Work-related activities that may contribute to musculoskeletal symptoms among dental students: validation study

Abstract

Background: Dentists are exposed to occupational hazards, such as musculoskeletal disorders, in which symptoms can manifest early in clinical practice. Objectives: To estimate the reliability and validity of the Portuguese version of the Questionnaire on work-related activities that may contribute to musculoskeletal symptoms when applied to dental students. Methods: 553 Brazilian students participated in the study. The one-factor structure originally proposed was tested through confirmatory factor analysis (CFA) using the indexes chi(2)/df, CFI, GFI, and RMSEA. After observing an inadequate good fit (chi(2)/df=7.140, CFI=0.791, GFI=0.778, RMSEA=0.129), the sample was subdivided into 3 groups and an exploratory factor analysis was conducted (EFA) (n=255). A CFA was then conducted using a subsample (n=113). To test the invariance of the obtained factorial solution, a multi-group analysis was performed using a third sample, independent from the others (n=185). The convergent (AVE) and discriminant (rho(2)) validity were assessed. The composite reliability (CR), Cronbach's alpha coefficient (a), and intra-class correlation coefficient (r) were calculated. Results: In the EFA, 3 factors were extracted: Repetitiveness, Working posture, and External factors (chi(2)=21895.154; p<0.001; KMO=0.905; lambda(1)=6.683; lambda(2)=1.464; lambda(3)=1.277) that explained 62.82% of the total variance. The three-factor model presented an adequate good fit (n=113 and n=185) (chi(2)/df=2.259; CFI=0.885; GFI=0.833; RMSEA=0.066). The convergent validity was compromised only for Repetitiveness (AVE=0.412-0.653). The discriminant validity (rho(2)=0.333-0.428), composite reliability, internal consistency, and reproducibility were adequate for all factors (CR=0.736-0.883; alpha=0.747-0.876; rho=0.729-0.940). Conclusion: The questionnaire proved to be reliable and valid for the sample of dentistry students if the three-factor model is used.