Method: A total of 235 undergraduate dental students at a public university from Brazil participated in the study. 72.8% were female and the mean age was 21.0±1.8 years. The portuguese version of the Copenhagen Burnout Inventory (CBI-S) was used. CBI-S enables the answers to be rated on the 5-point Likert-type scale and it is composed of the four following factors: personal burnout (PB), burnout related to studies (BRS), burnout related to colleagues (BRC), and burnout related to professors (BRP). To assess the psychometric properties of CBI-S the factorial, convergent, discriminant validities and reliability were estimated. Confirmatory factor analysis was performed. The items that presented factor weights (λ) lower than 0.50 and those that showed to be redundant by the inspection of modification index (Lagrange multiplier) (LM>11) were excluded. To assess adequacy of the model to the sample, the goodness-of-fit indexes χ2/gl, CFI, GFI and RMSEA were used. The convergent validity was calculated by average variance extracted (AVE) and the discriminant validity by correlational analysis (ρ2). The reliability of the instrument was assessed by composite reliability (CR) and the Cronbach’s alpha coefficient (α).
Result: No item presented λ<0.50 and the seven correlations among the items of the same factor needed to be inserted (LM>11). The factorial model of CBI-S showed good fit to the sample (λ=0.50-0.88; χ2/gl=1.794;CFI=0.949;GFI=0.864;RMSEA=0.058) with adequate convergent validity (AVE=0.50-0.69;CR=0.87-0.93), discriminant validity (ρ2=0.06-0.61) and internal consistency (α=0.86-0.93).
Conclusion: The Portuguese version for students of CBI-S showed adequate validity and reliability of the dental students sample.