Abstract
Though researchers have described psychosocial barriers to mental health care-seeking, limited research has examined ways in which gender and race-ethnicity are associated with individuals’ perceptions and attitudes. This study investigates correlates of psychosocial barriers to mental health care in a population of adults reporting unmet need for mental health care, focusing on gender and race-ethnicity. Data are from the 2002 National Survey on Drug Use and Health. Multivariate analyses show that non-Latino white male status is positively associated with stigma avoidance and mistrust/fear of the mental health care system. Persons of lower income or educational status are less likely to report negative attitudes towards care. Findings imply a need to reconsider the roles of gender, race-ethnicity, and socioeconomic status within investigations of psychosocial barriers to care. Future research should examine the relationships among social status, help-seeking behaviors, and attitudes toward mental health care.
Through behavioral models of care-seeking and health-decision making, sociological theory outlines how psychosocial variables can impact whether an individual seeks and receives health care (Andersen 1995; Mechanic 1962). Utilization of mental health services varies systematically by race-ethnicity and gender (Wang et al. 2005); males and racial-ethnic minorities receive less care than women and non-Hispanic whites. Applying behavioral models to our data, we propose that members of diverse racial-ethnic and gender groups differentially perceive or experience psychosocial barriers to mental health care, thereby reflecting variation in care-seeking attitudes and behaviors.
This study examines relationships between sociodemographic variables and attitudes underlying psychosocial barriers to care, using a nationally representative, population-based survey. We specifically investigate whether race-ethnicity and gender are differentially associated with self-reported stigma avoidance, mistrust/fear of the health care system, or negative attitudes toward treatment among adults who report unmet need for mental health care. By focusing on this population, we are better able to understand factors that result in diminished contact with mental health systems of care by clients or potential clients and which may result in racial and ethnic or gender differences in mental health service use. Moreover, examining the experiences of persons reporting unmet need is valuable for the development of policy and social interventions that can address these barriers. The resulting data may inform interventions that may narrow disparities in mental health service utilization among diverse populations.
Models of behavior surrounding illness and care-seeking, such as Andersen’s (1995) and Mechanic’s (1962), emphasize the social and behavioral underpinnings of health decision-making. Andersen identifies health beliefs as an important component of perceived need for health services and actual utilization. Attitudes and values about health and health services “provide one means of explaining how social structure might influence enabling resources, perceived need, and subsequent use” (p. 2). While beliefs about health and perceived need for services are insufficient to fully explain behavior, they are a potential source of information regarding behavioral variation in service utilization across populations.
Mechanic (1962) identifies “illness behavior” as an essential component of whether diagnosis and treatment will ever occur; this behavior is in part predicated upon whether the adoption of a “sick role” is consistent with an individual’s social setting and position within the social group. Situations of high “illness danger” (i.e., when the illness outcome is un-predictable and the amount of threat and loss likely to result from the illness is high) can negatively affect an individual’s inclination to seek medical care. Potential sources of danger include feared loss of social status or an in-ability to fulfill one’s social role. Gender, race-ethnicity, and other socioeconomic variables constitute salient markers of social role and status. Consequently, attitudes and perceived barriers may vary for persons subject to different normative social expectations or based on how much social status they perceive they may lose. These social variations have potential implications for the expression of illness behaviors, which in turn may have practical consequences within the provision of mental health care.
Race and ethnicity have been identified as correlates of attitudes, perceptions, and behaviors related to seeking mental health care (Alvidrez 1999; Ojeda and McGuire 2006). There are numerous reasons why attitudes can result in differential help-seeking behavior across ethnic communities. For example, Alegría et al. (2002) identify the importance of self-reliance among poor Latinos as one barrier to mental health care. Stigma and shame have been suggested as potential impediments to care-seeking among minority groups (Anglin, Link, and Phelan 2006; Antai-Otong 2002; Chiu 2004; Okazaki 2000; Wynaden etal. 2005).
Gender also has implications for help-seeking (Albizu-Garcia et al. 2001; Cleary, Mechanic, and Greenley 1982; Hinton et al. 2006; Kessler, Brown, and Broman 1981) and care pathways (Cleary, Burns, and Nycz 1990). Mental health-care seeking behaviors have been found to differ systematically between males and females (Ojeda and McGuire 2006). Stigma is considered a potential explanatory factor for the lower rate of mental health treatment among men, even those with severe mental illness (Wang et al. 2005). Though the National Institute of Mental Health has identified increased male help-seeking for depression as a goal (National Institute of Mental Health 2006), relatively little research has addressed the particular reasons that men do not seek care.
Finally, though few investigations have examined the intersection of race-ethnicity and gender in health attitudes and behaviors, the concept of intersectionality (Bograd 1999) suggests that overlapping social categories have important social consequences that extend beyond basic demographic variables. This emerging work suggests that interactions between categories are themselves important analytical entities.
BACKGROUND
Investigations examining psychosocial components of decision-making in mental health care, including help-seeking behaviors and treatment adherence, have identified other barriers to adequate care beyond issues pertaining to financial and physical access to services (Edlund et al. 2002). For example, mental health treatment dropout rates in the United States and Ontario were similar despite Canada’s nationalized health care; in both sites, the authors identified perceived stigma and other psychological barriers as important characteristics of those likely to drop out of care (Edlund et al. 2002). National survey statistics point to the importance of psychosocial barriers. A study using data from the 2002 National Survey on Drug Use and Health found that, among adults aged 18 or older with serious mental illness who received no mental health treatment in the previous year and perceived an unmet need for treatment, 28.2 percent identified stigma-related items among reasons for not receiving care (SAMHSA 2003a).
Stigma Avoidance
Link and Phelan (2001) provide a four-stage definition of stigma. First, categories are established to differentiate people, and these categories are assigned labels. Second, negative attributes are associated with particular labeled groups. Third, a dichotomy of “us” versus “them” is established. Finally, individuals associated with denigrated groups experience discrimination and status loss. In this conceptualization, it is implied that all four conditions must be met. Notably, opportunity and social status are jeopardized when groups become stigmatized.
Stigma has been studied as a major psychosocial impediment to care (Cooper, Corrigan, and Watson 2003; Corrigan 2004) because of its influence on attitudes and actions. Besides labels from external sources, self-labeling is an important process in an individual’s self-conception and treatment-seeking behavior (Thoits 1985; Thoits 2005). Seeking or accepting mental health treatment from a medical establishment could represent unwanted confirmation of a stigmatic status, thus stigma avoidance represents a barrier to the seeking of mental health care.
Perceived stigma can further impede care (Link 1987; Perlick et al. 2001; Sirey et al. 2001a; Sirey et al. 2001b; Struening et al. 2001) by impacting self-perception and shaping an individual’s help-seeking behaviors based on the anticipated reactions of others. Through stigma and perceived stigma, social expectations become internalized and manifest in social relations, behaviors, and the workings of institutions. Discrimination against the mentally ill is a possible outcome (Corrigan 2004; Link 1982). Individuals may fear stigma and act to prevent it by avoiding connections to mental health labels or services. Research has also examined the relationship between labeling theory and social status (Thoits 2005), showing that the connections among stigma, status, and treatment are further complicated by other factors including severity of symptoms, structural constraints, and willingness to seek assistance.
Negative Attitudes toward Treatment
Sociodemographic characteristics may significantly shape help-seeking attitudes and behaviors. Kessler et al. (2001) found that 55 percent of National Comorbidity Study respondents classified as having “severe mental illness” reported that they “did not believe they had a problem requiring treatment;” those with severe mental illness who perceived a need but did not receive treatment reported wanting to solve the problem themselves or thinking that the problem would improve on its own. Thoits (2005) found that persons of lower education or income were less likely to pursue professional assistance voluntarily. One interpretation of these findings is that, in addition to greater structural and financial barriers, individuals of lower socioeconomic status are also more likely to have negative attitudes toward professional treatment. Thoits (2005) also found that severity of disorder was related to voluntary care-seeking, where people with the most severe disorders were most likely to seek care voluntarily.
Mistrust or Fear of System
Trust is increasingly identified as an essential component of utilization of medical care. Specifically, mistrust can deter care-seeking, willingness to return for follow-up care, and compliance with treatment recommendations (Thom, Hall, and Pawlson 2004; Thom et al. 2002). Researchers have investigated mistrust towards medicine in minority communities, particularly within the African American community (Doescher et al. 2000; Takeuchi and Williams 2003). The aftermath of the Tuskegee study is often identified as a source of mistrust toward research, medicine, and health professionals generally (Freimuth et al. 2001; Thomas and Quinn 1991). Racial and ethnic discrimination within health care systems has also been identified as a source of mistrust or fear (Brown 2003). Finally, concerns about efficacy and side effects of psychiatric medications have been reported as an important source of African Americans’ skepticism toward voluntary psychiatric care (Schnittker 2003).
Research Questions
This article addresses two questions. First, what are the characteristics of adults reporting unmet need for mental health care? Second, among adults reporting unmet need for mental health care, are race-ethnicity and gender differentially associated with report of stigma avoidance, mistrust/fear of the health care system, or negative attitudes toward treatment?
METHODS
Based on theoretical premises and empirical findings from prior research, our analytic plan focused explicitly on the relationships between gender and race-ethnicity and report of psychosocial barriers to mental health services. We operationalized psychosocial barriers to care as three outcomes representing various dimensions of individual-level responses to in-ternally- and externally-situated stimuli. This study examines gender and race-ethnicity as correlates of (1) stigma avoidance, (2) mistrust or fear of the system, and (3) negative attitudes toward treatment.
Statistical Analyses
Analyses relied on nationally representative data and proceeded as follows. We first examined relationships between independent and dependent variables using simple descriptive methods. Applying multivariate logistic regression techniques, we next tested associations between the key independent variables and the three dependent variables, controlling also for other sociodemographic factors that may affect individuals’ relationships with the mental health system. A fourth logistic regression model examined the association between the independent variables and reports of any psychosocial barrier (i.e., stigma avoidance, mistrust or fear of the system, or negative attitudes toward treatment). Analyses were conducted with SUDAAN version 9 (Research Triangle Institute 2005), a software package that accounts for the survey’s complex sample design.
Data Source and Sample
Data were from the 2002 National Survey on Drug Use and Health (NSDUH), a survey sponsored by the Office of Applied Statistics in the U.S. Department of Health and Human Services Substance Abuse and Mental Health Services Administration. We used the National Survey on Drug Use and Health to calculate national and state-level estimates of the prevalence and correlates of drug, alcohol, and tobacco use by non-institutionalized persons ages 12 and older residing in the United States (including civilians residing on military bases). The survey is ideally suited for this study because it also collects sociodemographic, mental health status, self-reported information about respondents’ contact with mental health and substance use services, and other data. The sample is comprised of a 50-state design that includes an independent multistage probability sample for each of the 50 states and the District of Columbia. State identifiers are unavailable in public use files. Respondent’s privacy and confidentiality of sensitive behaviors are maximized through the use of computer-assisted interviewing methods. The Survey was administered in English, and for Spanish-speaking respondents, a Spanish language version of the questionnaire was available. A screening response rate of 91 percent was obtained, and the overall response rate for the survey is 79 percent. In multivariate analyses, two cases were lost due to missing data. The public use file we analyzed includes a total of 54,079 records (U.S. Department of Health and Human Services 2004).
This study focuses on non-institutionalized adults ages 18 and older who answered “yes” to the question, “During the past 12 months, was there any time when you needed mental health treatment or counseling for yourself but didn’t get it?” (unweighted n = 2,680). This item represents perceived unmet need for mental health care. We restricted this investigation to this population because questions addressing perceived barriers to care were asked only of individuals who reported unmet need. Analyses are based on an unweighted sample of 2,680 persons reporting unmet need for mental health care; when weighted, they represent 11,488,567 U.S. adults. All reported estimates are weighted to be nationally representative.
Variables
Independent variables.
The key independent variables are gender and self-reported race-ethnicity. We also created a variable for all combinations of gender and race-ethnicity. The unweighted distribution of study participants according to their race-ethnicity and gender is as follows: 2,033 non-Latino whites (1,449 females and 584 males, hereafter described as white males or white females), 231 African Americans (169 females and 62 males), 248 Latinos (176 females and 72 males), and 168 others (i.e., American Indians or Alaska Natives, Asian Americans or Pacific Islanders, or persons of multiple races-ethnicities; 115 females and 53 males). We also included age, marital status, and nativity as additional demographic controls.
A dichotomous indicator for mental health status is available in the National Survey on Drug Use and Health, and we included it as a covariate in multivariate logistic regression models. This indicator reflects the absence or presence, in the previous year, of a diagnosable mental, behavioral, or emotional disorder that met DSM-IV criteria. Adults meeting criteria for mental illness scored at least 13 points (out of 24 possible points) on the K6, a six-item scale included in the NSDUH (refer to Kessler et al. 2002 for validation data).1
Socioeconomic position was measured by current-year family income, educational attainment, and employment status; responses were provided as categorical variables and aggregated. The region’s metropolitan statistical area (MSA) status reflects the number of persons residing in a geographic area and was a control variable. Logistic regression models also included control measures of financial and access barriers to mental health care.
Dependent variables.
To operationalize the concept of psychosocial barriers to mental health care we first examined responses to eight survey statements describing various conditions responsible for adults’ unmet need for mental health care. Persons selecting a ninth “other reason” were asked: “You have indicated that during the past 12 months you did not get the mental health treatment or counseling you needed for some reason other than those listed. Please use the keyboard to type in the most important other reason you did not get the treatment you needed” (U.S. Department of Health and Human Services 2004). Respondents’ written responses were recoded by the NSDUH investigators into 34 separate barriers to care. We aggregated the eight standardized responses and written self-reported barriers into five indicators of barriers to mental health care. Three of these barriers are classified as psychosocial barriers (e.g., stigma avoidance, negative attitudes toward treatment, and mistrust of the system). The remaining items were grouped into two other barriers: (1) access barriers or (2) financial barriers to mental health care.2
We operationalized the three psychosocial barriers to mental health care as follows: Stigma avoidance summarizes responses to survey items including: “no mental health treatment in past year because of fear of neighbors’ negative opinion” [includes “Didn’t want others to find out you needed treatment”] and “because of fear of negative effect on job, confidentiality concerns, ashamed, embarrassed, or afraid—reason unspecified.” Negative attitudes toward treatment summarizes the following survey items: “didn’t think you needed treatment at the time”; “thought you could handle problem without treatment”; “didn’t think treatment would help”; “too stubborn/prideful to go”; and “unmotivated/depressed/confused/angry/unworthy.” Mistrust or fear of the system summarizes responses to survey items including: “afraid or concerned of being committed/having to take medications;” “didn’t like how treated/don’t like doctors, hospitals”; and “concerned about how court system would treat you.”
For each psychosocial barrier, respondents were coded 1 as having that barrier if they identified at least one item in that barrier category. Respondents who did not identify any item within each of the barrier categories were coded 0 since they did not report that particular barrier to mental health care. The NSDUH allowed for respondents to identify multiple barriers to mental health care. Thus, in this analysis, identification of one barrier did not preclude identification of other barriers and altogether they sum to greater than 100 percent. We hypothesized that racial-ethnic minority status and male status would individually be positively and significantly associated with report of each psychosocial barrier to mental health care (i.e., avoidance of social stigma, negative attitudes towards care, and mistrust of the system) when compared to non-minority whites and females. We also expected that minority males and minority females would be significantly more likely than white females to recognize psychosocial conditions as obstacles to receipt of mental health services.
RESULTS
Descriptive Analyses
Overall, 5.47 percent of adults reported unmet need for mental health care (Table 1). We present selected sociodemographic characteristics of adults by reported unmet need for mental health care status in Table 1. By gender, 7.12 percent of women and 3.69 percent of men reported unmet need for mental health care. Among whites and African Americans, about 5.7 percent reported unmet need for mental health care, as did 4.25 percent of Latinos. When data were examined by gender and race-ethnicity combined, 7.5 percent of white women reported unmet need, followed by African American women (6.48%), Latinas (5.93%), all others (4.70%), African American males (4.61%), white males (3.78%), and Latino males (2.64%). Adults born in the United States were nearly twice as likely as immigrants to report unmet need (5.87% vs. 3.07% of immigrants). Adults meeting criteria for mental illness were nearly ten times as likely to report unmet need for mental health care as adults not meeting mental illness criteria (31.06% vs. 3.14%, respectively). Reportage of unmet need was inversely related with age, declining from 8.39 percent among adults ages 18 to 25 to 2.40 percent among those 50 years and older. Adults with a college education were more likely to report unmet need than adults completing a high school education (6.12% vs. 4.83%). Unmet need for mental health care appeared to be inversely related to family income, declining from 7.42 percent among adults with family incomes below $20,000 per year to 4.41 percent among adults with incomes exceeding $50,000.
TABLE 1.
Selected Sociodemographic Characteristics of Adults Ages 18 and Older by Self-Reported Unmet Need for Mental Health Treatment in Past Year
| Reported Unmet Mental Health Care Need | No Reported Unmet Need for Mental Health Care | |
|---|---|---|
| % | % | |
| Characteristic | (95% CI) | (95% CI) |
|
| ||
| Overall Distribution | 5.47% | 94.53% |
| (5.15,5.87) | (94.19,94.85) | |
| Gender | ||
| Male | 3.69* | 96.31* |
| (3.27,4.16) | (95.84,96.73) | |
| Female | 7.12 | 92.88 |
| (6.64,7.63) | (92.37,93.36) | |
| Ethnicity | ||
| Non-Latino White | 5.71 | 94.29 |
| (5.35,6.10) | (93.90,94.65) | |
| Non-Latino African American | 5.66 | 94.34 |
| (4.35,7.33) | (92.67,95.65) | |
| Latino | 4.25* | 95.75* |
| (3.45,5.23) | (94.77,96.55) | |
| All Other/Multiple Ethnicities | 4.70 | 95.30 |
| (3.65,6.04) | (93.96,96.35) | |
| Gender by Ethnicity | ||
| Male, White | 3.78* | 96.22* |
| (3.33,4.29) | (95.71,96.67) | |
| Female, White | 7.50 | 92.50 |
| (6.96,8.08) | (91.92,93.04) | |
| Male, African American | 4.61* | 95.39* |
| (2.68,7.85) | (92.15,97.32) | |
| Female, African American | 6.48 | 93.52 |
| (4.95,8.45) | (91.55,95.05) | |
| Male, Latino | 2.64* | 97.36* |
| (1.81,3.85) | (96.15,98.19) | |
| Female, Latina | 5.93 | 94.07 |
| (4.56,7.68) | (92.32,95.44) | |
| Male/Female All Other/ | 4.70* | 95.30* |
| Multiple Ethnicities | (3.65,6.04) | (93.96,96.35) |
| Nativity | ||
| U.S.-Born | 5.87 | 94.13 |
| (5.51,6.25) | (93.75,94.49) | |
| Foreign-Born | 3.07* | 96.93* |
| (2.40,3.92) | (96.08,97.60) | |
| Mental Illness | ||
| Yes | 31.06* | 68.94* |
| (28.70,33.52) | (66.48,71.30) | |
| No | 3.14 | 96.86 |
| (2.87,3.44) | (96.56,97.13) | |
| Age | ||
| 18–25 | 8.39* | 91.61* |
| (7.90,8.90) | (91.10,92.10) | |
| 26–34 | 7.51 | 92.49 |
| (6.65,8.47) | (91.53,93.35) | |
| 35–49 | 6.70 | 93.30 |
| (6.05,7.41) | (92.59,93.95) | |
| 50+ | 2.40* | 97.60* |
| (1.91,3.02) | (96.98,98.09) | |
| Education | ||
| ≤ 12th Grade | 4.83* | 95.17* |
| (4.43,5.26) | (94.74,95.57) | |
| College or Higher | 6.12 | 93.88 |
| (5.62,6.66) | (93.34,94.38) | |
| Family Income, Current Year | ||
| < $20,000 | 7.42* | 92.58* |
| (6.62,8.30) | (91.70,93.38) | |
| $20,000–$29,999 | 5.87 | 94.13 |
| (5.08,6.77) | (93.23,94.92) | |
| $30,000–$49,999 | 5.52 | 94.48 |
| $50,000 or more | (4.80,6.35) 4.41* (3.98,4.89) |
(93.65,95.20) 95.59* (95.11,96.02) |
| Employment Status | ||
| Full-Time, Part-time, | 5.38 | 94.62 |
| or Absent Worker | (5.02,5.76) | (94.24,94.98) |
| Unemployed, Keeping House, Student, | 5.69 | 94.31 |
| Retired, Disabled, or Jobless | (5.02,6.44) | (93.56,94.98) |
| Identified Access Barriers | 29.54 | NA |
| (26.74,32.50) | ||
| Identified Financial Barriers | 44.53 | NA |
| (41.27,47.84) | ||
Notes: Original tabulations of the 2002 National Survey on Drug Use and Health public use files. Within each group by unmet need status, reference groups were: women, whites, white women, adults ages 35 to 49, persons with some college education or higher, persons with family incomes of $30,000–$49,9999, working adults, U.S.-born adults, and persons without mental illness. NA refers to “data not available” since questions on access and financial barriers were asked only of adults who reported an unmet need for mental health care. The term “CI” refers to the confidence interval.
Reported Unmet Need: Unweighted N = 2,680; Weighted N: 11,488,567
No Reported Unmet Need: N = 33,603; Weighted N: 198,411,961
p ≤ .05
One significant difference between adults with and without unmet need for mental health care merits discussion: 47.4 percent of adults reporting unmet need also met criteria for mental illness (vs. 6.1% of adults not reporting unmet need). Overall, 8.4 percent of adults met mental illness criteria in the entire sample, indicating that the indicator for self-reported unmet need successfully identified 34.4 percent of all individuals who met criteria for severe mental illness within the larger sample. Persons reporting unmet need may have had contact with the mental health system in the year prior to participating in the survey. Of persons reporting unmet need, 19 percent met criteria for mental illness in the previous year but did not obtain any mental health care; an additional 29 percent met mental illness criteria and obtained some mental health treatment (data not shown).
We examined characteristics of adults reporting unmet need for mental health care across the three psychosocial barriers, and we present these results in Table 2. Of adults indicating an unmet need for mental health care, 24.14 percent of adults identified stigma avoidance as a barrier to care, 19.45 percent identified negative attitudes toward treatment, and 9.14 percent reported mistrust/fear of the system as obstacles to care. Overall, nearly half of adults (47.08%, Table 2) reporting unmet need for mental health care identified any of three psychosocial barriers as obstacles to mental health care. Persons reporting unmet need also identified mental health services access barriers (29.5%) as well as financial barriers (44.5%) (Table 1).
TABLE 2.
Perceived Barriers to Care by Sociodemographic Characteristics, Adults Ages 18 and Older Reporting Unmet Need for Mental Health Treatment in Past Year
| Stigma Avoidance % |
Negative Attitudes Toward Treatment % |
Mistrust or Fear of System % |
Any Psychosocial Barriera | |
|---|---|---|---|---|
| Characteristic | (95% CI) | (95% CI) | (95% CI) | (95% CI) |
|
| ||||
| Total | 24.14 | 19.45 | 9.14 | 47.08 |
| (21.74, 26.73) | (16.78, 22.42) | (7.72, 10.78) | (43.83, 50.35) | |
| Gender | ||||
| Female | 21.60 | 18.55 | 8.33 | 43.93 |
| (19.04, 24.39) | (15.89, 21.55) | (6.80, 10.17) | (40.42, 47.51) | |
| Male | 29.48* | 21.31 | 10.82 | 53.66* |
| (24.39, 35.13) | (15.53, 28.51) | (7.86, 14.71) | (47.00, 60.20) | |
| Ethnicity | ||||
| Non-Latino White | 25.13 | 19.16 | 9.72 | 47.72 |
| (22.40, 28.06) | (16.65, 21.96) | (8.09, 11.64) | (44.21, 51.26) | |
| Non-Latino African American | 20.58 | 23.97 | 6.75 | 48.74 |
| (13.66, 29.78) | (12.01, 42.14) | (3.72, 11.91) | (36.04, 61.61) | |
| Latino | 21.83 | 19.70 | 8.49 | 37.99 |
| (14.82, 30.96) | (12.37, 29.89) | (4.16, 16.57) | (26.34, 51.21) | |
| Other/Multiple Ethnicities | 21.90 | 12.83 | 7.06 | 44.76 |
| (13.10, 34.29) | (5.80, 26.03) | (3.23, 14.74) | (34.88, 55.06) | |
| Gender by Ethnicity | ||||
| White Female | 21.44 | 19.29 | 8.14 | 44.58 |
| (18.63, 24.54) | (16.14, 22.89) | (6.49, 10.17) | (40.48, 48.75) | |
| White Male | 33.02* | 18.89 | 13.10* | 54.45* |
| (27.41, 39.15) | (14.59, 24.11) | (9.38, 17.99) | (47.88, 60.87) | |
| Male, African American | 17.25 | 35.28 | 5.54 | 57.43 |
| (7.34, 35.43) | (11.03, 70.56) | (1.55, 17.94) | (32.65, 78.96) | |
| Female, African American | 22.46 | 17.58 | 7.42 | 43.83 |
| (14.20, 33.63) | (10.45, 28.06) | (3.91, 13.65) | (32.24, 56.13) | |
| Male, Latino | 16.79 | 31.00 | 2.33* | 50.11 |
| (8.32, 30.99) | (14.97, 53.41) | (.32, 15.13) | (31.85, 68.34) | |
| Female, Latina | 24.17 | 14.47 | 11.35 | 42.28 |
| (15.26, 36.07) | (8.36, 23.87) | (5.32, 22.56) | (31.21, 54.18) | |
| Male/Female All Other/ | 21.90 | 12.83 | 7.06 | 37.99 |
| Multiple Ethnicities | (13.10, 34.29) | (5.80, 26.03) | (3.23, 14.74) | (26.34, 51.21) |
| Mental Illness | ||||
| Yes | 26.77* | 14.04* | 12.49* | 44.82 |
| (23.39, 30.45) | (11.22, 17.42) | (10.08, 15.38) | (40.41, 49.32) | |
| No | 21.78 | 24.31 | 6.12 | 49.11 |
| (18.45, 25.51) | (20.20, 28.96) | (4.58, 8.13) | (44.56, 53.67) | |
| Age | ||||
| 18–25 | 28.94* | 18.99 | 13.38* | 52.11* |
| (26.31, 31.73) | (16.68, 21.53) | (11.44, 15.59) | (49.08, 55.13) | |
| 26–34 | 26.94 | 22.42 | 11.16 | 54.15* |
| (22.01, 32.52) | (17.47, 28.30) | (7.77, 15.77) | (48.16, 60.02) | |
| 35–49 | 23.25 | 16.56 | 6.75 | 42.77 |
| (19.24, 27.81) | (13.01, 20.84) | (4.88, 9.26) | (37.50, 48.22) | |
| 50+ | 15.72 | 22.61 | 6.00 | 40.27 |
| (9.24, 25.46) | (12.71, 36.96) | (2.89, 12.06) | (28.79, 52.92) | |
| Education | ||||
| ≤ 12th Grade | 22.29 | 12.70* | 10.22 | 39.65* |
| (19.08, 25.88) | (10.09, 15.86) | (7.93, 13.08) | (35.32, 44.15) | |
| Some College and Above | 25.60 | 24.77 | 8.28 | 52.93 |
| (22.31, 29.20) | (20.74, 29.29) | (6.54, 10.43) | (48.72, 57.10) | |
| Family Income, Current Year | ||||
| < $20,000 | 20.92 | 10.91* | 11.25 | 36.08* |
| (16.87, 25.64) | (8.36, 14.13) | (8.26, 15.13) | (30.99, 41.50) | |
| $20,000–$29,999 | 20.64 | 16.27 | 9.89 | 41.93* |
| (15.85, 26.42) | (11.51, 22.49) | (6.48, 14.83) | (35.21, 48.96) | |
| $30,000–$49,999 | 24.15 | 23.75 | 9.89 | 51.66 |
| (19.28, 29.79) | (16.79, 32.47) | (6.95, 13.87) | (44.56, 58.69) | |
| $50,000 or more | 28.07 | 24.07 | 6.59 | 54.22 |
| (23.69, 32.91) | (19.73, 29.03) | (4.67, 9.23) | (48.74, 59.59) | |
| Nativity | ||||
| U.S.-Born | 24.33 | 19.01 | 9.27 | 46.79 |
| (21.86, 26.97) | (16.30, 22.04) | (7.78, 11.00) | (43.43, 50.18) | |
| Foreign-Born | 22.04 | 24.67 | 7.60 | 50.55 |
| (14.31, 32.36) | (14.46, 38.82) | (3.63, 15.25) | (39.16, 61.88) | |
| Employment Status | ||||
| Full-Time, Part-Time, | 26.18 | 19.53 | 7.66 | 48.34 |
| or Absent Worker | (23.27, 29.32) | (16.88, 22.48) | (6.23, 9.39) | (44.62, 52.07) |
| Unemployed, Keeping House, Student, | 19.97* | 19.31 | 12.19* | 44.55 |
| Retired, Disabled, or Jobless | (16.03, 24.59) | (13.51, 26.83) | (9.05, 16.23) | (38.07, 51.22) |
Notes: Original tabulations from public use files of the 2002 National Survey on Drug Use and Health. Questions on barriers to mental health care were asked only of persons reporting unmet need for mental health services. This table reports the percentage of respondents in each status group who endorsed each psychosocial barrier, defined as adults who identified at least one item within each barrier. Respondents could identify more than one barrier and therefore, totals do not add to 100 percent. Within each barrier, reference groups were: women, whites, white women, adults ages 35 to 49, persons with some college education or higher, persons with family incomes of $30,000–$49,999, working adults, U.S.-born adults, and persons without mental illness. The term “CI” refers to the confidence interval.
p ≤ .05
“Any Psychosocial Barrier” represents positive identification of “stigma avoidance,” or “negative attitudes toward treatment,” or “mistrust or fear of system.”
Across the three psychosocial barriers, a greater percentage of men than women reported each barrier, though gender differences were statistically significant for stigma avoidance only. Report of each barrier varied by race-ethnicity, but differences were statistically nonsignificant. One-quarter of whites reported stigma avoidance, and 23.97 percent of African Americans reported negative attitudes toward treatment, while 9.72 percent of whites reported mistrust or fear of the system, followed by Latinos (8.49%). A complex picture emerged when barriers were analyzed by gender and race-ethnicity combined. For example, 33.02 percent of white males reported stigma avoidance as an obstacle to mental health care, followed next by Latinas (24.17%). More than one-third (35.28%) of African American males indicated that negative attitudes resulted in unmet need. However, this estimate may not be reliable, as evidenced by the wide confidence interval. Mistrust/fear of the system was less commonly reported, though white males were significantly more likely and Latino men were significantly less likely to report mistrust as compared to white women. Other factors were also associated with report of the three barriers. For example,adults meeting criteria for mental illness were significantly more likely to report each barrier. Age, education, and income were inconsistent correlates of psychosocial barriers, though adults ages 18 to 25 and non-working adults were more likely to identify stigma avoidance and mistrust of the system than were adults ages 35 to 49 and working adults, respectively.
Multivariate Analyses
Table 3 presents the results of multivariate logistic regression models examining the independent relationships between race-ethnicity and gender and the three psychosocial barriers to mental health care and any psychosocial barrier. Models also controlled for other sociode-mographic characteristics, mental health status, financial and access barriers, and size of metropolitan area. Results indicate that male gender was positively associated with stigma avoidance (OR = 1.66, p < .05), mistrust/fear of the system (OR = 1.51, p < .05), and any psychosocial barrier (OR = 1.52, p < .05). Surprisingly, African American or Latino ethnicities were not significantly associated with any psychosocial barrier. Mental illness was associated with stigma avoidance (OR = 1.51, p < .05) and mistrust/fear of the system (OR = 2.13, p < .05) and was protective against negative attitudes toward treatment (OR = .68, p < .05). The relationship between age and psychosocial barriers was inconsistent. Adults ages 18 to 25 were significantly more likely to report stigma avoidance (OR = 1.44, p < .05), mistrust of the system (OR = 1.91, p < .05), and any psychosocial barrier (OR = 1.70, p < .05); adults ages 26 to 34 reported mistrust (OR = 1.78, p < .05) and any psychosocial barrier (OR = 1.69, p < .05). Education, income, and work status were also inconsistent correlates of psychosocial barriers. For example, adults with a high school education or less were about half as likely to report negative attitudes toward treatment and any psychosocial barrier as compared to adults with any college education. Adults with incomes below $29,999 were about half as likely as their peers in the $30,000 to $49,999 income group to report any psychosocial barrier. Access and financial barriers were less likely to be associated with stigma avoidance (OR ≤ .80, only access barriers were significant at p < .05), negative attitudes toward mental health care (OR ≤ .13, p < .05), and any psychosocial barrier (OR = .22, p <.05).
TABLE 3.
Logistic Regressions Predicting Perceived Barriers to Mental Health Care, Ages 18 and Older Reporting Unmet Need for Mental Health Care in Past Year
| Stigma Avoidance | Negative Attitudes Toward Treatment | Mistrust or Fear of System | Any Psychosocial Barriera | |
|---|---|---|---|---|
| Characteristic | OR (95% CI) | OR (95% CI) | OR (95% CI) | OR (95% CI) |
|
| ||||
| Gender | ||||
| Female | REF. | REF. | REF. | REF. |
| Male | 1.66* | 1.01 | 1.51* | 1.52* |
| Ethnicity | (1.22, 2.24) | (.67, 1.51) | (1.01, 2.27) | (1.11, 2.09) |
| Non-Latino White | REF. | REF. | REF. | REF. |
| Non-Latino African American | .94 | 1.07 | .66 | 1.08 |
| (.55, 1.61) | (.55, 2.07) | (.33, 1.33) | (.70, 1.67) | |
| Latino | .80 | 1.08 | .77 | .75 |
| (.47, 1.37) | (.59, 1.98) | (.32, 1.86) | (.46, 1.20) | |
| Other/Multiple Ethnicities | .91 | .79 | .75 | .73 |
| Mental Illness | (.49, 1.71) | (.27, 2.33) | (.32, 1.77) | (.41, 1.31) |
| Yes | 1.51* | .68* | 2.13* | 1.19 |
| No | (1.15, 1.98) REF. |
(.47, .97) REF. |
(1.43, 3.15) REF. |
(.91, 1.55) REF. |
| Age | ||||
| 18–25 | 1.44* | 1.25 | 1.91* | 1.70* |
| (1.04, 2.00) | (.84, 1.87) | (1.26, 2.90) | (1.23, 2.34) | |
| 26–34 | 1.24 | 1.42 | 1.78* | 1.69* |
| (.86, 1.78) | (.84, 2.40) | (1.04, 3.06) | (1.18, 2.43) | |
| 35–49 | REF. | REF. | REF. | REF. |
| 50+ | .60 | 1.07 | .66 | .64 |
| Education | (.32, 1.14) | (.55, 2.07) | (.29, 1.50) | (.39, 1.05) |
| ≤ 12th Grade | .82 | .43* | 1.00 | .54* |
| Some College and Above | (.62, 1.09) REF. |
(.29, .62) REF. |
(.69, 1.44) REF. |
(.41,.71) REF. |
| Family Income, Current Year | ||||
| < $20,000 | .80 | .40* | .89 | .47* |
| (.53, 1.21) | (.23, .68) | (.53, 1.50) | (.32, .68) | |
| $20,000–$29,999 | .76 | .61 | .93 | .60* |
| (.50, 1.16) | (.34, 1.10) | (.52, 1.67) | (.41, .86) | |
| $30,000–$49,999 | REF. | REF. | REF. | REF. |
| $50,000 or more | 1.14 | .76 | .67 | .89 |
| Nativity | (.78, 1.66) | (.46, 1.26) | (.40, 1.15) | (.61, 1.31) |
| U.S.-Born | REF. | REF. | REF. | REF. |
| Foreign-Born | 1.06 | 1.42 | .97 | 1.47 |
| Employment Status | (.61, 1.86) | (.67, 3.00) | (.39, 2.44) | (.88, 2.46) |
| Full-Time, Part-Time, or Absent Worker |
REF. | REF. | REF. | REF. |
| Unemployed, Keeping House, Student, | .83 | 1.22 | 1.79* | 1.10 |
| Retired, Disabled, or Jobless | (.60, 1.14) | (.79, 1.91) | (1.23, 2.59) | (.80, 1.51) |
| Access Barriers | ||||
| Yes | .80 | .08* | 1.04 | .22* |
| No | (.59, 1.07) REF. |
(.05, .15) REF. |
(.66, 1.66) REF. |
(.16, .30) REF. |
| Financial Barriers | ||||
| Yes | .74* | .13* | .74 | .22* |
| No | (.56, .99) REF. |
(.08, .21) REF. |
(.49, 1.10) REF. |
(.17, .30) REF. |
Notes: Original tabulations from public use files of the 2002 National Survey on Drug Use and Health. Questions on barriers to mental health care were asked only of persons reporting unmet need for mental health services. Models also controlled for marital status and metropolitan statistical area size. The term OR refers to “odds ratio” and the term “CI” refers to the confidence interval.
p ≤ .05
“Any Psychosocial Barrier” represents positive identification of “stigma avoidance,” or “negative attitudes toward treatment,” or mistrust or fear of system.”
Table 4 presents results of multivariate logistic regression analyses, which include a variable for race-ethnicity and gender combined. Findings indicate that only white male status was positively associated with stigma avoidance (OR = 1.94, p < .05), mistrust/fear of the system (OR = 1.89, p < .05), and any psychosocial barrier (OR = 1.59, p < .05). Notably, neither African American nor Latino male or female statuses were associated with report of the psychosocial barriers examined. Mental illness was positively associated with stigma avoidance (OR = 1.50, p < .05) and mistrust/fear of the system (OR = 2.15, p < .05); in contrast, adults meeting criteria for mental illness were less likely to report negative attitudes toward treatment (OR = .68, p < .05) than healthy adults. Being ages 18 to 25 years was positively associated with stigma avoidance (OR = 1.43, p < .05), mistrust/fear of the system (OR = 1.92, p < .05), and any psychosocial barrier (OR = 1.69, p < .05); adults ages 26 to 34 were significantly more likely to report mistrust/fear of the system (OR = 1.77, p < .05) and any psychosocial barrier (OR = 1.68, p < .05). The relationship between educational attainment and psychosocial barriers varied: adults completing up to a high school education were about half as likely to report negative attitudes to mental health care and any psychosocial barrier; the relationships between education and stigma avoidance and mistrust or fear of the system were nonsignificant. The relationship between income and psychosocial barriers was weak; adults with family incomes below $20,000 were less than half as likely as the reference group, adults with incomes in the $30,000–$49,999 range, to report negative attitudes toward mental health care or any psychosocial barrier, while adults in the $20,000 to $29,999 range were about two-thirds as likely as the reference group to report any psychosocial barrier. Non-working adults were significantly more likely to report mistrust of the system. Adults identifying financial and access barriers to mental health care were significantly less likely to report negative attitudes to care and any psychosocial barrier (OR ≤ .22); adults reporting financial barriers were less likely to identify stigma avoidance (OR = .74, p < .05).
TABLE 4.
Logistic Regressions Predicting Perceived Barriers to Mental Health Care, Ages 18 and OlderReporting Unmet Need for Mental Health Care in Past Year
| Stigma Avoidance | Negative Attitudes Toward Treatment | Mistrust or Fear of System | Any Psychosocial Barriera | |
|---|---|---|---|---|
| OR | OR | OR | OR | |
| Characteristic | (95% CI) | (95% CI) | (95% CI) | (95% CI) |
|
| ||||
| Gender by Ethnicity | ||||
| White Female | REF. | REF. | REF. | REF. |
| White Male | 1.94* | .88 | 1.89* | 1.59* |
| (1.38, 2.71) | (.57, 1.38) | (1.19, 2.99) | (1.11, 2.28) | |
| Male, African American | 1.12 | 1.28 | .86 | 1.67 |
| (.43, 2.94) | (.38, 4.27) | (.23, 3.20) | (.76, 3.67) | |
| Female, African American | 1.22 | .88 | .82 | 1.09 |
| (.69, 2.16) | (.45, 1.73) | (.39, 1.73) | (.67, 1.78) | |
| Male, Latino | .68 | 2.31 | .25 | 1.03 |
| (.29, 1.61) | (.75, 7.12) | (.03, 2.25) | (.44, 2.44) | |
| Female, Latina | 1.16 | .68 | 1.30 | .79 |
| (.64, 2.11) | (.35, 1.29) | (.53, 3.16) | (.46, 1.35) | |
| Male/Female All Other/ | 1.12 | .78 | .94 | .85 |
| Multiple Ethnicities | (.59, 2.15) | (.26, 2.31) | (.39, 2.22) | (.48, 1.51) |
| Mental Illness | ||||
| Yes | 1.50* | .68* | 2.15* | 1.19 |
| (1.15, 1.97) | (.47, .97) | (1.45, 3.18) | (.91, 1.55) | |
| No | REF. | REF. | REF. | REF. |
| Age | ||||
| 18–25 | 1.43* | 1.24 | 1.92* | 1.69* |
| (1.03, 1.99) | (.83, 1.84) | (1.27, 2.92) | (1.22, 2.33) | |
| 26–34 | 1.21 | 1.45 | 1.77* | 1.68* |
| (.84, 1.75) | (.87, 2.43) | (1.03, 3.03) | (1.17, 2.42) | |
| 35–49 | REF. | REF. | REF. | REF. |
| 50+ | .60 | 1.04 | .65 | .64 |
| (.32, 1.13) | (.54, 2.02) | (.29, 1.46) | (.39, 1.04) | |
| Education | ||||
| ≤ 12th Grade | .84 | .42* | 1.00 | .54* |
| (.63, 1.11) | (.29, .61) | (.70, 1.43) | (.41, .71) | |
| Some College and Above | REF. | REF. | REF. | REF. |
| Family Income, Current Year | ||||
| <$20,000 | .78 | .41* | .87 | .47* |
| (.52, 1.18) | (.24, .69) | (.51, 1.47) | (.32, .67) | |
| $20,000–$29,999 | .74 | .62 | .91 | .59* |
| – | (.48, 1.14) | (.35, 1.11) | (.50, 1.67) | (.41, .86) |
| $30,000–$49,999 | REF. | REF. | REF. | REF. |
| $50,000 or more | 1.13 | .77 | .67 | .89 |
| (.78, 1.63) | (.47, 1.27) | (.39, 1.14) | (.61, 1.30) | |
| Nativity | ||||
| U.S.-Born | REF. | REF. | REF. | REF. |
| Foreign-Born | 1.13 | 1.30 | 1.02 | 1.51 |
| (.65, 1.96) | (.65, 2.60) | (.42, 2.50) | (.90, 2.51) | |
| Employment Status | ||||
| Full-Time, Part-Time, | REF. | REF. | REF. | REF. |
| or Absent Worker | ||||
| Unemployed, Keeping House, Student, | .82 | 1.23 | 1.77* | 1.09 |
| Retired, Disabled, or Jobless | (.60, 1.13) | (.80, 1.91) | (1.23, 2.56) | (.80, 1.50) |
| Access Barriers | ||||
| Yes | .79 | .08* | 1.04 | .22* |
| (.59, 1.06) | (.05, .15) | (.65, 1.66) | (.16, .30) | |
| No | REF. | REF. | REF. | REF. |
| Financial Barriers | ||||
| Yes | .74* | .13* | .72 | .22* |
| (.55, .98) | (.08, .21) | (.48, 1.09) | (.16, .29) | |
| No | REF. | REF. | REF. | REF. |
Notes: Original tabulations from public use files of the 2002 National Survey on Drug Use and Health. Questions on barriers to mental health care were asked only of persons reporting unmet need for mental health services. Models also controlled for marital status and metropolitan statistical area size. The term OR refers to “odds ratio” and the term “CI” refers to the confidence interval.
p ≤ .05
“Any Psychosocial Barrier” represents positive identification of “stigma avoidance,” or “negative attitudes toward treatment,” or “mistrust or fear of system.”
Using models shown in Table 4, we included indicators for alcohol and drug use, mental health services use, and social status as measured by occupation type (i.e., blue collar, white collar professional, white collar support; data not shown). After controlling for occupational status, only white male status was positively associated with stigma avoidance (OR =1.55; 95% CI = 1.03, 2.33); occupational status was not significantly associated with any barrier examined. After controlling for drug and alcohol use, white male status was significantly associated with stigma avoidance (OR = 1.85; 95% CI = 1.32, 2.60), mistrust (OR = 1.80; 95% CI = 1.13, 2.87), and any psychosocial barrier (OR = 1.58; 95% CI = 1.10, 2.27). Drug use was independently, positively associated with stigma avoidance (OR = 1.63; 95% CI = 1.04, 2.56). In models controlling for mental health service use, white male status was positively associated with any psychosocial barrier (OR = 1.50; 95% CI = 1.04, 2.16), stigma avoidance (OR = 1.94; 95% CI = 1.38, 2.72), and mistrust of system (OR = 1.97; 95% CI = 1.24, 3.14), while consumer status was associated with roughly two-thirds odds of reporting any psychosocial barrier (OR = .68; 95% CI = 0.51, 0.89) and negative attitudes (OR = .60; 95% CI = .41, .87; data available upon request).
DISCUSSION
In addition to financial and access barriers, psychosocial barriers have been identified as possible factors underlying racial-ethnic and gender disparities in use of mental health services. Relying primarily on a large, nationally representative data base that included qualitative textual data provided by respondents, this study examined the relationships between measures of race-ethnicity and gender and outcome measures of stigma avoidance, mistrust, and negative attitudes toward treatment in a population of adults reporting unmet need for mental health care. Psychosocial barriers have been proposed as important factors affecting mental health service use by racial and ethnic minority or immigrant communities and males. For these reasons, significant positive associations between white male status and stigma avoidance and mistrust/fear of the system were contrary to our expectations.
Although disparities in mental health services use have been identified in the research literature (e.g., Alegria et al. 2002), detailed information on reasons for delays or lack of service use by different population subgroups from nationally representative samples is rarely available. One important feature of the National Survey on Drug Use and Health is that it provides respondents with the option of reporting on factors that they feel influence their level of contact with the mental health system. For this study, we drew on this qualitative information derived from a write-in response category for the item on unmet need for mental health care. An advantage of the write-in process is that it permits respondents to identify important factors that may not have been identified by the survey’s designers. These data permitted examination of diverse psychosocial barriers to mental health care from a client’s or potential client’s perspective. The admission of a stigmatized condition such as mental illness, or moving from “discreditable” to “discredited” status (Goffman 1963), implies a greater risk of social loss for individuals already in positions of higher perceived status (e.g., higher “illness danger”) or for individuals within certain social status groups. A potential explanation for the association between stigma avoidance and non-Latino white male status hinges upon theorized relationships between social stigma and the subsequent loss of social status (Link and Phelan 2001); non-Latino white males may feel as though they have more to lose. Possibly, white male identity itself confers social value above and beyond other socioeconomic variables typically available in population-based surveys. In logistic regression analyses, we operationalized the concept of social status using various measures, including education, income, labor market activity, and occupational status. Nevertheless, the independent “non-Latino white male” effect persisted for stigma avoidance only. Additionally, controlling for alcohol and drug dependence did not substantively change our results.
The ways in which psychosocial factors introduce variation into mental health services care-seeking and utilization behaviors are not yet fully understood. Specifically, Thoits’s (2005) findings introduce uncertainty into the assumption that members of disadvantaged social groups are systematically less willing to seek mental health care voluntarily. Additional research may clarify the complex relationships between social status and health-seeking behaviors. For example, non-Latino white males were most likely to report mistrusting the system. Thus, new research should specifically explore the connections between social status and trust of medicine and other larger social institutions across all racial-ethnic subgroups.
Our findings also have implications for the assumption that use of mental health services is a highly stigmatized behavior within minority communities. Our findings, based on a large, nationally representative survey data base, suggest that some unquestioned assumptions about stigma require additional empirical testing. Notably, these findings suggest that stigma avoidance appears largely to be a feature of the non-minority white population. We recognize, however, that a limitation of this study relates to the sample size: We may have lacked a sufficiently large sample of Latinos and African Americans to detect between-group differences, particularly when data were disaggregated by gender. Although we considered aggregating data for African Americans and Latinos, both populations vary significantly in their composition (i.e., nativity, citizenship status, geographic dispersion) making aggregation of data undesirable. For this reason, we suggest that future studies examining issues of stigma avoidance and other psychosocial barriers would benefit from larger samples of racial-ethnic subgroups, including Latinos, African Americans, and Asians. In this manner, the relevance of stigma avoidance, trust of the system, and attitudes toward treatment for racial-ethnic minority populations can be assessed more precisely and with greater confidence.
While the influence of psychosocial factors on mental health services use cannot and should not be overlooked, overemphasizing the role of psychosocial factors as a significant mechanism underlying disparities in care may also inadvertently mask the contributions of other factors that also result in minorities’ and men’s unequal contact with mental health care providers or substandard care. Our data demonstrate that financial and access barriers shape mental health services utilization, suggesting that policy interventions may be beneficial in improving economic access to mental health services. Alternatively, systematic differences by gender and race-ethnicity in perceived need for mental health care may play important roles in the outcomes observed here. Our data suggest that assuming that psychosocial barriers are more prevalent among minority populations could result in the exclusion of other strategies that may eliminate disparities in access to mental health services.
Solely identifying psychosocial factors as potential barriers to mental health care is no longer a sufficient research goal. Rather, the specific connections between attitudes, social status, and help-seeking behaviors must also be investigated within specific gender, social, and cultural contexts. Our results identified the combination of gender and race-ethnicity as an important predictor of attitudes towards mental health care-seeking, illustrating that differences can be masked within analyses focused solely on gender or race-ethnicity. The surprising finding that non-Latino white males are significantly more likely to report stigma than other groups suggests that the role of stigma avoidance, negative attitudes towards care, and mistrust or fear of the system have not yet been adequately understood in relation to real-world mental health care problems. Overall, these findings point to a greater need for additional consideration of both gender and race-ethnicity in future research on attitudes, perceived need, care-seeking, and utilization of mental health services. Furthermore, data indicating that young adults are likely to report stigma avoidance suggests that existing programs designed to destigmatize mental illness have not been diffused within the population or been as effective as we may have anticipated. Prevention and intervention programs may consider targeting youths more precisely to break down psychosocial barriers and improve mental health care-seeking among this sub-group.
This study is limited by features of the sub-sample and questionnaire. Questions about perceived barriers to care were asked only of individuals who answered “yes” to a prior question addressing self-report of unmet need, leading to potential sample selection bias based on differential perception of need. Findings may also be influenced by our ability to adequately measure the psychosocial under-pinnings of decisions to reject mental health care. Few psychosocial items were included in the survey, and we relied on respondents’ written responses to a residual “other barrier” category for additional data. We could not assess the availability of providers in the community. However, our analyses did control for prior mental health service utilization and results were substantively unchanged. The cross-sectional nature of the survey’s design limits us from examining changes in behaviors or attitudes over time. Additionally, results from this study are generalizable to noninstitutionalized populations and do not represent the experiences of active military persons or individuals residing in groups quarters (e.g., hospitals, prisons, nursing homes). Thus, our data may underestimate the prevalence of perceived unmet need in the community due to the exclusion of highly vulnerable populations (e.g., homeless, incarcerated, or those hospitalized for mental or physical illnesses) from the survey. Expression of physical health symptoms, such as in “nervios,” a culture-bound syndrome (Salgado de Snyder, Diaz-Perez, Ojeda 2000), may have contributed to underrecognition of mental illness or need for mental health care in ethnic minority or immigrant communities, a condition which may also have impacted our study sample. Finally, we lacked additional measures of mental illness. A potential limitation of the K6 is its emphasis on symptoms of depression and anxiety (Kessler et al. 2002); therefore, we may have underestimated the prevalence of mental illness in the population. However, it is worth mentioning that persons were not selected into the study sample based on their mental health status. Although respondents were not required to overtly rate their mental health status and service needs, this was implicit within the screening question that defined the study sample.
A survey written within the context of the general population of the United States may itself be culture-bound; therefore, survey items addressing social stigma, mistrust, or attitudes towards care may be phrased in an idiom that speaks directly to the cultural experience of certain segments of society. Population sub-groups, particularly racial-ethnic minorities, immigrants, or men, may experience or report a different manifestation of stigma or other psychosocial barriers to mental health care. We also examined other psychosocial barriers to care, including negative attitudes towards treatment and fear or mistrust of the system. Little is known about the specific mechanisms and manifestations of mistrust/fear of system as a barrier to mental health care. Additional research is needed to elucidate this topic, particularly in relation to the African American community and other groups that have historically had negative interactions with medical institutions. We agree with Thom et al.’s (2004) suggestion that measures of trust should be included in more surveys, especially population-based studies. Particular attention may be paid to studying the reasons that people cite for mistrusting social and health care systems, whether mistrust is connected to personal past experiences or larger social or historical attitudes, and whether general mistrust actually translates into avoidance of mental health care. Our assessment of mistrust and rejection of formal treatment depended heavily on respondents’ own write-in answers, since the survey did not explicitly ask about these items. We expect that the results reported here likely underestimate the percentage of people who would have agreed in response to a direct question on this issue.
This study provides an initial assessment of gender and race-ethnicity in relation to attitudes, perceptions, and the psychosocial barriers of stigma avoidance, negative attitudes towards treatment, and mistrust or fear of the system. Additional research is needed to test the robustness of results reported here. With a greater understanding of psychosocial barriers to care, targeted interventions could be designed to promote care-seeking for both genders and among ethnically diverse populations. In cases where psychosocial factors have been identified as a major barrier to mental health service use, more focus could be placed on developing interventions for overcoming them, including both person-level and systems-level interventions.
Acknowledgments
This research was conducted while both authors were National Institute of Mental Health Policy Post-Doctoral fellows at the Harvard Medical School, Department of Health Care Policy. The authors are grateful to the Substance Abuse and Mental Health Services Administration for making available the public use files of the National Survey on Drug Use and Health. This study was supported by the National Institute of Mental Health Post-Doctoral Traineeship, grant T32-MH 019733 and the National Center for Minority and Health Disparities, grant P60MD002261. The authors would like to thank Richard Frank, Thomas McGuire, and Todd Gilmer for their helpful suggestions on earlier drafts of this manuscript.
Biographies
Victoria D. Ojeda is an assistant professor at the University of California at San Diego, Department of Family and Preventive Medicine. Her research focuses on the health of vulnerable populations, including immigrants, Latinos, and women. Dr. Ojeda has published on health services issues including access to health insurance coverage, utilization of health and mental health services, and the financing of health care. Her current research addresses migration, substance use and HIV/AIDS in injection drug users and female sex workers.
Sara M. Bergstresser is a research scientist at the Nathan S. Kline Institute for Psychiatric Research in Orangeburg, New York. Her research focuses on the anthropology of mental health and illness, the integration of qualitative and quantitative methodologies in public health, and international mental health policy. Her dissertation investigated the history of deinstitutionalization and the current status of nationalized community mental health care in Italy in historical and comparative perspectives.
Footnotes
We assessed mental health status, operationalized as the presence or absence of mental illness, with the respondent’s score to the K6 scale, a validated scale of serious psychological distress (see Kessler et al. 2002 for validation data). In adults, serious psychological distress is defined as having, at some time in the prior year, a diagnosable mental, behavioral, or emotional disorder that met DSM-IV criteria (excluding substance use disorder) and that resulted in functional impairment or significant interference with/limitation of at least one major life activity. The K6 score, however, does not permit an assessment of the severity of the illness. The K6 was implemented in the National Survey on Drug Use and Health based on results from a study that evaluated several brief screening scales for measuring serious psychological distress in the population. Kessler and colleagues (2003) conducted receiver operating characteristic (ROC) curve analyses to evaluate the precision of the scales; results indicated that, of the scales analyzed, K6 was the best predictor of mental illness (Substance Abuse and Mental Health Services Administration 2003b). Investigators (Kessler et al. 2003) reported the area under the ROC curve for the K6 to be .865, with a sensitivity of .36 (SE = .08), and a specificity of .96 (SE = .02) in predicting severe mental illness. Cronbach alpha, a measure of internal consistency reliability, is reported as high at .89 (Kessler et al. 2003:186). Serious psychological distress status is determined by responses to a six-item scale asking respondents to report how frequently they experienced symptoms of distress during the one month in the previous year when they were at their worst, emotionally. The six-symptom items were: (1) “During that month, how often did you feel nervous?”; (2) “How often did you feel hopeless?”; (3) “How often did you feel restless or fidgety?”; (4) “How often did you feel so sad or depressed that nothing could cheer you up?”; (5) “How often did you feel that everything was an effort?”; and (6) “How often did you feel down on yourself, no good, or worthless?” Possible response categories were: “all of the time”; “most of the time”; “some of the time”; “a little of the time’; “none of the time”; or “don’t know/refused.” Respondents scoring 13 or more (out of a possible 24 points) were classified as experiencing serious psychological distress. Persons scoring fewer than 13 points were classified as not experiencing serious psychological distress (Kessler et al. 2003; SAMHSA 2002). The National Survey on Drug Use and Health provides a dichotomous indicator variable that reflects the absence (coded 0) or presence (coded 1) of serious psychological distress, based on the score obtained by the respondent.
The “financial barrier” category includes the following items: “couldn’t afford cost”; “insurance did not cover at all”; “insurance didn’t cover enough”; “health insurance does not cover”; “health insurance does not pay enough”; and “do not have any health insurance coverage.” The “availability/ knowledge/time” barrier includes the following items: “didn’t know where to go”; “did not know where to get services”; “didn’t have time (because of job or childcare or other)”; “no transportation/too far/hours inconvenient”; “services unavailable or limited in your area”; “tried to find help but unsuccessful”; “couldn’t find program/counselor comfortable with”; “too much red tape/hassle”; and “no openings/wait list too long.”
Contributor Information
VICTORIA D. OJEDA, University of California, San Diego
SARA M. BERGSTRESSER, Nathan S. Kline Institute for Psychiatric Research, Orangeburg, New York.
REFERENCES
- Albizu-Garcia Carmen E., Alegría Margarita, Freeman Daniel, and Vera Mildred. 2001. “Gender and Health Services Use for a Mental Health Problem.” Social Science and Medicine 53:865–78. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Alegría Margarita, Canino Glorisa, Ríos Ruth, Vera Mildred, Calderón José, Rusch Dana, and Ortega Alexander N.. 2002. “Inequalities in Use of Specialty Mental Health Services Among Latinos, African Americans, and Non-Latino Whites.” Psychiatric Services 53:1547–55. [DOI] [PubMed] [Google Scholar]
- Alvidrez Jennifer. 1999. “Ethnic Variations in mental Health Attitudes and Service Use among Low-Income African American, Latina, and European American Young Women.” Community Mental Health Journal 35:515–30. [DOI] [PubMed] [Google Scholar]
- Andersen Ronald M. 1995. “Revisiting the Behavioral Model and Access to Medical Care: Does It Matter?” Journal of Health and Social Behavior 36:1–10. [PubMed] [Google Scholar]
- Anglin Deidre M., Link Bruce G., and Phe-lan Jo C.. 2006. “Racial Differences in Stigmatizing Attitudes toward People with Mental Illness.” Psychiatric Services 57:857–62. [DOI] [PubMed] [Google Scholar]
- Antai-Otong D 2002. “Culturally Sensitive Treatment of African Americans with Substance-Related Disorders.” Journal of Psychosocial Nursing and Mental Health Services 40:14–21. [DOI] [PubMed] [Google Scholar]
- Bograd Michele. 1999. “Strengthening Domestic Violence Theories: Intersections of Race, Class, Sexual Orientation, and Gender.” Journal of Marital and Family Therapy 25:275–89. [DOI] [PubMed] [Google Scholar]
- Brown Tony N. 2003. “Critical Race Theory Speaks to the Sociology of Mental Health: Mental Health Problems Produced by Racial Stratification.” Journal of Health and Social Behavior 44:292–301. [PubMed] [Google Scholar]
- Chiu Marcus Y. 2004. “Why Chinese Women Do Not Seek Help: A Cultural Perspective on the Psychology of Women.” Counseling Psychology Quarterly 17:155–66. [Google Scholar]
- Cleary Paul D., Burns Barbary J., and Nycz Gregory R.. 1990. “The Identification of Psychiatric Illness by Primary Care Physicians: The Effect of Patient Gender.” Journal of General Internal Medicine 5:355–60. [DOI] [PubMed] [Google Scholar]
- Cleary Paul D., Mechanic David, and Greenley James R.. 1982. “Sex Differences in Medical Care Utilization: An Empirical Investigation.” Journal of Health and Social Behavior 23:106–19. [PubMed] [Google Scholar]
- Cooper Amy E., Corrigan Patrick W., and Watson Amy C.. 2003. “Mental Illness Stigma and Care Seeking.” The Journal of Nervous and Mental Disease 191:339–41. [DOI] [PubMed] [Google Scholar]
- Corrigan Patrick W. 2004. “How Stigma Interferes with Mental Health Care.” American Psychologist 59:614–25. [DOI] [PubMed] [Google Scholar]
- Doescher Mark P., Saver Barry G., Franks Peter, and Fiscella Kevin. 2000. “Racial and Ethnic Disparities in Perceptions of Physician Style and Trust.” Archives of Family Medicine 9:1156–63. [DOI] [PubMed] [Google Scholar]
- Edlund Mark J., Wang Philip S., Berglund Patricia A., Katz Stephen J., Lin Elizabeth, and Kessler Ronald C.. 2002. “Dropping out of mental Health Treatment: Patterns and Predictors among Epidemiological Survey Respondents in the United States and Ontario.” American Journal of Psychiatry 159:845–51. [DOI] [PubMed] [Google Scholar]
- Freimuth VS, Quinn SC, Thomas SB, Cole G,Zook E, and Duncan T. 2001. “African Ameri-cans’ Views on Research and the Tuskegee Syphilis Study.” Social Science and Medicine 52:797–808. [DOI] [PubMed] [Google Scholar]
- Goffman Erving. 1963. Stigma: Notes on the Management of Spoiled Identity. New York: Simon & Schuster, Inc. [Google Scholar]
- Hinton Ladson, Zweifach Mark, Oishi Sabine, Tang Lingqi, and Unützer Jürgen. 2006. “Gender Disparities in the Treatment of Late-Life Depression: Qualitative and Quantitative Findings from the IMPACT Trial.” American Journal of Geriatric Psychiatry 14:884–92. [DOI] [PubMed] [Google Scholar]
- Kessler Ronald C., Andrews G, Colpe LJ, Hiripi E, Mroczek DK, Normand SLT, Walters EE, and Zaslavsky AM. 2002. “Short Screening Scales to Monitor Population prevalences and Trends in Non-Specific Psychological Distress.” Psychological Medicine 32:959–76. [DOI] [PubMed] [Google Scholar]
- Kessler Ronald C., Barker Peggy R., Colpe Lisa J., Epstein Joan F., Gfroerer Joseph C., Hiripi Eva, Howes Mary J., Normand Sharon-Lise T., Manderscheid Ronald W., Walters Ellen E., and Zaslavsky Alan M.. 2003. “Screening for Serious Mental Illness in the General Population.” Archives of General Psychiatry 60:184–89. [DOI] [PubMed] [Google Scholar]
- Kessler Ronald C., Berglund Patricia A., Bruce Martha L., Randy Koch J, Laska Eugene M., Leaf Philip J., Manderscheid Ronald W., Rosenheck Robert A., Walters Ellen E., and Wang Philip S.. 2001. “The Prevalence and Correlates of Untreated Serious Mental Illness.” HSR: Health Services Research 36:987–1007. [PMC free article] [PubMed] [Google Scholar]
- Kessler Ronald C., Brown RL, and Broman CL. 1981. “Sex Differences in Psychiatric Help-Seeking: Evidence from Four Large-Scale Sur-veys.” Journal of Health and Social Behavior 22:49–64. [PubMed] [Google Scholar]
- Link Bruce G. 1982. “Mental Patient Status, Work and Income: An Examination of the Effects of a Psychiatric Label.” American Sociological Review 47:202–15. [PubMed] [Google Scholar]
- ———. 1987. “Understanding Labeling Effects in the Area of Mental Disorders: An Assessment of the Effects of Expectations of Rejection.” Amer-ican Sociological Review 52:96–112. [Google Scholar]
- Link Bruce G. and Phelan Jo C.. 2001. “Conceptualizing Stigma.” Annual Review of Sociology 27:363–85. [Google Scholar]
- Mechanic David. 1962. “The Concept of Illness Behavior.” Journal of Chronic Diseases 15:189. [DOI] [PubMed] [Google Scholar]
- National Institute of Mental Health. 2006. “Real Men, Real Depression.” Retrieved April 24, 2008 (http://www.nimh.nih.gov/health/topics/depression/men-and-depression/). [Google Scholar]
- Ojeda Victoria D. and McGuire Thomas G.. 2006. “Gender and Racial/Ethnic Differences in Use of Outpatient Mental Health and Substance Use Services by Depressed Adults.” Psychiatric Quarterly 77:211–22. [DOI] [PubMed] [Google Scholar]
- Okazaki Sumie. 2000. “Treatment Delay among Asian-American Patients with Severe Mental Ill-ness.” American Journal of Orthopsychiatry 70:58–64. [DOI] [PubMed] [Google Scholar]
- Perlick Deborah A., Rosenheck Robert A., Clarkin John F., Sirey Jo Anne, Salahi Jamelah, Struening Elmer L., and Link Bruce G.. 2001. “Stigma as a Barrier to Recovery: Adverse Effects of Perceived Stigma on Social Adaptation of Persons Diagnosed with Bipolar Affective Disorder.” Psychiatric Services 52:1627–32. [DOI] [PubMed] [Google Scholar]
- Research Triangle Institute. 2005. SUDAAN 9.0.1. Research Triangle Park, NC: Research Triangle Institute. [Google Scholar]
- Salgado de Snyder V. Nelly, Diaz-Perez Maria de Jesus, Ojeda Victoria D.. 2000. “The Prevalence of Nervios and Associated Symptomatology among Inhabitants of Mexican Rural Communities.” Culture, Medicine, and Psychiatry 24:453–70. [DOI] [PubMed] [Google Scholar]
- Schnittker Jason. 2003. “Misgivings of Medicine?: African Americans’ Skepticism of Psychiatric Medication.” Journal of Health and Social Behavior 44:506–24. [PubMed] [Google Scholar]
- Sirey Jo Anne, Bruce Martha L., Alex-opoulos George S., Perlick Deborah A., Fried-man Steven J., and Meyers Barnett S.. 2001a. “Stigma as a Barrier to Recovery: Perceived Stigma and Patient-Rated Severity of Illness as Predictors of Antidepressant Drug Adherence.” Psychiatric Services 52:1615–20. [DOI] [PubMed] [Google Scholar]
- Sirey Jo Anne, Bruce Martha L., Alex-opoulos George S., Perlick Deborah A., Raue Patrick, Friedman Steven J., and Meyers Barnett S.. 2001b. “Perceived Stigma as a Predictor of Treatment Discontinuation in Young and Older Outpatients with Depression.” American Journal of Psychiatry 158:479–81. [DOI] [PubMed] [Google Scholar]
- Struening Elmer L., Perlick Deborah A., Link Bruce G., Hellman Fredric, Herman Daniel, and Sirey Jo Anne. 2001. “Stigma as a Barrier to Recovery: The Extent to Which Caregivers Believe Most People Devalue Consumers and their Families.” Psychiatric Services 52: 1633–38. [DOI] [PubMed] [Google Scholar]
- Takeuchi David T. and Williams David R.. 2003. “Race, Ethnicity and Mental Health: Introduction to the Special Issue.” Journal of Health and Social Behavior 44:233–36. [Google Scholar]
- Thoits Peggy A. 1985. “Self-Labeling Processes in Mental Illness: The Role of Emotional Deviance.” American Journal of Sociology 91:221–49. [Google Scholar]
- ———. 2005. “Differential Labeling of Mental Illness by Social Status: A New Look at an Old Problem.” Journal of Health and Social Behavior 46:102–19. [DOI] [PubMed] [Google Scholar]
- Thom David H., Hall Mark A., and Gregory Pawlson L. 2004. “Measuring Patients’ Trust in Physicians When Assessing Quality of Care.” Health Affairs 23:124–32. [DOI] [PubMed] [Google Scholar]
- Thom David H., Kravitz Richard L., Bell Robert A., Krupat Edward, and Azari Rahman. 2002. “Patient Trust in the Physician: Relationship to Patient Requests.” Family Practice 19:476–83. [DOI] [PubMed] [Google Scholar]
- Thomas Stephen B. and Quinn Sandra Crouse. 1991. “The Tuskegee Syphilis Study, 1932 to 1972: Implications for HIV Education and AIDS Risk Education Programs in the Black Community.” American Journal of Public Health 81:1498–1505. [DOI] [PMC free article] [PubMed] [Google Scholar]
- U.S. Department of Health and Human Services, SAMHSA, Office of Applied Studies. 2004. Data CD of the National Survey on Drug Useand Health, 2002 [MRDF]. Research Triangle Park, NC: Research Triangle Institute [producer]. Ann Arbor, MI: Inter-university Consortium for Political and Social Research [distributor]. [Google Scholar]
- U.S. Department of Health and Human Services, Substance Abuse and Mental Health Services Administration. 2002. Results From the 2001 National Household Survey on Drug Abuse: Volume II. Technical Appendices and Selected Data Tables. Rockville, MD: Office of Applied Studies, SAMHSA. [Google Scholar]
- ———. 2003a. NSDUH Report: Reasons for Not Receiving Treatment among Adults with Serious Mental Illness. Rockville, MD: Office of Applied Studies, SAMHSA. [Google Scholar]
- Wang Philip S., Lane Michael, Olfson Mark, Pincus Harold A., Wells Kenneth B., and Kessler Ronald C.. 2005. “Twelve-Month Use of Mental Health Services in the United States.” Archives of General Psychiatry 62:629–40. [DOI] [PubMed] [Google Scholar]
- Wynaden Dianne, Chapman Rose, Orb Angelica, McGowan Sunita, Zeeman Zenith, and Yeak SiewHo. 2005. “Factors that Influence Asian Communities’ Access to Mental Health Care.” International Journal of Mental Health Nursing 14:88–95. [DOI] [PubMed] [Google Scholar]