Accuracy and validity of observational estimates of wrist and forearm posture

Numerous observational methods for analysis of working posture of the wrist/forearm have been reported in the literature yet few of these methods have been validated for the accuracy of their posture classification. The present study evaluated the accuracy of estimates of working posture made by 28 experienced ergonomists using methods of scaling upper limb posture typical of those reported in the literature. Observational estimates of wrist/forearm posture of four jobs presented on video-recording were compared with posture levels measured directly with an electrogoniometer system. Ergonomists using a visual analogue scale tended to underestimate peak and average wrist extension with mean errors of -29.4% and -10.5% of the joint ROM, respectively (p<0.05). While estimates of wrist flexion, pronation and supination resulted in less bias, variability in observer error was large for all wrist postures. The probability of an analyst misclassifying the most frequently occurring posture using a three- and a six-category scale was 54 and 70%, respectively. The probability of misclassifying peak posture was 22 and 61% using a three- and a six-category scale respectively. This suggests a trade-off between the degree of precision afforded by the categorical scale and the likelihood of posture misclassification. Estimates of the temporal distribution of posture among the categories appeared to be biased towards more neutral postures than were measured for the jobs. This indicated the possibility of a trend towards underestimation of posture duration severity by the ergonomists.