Validation of a self-administered questionnaire for assessing exposure to back pain mechanical risk factors

Objective: To validate a self-administered questionnaire assessing exposure to mechanical risk factors, developed for a cohort study aiming at assessing the influence of physical and psycho-social factors on the incidence of low back pain (LBP). Methods: The study first involved a criterion validity test. A sample of the cohort workers (n=152) was observed at the workplace during four 30 min periods randomly distributed along the shift. At the end of the work shift, the questionnaire was filled in both by the worker and the observer. Agreements were tested between self-reports and observations, and between self-reports and observer opinion. Secondly, a comparison of exposure-effect relationships based on self-reports to those based on observations was carried out on the whole study cohort (n=716). Both sets of Relative Risks of being an incident case (LBP lasting at least 7 consecutive days in the follow-up year) were tested for heterogeneity. Results: Self-reports agreement levels were better with observer opinion than with observational data and were higher for answers at a dichotomous level. Vehicle driving, manual handling without estimation of weight and frequencies, or trunk bending without rotation showed a fair to good agreement with the external criteria. Limits in the validation procedure did not allow validating the sitting and standing durations. As regards the health outcome comparison, questionnaire and observations led to homogeneous Relative Risks for the variables tested. Conclusions: Results show that self-reports provide a limited accuracy to assess actual frequencies and durations of work activities. Using a questionnaire, classifying the workers into exposure categories is rather relative, but questionnaire and observations seem similar in their relationships to outcome.