Culturally sensitive and universal measure of resilience for Japanese populations: Tachikawa Resilience Scale in comparison with Resilience Scale 14-item version
Correspondence: Daisuke Nishi, MD, PhD, National Institute of Mental Health, National Center of Neurology and Psychiatry, 4-1-1, Ogawahigashi-cho, Kodaira, Tokyo 187-8553, Japan. Email: firstname.lastname@example.org
Although scales specific to resilience are available and widely used, qualities of resilience could be culturally sensitive. This study aimed to develop a concise scale of resilience for Japanese populations, and compare its validity to that of the Resilience Scale 14-item version (RS-14), one of the most widely used scales for measuring resilience.
The Tachikawa Resilience Scale (TRS) was developed on the basis of data obtained from unstructured interviews with Japanese motor vehicle accident survivors without psychiatric disorder. The reliability and validity of the TRS and RS-14 were then examined in cross-sectional studies performed with 523 company workers and 140 psychiatric outpatients.
The TRS and RS-14 were negatively correlated with depressive symptoms in company workers and psychiatric outpatients and with anxiety in psychiatric outpatients, and were positively correlated with social support in company workers. Internal consistency and test–retest reliability of the TRS were high. Construct validity of the TRS was equivalent to that of the RS-14 in company workers, and higher than that of the RS-14 in psychiatric outpatients.
The reliability and validity of the TRS and RS-14 in Japanese company workers and patients with psychiatric disorders were acceptable. The validity of the TRS was equivalent to or better than that of the RS-14. Although the TRS cannot be regarded as an established scale due to a lack of theoretical rationale, the results of this study suggest that scales measuring resilience that cover cultural aspects might be more relevant in given populations.
RESILIENCE HAS GAINED considerable attention as one of the key concepts for prevention of and recovery from mental disorders. Although resilience is variously defined, most of the definitions mention two common factors: the experience of adversity or stress, and the achievement of positive outcomes. Generally, it has been viewed as a stress-coping ability in the face of adversity.
One of the most popular scales for measuring resilience is the Resilience Scale (RS) and its 14-item version (RS-14).[2, 3] The RS was developed based on a qualitative study of 24 older women in the USA who had experienced a recent loss, such as loss of a spouse, health, or employment, and adapted successfully.[2-7] Its conceptual framework is composed of five characteristics. Items are categorized by (i) self-reliance (e.g. ‘I feel I can handle many things at a time’); (ii) meaning (e.g. ‘I feel proud that I have accomplished things in life’); (iii) equanimity (e.g. ‘I usually take things in stride’); (iv) perseverance (e.g. ‘I am determined’); and (v) existential aloneness (e.g. ‘My belief in myself gets me through hard times’). We previously examined the reliability and validity of the Japanese version of the RS and RS-14, and the results suggested that the validity might be relatively low compared with the original English version. The reasons for this may not only include the fact that participants of the study were nursing and university students, but also the nature of Japanese culture, as it is known that the Japanese have a general tendency to suppress the expression of positive affects.[9, 10]
Moreover, the qualities of resilience could be culturally sensitive. For example, many Japanese might attach a high value to the notion behind the well-known Morita therapy, which leads patients toward accepting anxiety as natural (arugamama) rather than toward optimism or positive thinking. Although one scale called the Adolescent Resilience Scale was developed in Japan, a recent review pointed out that unfortunately no clinical applications of that scale have yet been reported.
Thus, the aim of this study was to develop a concise measure of resilience for the Japanese, and compare its validity to that of the RS-14. Because the scores of scales measuring resilience should be negatively correlated with depressive symptoms and anxiety and positively correlated with social support,[8, 14, 15] we examined how well the new scale and the RS-14 correlated with scales measuring those concepts.
We named the new measure the Tachikawa Resilience Scale (TRS). We previously conducted a cohort study of motor vehicle accident survivors, and in the present study we conducted a structured diagnostic interview and an unstructured interview with them 3 years after the accident. We regarded accident survivors who had no psychiatric disorder as resilient because despite experiencing a major accident, they achieved a positive outcome. Of 63 participants assessed by structured interview 3 years after their accident, 15 had a psychiatric disorder and 48 did not. Forty-two of these 48 participated in the unstructured interview. Of these 42, 28 had not developed a psychiatric disorder and 14 had recovered from it after their accident. The mean age of the 42 accident survivors was 40.5 (SD = 14.9) years and 10 were women. We used unstructured interviews, which allow participants to speak freely. We audiotaped the interview and transcriptions of audiotape were checked by psychiatrists and a psychologist. We drew items, which were thought to be helpful for prevention or recovery. We also referred a number of scales measuring resilience, such as RS and Connor–Davidson Resilience Scale, reviews on resilience,[15, 18] and what was thought to be Japanese-culturally sensitive, such as Morita therapy. Originally, 11 items were derived from the interview data and we conducted a preliminary study. The original aim of a preliminary study was to demonstrate the reliability and validity of the Japanese version of RS and RS-14, and we published the results elsewhere. In this study, collaterally, the preliminary TRS results were examined and one item was excluded because a preliminary study showed a factor loading <0.40. Thus, the TRS was developed as a 10-item measure. Item descriptions are shown in Table 1. Because items were derived from Japanese accident survivors, at least some items seem to be closely related to Japanese culture. For example, acceptance represented by the items such as ‘I accept things as they are when there are no alternatives’ and ‘I try not to worry about what is beyond my capabilities’ can be regarded as Japanese-culturally sensitive, because these items reflect the idea of Morita therapy (arugamama) to some extent. Although one of the factors of the original RS is also called ‘Acceptance of Self and Life’, the items of the factors, such as ‘I seldom wonder what the point of it all is’ and ‘My life has meaning’ in the RS are very different from the contents of the TRS.
Table 1. Demographic characteristics of workers in a company and psychiatric outpatients
|Age|| ||38.2 (9.0)|| ||47.9 (12.4)|
|Sex, women||102 (19.5)|| ||67 (47.9)|| |
|Married||315 (60.2)|| ||83 (59.3)|| |
College or university
|453 (86.6)|| ||77 (55.0)|| |
When completing the TRS, respondents were asked to circle the number best representing their level of agreement with each statement in the face of adversity or stressful life events. The instructions of TRS and its items clearly relate to experiences of adversity or stress, which coincides with the notion that a resilience scale should measure an individual's reactions to stressful life events. All items were scored on a 7-point scale from 1 (strongly disagree) to 7 (strongly agree), with a total score ranging from 10 to 70. Higher scores reflect higher resilience.
Participants and procedures
A cross-sectional study was carried out at three separate worksites of a large company located in an urban area between August and November 2010. The inclusion criteria were as follows: company worker, age 18 years or older, and capable of understanding and providing consent for study participation. The company's occupational health staff provided participants with a written explanation of the research, a consent form, and the self-reporting questionnaires. Workers who agreed to participate in this study provided consent by returning the consent form and questionnaires by postal mail. A subsample was retested 3 months after initial assessment.
The study protocol was approved by the institutional review boards of both the National Disaster Medical Center (NDMC) and the participating company.
A cross-sectional study was carried out at a hospital in Hiroshima prefecture between January and May 2011. The inclusion criteria were as follows: psychiatric outpatient with mood disorder or with neurotic, stress-related, or somatoform disorder diagnosed according to the ICD-10 classification of mental and behavioral disorders by psychiatrists in clinical settings; aged ≥18 years; and capable of understanding and providing consent for study participation as judged by the psychiatrist in charge of the outpatient. Doctors or nurses at the hospital provided the outpatients with a written explanation of the research, a consent form, and self-reporting questionnaires. Outpatients who agreed to participate in this study provided consent by returning the consent form and questionnaires by postal mail.
The study protocol was approved by the institutional review boards of both the NDMC and the participating hospital.
Information on age, sex, marital status, and education level was gathered alongside data obtained using the following measures.
Short version of the Resilience Scale
The 25-item RS measures the degree of individual resilience and a recent review determined this was one of the best instruments to study resilience in adolescent populations. The short version, the RS-14 which consists of 14 items, strongly correlates with the RS, and each item is rated on a 7-point Likert scale (range, 14–98). We previously developed the Japanese versions of the RS and RS-14, and the respective Cronbach's alpha coefficients were 0.90 and 0.88, test–retest correlation coefficients were 0.83 and 0.84, and correlation coefficients with depressive symptoms were −0.30 and −0.28. Both scales had psychometric properties with high degrees of internal consistency and test–retest reliability, but relatively low concurrent validity.
Depressive symptoms are well known to be negatively correlated with resilience.[8, 14] For company workers, the Center for Epidemiologic Studies Depression Scale (CES-D) was administered. The CES-D, one of the most widely used scales to assess depressive symptoms in the general population, measures the level of depressive symptoms in the past 1 week. Reliability and validity of the Japanese version have been verified. For psychiatric outpatients, the Beck Depression Inventory-II (BDI-II) was administered. It is one of the most widely used scales to assess depressive symptoms in patients with psychiatric disorders and measures the level of depressive symptoms in the past 2 weeks. Reliability and validity of the Japanese version have been verified.
Anxiety is also well known to be negatively correlated with resilience. For psychiatric outpatients, the State–Trait Anxiety Inventory (STAI) was used. The STAI is a questionnaire that measures anxiety in adults and it clearly differentiates between the temporary condition of ‘state anxiety’ and the more general and long-standing quality of ‘trait anxiety’. Reliability and validity of the Japanese version have been verified.
Resilient individuals are thought to have the ability to extract and enhance social support from others, and the Social Support Questionnaire (SSQ) was shown to be positively correlated with resilience. For company workers, the SSQ was administered to assess the perceived availability of and satisfaction with social support, namely, the existence or availability of people on whom they could rely.[27, 28] The short version of the SSQ consists of 12 items, six of which measure perceived number of social supports and the remainder of which measure satisfaction with social support. Reliability and validity of the Japanese version have been verified.
We calculated the mean and standard deviation for each item of the TRS. Cronbach's alpha coefficient and test–retest correlation of the TRS were determined to assess reliability.
Confirmatory factor analysis was conducted to examine the factorial validity of the TRS in company workers. Based on observations made on resilience in our previous study, we tested a unidimensional model. The goodness-of-fit indices the root mean square error of approximation (RMSEA), goodness-of-fit index (GFI), adjusted goodness-of-fit index (AGFI), and the Comparative Fit Index (CFI) were examined to indicate how well the model fit the data. Values of RMSEA up to 0.08, GFI > 0.90, AGFI > 0.85, and CFI > 0.90 were taken to indicate proper fit.
The correlation between the TRS and RS-14 was examined in relation to concurrent validity. Difference in the mean TRS score between workers and psychiatric outpatients and the correlation of the TRS with other measures, such as depressive symptoms, anxiety, and social support, were examined to assess construct validity. Multivariate regression analysis was used to examine the relation of the RS-14 and TRS with depressive symptoms and anxiety adjusted for age and sex. In addition, correlation of the RS-14 with other measures was also examined to assess its validity. All statistical analyses used two-tailed tests and P-values less than 0.05 were considered to indicate significant differences. All data analyses were performed using the statistical software package spss, version 19.0J for Windows (spss Japan Inc., Tokyo, Japan) and amos 18.0 (Chicago, IL, USA).
Of the 807 workers approached, 523 (64.8%) agreed to participate in the study. The 284 workers who did not participate did not differ significantly from the participants in terms of age or sex. Demographic characteristics are shown in Table 1.
Mean score for each item on the TRS is shown in Table 2. The overall Cronbach's alpha coefficient of the scale was 0.82. Of the 239 retests issued, 151 (63.2%) responses were received. The test–retest correlation coefficient of the scale was 0.84.
Table 2. Mean score and standard deviation of each item of the Tachikawa Resilience Scale
| 1. Even during hardships, I think I will be able to manage.||5.2 (1.2)|
| 2. I accept things as they are when there are no alternatives.||5.2 (1.2)|
| 3. I am good at changing my way of thinking.||4.0 (1.5)|
| 4. During hardships, I think things could be even worse so I should feel lucky.||4.3 (1.4)|
| 5. Even if I am angry with someone who caused serious problems for me, my anger does not last long.||4.2 (1.5)|
| 6. Even during hardships, I can maintain a normal state of mind.||4.3 (1.6)|
| 7. I try not to worry about what is beyond my capabilities.||4.2 (1.5)|
| 8. Even during hardships, I can appreciate the situation to some extent.||3.8 (1.5)|
| 9. I know several ways to cope with stress that are effective for me.||4.7 (1.4)|
|10. Even though there are many things I cannot achieve, I think it is important for me to do my best in the things I can manage.||5.2 (1.3)|
Confirmatory factor analyses showed that the goodness-of-fit indices for the unidimensional model were sufficient (RMSEA = 0.063, GFI = 0.96, AGFI = 0.94, CFI = 0.94). Standardized coefficients (factor loadings) of the items were all > 0.44, except that of item 10, which was 0.35.
Mean score and concurrent and construct validity are shown in Table 3. The scales were highly correlated, were negatively correlated with depressive symptoms, and were positively correlated with social support. On multivariate regression analysis, the RS-14 and TRS showed a negative relation with the CES-D. Regression coefficients (beta weights) of the two scales were roughly equivalent (Table 4).
Table 3. Correlations between RS-14, TRS and other measures in company workers
|RS-14||63.4 (11.0)|| || ||0.69||<0.01|
|TRS||45.1 (8.5)||0.69||<0.01|| || |
|SSQ|| || || || || |
Table 4. Results of multivariate regression analysis adjusted for age and sex
|RS-14||−0.21 (−0.28, −0.15)||−0.24 (−0.38, −0.09)||−0.21 (−0.37, −0.06)||−0.07 (−0.08, 0.22)|
|TRS||−0.27 (−0.35, −0.18)||−0.54 (−0.77, −0.32)||−0.41 (−0.66, −0.16)||−0.88 (−1.12, −0.65)|
Of the 270 psychiatric outpatients approached, 140 (51.9%) agreed to participate in this study. The 130 outpatients who did not participate did not differ significantly from the participants in terms of age or sex. Demographic characteristics are shown in Table 1.
From the 140 respondents, 107 (76.4%) had mood disorder and 33 (23.6%) had neurotic, stress-related, or somatoform disorder. In detail, 27 had bipolar affective disorder (F31), 72 had depressive episode (F32), eight had persistent mood disorder (F34), one had phobic anxiety disorder (F40), five had other anxiety disorder (F41), one had obsessive–compulsive disorder (F42), 13 had reaction to severe stress and adjustment disorders (F43), one had dissociative disorder (F44), four had somatoform disorders (F45), and eight had other neurotic disorders (F48).
Mean score and concurrent validity of the TRS and RS-14 are shown in Table 5. Both scales were highly correlated and negatively correlated with depressive symptoms and anxiety, respectively. Mean TRS score was significantly lower in psychiatric outpatients than in company workers (P < 0.01). On multivariate regression, the RS-14 and TRS both showed a negative relation with the BDI-II and STAI state and trait measures. In general, the regression coefficient of the TRS was higher than that of the RS-14 (Table 4). Regarding the relation of RS-14 with trait anxiety, correlation coefficients between the following three items of the RS-14 and trait anxiety were smaller than −0.40, namely, ‘I have self-discipline’ (−0.31), ‘In an emergency, I'm someone people can generally rely on’ (−0.33), and ‘I am determined’ (−0.39).
Table 5. Correlations between RS-14, TRS and other measures in psychiatric outpatients
|RS-14||53.6 (17.8)|| || ||0.68||<0.01|
|TRS||38.4 (11.1)||0.68||<0.01|| || |
|STAI|| || || || || |
The results demonstrated that the TRS has a high degree of internal consistency, test–retest reliability, and concurrent and construct validity in both company workers and psychiatric outpatients. In addition, the mean TRS score was significantly higher for workers than for psychiatric outpatients. These results are consistent with the concept of resilience. The reliability and validity of the TRS were at least equivalent to that of the RS-14 in the Japanese.
On multivariate regression, the TRS was more strongly associated with depressive symptoms and anxiety in psychiatric outpatients than the RS-14, especially regarding trait anxiety. Although some items of the TRS (such as ‘Even during hardships, I think I will be able to manage’) are similar to items on the RS-14 (such as ‘I usually manage one way or another’) the TRS contains items different from those of the RS-14 (such as ‘I accept things as they are when there are no alternatives’ and ‘During hardships, I think things could be even worse, so I should feel lucky’). Previous studies showed that two different coping styles appear to predict a resilient outcome, pragmatic coping and flexible adaptation.[33, 34] It seems that RS reflects rather pragmatic coping and TRS reflects more flexibility, which is the capacity to modify behavior to meet adversity. The results of multivariate regression suggested that the TRS correlated with depressive symptoms and anxiety independently from the RS-14, and seems to have a more important function in the face of adversity. It may be possible that some items, such as ‘Even if I am angry at someone who caused serious problems for me, my anger does not last long’ and ‘Even during hardships, I can maintain a normal state of mind’, can be enhanced by some types of psychotherapy, such as well-being therapy which can promote resilience,[35, 36] or mindfulness. However, this should be clarified in the future.
The relatively low correlation between the RS-14 and trait anxiety in patients with psychiatric disorder might be partly explained by rigidity. A person with low resilience might stick to a given identity in a rigid fashion in attempts to attenuate terrible anxiety. Perhaps some of the RS-14 items, such as ‘I have self-discipline’ and ‘I am determined’, reflect rigidity, and therefore he or she might have high scores on these items. Moreover, relatively resilient Japanese might have low scores on some RS-14 items, such as ‘In an emergency, I'm someone people can generally rely on’, because they might feel it is too positive a trait to apply to themselves. Resilience seems to depend on both a state and a fixed personality trait, as Hoge proposed that a resiliency factor is intrinsic to the individual and might be modifiable.
We acknowledge, however, that the TRS is based on little theoretical rationale and it cannot reflect the multi-dimensional nature of resilience fully. Although one of the definitions of resilience is that it is a dynamic process of adaptation to challenging life conditions, the study was cross-sectional and did not include a dynamic process. Thus, we cannot regard it as an established scale. Rather, the value of the findings of this study might be to indicate the possibility that widely used scales measuring resilience are not the best tools for Japanese populations. It would be important to measure an individual's reaction to an experimental stress or to stressful life events or traumas over time biologically and physiologically, and examine the relation between TRS and these variables in the future.
The correlation between depressive symptoms and RS-14 score in the present study was equivalent to or better than that reported in previous studies.[2, 38, 39] The relatively low correlation between depressive symptoms and RS-14 score in our previous study was likely caused by the restricted sample population of nursing and university students. Thus, the RS-14 also seems to be useful for Japanese populations in terms of reliability and validity. However, the mean RS-14 score of the company workers was almost the same as that of the students in our previous study, and much lower than that of the healthy population in the USA. It seems impossible that the majority of the participants in Study 1 would have very low resilience levels considering their CES-D scores. Iwata et al. pointed out that the Japanese show a general tendency to suppress the expression of positive affects,[9, 10] which could explain this finding. This requires consideration when scales for resilience are used in Japan.
This study has some limitations. The items we draw did not cover diverse aspects of the resilience, nor were they based on theoretical rationale. The first group of participants were mainly men, highly educated and did not include persons aged 18 years or younger or older people who had retired from work. Therefore, they might not be representative of the Japanese general population. The number of participants in the second group was modest and psychiatric disorders were diverse. Duration of illness was not taken into account. Reliability and validity should be examined with respect to each psychiatric disorder and severity, as well as in other countries. Lastly, the cross-sectional design could not reveal causality. Whether TRS predicts depression or anxiety should be elucidated in longitudinal studies.
In conclusion, the reliability and validity of the TRS and RS-14 in company workers and patients with psychiatric disorders in Japan were acceptable. Moreover, the validity of the newly developed TRS was equivalent to or better than that of the RS-14. The results of this study suggest that scales measuring resilience that cover cultural aspects might be more relevant in given populations.
We thank Ms Kyoko Akutsu and Ms Yumiko Kamoshida for data management. This work was supported by grants from the Foundation for Total Health Promotion, and CREST, Japan Science and Technology Agency.
Dr Nishi has received research support from Toray Industries, Inc. and the Foundation for Total Health Promotion, and lecture fees from Qol Co., Ltd, DHA & EPA Association, and NTT DoCoMo, Inc. Dr Matsuoka has received research support from the Japan Science and Technology Agency, CREST, and the Ministry of Health, Labor, and Welfare of Japan, an Intramural Research Grant for Neurological and Psychiatric Disorders from NCNP, and lecture fees from Suntory Wellness Ltd, Eli Lilly Japan KK, and Otsuka Pharmaceutical Co., Ltd. All other authors declare that they have no competing interests in this work. Funding sources did not play any other role.