Background: General population job-exposure matrices (JEMs) based on expert assessments of physical exposures may be valuable tools for studying occupationrelated musculoskeletal disorders. Wrist-hand JEMs are few and the reliability and validity of expert assessments of wrist-hand exposures is uncertain.Methods: We examined intra-and inter-rater reliability of ratings of five experts of hand-wrist repetition, deviation, force, vibration, and computer work in 33 jobs selected to represent a large exposure variation. The validity of ratings of handwrist repetition was examined by comparison with electro-goniometer measurements of wrist angular velocity and mean power frequency (MPF), and the validity of hand-wrist deviation by comparison with goniometer measurements of range of motion (ROM).Results: Intra-rater test-retest and inter-rater Spearman correlation coefficients controlling for rater effects, varied between 0.70 and 0.87. Corresponding kappa statistics of overall agreement showed similar high values, except for wrist deviation (kappa = 0.50). Regression analyses showed strong positive associations between expert assessments of repetition and goniometer measurements of wrist angular velocity (R 2 = 0.56, p < 0.0001) and MPF (R 2 = 0.37, p < 0.0003), while expert ratings of wrist deviation showed a weak statistically nonsignificant association with goniometer measurements of ROM (R 2 = 0.032, p = 0.34).
Conclusion:The reliability of expert assessments of wrist-hand physical exposures was high. Compared to goniometer measurements, the validity of assessments of wrist-hand repetition was also high, but it was low for assessments of wrist-hand deviation. The results are encouraging for establishing a hand-wrist JEM, but the results for wrist deviation emphasize that expert assessments should be validated against objective measurements.