Validation of a French version of the empowerment scale for mental health service users.

Background: As the concept of empowerment is increasingly adopted across various mental health care contexts, there is a growing need for standardized measures to assess the effectiveness of empowerment approaches. The Empowerment Scale is widely utilized and translated within the field of mental health, despite its varied psychometric properties. This study aimed to translate the Empowerment Scale into French and assess its internal consistency, validity, and responsiveness.

Methods: This study was part of a larger research project involving 394 participants. The Empowerment Scale was translated into French following cross-cultural adaptation guidelines, with a translation committee consisting of experts and a professional translator. Psychometric properties were assessed using classical test theory. The factor structure was determined through principal component analysis, exploratory factor analysis, and confirmatory factor analysis. Internal consistency was measured using Cronbach's alpha, while validity was evaluated through convergent, discriminant, and concurrent validity analyses. Responsiveness was assessed by comparing empowerment scores to changes in recovery rates.

Results: The factor analyses supported a four-factor, 18-item model, showing good fit indices (CFI = 0.97, TLI = 0.97, AGFI = 0.97, SRMR = 0.07, RMSEA = 0.07). Internal consistency was acceptable for the overall scale (alpha = 0.84) and the "self-esteem-self-efficacy" dimension (alpha = 0.88) but lower for the other dimensions. The scale demonstrated moderate correlations with recovery (r = 0.47) and quality of life (r = 0.28). The Empowerment Scale demonstrated low or insignificant responsiveness, except for the "self-esteem-self-efficacy" dimension, which showed moderate responsiveness.

Conclusions: The French version of the Empowerment Scale has a good factor structure with 4 factors and 18 items. The "self-esteem" dimension demonstrates good concurrent validity and reliability, and moderate responsiveness, while other dimensions require additional validation.