Mental Health Care and the Cultural Toolboxes of the Present-day Japanese Population: Examining Suggested Patterns of Care and Their Correlates

Abstract

An extensive body of literature has documented the under-utilization of mental health services among Asian populations, regardless of where they live. Japan is one Asian country where the mental health care system has improved substantially in the recent decades. Yet, Japan continues to report greater under-utilization of mental health services than other developed countries. One primary reason for this is hypothesized to be the cultural climate, which includes the stock of social, cultural and medical knowledge surrounding mental illness. Previous studies have not examined the cultural toolkit (Swidler, 2001) of the Japanese public, nor have they linked these to public attributions or assessments surrounding mental health. The Stigma in Global Context -Mental Health Study (SGC-MHS), a multi-stage probability sample of Japanese residents aged 18–64 years (N = 994) in 2006 provides data to describe the unprompted care suggestions from the general population regarding case scenarios meeting psychiatric criteria for schizophrenia and depression. Guided theoretically by the Network Episode Model, we analyze if and how respondents socially organize solutions into patterns of culturally acceptable care for mental health problems in Japan. Six cultural utilization patterns appear to be shaped by beliefs about underlying attributions, perceived severity of conditions, and type of mental health problem. Further, women, older individuals and those with differing education levels suggest different care patterns than their counterparts. In sum, the findings indicate unique patterns of mental health care recommendations among the Japanese public, which are culturally and socially constrained.

1. Introduction

Asian populations underutilize mental health services (Chen, 2012; Lee, 2011; WHO, 2004). In Japan, even as mental health care provision has improved substantially over the last few decades, most individuals meeting criteria for DSM-IV disorders report no treatment (Naganuma et al., 2006; Kawakami et al., 2016). The estimated level of unmet need in the Japanese population continues to be considerably higher than what is reported in other developed countries (WHO, 2004). This poses an obvious question: What explains the disconnect in Japan between substantial need and use, particularly in light of substantial national efforts to improve access to mental health services?

From a social science perspective, cultural context has been seen as key in understanding people’s choice of mental health care both empirically and theoretically (Pescosolido, 1991, 1992; Olafsdottir and Pescosolido, 2009; Alegria et al., 2004; Takeuchi et al., 1999). Specifically, cultural climate has two unique components. First, individuals’ ideas about appropriate care constitutes the health or medical “cultural toolkits” (Swidler, 2001) held by the public. Second, attitudes and beliefs about mental illness, including attributions and assessments, constitute contingencies that shape who suggests different patterns of care and why they do so. Previous studies in Japan have not fully examined the nature and correlates of the public’s socio-cultural construction of mental health problems and appropriate solutions. This study aims to fill part of this gap by investigating patterns and correlates of care that Japanese respondents suggest, on their own, for mental health problems. More specifically, we ask a series of questions: 1) What kinds of care did the respondents recommend on their own, without prompting?; 2) Are there any patterns to those suggestions?; and 3) What factors are associated with the suggested patterns of care?

This approach expands on the idea of “cultural mapping” suggested by Olafsdottir and Pescosolido (2009), which involves demarcating “the cognitive schema and cultural scripts that individuals possess to define the potential universe of providers” (p. 231). Work on patterns of care have been done, ironically, mostly on the U.S. case and other Western countries where modern medicine is thought to be more dominant and challenged by fewer, well-developed alternatives. This study examines the public’s view of mental health care options in Japan, a country with a well-developed traditional alternative medical system (Lock, 1980) as well as allopathic options. As such, Japan provides a unique view into lay culture. In order to examine the public’s view in Japan, we used a “bottom-up method” which asks, in an open-ended fashion, what a person, described with symptoms that meet DSM-IV criteria of depression or schizophrenia should do, if anything. The distinction between suggestions (i.e., individuals’ unprompted recommendations to problems) versus endorsements (i.e., individuals’ response to an investigator-derived list of providers or practices) is important to capture the public’s perspective. As Pescosolido and Olafsdottir (2010) note, only the former mark what individuals’ can readily access in times of trouble. The latter tends to reflect larger cultural values, particularly of the dominance of allopathic medicine, but not what is actually in individuals’ stock of knowledge and belief. As a result, we offer a general look at the complexities of care options outside of the West from individuals’ point of view.

1.1. Theoretical Background: Culture and Mental Health Care Options

Theories of health care utilization, both within and across countries have always posited the role for beliefs, attitudes and lay assessments (Pescosolido, 1991). More recently, studies of health care disparities, including those targeting mental illness, have documented cultural facilitators of and barriers to service use (Algeria et al. , 2004; Pescosolido et al. 1998; Takeuchi et al. 1999). The Network Episode Model (NEM), for example, goes beyond the usual idea that there are social and cultural correlates of use to the more basic notion that health care options are culturally embedded into patterns of use, whether actual or recommended. More recently, Olafsdottir and Pescosolido (2009) argued for the importance of directly capturing “the whole cultural systems,” which include “deeply embedded and collectively shared norms, values, expectations, and beliefs” (p.229) that shape the public’s tacit knowledge of options and reliance on them. More specifically, a cultural systems approach reconceptualizes the importance of cultural contexts. This approach requires a shift from simplistic dichotomies (e.g., use/no use) to complex constructions described in the NEM. That is, respondents’ patterns of care reveal a cultural map regarding mental health options from the lay perspective. In addition, examining sociocultural factors associated with patterns of care requires going past, but not excluding, individuals’ social location (e.g., gender, education) to direct cultural measures (e.g., values, expectations, and beliefs), including perceived risks (e.g., perceived seriousness of conditions) and attributions (i.e., chemical imbalance in the brain).

This study expands on the NEM’s “cultural map” in two ways. First, we examine how people discriminate across a variety of sources of care, not just potential “providers,” using an emic approach rather than an etic one. That is, drawing on the distinction between “suggestions” (i.e., asking individuals to suggest “what should be done” in an open-ended frame) and “endorsements” (i.e., asking individuals to respond to a standard, usually an allopathically-oriented list of providers), we use the former approach, which requires more cognitive work and allows for cultural resistance on the part of individuals. While this is important everywhere, in a society with a long and deep history of alternative medical systems and practices, such as Japan, this may include professional consultation, consultation with laypeople (e.g., family and friends), traditional and indigenous practices and providers (e.g., kampo, acupuncture; see Lock, 1980) and various other actions (e.g., “deflection”, defined as cognitively focusing away from the issue; see, for example, Canino et al., 1997).

Second, we do not look at a series of responses in isolation. Rather, we ask whether there are unique patterns (i.e., combinations) of respondents’ suggestions that comprise the cultural toolkits from which individuals tend to draw from in response to mental health problems. Further, we examine whether their construction differs for different kinds of individuals (i.e., sociodemographic characteristics) and cultural predisposition (e.g., attributions, assessments of problems). Facing their own or others’ health problems, some people may evoke only one care option (e.g., talking to their family), while others may evoke multiple options (e.g., talking to a doctor and their family). These categorizations will allow us to identify groups who do not favor professional care and explore how they differ culturally from one another, and from those who suggest such care. As such, this analysis has the potential to contribute to the specific understanding of under-utilization in Japan, as well as to the general understanding of the role of culture in the public understanding of and response to mental illness.

1.2. The Case of Japan

The concept of “cultural mapping” may help explain the continued low utilization of professional mental health care in Japan as it has in the U.S. (Pescosolido and Olafsdottir, 2010). In particular, three factors are relevant. First, Japanese health beliefs and practices are considered to be influenced by both Chinese and Western-style medicine. Chinese-style medicine was introduced to Japan and has been in use over centuries; Western-style medicine was officially adopted only at the end of the 19th century and gained influence during the process of industrialization (Ohnuki-Tierney, 1984; Lock, 1980). According to more traditional East Asian health beliefs, health problems occur when the body loses its balance. Treatment aims at restoring balance and harmony through holistic methods (Haug et al., 1991; Lock, 1980). Further, self-reliance is more encouraged under this tradition than in Western medicinal practices (Haug et al., 1991). In contemporary Japan, Western-style medicine dominates. Its national health insurance system enables the public to access low cost Western-style medicine while limiting the coverage for Chinese-style medicine. Nevertheless, studies suggest that East Asian health beliefs and practices continue to be observed (especially regarding self-care) in Japan (Haug et al., 1991; Ohnuki-Tierney, 1984).

Second, social support networks may be more narrowly kin-centered in Japanin comparison to the other countries, reflecting the cultural atmosphere of its “familialistic welfare regime” (Esping-Andersen, 1999). Under this regime, norms emphasize that families are responsible for the care of their members. As a potential corollary, people tend not to depend on people outside the family, partly because seeking help from those who do not belong to the family is considered shameful (Kikuzawa, 2006) and doing so impose “obligations that one is morally obliged to repay” (Lock, 1980). Indeed, according to a recent comparative survey on the elderly in four countries (Japan, U.S., Germany, and Sweden), 66.2% of the elderly in Japan reported that they have family or relatives who live separately, but on whom they can rely in times of sickness. Only 18.5% reported that they have such friends. The corresponding percentages in other countries are 59%−69% for family or relatives who live separately and 43%−45% for friends (Japan Cabinet Office, 2015).

Third, Japan is an Asian country, whose culture emphasizes harmonious social relationships, which in turn affects the use of support from close others. According to Kim et al. (2008), cultural differences exist between Asians, Asian Americans and European Americans on how to seek social support for coping with stress. These differences, rooted in cultural dissimilarities, find that individuals with an Asian cultural background are reluctant to use “explicit social support” which is defined as “people’s specific recruitment and use of their social networks in response to specific stressful events that involves the elicitation of advice, instrumental aid, or emotional comfort” (Kim et al., 2008: 522). Their reluctance for seeking support stems from a concern that doing so may reverberate on their relationships with others, including disturbing group harmony or inviting criticism. As a result, Kim et al. (p. 523) suggest that “in a culture in which maintenance of harmonious social relationships is emphasized,” people may seek more “implicit social support,” or “the emotional comfort one can obtain from social networks without disclosing or discussing one’s problems vis-à-vis specific stressful events,” (p. 522). Implicit social support does not introduce the “relational risks” inherent in explicit social support.

More recently, and as noted earlier, the provision of formal care services for people with mental illness in Japan has substantially improved, although challenges remain. Psychiatric clinics in communities have rapidly increased in number (Enomoto, 2008; Japan Ministry of Health, Labour and Welfare, 1996, 2005, 2014), and there is greater variety in legally accredited medical departments where the public can consult about mental health problems. Shinryo-naika (psychosomatic medicine) clinics were legally accredited in 1996 (Ebana, 2002). “Counselors” have become familiar to the public through the school counselor system developed by the Ministry of Education in 1995. Many universities now provide training programs for clinical psychologists certified by the Foundation of the Japanese Certification Board for Clinical Psychologists founded in 1988 (Yagi, 2008). Some counseling clinics are run by psychologists alone, although they are still limited in number. Finally, as part of a national effort to prevent suicide, official consultation sites for those with “worries”, as well as their families, were set up by local governments, and the Center for Suicide Prevention was created in 2006 to support these efforts.

There have also been efforts to increase awareness about mental illness and eradicate stigma. For example, following a 2002 proposal by the Japanese Society of Psychiatry and Neurology, the Japanese translation for “schizophrenia” was revised from seishin-bunretsu-byo (mind-split disease) to togo-shitcho-sho (integration disorder) in official documents and mass media. This was expected, and has been documented, to reduce the negative connotation and prejudice associated with the old term (Koike et al., 2017). The government has also disseminated information about mental illnesses, especially depression, in an effort to reduce Japan’s high suicide rate. In 2010, the Ministry of Health, Labor, and Welfare launched a website called “Minna no mentaru herusu sogo saito” (translated as “Comprehensive site for everyone’s mental health”) to provide information on mental illness, services and policies.

This is a critical time to examine individuals’ cultural toolboxes for mental health problems in Japan. The building of the mental health service system in Japan has been met with recent reports indicating changes in the public’s use of mental health services. For example, a recent epidemiological survey documented increases in service use among those with mental disorders (e.g., mood disorder, anxiety disorder) over the past decade (Kawakami et al., 2016). Yet, the same survey found that about 70% of those with mental disorders in Japan still do not receive any treatment. There is little evidence to account for this phenomena. Taking a cultural mapping approach offers one possible avenue to explain this continuously low level of utilization.

1.3. Utilization Research in Japan

Studies on mental health care use, including studies of help-seeking for mental health problems and studies of attitudes toward mental health care, was quite limited before 2000 in Japan compared to that of Western countries (Kido et al., 2013; Mizuno and Ishikuma, 1999; Nagai, 2017). The number of empirical studies on help-seeking has increased in recent decades, but most studies, using small samples of a specific population (e.g., students, teachers), have focused on psychological factors associated with help-seeking attitudes or behaviors (see Nagai, 2017 for a review).

Only a few studies have investigated attitudes toward mental health care targeting more general populations. For instance, Jorm et al. (2005) conducted a general population, though not nationally representative, survey in Japan on public beliefs relating to four vignettes describing depression, depression with suicidal thoughts, early schizophrenia, and chronic schizophrenia. Respondents endorsed a strong belief in helpfulness of counselors, although interventions that were commonly rated as “helpful” varied by vignette. For depression, the most commonly endorsed interventions were counselors, family and friends, and for schizophrenia, counselors, family, and psychiatrist were most commonly endorsed. Using data from the 2005 Japanese General Social Survey, Yamawaki et al. (2011) examined the choice of confidents respondents might talk to if they were “suffering from great personal worry or stress” and anticipating a possible “nervous breakdown.” Women, married people, and younger individuals were more likely to endorse mental health professionals than others. In addition, women and married people were more likely to endorse family, while those who were younger, single, and women were more likely to talk to friends than their counterparts.

While informative, prior studies on individual’s endorsement of mental health care options in Japan have been limited in several important respects. First, the response categories have been mostly closed-ended or forced choice rather than open-ended. As previous research has shown, this approach is more likely to produce responses conforming to larger cultural values instead of individuals’ actual repertoires (Pescosolido and Olafsdottir, 2010). Second, research on Japan has typically focused on “use or nonuse” of specific care options. However, such decisions are unlikely to be isolated. Rather, according to the Network Episode Model, the decision is likely to be “enmeshed in a wider pattern of help-seeking” (Pescosolido, 1991, 1992). Eliciting culturally appropriate combinations (or patterns) of care from the respondents’ point of view provides a better window into the whole medical-cultural system.

1.4. The Current Study

Taking all of these factors into account, we analyze the patterns of care that Japanese residents suggest when confronted with a scenario describing a mental health problem. We hope to provide new insights into the cultural context of low utilization of mental health care on three issues.

First, as part of the Stigma in Global Context—Mental Health Study (SGC-MHS), a survey was conducted on a nationally representative sample of Japan residents. As in all countries, respondents were randomly provided with one vignette describing an individual with symptoms meeting diagnostic criteria for depression or criteria for schizophrenia. Immediately following the scenario, respondents were asked, using open-ended format, to suggest what that individual should do, if anything. Their unprompted list, as we argued earlier, reflects available cultural repertoires for care among individuals in contemporary Japan (Pescosolido and Olafsdottir, 2010).

Second, we use these responses to examine the types and patterns of mental health care that Japanese residents suggest, revealing their “cultural toolkits” regarding mental health care. This broadens the typical narrow and formal categorization of mental health care found in help-seeking studies. As such, it is one of the few studies to do so.

Third, we investigate whether and how those cultural toolkits are associated with individuals’ social locations, attitudes about mental health, and beliefs about mental health problems (Olafsdottir and Pescosolido, 2009). While previous US-based research suggest that the influence of sociodemographic cleavages on mental health care options are diminishing as the society proceeds from modern to post-modern (Pescosolido and Rubin, 2000; Olafsdottir and Pescosolido, 2009), this has yet to be tested in sociocultural contexts like Japan.

2. Materials and Methods

2.1. Data Source

Data for this study were obtained from a nationwide population-based survey conducted by the Japanese SGC-MHS team between September and December, 2006. SGC-MHS, an international project, aims to “attempt to understand the extent to which mental illness is understood and stigmatized across countries” (http://www.indiana.edu/~sgcmhs/). The SGC-MHS survey was originally designed to create a dataset for cross-cultural analyses. However, the Japanese team decided on significant design differences including respondent choice of interview or self-administered questionnaire, elimination of an in-group/out-group variant in the vignettes, and the elimination of a control condition (asthma). These changes make the Japan data unique, not strictly comparable to other SGC-MHS data, but critical in its coverage of East Asia.

A two-stage stratified random sampling strategy produced a nationally representative sample of Japan residents aged 18–64 (N = 1,800). Respondents were visited at their home by enumerators between September and October 2006, after a written request for participation was mailed. For respondents who gave verbal consent, the survey was conducted through face-to-face structured interviews (10% of completers) or self-administered questionnaires, depending on respondent preferences. Self-administered questionnaires were picked up by enumerators or mailed in after completion. For respondents absent at the time of the enumerators’ visits, questionnaires were mailed and collected by post between November and December 2006. A second mail wave was sent to those giving consent but who did not send questionnaires back by October. The survey yielded a total of 994 respondents (response rate: 55.2%). Respondent characteristics were not significantly different between the survey modes except that self-administered respondents were somewhat younger and more educated than interviewed respondents. Ethics approval was granted by the Department of Medicine at Tokyo University, as well as Indiana University’s Institutional Review Board.

2.2. Measurement

2.2.1. Dependent Variables

Dependent variables were respondents’ suggestions for types and patterns of care for the hypothetical individual in the mental health vignette. The presentation of a vignette (see Appendix) describing a person meeting DSM-IV criteria for schizophrenia or depression was followed immediately by an open-ended question asking respondents what the hypothetical individual should do (three open-ended suggestions were recorded). Schizophrenia and depression were selected because these conditions represent the most severe and the most common types of mental disorders, respectively, that generally elicit considerable concern (WHO, 2001). Respondents’ suggestions were carefully read and coded into 12 dummy variables for the recommended types of mental health care (1 = recommended, 0 = not recommended): (1) do nothing, (2) talk to family, (3) talk to friends, (4) talk to a religious leader, (5) general medicine, (6) specialty medicine, (7) alternative medicine (e.g., acupuncturist), (8) take non-prescription medication, (9) deflect (e.g., get involved in other activities, go out or travel, start exercising, try spending time with others), (10) care for oneself (e.g., reflect on oneself, rest well, (11) talk to some unspecified person (e.g., someone close, someone he/she can trust), or (12) talk to a boss or colleagues. Category (5) includes suggestions to talk to a professional in general medicine or general public staff (e.g., doctor, hospital, local health center staff, public consultation office staff, district welfare commissioner), and category (6) includes suggestions to talk to a professional in specialty medicine or a related profession (e.g., psychiatrist, specialized hospital, counselor, staff of the department of psychosomatic medicine). We abbreviated (5) as “general medicine” and (6) as “specialty medicine” because they best capture the feature of the suggestions. For example, almost all respondents who gave responses that fell into category 5 suggested talking to a doctor, while less than 10 respondents suggested talking to a public staff member without suggesting to going to a doctor/hospital.

Using these variables, non-hierarchical cluster analysis (FASTCLUS, SAS, 2011) was used to examine the patterns of public recommendations for mental health care. To ensure stability in the cluster analysis, variables that were not selected by at least 10 individuals were excluded. Peaks on the Cubic Clustering Criterion (CCC) plot with values greater than three generally indicate appropriate clusterings (Sarle, 1983, p. 49). The relatively large values of the pseudo F statistic also indicate good clusterings. The density and image matrices for the six recommendation pattern solution are presented in Figure 3, described in the Analysis section below. The numbers in the cells of the density matrix represent the percentages of respondents in each pattern who chose a particular option. In the image matrix, those frequencies are translated into a dummy variable (1 = recommended, 0 = not recommended) using a cut-off of 50%. To measure whether the respondents discriminated between those six options, we used this number of clusters as a second multinomial dependent variable (suggested patterns of mental health care).

Figure 3.

Density and Image Matrices for Recommended Patterns of Mental Health Care, Japanese SGC-MHS (N = 889)

2.2.2. Independent Variables

The independent variables, based on previous research (Pescosolido and Olafsdottir, 2010; Furnham and Henley, 1988; Mechanic et al., 1991; Angermeyer et al., 1999), were sociodemographic factors (e.g., respondent gender, marital status, age, education, rural/urban residence, and economic condition), vignette types coded as dummy variables (e.g., gender of vignette character, schizophrenia), and attitude toward the cause and consequences of the symptoms described in the vignette. The descriptive statistics and details of the measurements are provided in Table 1.

Table 1.

Sample Characteristics, Japanese SGC-MHS^*

Variable	Coding	Mean	SD
Causes and Consequences
Seriousness	Respondents’ evaluation of seriousness: 1 = “not at all serious” to 4 “very serious”	3.64	0.55
Bad Character	Caused by bad character: 1 = “not at all likely” to 4 “very likely”	1.93	0.77
Brain Disease	Caused by brain disease: 1 = “not at all likely” to 4 “very likely”	2.39	0.91
Way Raised	Caused by the way she/he was raised: 1 = “not at all likely” to 4 “very likely”	2.62	0.81
Stress	Caused by stress: 1 = “not at all likely” to 4 “very likely”	3.71	0.54
Genetic Transmission	Caused by genetic or inherited problem: 1 = “not at all likely” to 4 “very likely”	2.01	0.79
God’s Will	Caused by God’s will: 1 = “not at all likely” to 4 “very likely”	1.36	0.60
Bad Luck	Caused by bad luck: 1 =“not at all likely” to 4 “very likely”	1.77	0.83
Considers Violent Toward Others	Respondents’ evaluation of dangerousness toward others: 1 = “not likely at all” to 4 “very likely”	2.42	0.72
Considers Violent Toward Self	Respondents’ evaluation of dangerousness toward self: 1 = “not likely at all” to 4 “very likely”	3.23	0.69
Vignette Characteristics
Schizophrenia	Dummy variable: 1 = schizophrenia vignette, 0 = depression vignette	0.51	0.50
Female	Dummy variable: 1 = female vignette, 0 = male vignette	0.51	0.50
Respondents’ Characteristics
Female	Dummy variable: 1 = female, 0 = male	0.53	0.50
Married	Dummy variable: 1 = currently married, 0 = currently not married	0.72	0.45
Age	18–64 years	45.01	12.10
Education	Dummy variable: 1 = university or higher degree, 0 = others	0.25	0.43
Rural Residence	Dummy variable: 1 = live in rural area, 0 = others Respondents’ judgment of his/her economic	0.28	0.45
Economic Condition	condition: 1 = “Has just about to live on but cannot afford more than that” to 4 = “Is able to lead an affluent life”	2.61	0.71

^*Based on effective sample size for the multivariate analysis (N=748)

2.3. Analysis

Our analyses were motivated by a series of questions: 1) What kinds of care did the respondents suggest? 2) Are there patterns to those suggestions? 3) What factors are associated with patterns of suggestions? To answer the first question, we examined the frequency distributions of the original dummy variables showing public recommendation of the types of care for the vignette person. To answer the second question, we conducted cluster analyses as described above to derive our second dependent variable (cluster) and examined the frequency distributions to reveal patterns of suggestions. To answer the final question, we conducted a multinomial logistic regression in order to explore the factors associated with those pattern sets. SAS University Edition (SAS Institute Inc., 2017) was used for cluster analyses, and Stata 14.2 (Stata Corp., 2015) was used for the multinomial logit analysis.

3. Results

3.1. Public Recommendations of Types of Mental Health Care

Figure 1 shows the frequencies for our first set of dependent variables. The most frequently recommended action was consulting professionals in “general medicine” (38.2%), followed by “deflect” (e.g., get involved in other activities, go out or travel, start exercising, try spending time with others, 37.8%) and talking to professionals in “specialty medicine” (36.1%). As expected, two of the three actions were related to the use of professionals, although the other was unrelated. Actions such as “talk to family” (28.3%) and “care for oneself” (20.8%) were also recommended, but not as much as might be expected given data from other countries such as the U.S. (Swindle et al., 2000). Further, 21% of respondents recommended talking to “some unspecified person,” 16.9% recommended talking to “friends,” and fewer than 10 of individuals recommended the following: “talk to a boss or colleagues,” talk to professionals in “alternative medicine,” “do nothing,” “take non-prescription medication,” or “talk to a religious leader.” Those small categories were excluded from the subsequent cluster analyses to ensure stability.

Figure 1.

Recommended types of Mental Health Care, Japanese SGC-MHS (N=890)

3.2. Patterns of Mental Health Care Recommendations

Figure 2 shows the values of the CCC (2a), and the values of the pseudo F statistics (2b) for a range of partitions (2–30 clusters). Although there are several jumps in the CCC plots (e.g., 6, 12), the value of the pseudo F statistic drops after six clusters. Considering both CCC and the pseudo F statistic, the six clusters solution represent a “best” cluster set (Figure 3). Three are singleton patterns, while the other three reveal “dual use” (Press, 1969). As expected, several patterns included the use of “general medicine” (Patterns 2 and 3) and “specialty medicine” (Patterns 5 and 6), while two patterns did not include the use of professionals. Pattern 1 (N = 215), the most frequently suggested, was “only deflect,” while Pattern 4 (N = 161) was “only talk to some unspecified person.” Among the patterns that included the use of professionals, Pattern 2 (N = 123) represented a “single” strategy of consulting a general professional, the strategy that has been the focus of the traditional rational choice-based utilization research (see Pescosolido, 1992 for a review). Nevertheless, this singleton pattern was an exception. Dominant patterns that included the use of professionals tended to be combined strategies such as “talk to family” and “general medicine” (Pattern 3, N = 136); “talk to family” and “specialty medicine” (Pattern 6, N = 81); and “specialty medicine” and “deflect” (Pattern 5, N = 173). These results reflect variations among the individuals who recommended professional consultation.

Figure 2.

Values of Cubic Clustering Criterion by Number of Clusters Specified (2a) and Pseudo F Statistic by Number of Clusters Specified (2b), Japanese SGC-MHS

3.3. Factors Associated with Patterns of Mental Health Care Recommendations

Table 2 shows the results of the multinomial logit analysis of patterns of care using one contrast Pattern 1 (“deflect only”) as a reference category. The results reveal that endorsed causes and consequences of the vignette condition was associated with different toolbox patterns. For example, people who evaluated the condition as more serious were more likely to recommend consulting others (Patterns 2–6), especially the combined strategies of consulting family and professionals than “deflect only.” Those who thought that “the way the person was raised” or “bad luck” were the cause of the condition were less likely to recommend the strategy of talking to general professionals than “deflect only.” People who considered brain disease to be the cause were more likely to recommend medicalized strategies that included “general/specialty medicine.” Those who endorsed stress as the underlying cause were more likely to suggest only talking to some unspecified person than “deflect only.” Finally, respondents who considered genetic transmission as the cause were more likely to suggest other strategies (e.g., “talk to some unspecified person”, “specialty medicine + deflect”), relative to “deflect only.”

Table 2.

Multinomial Logit Analysis of Social Correlates of the Recommended Strategy of Mental Health Care, Japanese SGC-MHS (N=748)

	vs. (1) Deflect Only
	(2)	(3)	(4)	(5)	(6)
	General Medicine Only	Family + General Medicine	Unspecified Only	Specialty Medicine + Deflect	Family + Specialty Medicine
Causes and Consequences
Seriousness	0.64 (0.24)**	1.33 (0.27)**	0.46 (0.22)*	0.95 (0.24)***	1.50 (0.38)***
Bad Character	−0.35 (0.19)	−0.35 (0.18)	−0.12 (0.17)	−0.16 (0–17)	−0.28 (0.23)
Brain Disease	0.34 (0.16)*	0.20 (0.15)	−0.03 (0.15)	0.21 (0.15)	0.48 (0.19)*
Way Raised	−0.72 (0.18)***	−0.22 (0.18)	−0.17 (0.17)	−0.22 (0–17)	0.11 (0.23)
Stress	0.09 (0.26)	0.26 (0.25)	0.54 (0.24)*	0.35 (0.25)	0.01 (0.30)
Genetic Transmission	0.17 (0.20)	0.27 (0.19)	0.42 (0.18)*	0.42 (0.18)*	0.11 (0.23)
God’s Will	0.05 (0.25)	0.01 (0.23)	−0.08 (0.22)	0.26 (0.20)	−0.05 (0.28)
Bad Luck	−0.34 (0.17)*	−0.29 (0.16)	−0.06 (0.15)	−0.22 (0.15)	−0.06 (0.20)
Considers Violent Toward Others	−0.03 (0.21)	−0.29 (0.20)	−0.11 (0.19)	−0.22 (0.19)	−0.34 (0.26)
Considers Violent Toward Self	0.20 (0.22)	0.11 (0.20)	0.15 (0.20)	0.33 (0.20)	0.32 (0–27)
Vignette Characteristics
Schizophrenia	0.26 (0.29)	0.88 (0.28)**	1.31 (0.28)***	0.44 (0.27)	0.88 (0.36)*
Female	0.18 (0.27)	0.19 (0.25)	0.02 (0.25)	0.01 (0.24)	−0.35 (0.32)
Respondents’ Characteristics
Female	−0.37 (0.28)	0.48 (0.27)	0.76 (0.27)**	0.43 (0.26)	1.28 (0.37)***
Married	0.58 (0.33)	0.32 (0.31)	0.86 (0.30)**	0.49 (0.29)	1.21 (0.44)**
Age	−0.01 (0.01)	0.03 (0.01)*	−0.02 (0.01)	−0.01 (0.01)	0.02 (0.02)
Education (University)	−0.27 (0.35)	0.23 (0.33)	0.61 (0.31)	0.48 (0.30)	1.29 (0.39)**
Rural Residence	−0.23 (0.29)	−0.52 (0.29)	−0.45 (0.28)	−0.49 (0.27)	−0.85 (0.39)*
Economic Condition	0.05 (0.19)	0.00 (0.18)	0.01 (0.18)	0.06 (0–17)	−0.02 (0.23)
Constant	−1.90 (1.60)	−7.13 (1.65)***	−4.99 (1.48)**	−6.26 (1.53)***	−10.46 (2.24)***
LR chi-square	259.33***
df	90

^*p < .05;

^**p < .01;

^***p < .001

Besides the evaluation of conditions, vignette characteristics were also associated with care suggestions. Compared to those who received the depression vignette, those who read the schizophrenia one were more likely to recommend either “only talk to some unspecified person” or combined strategies of “family + general/specialty medicine,” relative to “deflect only.”

Finally, some respondents’ sociodemographic characteristics were significant. Females and married respondents were more likely to recommend the combined strategy of “family + specialty medicine” or “only talk to some unspecified person,” relative to “deflect only,” compared to counterparts (male, unmarried). University graduates were also more likely to recommend the combined strategy of “family + specialty medicine,” while those who lived in rural areas were less likely to do so (compared to those with less education, the urban residents). Older respondents were more likely to suggest cultural toolboxes that combined strategy of “family + general medicine” relative to “deflect only” than younger respondents.

While Table 2 is limited to a specific contrast, Figure 4 graphically summarizes the effects across all contrasts for those variables that had at least one statistically significant contrast. The numbers in the Figure represent the pattern numbers in Table 2. The distance between the numbers represents the magnitude of the effects of each variable. The lines between the numbers represent the lack of statistical significance (Long, 1997). Therefore, attention should be paid to both the distance and the absence of lines between the numbers to interpret variable effects. For example, in the first row (serious) in Figure 4, the number 1 (Pattern 1) is farther apart and disconnected from all the other numbers, indicating that the recommendation patterns were significantly affected by the respondents’ evaluation of seriousness. More specifically, perceived severity significantly increased the odds that respondents would recommend all the other strategies (Patterns 2 through 6) relative to “deflect only” (Pattern 1). However, among Patterns 2–6, the numbers 2 and 4 are apart and disconnected from the numbers 3 and 6, suggesting that the evaluation of seriousness significantly increases the odds respondents would recommend combined strategies of family and medicine (Patterns 3 and 6), relative to a single strategy (Patterns 2 and 4). Examining the other rows, we find that Pattern 2 (“general medicine only”) was differentiated from all the other recommendations when respondents attributed the person was raised as the cause of the condition. This attribution decreased the odds that respondents would suggest General Medicine alone relative to any other strategy.

Figure 4.

Graphic Representations of All Contrasts for Selected Variables in Multinomial Logit Analyses

Notes: Numbers represent the following patterns of mental health care strategies:

1 = Deflect only, 2 = General medicine only, 3 = Family + General medicine, 4 = Unspecified person only, 5 = Specialty medicine + Deflect, 6 = Family + Specialty Medicine

The following marks represent the type of discrete-change coefficient:

UnStd. Coef: Coefficients for a change of one unit, Std. Coef: Coefficients for a change of one standard deviation, 0/1: Changes from 0 to 1

The absence of lines between numbers indicates statistical significance at the .05 level. Only the horizontal distances are interpretable.

The results reveal that receiving the schizophrenia vignette increased the odds that respondents would suggest the strategy of talking to “some unspecified person only” (Pattern 4), relative to the other strategies (Patterns 1, 2, and 5) that include medical professionals. Consistent with the prior knowledge (Yamawaki et al., 2011), being female also increased the odds of suggesting combined strategies of “family and specialty/general medicine” (Patterns 6 &3) relative to a single strategy of “general medicine” (Pattern 2). Similarly, university education and marital status increased the odds of recommending combined strategy of “family + specialty medicine” (Pattern 6), relative to a single strategy of “Deflect only” (Pattern 1).

4. Discussion and Conclusions

Drawing on the concept of “cultural mapping” derived from the Network Episode Model, we explore the unprompted suggestions of a nationally representative sample of Japanese residents regarding mental health problems. Offering respondents an unlabeled scenario describing individuals with depression or schizophrenia, we find six discreet cultural toolkits from which individuals draw. Specifically, despite meeting the psychiatric criteria for mental illness, one clear pattern to “deflect” the issue was suggested by nearly one-quarter of Japanese respondents, and another 18% suggested only lay consultation. Two cultural toolbox patterns suggested general medical care (one singleton and one combined with family) while another two (both more complex patterns) suggested specialty care. Curiously, no respondents combined both general and specialty care as a culturally normative pattern.

Given that more than 40% of respondents suggested combined strategies that comprised professional care and other options, taking the concept of cultural toolboxes and the research strategy of cultural mapping into research designs suggest future directions. Knowing how those strategies are put into effect (e.g., the order of action and whether it results in the use of medical care) when people face mental health problems can provide a further understanding of the complexities of the public response to the onset of mental health problems. Further, why more individuals did not include care options from traditional medical systems needs to be pursued. One possibility is that some traditional medicine (e.g., kampo) has been integrated in modern medicine in Japan (Motoo, 2011), so that present population may be used to using it through consultation with medical doctors as a part of their treatment. If the public perceives traditional medicine as a part of modern medicine and does not consider it separate or special, a specific response of using “traditional medicine” would be less likely to be evoked. Moreover, the public conception of the usefulness of other traditional medicine (e.g., acupuncture) may be narrowly limited to specific illnesses (e.g., muscular or joint problems) (Lock, 1980: 164), so that they may not be evoked as an option to take care of the mental health problems described in the vignette.

On the other hand, some behaviors of “deflect” (e.g., get involved in other activities, go out or travel, start exercising) and “self care” (e.g., reflect on oneself, rest well) may be understood as attempts to restore the balance of the body based on holistic views of the human body. According to Lock (1980: 217), “holism” in traditional East Asian medicine has two meanings: “that of considering all the parts of the body to be interconnected and mutually affecting each other” and “that of viewing the human body as being in constant interaction with, and affected by, the environment, both social and physical.” Other behaviors of “deflect,” such as try spending time with others (not for consultation), appear to reflect Asian ways of seeking “implicit social support” to cope with stress (Kim et al., 2008). This behavior generally does not involve disclosing their problems or asking for advice because they are concerned that explicit support-seeking acts may negatively affect their harmonious relationships with others. However, such behavior may not be met with advice to use formal treatment because the nature of the problem (mental illness) is less likely to be recognized without disclosing or discussing what they are really going through.

Using the open-ended research strategy in the survey format does offer a deeper look at why utilization rates are low. The “endorsement” strategy (i.e., asking respondents to agree or not agree that the individual should use a provided list of sources of care) artificially inflates public responses and mistakenly suggests a gap between high levels of predispositions to use reported in surveys and low utilization behavior seen in community studies of people actually reporting mental health problems, including those in Japan. “Talking to family” was also frequently suggested, but not as much as we expected, given the research on other countries cited earlier. However, nearly one-quarter of the respondents suggested talking to “some unspecified person” (e.g., someone trustworthy). This suggestion may have been made because the person in the vignette showed some relational problems with his/her family or friends (e.g., pulling away from family and not feeling like talking, or thinking that people around him/her were making disapproving comments), one of the typical symptoms in cases of depression or schizophrenia. Some respondents may have considered it appropriate to avoid direct confrontations with family or close others because it may worsen their relationships (avoidance of seeking explicit social support); thus, they suggested talking to some unspecified person. The question is whether the person could find someone that he/she could talk to, other than family, under the circumstances; if not, this strategy is likely to be of no use when individuals face mental health problems in the real world. This topic may be suitable for consideration in further research in Japan because, compared with major advocacy strategies in the West that focus on tag lines such as “Let’s Talk Mental Health” or “End Stigma by Starting the Conversation,” Japanese efforts may have continued to focus on mental health literacy and stigma associated with formal labels. Notably, these strategies have been shown to no longer be effective in the West (Pescosolido et al., 2010; Angermeyer et al., 1999).

Of course, cultural toolboxes varied, at least to some degree, for groups in difference sociocultural locations. For example, women and university graduates were more likely to piece together “family + specialty medicine,” and older respondents were more likely to combine “family + general medicine,” relative to “deflect only.” Their counterparts, then,—those who are male, unmarried, less educated, or younger—tend to have virtually empty toolboxes relevant to mental health problems. Cleavages in mental health care recommendations continue to reflect a cultural divide based on social location in contemporary Japan, despite the social network discussions on diminishing power of sociodemographic cleavages (Pescosolido and Rubin, 2000). Notably, the gender division of labor remains prevalent in Japan compared with Western countries (Fuwa, 2004). As such, gender continues to be a good proxy for cultural characteristics in this context because the structure of networks and the information or advice those networks provide are likely to depend on the gender of individuals. Indeed, studies have illustrated that male and less educated individuals continue to engage in significantly less diverse social networks than their counterparts in Japan do (see Shishido, 2012 for review).

Japanese residents’ cultural maps have lines drawn based on their evaluations of causes and consequences of conditions, something observed across a number of studies, including those in Japan (Haug et al., 1991). In particular, severity matters. Seeing the problem as serious increased the odds respondents would suggest strategies that include talking to someone, especially the combined strategy of talking to both family and professionals in “general/specialty medicine,” relative to “deflect only.” Haug and their colleagues (1991) provide an explanation for Japan that may be more generally applicable. Japanese tend to rely on self-care and avoid using medical doctors except for serious conditions partly because becoming patients involves losing autonomy in the doctor-patient relationships. While the doctor-patient intersection may be more hierarchical in Japan compared with its Western counterparts, issues of power may be relative to expectations everywhere.

In addition, stigmatizing attributions result in cultural toolboxes that deny the relevance of medical resources (e.g., “the way one was raised”; “bad luck”). The results are consistent with the prior study in the U.S. that showed that attributing the problem to chemical imbalance in the brain is positively associated with endorsing professional care (Olafsdottir and Pescosolido, 2009). Yet, given this, the findings on mental health problem type are curious. Specifically, those who received the schizophrenia (vs. depression) vignette were more likely to suggest the strategy “only talk to some unspecified person” relative to various other strategies (“deflect only,” “general medicine only,” “specialty medicine + deflect”) compared with those who read the vignette about depression. Only the option to “talk to someone (e.g., someone trustworthy, someone close)” was likely to be the culturally appropriate strategy in Japan. This does not coincide with findings in the West where schizophrenia repeatedly demands a more serious, often medical response. We return to the higher levels of stigma documents in Eastern countries (Krendl and Pescosolido, 2019). If the use of medical care depends on not only the size of informal networks, but also their cultural contexts (Pescosolido et al., 1998), the prejudice and discrimination of others may play a role here. Further, research in Japan has documented that people are less knowledgeable about schizophrenia than depression (Yamazaki et al., 2012), making the cultural toolboxes relatively empty.

This study has limitations, some of which may be dealt with in future research. First, some time has passed since our survey was conducted in 2006; therefore, some positive changes may have occurred in the public recognition of mental illness. Since 2015, as a political effort to prevent karoshi (death caused by overwork), workplaces with 50 or more workers are obligated to administer job stress questionnaires to their regular employees. Further, since 2018 people with mental disabilities have been added to the basis calculation of the employment quota for persons with disabilities. These efforts, along with the continuing effort at the local level (Nakamura, 2013), may have facilitated public understanding of mental health and improved care options to some extent. Second, our analyses are limited to the most common of the serious mental illnesses. While these represent an obvious first start, other mental health problems may present even more challenges to the public’s construction of problem recognition and appropriate cultural responses.

Medicine too often considers effectiveness by narrowly focusing on professional care and utilization as a function of appropriate scientific understanding on the part of the public. Yet these findings suggest that formal providers may or may not have “partners” out in the community. Future research should examine the effectiveness of various mental health care options beyond the single use of formal care because those evidence could help us to enrich our tools to take care of mental health problems in clinical settings.

Our findings reveal substantial variations in the type and patterns of mental health care recommendations among Japanese residents. Those variations are likely to be culturally and socially constrained. They reveal, at least in part, the underlying cultural context of low utilization of formal care in cases of mental health issues in Japan. Comparing the patterns of mental health care recommendations at the onset of mental health problems across cultures and investigating their sociocultural correlates will enhance our broader understanding of the underlying mechanisms affecting the low utilization for mental illness.

Research Highlights.

• Explores culturally acceptable suggestions for mental health care in Japan.

• Finds substantial variations in the patterns of mental health care recommendations.

• Suggests those variations are likely to be culturally and socially constrained.

APPENDIX

Vignette Wording

Depressive Disorder

NAME is MAN/WOMAN. For the last several weeks NAME has been feeling really down. S/he wakes up in the morning with a sad mood and heavy feeling that stick with her/him all day long. S/he isn’t enjoying things the way s/he normally would. In fact nothing seems to give him/her pleasure. Even when good things happen, they don’t seem to make NAME happy. The smallest tasks are difficult to accomplish. S/he finds it hard to concentrate on anything. S/he feels out of energy, out of steam and cannot do things s/he usually does. And even though NAME feels tired, when night comes s/he can’t go to sleep. NAME feels pretty worthless, very discouraged, and guilty. NAME’s family has noticed that s/he has lost appetite and weight. S/he has pulled away from them and just doesn’t feel like talking.

Schizophrenia

NAME is MAN/WOMAN. Up until a year ago, life was pretty okay for NAME. But then, things started to change. He thought that people around him/her were making disapproving comments, and talking behind his/her back. NAME was convinced that people were spying on him/her and that they could hear what she/he was thinking. NAME lost his/her drive to participate in his/her usual work and family activities and retreated to his/her home, eventually spending most of his/her time on his/her own. NAME became so preoccupied with what s/he was thinking that s/he skipped meals and stopped bathing regularly. At night, when everyone else was sleeping, s/he was walking back and forth at home. NAME was hearing voices even though no one else was around. These voices told him/her what to do and what to think. S/he has been living this way for six months.