Personal agency is a variable which potentially facilitates personal recovery in people with serious mental illness. This study aimed to develop a new brief measure for subjective personal agency that can be completed by people with serious mental illness.

Two focus group interviews were first conducted with 11 people with schizophrenia to understand the fundamental components of subjective personal agency for people with serious mental illness living in the community. One group comprised six people with schizophrenia living in the community, while the other consisted of five people with schizophrenia working as peer-support workers. We then developed scale items through collaboration with people with schizophrenia and qualitative analysis (stage 1). A cross-sectional survey was then conducted to test the psychometric properties of the new scale among service users with schizophrenia in 18 assertive community treatment teams (stage 2). Factor validity was tested via exploratory factor analysis (EFA) and confirmatory factor analysis (CFA). We evaluated convergent validity with the Boston University Empowerment Scale (BUES), divergent validity with the global assessment of functioning (GAF), internal consistency, and test–retest reliability.

Seven items were included in the scale at stage 1. In stage 2, 195 participants completed this scale. EFA revealed a one-factor model with five items. CFA indicated good model fit (χ2 statistics [CMIN] = 8.445, df = 5 (CMIN/df = 1.689), p = 0.133, comparative fit index = 0.974, Tucker–Lewis fit index = 0.949, root mean square error of approximation = 0.077 and standardised root mean squared residual = 0.042). The new scale was significantly correlated with total BUES score (r = 0.526, p < 0.001), but not with GAF score. Cronbach's α for internal consistency was 0.79, and intra-class correlation coefficient for test–retest reliability was 0.70.

We developed a new, five-item Subjective Personal Agency scale (SPA-5) that can be completed by people with serious mental illness. Further studies are needed to confirm the results outside Japan.

Mental health-related stigma is a major challenge associated with the huge mental health treatment gap. It has remained unclear what kind of educational content is effective in reducing the stigma. Whether biomedical messages (BMM) about mental illness are effective or harmful in decreasing stigma is controversial. To investigate whether BMM can improve practically useful knowledge of mental illness, comparably to recommended messages (RCM) advocated by experts, of types such as ‘recovery-oriented’, ‘social inclusion/human rights’ and ‘high prevalence of mental illnesses’ through a randomised controlled trial (RCT).

This study is an individual-level RCT with a parallel-group design over 1 year, conducted in Tokyo, Japan. A total of 179 participants (males n = 80, mean age = 21.9 years and s.d. = 7.8) were recruited in high schools and universities, and through a commercial internet advertisement in June and July 2017, without any indication that the study appertained to mental health. Participants were allocated to the BMM and RCM groups. They underwent a 10-min intervention, and completed self-report questionnaires during baseline, post-test, 1-month follow-up and 1-year follow-up surveys. The primary outcome measures were practically useful knowledge of mental illness at the post-test survey using the Mental Illness and Disorder Understanding Scale (MIDUS). Analysis was conducted in October 2018.

Both groups demonstrated improved MIDUS score in the post-test survey, and showed similar intervention effects (F(1, 177) = 160.5, p < 0.001, η2 = 0.48). The effect of the interventions continued until the 1-year follow-up survey (B [95% CI] = −2.56 [−4.27, −0.85], p < 0.01), and showed no difference between groups. The reported adverse effect that BMM increase stigma was not confirmed.

BMM may have a positive impact on stigma, comparable to RCM. These findings may encourage reconsideration of the content of messages about mental health, as it is indicated that combining BMM and RCM might contribute to an effective anti-stigma programme.