A Cross-cultural Study of Recovery for People with Psychiatric Disabilities Between U.S. and Japan
- First Online:
- Cite this article as:
- Fukui, S., Shimizu, Y. & Rapp, C.A. Community Ment Health J (2012) 48: 804. doi:10.1007/s10597-012-9513-2
- 370 Views
The concept of recovery has been expanding overseas with remarkable speed. The Recovery Assessment Scale (RAS) is one of the measures widely used to capture self-perceptions of a sense of recovery for people with psychiatric disabilities. The current study tested measurement invariance of RAS between the US and Japanese samples for people with psychiatric disabilities, which is a precursor of further cross-cultural comparisons without any contamination of systematic cultural bias. A multiple-group confirmatory factor analysis was applied to US (N = 446) and Japanese (N = 214) participants for testing configural, loading, and intercept invariance. The results revealed that RAS items equally captured their associated recovery domains between American and Japanese participants. For two domains, “personal confidence and hope” and “reliance on others,” the two groups systematically responded with different patterns. Different cultural environments may have additive influences toward people’s response patterns to their recovery across countries.
KeywordsRecoveryMental illnessCross-cultural studyRecovery ScaleMeasurement invarianceJapaneseConfirmatory factor analysis
Recovery speaks primarily to the person’s rights for social inclusion and self-determination, irrespective of the nature or severity of his or her psychiatric condition. In this sense, recovery refers to the rights to access and to join in those elements of community life the person chooses, and to be in control of his or her own life and destiny, while remaining disabled (p. 569).
Particularly influential research included the International Pilot Study of Schizophrenia (Birchwood et al. 1992; Jablonsky 1989; Warner 1992) and the results of seven longitudinal studies conducted in five countries showing high rates of improvement (Bleuler 1978; Ciompi and Muller 1976; DeSisto et al. 1995a, b; Harding et al. 1987; Huber et al. 1975; Ogawa et al. 1987; Tsuang et al. 1979). This body of research conspired to demand reconsideration that schizophrenia and other major mental illnesses may neither be chronic nor progressive. Rather, recovery should be the central goal of mental health services fostering “a journey of healing and transformation enabling a person with a mental health problem to live a meaningful life in a community of his or her choice while striving to achieve his or her full potential” (U.S. Department of Health and Human Services 2005).
The concept of recovery, which was originally developed in the US, has been expanding overseas with remarkable speed (Slade et al. 2008). One example is Japanese mental health rehabilitation which explicitly introduced the concept in 2003 (Tanaka 2010). It is possible that people experience common or shared phenomena, such as hopefulness and self-esteem under the life restrictions of mental illness regardless of the cultural context to which they belong. Japanese professionals have found common applications in the concept and have introduced recovery-oriented programs, including the Wellness Recovery Action Plan (Copeland 2002), Illness Management and Recovery (Corrigan et al. 2008), Assertive Community Treatment (Corrigan et al. 2008), and Strength Model Case Management (Rapp and Goscha 2006), into their cultural context. Well accepted definitions of recovery in Japan are consistent with the definition used in the US (President’s New Freedom Commission on Mental Health 2003).
“Without reliable measurement, science cannot advance” (Liberman and Kopelowicz 2005). The Recovery Assessment Scale (RAS) is one of the measures widely used when capturing self-perceptions of a sense of recovery for consumers with psychiatric disabilities. It has been tested for validity and reliability targeting the US (Corrigan et al. 2004), Australian (McNaught et al. 2007), and Japanese populations (Chiba et al. 2010), so it is useful for cross-cultural comparisons of recovery. RAS consists of five domains which include; personal confidence and hope, willingness to ask for help, goal and success orientation, reliance on others, and no domination by symptoms (Corrigan et al. 2004). Chiba et al. (2010) found similar factor structures of RAS when applying it to the Japanese population. These studies have found the evidence of similar conceptual components of recovery within RAS across countries, but there is still lack of evidence for the usage of cross-cultural comparisons. This is mainly because the analyses were conducted individually by separate countries. Currently, a multiple-group confirmatory factor analysis method is recommended to test measurement comparability across countries because it allows for simultaneous estimations and factor models fitting samples across countries (Gregorich 2006). The goal of this study was to test the measurement invariance of RAS between the US and Japanese samples for people with psychiatric disabilities.
Participants were recruited from Consumer-Run Organizations (CRO) and at a statewide Recovery Conference for consumers held in Kansas. The survey participants were adult individuals, over the age of 18, and who had a self-reported diagnosis of severe and persistent mental illness. Twenty CROs in Kansas were invited to participate in this study and sixteen agreed to recruit their members for the survey through word of mouth, flyers, and newsletters. The data was collected through internet surveys from participants during their participation at a local CRO (January 15–February 15, 2010). One hundred fifty-one individuals answered the survey. In addition, participants at the Recovery Conference were handed self-administered questionnaires at the conference registration and they returned the questionnaires to a KU research booth (June 16–17, 2010). Four hundred forty-four people, with a combination of both consumers and staff, participated in the conference and 325 consumers returned the questionnaires. Socio-demographic characteristics of participants were obtained through self-report. A $5 incentive for each participant was supplied. Written informed consent approved by the University of Kansas Institutional Review Board was obtained from all participants.
All data was collected in Osaka City (metropolitan area). Twenty-four city districts and one branch which offered weekly group activities (social and recreational) and one mental health day care (psychosocial program) agreed to participate in the survey. Researchers visited the sites, delivered the self-administered questionnaires, and collected them between December 17, 2009 and March 23, 2010. The survey participants were adult individuals, over the age of 18, and who had a self-reported diagnosis of severe and persistent mental illness. A total of 226 consumers answered the questionnaires. Socio-demographic characteristics of participants were obtained through self-report. A $5 value token was provided for each participant. Written informed consent approved by Osaka City University Ethic Committee was obtained from all participants.
The Recovery Assessment Scale (RAS)
The Recovery Assessment Scale (RAS) (Corrigan et al. 2004) was developed in the US and its validity and reliability were tested targeting US, Australian, and Japanese populations using the 24-item version (Chiba et al. 2010; Corrigan et al. 2004; McNaught et al. 2007). Forty-one RAS items were generated through consumer narratives and item reviews, which captured self-perceptions of a sense of recovery for people with psychiatric disabilities (Corrigan et al. 2004). Five factors with 24 items were identified by exploratory and confirmatory factor analyses, which included personal confidence and hope, willingness to ask for help, goal and success orientation, reliance on others, and no domination by symptoms. These tap distinct domains and are correlated with psychosocial (including empowerment, hope, and quality of life) and symptom variables. The items use a five-point scale ranging from “Strongly Disagree (=1)” to “Strongly Agree (=5).” The Japanese instrument used in this study was validated in translation by following a blind back-translation (Brislin et al. 1973). Total Cronbach’s alpha of this scale for US was 0.937 and for Japanese was 0.910 in this study.
After removing the incomplete questionnaires (e.g., questionnaires missing all RAS items), 446 (151 from CROs and 295 from the Recovery Conference) for the US sample and 214 out of 226 for the Japanese sample were included in the analyses. Compatibility between the two datasets from CROs and the Recovery Conference was confirmed by demographics (gender, age, and diagnosis) and all 24 RAS items, which had no statistically significant differences between the samples (p > .05), so they were integrated as a US sample.
A multiple-group confirmatory factor analysis (CFA; Brown 2006) was conducted to assess measurement invariance of RAS between US and Japanese samples (Cheung and Rensvold 2002; Little 1997). Establishing measurement invariance assures that the component domains underlying the recovery concept measured by RAS are the same across two countries. The invariance testing was conducted by the following three steps: (1) configural invariance indicates that item clusters are identical (each common factor is associated with identical sets of measurement items) across countries, (2) loading (i.e., metric or weak factorial) invariance indicates that the common factors have the same meaning (corresponding factor loadings are equal) across countries, and (3) intercept (i.e., scale or strong factorial) invariance indicates no different additive influences of cultural norms which may systematically act to raise or lower item response patterns across countries (comparability of group means) (Gregorich 2006). The invariance for specific items can be relaxed to obtain partial invariance (Borsboom 2006; Byrne et al. 1989; Cheung and Rensvold 2002; Gregorich 2006; Schmitt and Kuljanin 2008; Steenkamp and Baumgartner 1998; Meredith and Teresi 2006; Wicherts and Dolan 2010).
The appropriateness of comparability for theory-driven assumptions about the latent structure of a set of RAS across countries can be assessed by the Chi-square difference test between nested models (the Likelihood Ratio Test) (Bollen 1989) as well as a modeling rationale using practical fit indices, specifically the Comparative Fit Index (CFI), Tucker Lewis Index (TLI), Root Mean Square Error of Approximation (RMSEA), Standardized Root Mean Square Residual (SRMR) (Steenkamp and Baumgartner 1998). The modeling rationale is favorable especially when the analysis involves large sample sizes and numerous constrained parameters because the Chi-square statistic is well-known to be overly sensitive to sample size (Cheung and Rensvold 2002; Little 1997; Little and Slegers 2005; MacCallum et al. 2006; Vandenberg and Lance 2000).
We evaluated invariance based on whether: (1) change in CFI is less than 0.01 and (2) the RMSEA falls within 90 % confidence interval of comparison models (Cheung and Rensvold 2002; Little 1997; MacCallum et al. 2006). We used a fixed-factor method of scaling to identify the model (Lee et al. 2010). Mplus version 6 with full information maximum likelihood estimation was used for the analyses.
43.8 (SD: 10.87, RG: 20–73)
45.7 (SD:11.73, RG: 19–74)
t (434.839) = 1.995, p = .047
Diagnosis age (years)
28.65 (SD: 9.84, RG: 10–55)
25.60 (SD: 13.60, RG: 0-62)
t (461.356) = –3.012, p = .003
χ2 (1, N = 649) = 14.912, p < .001
No high school or GED
No high school/GED versus more education
High school diploma or GED
1 Year of college
χ2 (1, N = 645) = 12.502, p < .001
Two or more years of college
Master’s, Ph.D, M.D.or J.D.
Schizophrenia/schizoaffective disorder versus other diagnoses
Bipolar disorder/manic depression
Major depressive disorder
χ2 (1, N = 616) = 112.022, p < .001
Single/never married versus married/committed relationship
χ2 (1, N = 654) = 90.725, p < .001
Live with (multiple responses)
Live alone versus with somebody
χ2 (1, N = 658) = 24.983, p < .001
Friends or roommate
Partner or spouse
In an apartment/home versus all others
In an apartment/home
In a boarding home
χ2 (1, N = 640) = 0.929, p = .335
In a group home or halfway house
In an institution
Hospitalization in the past year
χ2 (1, N = 644) = 1.176, p = .278
Means and standard deviations of the Recovery Assessment Scale
Recovery Assessment Scale
1) Strongly disagree, 2) Disagree, 3) Not sure, 4) Agree, 5) Strongly agree
Japan (N = 214)
USA (N = 446)
Domain 1: Personal confidence and hope
Fear doesn’t stop me from living the way I want to
I can handle what happens in my life
I like myself
If people really knew me, they would like me
I have an idea of who I want to become
Something good will eventually happen
I’m hopeful about the future
I continue to have new interests
I can handle stress
Domain 2: Willingness to ask for help
I know when to ask for help
I am willing to ask for help
I ask for help when I need it
Domain 3: Goal and success orientation
I have a desire to succeed
I have my own plan for how to stay or become well
I have goals in life that I want to reach.
I believe I can meet my current personal goals
I have a purpose in life
Domain 4: Reliance on others
Even when I don’t care about myself, other people do
I have people I can count on
Even when I don’t believe in myself, other people do
It is important to have a variety of friends
Domain 5: No domination by symptoms
Coping with my mental illness is no longer the main focus of my life
My symptoms interfere less and less with my life
My symptoms seem to be a problem for shorter periods of time each time they occur
Multiple-Group Confirmatory Factor Analysis (MG-CFA)
Relevant fit information for measurement invariant models
90 % CI
Partial intercept invariance
This is the first study testing measurement invariance for RAS cross-culturally. The study results are congruent with Chiba et al.’s study (2010). US and Japanese participants responded to similar components of recovery within RAS. We further examined the measurement invariance using a more rigorous method: multiple group confirmatory factor analysis.
Observed means consistently showed lower scores of RAS for Japanese than US participants (Table 2). This may reflect the different sample demographic characteristics. For example, the majority of Japanese participants reported either a schizophrenia or schizoaffective disorder diagnosis, while more of the American participants reported either a bipolar disorder or major depressive disorder diagnosis. In addition, US participants were actively involved in recovery related activities, including CROs and a Recovery Conference. Some of the US participants were leading recovery-oriented groups. On the other hand, Japanese participants were recruited at primary care settings. Most of the participants had just started to be exposed to the community resources that would motivate them to move from their isolated situations. In this context, they were still premature when cultivating hope and recovery.
In addition, observed responses (manifest variable means and standard deviations) can be contaminated by several biases and errors, including cultural bias, differential interpretation of items, different data collection administration conditions, and translation errors (Borsboom 2006; Little and Slegers 2005). For example, the item response style might differ between countries. One group might value decisiveness or certainty while another group might avoid strong statements because they value humility (Gregorich 2006). Cheung and Rensvold (1999) picked one item example regarding self-esteem, such as “I am a person of worth, at least as good as other people.” The question might be an indication of self-esteem in Western culture, but it may mean that agreement with the question can reflect “a grandiose” and “socially unacceptable sense of self-importance” in Asian culture (Cheung and Rensvold 1999). If the indicators are dependent on cultural influences, observed mean comparison is biased, so the relevant constructs measured using an instrument need to be cross-culturally invariant to understand true perceptions without any contamination of systematic bias, which can be tested by measurement invariance testing (Cheung and Rensvold 1999; Steenkamp and Baumgartner 1998).
The results showed configural and loading invariance as well as partial intercept invariance. This indicates that RAS captured the same general and basic components of recovery and had the same conceptual meaning for both Japanese and American participants. In addition, the indicators’ properties, except the non-invariant seven items, were transportable and generalizable between the two countries (Vandenberg and Lance 2000). Most of the non-invariant items were found to be associated in two recovery domains: (1) personal confidence and hope and (2) reliance on others. In summary, while all RAS items equally captured their associated recovery domains between American and Japanese participants, the participants systematically responded to some of the items associated with these two domains with different patterns.
Both American and Japanese participants identified hope and personal confidence as components of recovery. However, the emphasis on hope and personal confidence seems slightly different between Western and Eastern cultures in the literature. The Eastern culture emphasizes hope more in the context of self-cultivation (including refinement of the person as a whole as opposed to sub-personal attributes, courage, and fortitude) while hope is expressed more in the context of mastery over events (including personal control, self-efficacy, optimism, and problem solving skills) in Western culture (Averill and Sundarajan 2005). Japanese may also be more self-effacing about future events (Chang et al. 2001).
Finally, reliance on others is important for both Japanese and Americans in the context of recovery, but they may respond differently. Eastern cultures have been considered collectivist, given their focus on fostering a view of self as being fundamentally interrelated with significant others (Markus and Kitayama 1991; Muramoto 2003). For example, Japanese may maintain their positive self-regard in a reciprocal support relationship with others while Americans might view their self-enhancement more autonomously (Muramoto 2003) and find importance in associating with others in a support system or in using coping resources. Japanese, who are considered as having “an interdependent construal of self,” recognize their well-being when associating with feelings of belongingness instead of with feelings of individual competence among others (Hein and Lehman 1995).
These factors might cause different additive influences of cultural norms which could systematically act to raise or lower item response patterns in the two recovery domains across countries. Qualitative research conducted in Italy, Norway, Sweden, and the US argued there were no major social or cultural differences in the process of recovery (Davidson et al. 2005). However, they found that a recovery-enhanced environment might differ in different cultural settings. While our comparison was made between American and Japanese samples, our research is congruent with their insights that different cultural environments may have additive influences toward people’s response patterns to their recovery across countries.
The preliminary research results have limitations when identifying factors impacting on the cultural differences. Limitations include: (1) convenience samples were used; (2) RAS was administrated using different data collection methods; (3) the participants’ demographic characteristics were different; and (4) the level of exposure to the recovery concept was different between countries. Study participants in the US were recruited at CROs and a statewide Recovery Conference, while Japanese participants were recruited at primary care settings, which could make the samples different in terms of consumer initiatives toward recovery. These differences may have confounded the results and interpretations should be made with caution. It is uncertain how the differences in the person’s recovery stage, diagnosis, age, gender, etc., in addition to cultural aspects affect the conceptualization of recovery. Assurance of comparability is challenging in cross-cultural studies. Despite the limitations, however, the globalization of the recovery movement encourages people to share recovery as the central goal of mental health services. Recovery movement has brought a new culture into the traditional mental health system and confronts our cultural sensitivities. Exposure to different cultures gives us insight and awareness about our own cultural context. We need continued effort to promote recovery in culturally specific contexts within and between countries.
The study showed that RAS captured the same general and basic components of recovery and had the same conceptual meaning for both Japanese and American participants. For two domains, “personal confidence and hope” and “reliance on others,” the two groups systematically responded with different patterns. As Davidson et al. (2005) argued, although there may not be major social or cultural differences in the process of recovery in individuals, different cultural environments may have additive influences toward people’s response patterns to their recovery across countries. These cultural influences could be explored by identifying factors affecting the recovery process for people with psychiatric disabilities in their cultural context. Recovery has become an international movement and the concept has been expanding overseas with remarkable speed. Given the current globalization movements in recovery, future studies will be needed to further interpret the cultural differences in recovery.
This study was supported by the Japan Society for the Promotion of Science: Grant-in-Aid for Scientific Research (C: 21530593) and developed by the University of Kansas School of Social Welfare Office of Mental Health Research and Training through a contract with the Kansas Department of Social and Rehabilitation Services.