Background and Conceptual Framework

Asian Americans (AA) are the fastest growing racial group in the United States and expected to become the largest immigrant group by 2065 [1]. Despite their growth and presence in many communities, AAs and their mental health needs are seriously understudied [2]. This is troubling given disturbingly high prevalence of mental health problems [3, 4] but the lowest mental health service use among AAs [5]. Certain subgroups, such as Filipino Americans (FAs) and Korean Americans (KAs), may be particularly more vulnerable. A meta-analysis of 58 studies with 21,731 AA adults revealed notably higher rates of depression among FAs (34.4%) and KAs (33.3%) than other AA subgroups [4]. However, mental health service use among FAs and KAs are lower than or similar to those of other AA subgroups. For example, a study with 2285 FAs found that the rate of past-year mental health service use was 2.9% [6]. Another study with a population-based sample of Californian AAs found that only 3.8% of KAs used any type of mental health service in the past year [3].

Perceived public stigma toward mental health service use (herein referred to stigma) refers to the extent to which individuals perceive to feel stigmatized by the general public for using mental health services. It is one of the major barriers to mental health care seeking among AAs [7], and existing studies [8, 9] have identified some determinants of stigma. These factors are likely to both independently and jointly shape stigma. However, they have been rarely accounted for simultaneously in a single study to demonstrate both independent and joint contributions of each predictor to stigma. Previous efforts were also seriously hampered by aggregating diverse AA subgroups [10], despite their significant differences among these subgroups that can variably determine stigma.

Guided by the Culturally Infused Engagement (CIE) model [11], this study aims to illustrate how cultural and contextual influences from multiple systemic levels shape help-seeking and treatment engagement among FA and KA mothers with adolescent children. Specifically, the CIE model purports that the individual process of mental health help seeking is guided by the influences at the (a) familial level (e.g., familial values, beliefs about mental health and help seeking in the family), (b) meso level (e.g., ethnic culture or community values, beliefs, and practices), and (c) macro level (e.g., racial discrimination and health policies). The infusion of these multilevel influences in turn shape culturally determined patterns of help seeking for racial and ethnically diverse individuals.

Following the CIE model, at the individual level, this study examined participants’ demographic characteristics (including age, years living in the U.S., and household income), English language proficiency, mental health status, and previous mental health service use. Extant studies suggest that AAs with longer residence in the U.S. [8], mental health concerns [9], and a history of mental health service use [12] may have more favorable attitudes toward mental health service use than their counterparts without those characteristics. In addition, AAs who are older [9], have higher income [9], and have higher English language proficiency [9] are more likely to seek mental health services.

Familial factors that may influence stigma included child’s mental health, intergenerational cultural conflict (ICC), and gendered family process. AA parents with children experiencing mental struggles were reported to have more favorable attitudes toward mental health service use than their counterparts [9]. ICC (i.e., tensions between parent and children due to acculturation gap) has been shown to predict more mental health service use among AAs [13]. Gendered family process has not been explored much to understand how it may be related to mental health help-seeking attitudes. However, literature on AA family process has identified gendered AA family process (e.g., parental endorsement of gendered norms) as a source of mental distress [14]. Furthermore, less acculturation to the host society has been shown to predict negative attitudes toward mental health help-seeking [8].

This study also included the constructs of “saving face“ and fatalism as ethnic cultural factors of stigma [15]. Saving face (chemyun in KAs and hiya in FAs) indicates a sense of shame and propriety that gravely concerns how others think of oneself and one’s family and creates conformity in the family [15]. Fatalism is bahala na (come what may) [15] for FAs and palja (everything is decided by fate) [16] for KAs. Both ethnic cultural values are associated with more negative attitudes toward mental health service use among AAs [8, 9].

Finally, as the macro level factors, this study included the impact of racial discrimination and colonialism. Experiences of racial discrimination [8] and internalized colonial mentality (i.e., internalized sense of inferiority due to the history of colonization or post-colonialism) [9, 17], have been identified as important contributors to less favorable attitudes toward mental health service use.

Using a unique AA subgroup data and the comprehensive contextual CIE model, this study examined how multi-layered individual and contextual factors explain stigma in two distinct large AA subgroups. The findings should inform the development of group-specific guidelines to reduce stigma with an ultimate goal of improving mental health care use among underserved AAs.

Methods

Participants and Data Collection

This study used data from the Midwest Longitudinal Study of Asian American Families (MLSAAF), a 4-wave panel study of FA and KA families. The eligibility at baseline (2014) was families with self-identified Filipino or Korean mothers with children between 12 and 17 (or in middle or high school), living in 4 major counties near Chicago. The families were recruited from multiple sources, including phone book, ethnic community organization, public and private schools. Parent–child dyads were interviewed in Wave 1 and Wave 2, and only child participants in subsequent waves. This study used Wave 1 parent data because the measure of stigma was available only in Wave 1 parent data. In addition, Wave 1 child report of their own depressive symptoms was used to investigate its association with parent report of their perceived public stigma toward mental health service use. At baseline, 376 FA parents and 378 FA children and 412 KA parents and 408 KA children participated. The survey questionnaires were available in English, Tagalog, and Korean, and in paper and online formats. Survey was collected mostly by trained bilingual interviewers (84%) at Wave 1 and self-administered in subsequent waves. Participants received incentives ($40 for parents and $20 for children at Wave 1). The study was approved by the University’s IRB.

Parent participants were predominantly female (95.39% FAs and 97.8% KAs) and foreign-born (91.44% FAs and 99.51% KAs) and on average, in their mid-forties [M = 46.19 (SD = 5.78) for FAs; M = 45.31 (SD = 3.76) for KAs]. FAs have been living in the U.S. longer than KAs (22.23 vs. 16.15 years) and reported significantly higher annual household income and English proficiency than KAs. Average age of youth participants were 15.24 (SD = 1.87) for FAs and 14.74 (SD = 1.90) for KAs.

Measures

Unless noted, response options were a five-point Likert scale (e.g., 1 = not at all to 5 = very likely) and a higher score indicates a higher level of the measured construct.

Individual Level Factors

Demographic Variables

Age and years of living in the U.S. of participants, and annual household income (1 =  < $25 K; 2 = $25–$49 K; 5 = $100–$149 K; 6 =  > $150 K) were measured.

English Language Proficiency

Two items from the Language, Identity, and Behavior [18] measured English language proficiency, e.g., how well they understand and speak English. (r = 0.81 for FAs and 0.87 for KAs).

Parental Depression

The 20-item modified version of the Center for Epidemiologic Studies Depression (CES-D) scale [19, 20] was used to measure depression during the week prior to the survey. Participants who scored a total of 16 or above were classified as depressed [21] (α = 0.85 for FAs; α = 0.90 for KAs).

Lifetime Mental Health Service Use

Parent participants were asked whether they had ever been treated for mental health problems. Response options were No (0) or Yes (1).

Familial Factors

Child’s Depressive Symptoms

Survey asked child participants about their mental health status during the two weeks prior to the survey, using a 13-item scale from the Children’s Depressive Inventory [22] (α = 0.93 for both groups).

Intergenerational Cultural Conflict

Ten items from the MLSAAF asked to measure parental perception of the level of cultural gap between them and their adolescent children. Examples included “My child talks back and I find it disrespectful” and “I feel my child is too Americanized” (α = 0.77 for FAs; α = 0.75 for KAs).

Gendered Norms

To measure parental gendered norms, 8 items from the MLSAAF were used. Examples included “Girls should not date while in high school” and “Boys should avoid anything girlish or feminine” (α = 0.79 for FAs; α = 0.80 for KAs).

Ethnic Cultural Factors

Saving Face

A total of seven items from the MLSAAF, del Prado and Church [15], and Kim, Atkinson [23] were adopted to measure the concept of saving face among parents. Examples included “Family reputation is important” and “A personal failure is a letdown for the entire family” (α = 0.79 for FAs; α = 0.81 for KAs).

Fatalism

Fatalism was measured by one item, differently worded for each subgroup. Specifically, parents were asked the extent to which they believe in bahala na for FAs [15] and palja for KAs [16].

Macro Level Factors

Racial Discrimination

A total of four items from the MLSAAF and Phinney, Madden [24] were adopted. Examples included “I have felt discriminated [against] by Whites” or “by racial and ethnic minorities like Blacks or Hispanics.” (α = 0.79 for FAs; α = 0.71 for KAs).

Colonial Mentality

Ten items from the Colonial Mentality Scale [25] measured participants’ internalized sense of inferiority due to the legacy of colonization and neocolonialism [25]. Examples included “I think that a person that is part White and part Filipino/Korean is more attractive than a full-blooded Filipino/Korean” (α = 0.76 for FAs; α = 0.81 for KAs).

Dependent Variable

Stigma Toward Mental Health Service Use

A total of three items from the Indifference to Stigma subscale [26] was used to measure service-use stigma. Items included feeling uneasy going to mental health professionals because of what some people would think of them and believing that having been a client at the mental health services is a blot on a person’s life and that having been mentally ill carries a burden of shame (α = 0.81 for FAs; α = 0.72 for KAs).

Analyses

STATA version 16.1 was used for data analyses. Descriptive statistics were generated by ethnic groups and examined for statistical differences. Bivariate correlations were also generated for each subgroup.

To examine both unique and joint contributions of each cluster of predictors of stigma, hierarchical ordinary least squares (OLS) regression was estimated. Step 1 included individual level factors (three demographics, English proficiency, parental depression, and lifetime mental health service use). Each cluster was examined in subsequent steps (familial variables in Step 2, ethnic cultural variables in Step 3, and macro level variables in Step 4). Steps 2 through 4 included individual level factors of Step 1. All clusters were examined in Step 5.

The missing rates in completed survey were less than 5%. To handle missing data, we conducted multiple imputation using chained equations with 20 iterations. An additional analysis was conducted with final model (Step 5) to investigate whether the study results are sensitive to the exclusion of a small proportion of male and/or U.S.-born participants among KAs. The study findings remained the same with and without male and U.S.-born samples. For FAs with a larger proportion of U.S.-born samples than KAs (8.56% vs. 0.49%), we conducted an analysis without male participants, while controlling for place of birth, and the study results did not change with this restricted analysis.

Results

Descriptives

Tables 1 and 2 summarize descriptive statistics. Of note, rates of lifetime mental health service use at Wave 1 were 5% for both groups and past year mental health service use at Wave 2 (not included in this study) were 2% for FAs and 1% for KAs. KAs showed significantly lower rates of stigma (1.89 vs. 2.36), but higher rates of depression than FAs (23.59% vs. 11.14%).

Table 1 Sample characteristics
Full size table
Table 2 Correlations between study variables
Full size table

Hierarchical Regression Analyses

Table 3 summarizes the regression results. We first included individual level factors (Step 1; R2 = 0.03 for both groups), which explained 3% of the variations in stigma for both groups. Only English proficiency was significantly associated with less stigma among KAs. Familial (Step 2; R2 = 0.08 for FAs; R2 = 0.16 for KAs) and ethnic cultural (Step 3; R2 = 0.10 for FAs; R2 = 0.20 for KAs) factors explained more variations in stigma among KAs than FAs. In Step 2, ICC and gendered norms were significantly associated with stigma in both groups. In Step 3, saving face was associated with stigma in both groups, while fatalism was significant only among KAs. Conversely, macro level factors (Step 4; R2 = 0.09 for FAs; R2 = 0.07 for KAs) explained 9% and 7% of the variations in stigma for FAs and KAs, respectively. Racial discrimination and colonial mentality were associated with stigma among FAs, and colonial mentality alone was significant among KAs. In the final full model (Step 5; R2 = 0.15 for FAs; R2 = 0.26 for KAs), saving face remained significant in both groups. Gendered norms and fatalism remained significant among KAs, and all macro level factors remained significant among FAs.

Table 3 Hierarchical OLS regression results [b (95% CI)]
Full size table

Discussion

To address the serious public health concern, a combination of alarming rates of mental distress and the lowest use of mental health services among AAs, this study aimed to examine the multilevel influences on stigma toward mental health service use that may deter FAs and KAs from using mental health services when they need them. The findings of this study demonstrate how each cluster of predictors of stigma independently and collectively may influence stigma. The findings also highlight several important similarities and differences in how stigma may be shaped among two major AA ethnic subgroups, signifying the importance of subgroup specific interventions to be effective in addressing unmet mental health care needs in distinct subgroups of AAs.

For example, saving face, a keen attention to how others think of oneself and one’s family, was a significant correlate of stigma in both groups, emerging as the robust and shared determinant of stigma. This finding is not surprising given the nature of saving face and a higher level of stigma against mental health problems in AA communities. Nevertheless, this finding highlights the importance of community level education to ease stigma. In addition, the importance of macro level factors among FAs is notable. Consistent with the existing literature [8, 9, 17], racial discrimination and colonial mentality were significant correlates of stigma, particularly among FAs for whom they explained more variance and remained significant after accounting for all other clusters of factors. Conversely, familial and ethnic cultural factors (e.g., gendered norms, saving face and fatalism) explained twice the variance in stigma for KAs than FAs.

These findings may be explained by prior research on differences in acculturation, patterns of integration in U.S. society and immigrant/colonial history among AA subgroups [27, 28]. That is, FAs, whose country of origin is a former colony of the U.S., are highly acculturated into American society (e.g., occupational integration) [28] and may be more likely to be influenced by racial minority status. On the other hand, KAs, one of the least acculturated groups and living in neighborhoods of high co-ethnic density [27], may be more sensitive to familial and ethnic cultural factors (cultural family process and traditional values) than macro factors. Racial discrimination significantly increases mental distress [29]. It is concerning that racial discrimination can also increase stigma, which means AAs can be distressed because of racial discrimination but because of it, they may not seek needed help.

It is also important to note that gendered norms and colonial mentality were no longer significant for FAs and KAs, respectively, and ICC for both groups, when all predictors were accounted together. It is possible that other factors are more powerful than ICC, gendered norms, and colonial mentality. Alternatively, this could indicate a potential mechanism for mediation, in which familial (e.g., ICC and gendered norms) and macro level (e.g., colonial mentality) factors may enhance ethnic cultural factors such as saving face to indirectly influence stigma, which warrants further investigation in future research.

The findings have some limitations. First, we would have liked to directly examine correlates of mental health service use. However, the actual service utilization was too low to use as an outcome. This study used regional data, also from predominantly immigrant and female adults, thus has limited generalizability. Immigrants and men are less likely to seek mental health service and have more stigma than their U.S.-born and female counterparts [30, 31]. Future studies should expand in gender and immigrant status as well as regional representations.

New Contribution to the Literature

Despite these limitations, this study is one of few to examine correlates of stigma toward mental health service use among AA subgroups, using one of the largest studies on FAs and KAs in the Midwest. By examining multilevel influences on stigma while focusing on ethnic subgroup differences, this study provides empirical evidence for the complexity of influences on mental help-seeking attitudes. The comprehensive approach in this study points to the need for intervention efforts to go beyond targeting individual or familial level factors, which has largely been of focus in the field. Indeed, a handful of new interventions for AAs have been integrating the salient roles of culture and race [32]. For example, a culturally grounded intervention that addressed the impact of culturally specific stressors (e.g., culturally gendered parenting norms) found reduction in depressive symptoms among AA women [33]. A pilot intervention that centered on reducing race-related stress among AA young adults reported symptom reductions in general distress, depression, anxiety, and PTSD symptoms [34]. The efficacy of these studies points to the greater need for future interventions that are additionally tailored to the cultural and racial contexts of AAs. This study also provides a unique contribution by studying AAs in the Midwest, as most studies of AA mental health and service use are from the West Coast [9]. The findings on the importance of racial minority status are particularly crucial given that Midwestern AAs report the highest incidence of racial microaggressions among AAs in the U.S. [35].