Validation of the Maslach burnout inventory-general survey 9-item short version (MBI-GS9) among care aides in Canadian nursing homes

Background: The Maslach Burnout Inventory-General Survey (MBI-GS) is the leading measure of burnout for all occupations. The MBI-GS9, the 9-item version of the MBI-GS, was formulated based on the MBI-GS and has been used for several years. However, very few studies have systematically tested its psychometric properties, and none have focused on care aides working in nursing homes who are susceptible to burnout.

Methods: Following the Standards for Educational and Psychological Testing, this study validated the MBI-GS9 among 3,765 care aides from 91 Canadian nursing homes, using data collected between September 2019 and February 2020 by the Translating Research in Elder Care (TREC) program.

Results: The Exhaustion subscale had good reliability with coefficients around 0.66-0.74. The Cynicism subscale had medium reliability with coefficients around 0.60-0.66, and the Efficacy subscale also had medium reliability with coefficients around 0.51-0.58. The MBI-GS9 was significantly correlated with various conceptually related constructs, such as health status, working environment, job satisfaction, psychological empowerment, work engagement, and organizational citizenship behaviors. The MBI-GS9 had a three-factor structure in the full sample and showed equivalent factor structure, factor loadings, latent values, factor variance and error variance across different sex and age groups. Care aides with English as their first language showed higher latent values of the Exhaustion subscale compared to those with English as a second language.

Conclusion: Overall, the MBI-GS9 exhibited acceptable psychometric properties, but medium reliability of cynicism and efficacy subscales, for measuring burnout among care aides in nursing homes, demonstrating equivalence across sex and gender groups. When comparing across different languages or racial or ethnic groups among care aides, it is important to consider inequivalent latent values on Exhaustion before comparing scores on the measure.