Psychometric properties of the Japanese version of the Recovery Attitudes Questionnaire (RAQ) among mental health providers: a questionnaire survey

Abstract

Background: "Recovery" is a central concept in mental health, particularly for mental health services and policy-makers. The present study examined the factorial and concurrent validity, internal consistency reliability, and test-retest reliability of the Japanese version of the 7-item Recovery Attitudes Questionnaire (RAQ) among mental health service providers in community and inpatient settings in Japan.

Methods: We conducted a cross-sectional questionnaire with a number of eligible professional groups, including psychiatrists, registered/assistant nurses, public health nurses, clinical psychologists, pharmacists, occupational therapists, and social workers. Participants were drawn from two psychiatric hospitals and 56 psychiatric clinics or community service agencies. In total, 331 participants completed the questionnaire. After excluding those with missing RAQ values, 307 participants were included in the analysis; the participants' mean age was 40.2 years and 29.6 % were men. The questionnaire comprised the Japanese version of the 7-item RAQ developed by the present authors, the revised scale of the positive attitudes of staff toward persons with mental disorder (the positive attitudes scale), and the Japanese-language version of the Social Distance Scale (SDSJ). Confirmatory factor analyses were used to examine factorial validity of a two-factor structure reported in a previous study (Borkin et al., 2000) as well as a single-factor structure. Concurrent validity was determined by calculating correlations between RAQ and the other two scales. Internal consistency reliability was assessed with Cronbach's alpha coefficients and inter-item correlations. Test-retest reliability was assessed by the intraclass correlation coefficient (ICC), with a weighted kappa in a subsample of participants (n = 13).

Results: The two-factor structure showed acceptable factorial validity. RAQ scores were significantly and positively correlated with the positive attitudes scale, and there was a significant inverse correlation with the SDSJ (p < 0.01). The RAQ had an overall Cronbach's alpha coefficient of 0.64. Four inter-item correlations were not significant. The ICC and weighted kappa values indicated unsatisfactory test-retest reliability.

Conclusion: The Japanese RAQ showed acceptable factorial validity, reasonable concurrent validity, and unsatisfactory reliability in community and inpatient mental health settings in Japan. Further large-scale research is required to ensure robust verification.

Fig. 1

Path diagram of the Japanese version of RAQ, showing standardized coefficients from confirmatory factor analysis