, and West), educational attainment (<12 years, 12 years, 13–15 years, and ≥16 years), and income as a percentage of the federal poverty level for 2001 (low income, <150% of the poverty level; low-average, 150%–299%; high-average, 300%–599%; and high, ≥600%) (16). Other sociodemographic variables included marital status (married or cohabiting; divorced, separated, or widowed; and never married), employment status (employed, unemployed, and not in the labor force), and religious affiliation (Protestant, Catholic, and no identified religious affiliation). Regarding the latter variable, responses of "other" and "don't know" and refusals to answer were excluded because of small samples (552, nine, and 14, respectively).
Using independent-samples Student's t tests and chi square tests, we investigated bivariate associations between race-ethnicity and attitudes about mental health treatment seeking. Multiple logistic regression models were then constructed to assess the independent significance of race on attitudes and beliefs about mental health treatment seeking, with controls for key socioeconomic variables.
NCS-R used complex stratified sampling designs. Survey weights, stratification, and clustering (primary sampling units) were included in the analysis to adjust for their effects and make samples representative of the U.S. population. All percentages reported here are weighted. Survey documentation and analysis software, available online, were used for all analyses (18).
Table 1 shows sociodemographic characteristics of the sample by race-ethnicity. Considerable differences were found across racial-ethnic groups. Age, gender, region of the country, educational attainment, income, marital status, employment status, and religious affiliation each differed across racial-ethnic groups. Respondents in the Hispanic or Latino group were younger than the non-Hispanic white group. Hispanics or Latinos and African Americans had lower income levels than non-Hispanic whites. For example, 258 of 667 (39%) African Americans and 221 of 630 (35%) Hispanics or Latinos had low income, compared with 677 of 4,141 (16%) non-Hispanic whites. A higher percentage of Hispanics or Latinos were unemployed compared with non-Hispanic whites and African Americans.
Race-ethnicity and attitudes toward mental health treatment seeking
Table 2 shows bivariate (unadjusted) associations between race-ethnicity and attitudes toward mental health treatment seeking. Differences between the three racial-ethnic groups were examined in relation to whether respondents provided a positive response or negative response to the question, "If you had a serious emotional problem, would you … go for professional help?" As shown in Table 2, compared with non-Hispanic white respondents, African-American respondents reported being more likely to go for professional help (87% versus 84%), whereas Hispanic or Latino respondents were less likely to go for professional help (81%). In response to the question, "How comfortable would you feel talking about personal problems with a professional?" African Americans (76%) were less likely than non-Hispanic whites (80%) to report feeling comfortable. Finally, African Americans (29%) and Hispanics or Latinos (30%) were less likely than non-Hispanic whites (35%) to report feeling embarrassed when asked, "How embarrassed would you be if your friends knew you were getting professional help for an emotional problem?"
Expressed in terms of odds ratios (ORs) and corresponding 95% confidence intervals (CIs), Hispanics or Latinos were less likely than non-Hispanic whites to report being likely to go for professional help (OR=.79, CI=.64–.98) (Table 2). African Americans had a 23% decrease in the odds of feeling comfortable talking to a professional compared with non-Hispanic whites (OR=.77, CI=.63–.93). African Americans had greater odds of not feeling embarrassed if friends knew they were getting help compared with non-Hispanic whites (OR=1.32, CI=1.10–1.58). Similar results were found in comparisons of Hispanics or Latinos with non-Hispanic whites (OR=1.23, CI=1.03–1.48) on this latter variable.
Table 2 also shows adjusted ORs and corresponding CIs from regression analyses that controlled for three key socioeconomic status variables (education, employment, and income) based on an a priori analytic decision. After analyses adjusted for socioeconomic status, we found that African-American respondents were more likely than non-Hispanic white respondents to report that they would definitely or probably seek professional help (OR=1.36, CI=1.06–1.75). Being Hispanic or Latino was not an independently significant predictor of seeking professional help for a serious emotional problem. After we adjusted for socioeconomic status, we found that racial-ethnic minority status did not independently significantly predict reports of being comfortable with talking to a professional about personal problems. However, being African American increased the odds of reporting not being embarrassed that friends knew about getting professional help compared with non-Hispanic whites (OR=1.29, CI=1.06–1.56). Being Hispanic or Latino was not found to be an independent significant predictor of level of embarrassment.
In contrast to the original hypothesis, African Americans were more likely than Hispanics or Latinos and non-Hispanic whites to report that they would seek professional help for serious emotional problems and were less likely than non-Hispanic whites to report that they would feel embarrassed if friends found out about their treatment seeking. Although these findings are supported by recent studies (14,15), they contradict the prevailing assumptions that African Americans are more likely than non-Hispanic whites to avoid professional treatment for serious emotional problems (11,12,13,14,15). In fact, the results suggest that at least in terms of self-reported attitudes, African Americans may be less impeded by stigmatizing beliefs (such as embarrassment if others found out about their mental health treatment) than non-Hispanic whites. Such findings, from a large, representative data set, have important implications for understanding well-documented health disparities based on race-ethnicity.
As reported in the Surgeon General's Mental Health: Culture, Race, and Ethnicity and the President's New Freedom Commission on Mental Health, racial-ethnic minority populations have lower rates of accessing mental health treatment (1,3). If African Americans have more positive attitudes toward mental health treatment seeking than Hispanics or Latinos and non-Hispanic whites, then additional factors must play a role in the decreased use of services among African-American and Hispanic or Latino populations compared with non-Hispanic whites. The Institute of Medicine has outlined a range of patient-, provider-, and system-level factors that underlie disparities in general health care (19). A similar framework may be valuable in understanding factors contributing to mental health care disparities (20). The findings presented here suggest a need to move beyond assumptions that race-based disparities are mediated by unfavorable attitudes toward mental health care treatment seeking. Individuals' own reluctance to seek care may not sufficiently explain inequality in service use. Other patient-, provider-, and system-level mediators should be explored in similar large-scale U.S. data sets.
Patient-level factors may contribute to disparities in care when African Americans seek treatment in settings other than traditional psychiatric services. For example, African Americans may choose to obtain mental health care from primary care clinicians, clergy, and individuals within their social support network. In fact, research has shown that African Americans are more likely to discuss mental health problems in primary care settings without seeking treatment from mental health specialists (21). Even when patients have favorable attitudes toward mental health care, provider bias might reduce patients' access to care. For example, a provider may erroneously assume that persons from specific minority groups would be resistant to therapy or other mental health treatments, which may lead to lack of referral to appropriate treatment settings. Finally, system-level factors, such as insurance status, may play a role in the decreased rates of utilization despite the findings of apparently more favorable attitudes toward help seeking.
Despite the overall positive attitudes documented here among African Americans toward mental health care, cultural and linguistic barriers may create a greater degree of discomfort than non-Hispanic whites feel in talking to a mental health professional. Cultural and linguistic barriers may create system-level barriers that can impede the development of an effective therapeutic alliance between mental health professionals and patients. People often report less comfort with sharing personal information about themselves with people of different cultural backgrounds. Discrimination and unethical standards, exemplified by historical events such as the Tuskegee Syphilis Study, contribute to cultural mistrust, feelings of powerlessness, and discomfort with health care professionals that may continue to exist in the African-American community (22,23).
Several methodological limitations of this study should be considered. Responses to the questions posed in the NCS-R are self-reported and may be confounded by social desirability bias. However, there is no apparent reason to believe that such bias would have influenced sociodemographic groups differentially. In this study, embarrassment about friends finding out about mental health treatment was used as a proxy for stigma. However, stigma is a complex construct that cannot be so easily simplified to this individual survey question. Similarly, within the NCS-R, there are relatively few questions that assess attitudes about mental health treatment seeking; therefore, further assessment of attitudes related to discrimination and perceptions of treatments received could provide greater insight into overall attitudes about treatment acceptability. Finally, it was not possible to examine actual use of services in this analysis. Future studies should examine how attitudes toward mental health treatment predict disparities in service utilization among racial and ethnic minority populations.
Contrary to conventional wisdom and the investigative team's initial expectations, African Americans and Hispanics or Latinos may have more positive attitudes toward mental health treatment seeking than non-Hispanic whites. Although these findings contradict previous research showing that minority populations have less positive attitudes toward mental health treatment seeking than the non-Hispanic white population, this research expands the knowledge base for this extremely complex issue. Given documented low rates of mental health services use in these racial-ethnic minority groups, it is imperative to gain a more sophisticated and comprehensive understanding of the barriers to mental health services in these populations.
The NCS-R is supported by grant U01-MH 60220 from the National Institute of Mental Health, with supplemental support by grants from the National Institute on Drug Abuse and the Substance Abuse and Mental Health Services Administration, grant 044708 from the Robert Wood Johnson Foundation, and by the John W. Alden Trust.
The authors report no competing interests.