INTERRATER RELIABILITY OF A VIDEO-BASED FIXED INTERVAL SAMPLING METHOD TO ASSESS HAND FORCE AND WRIST POSTURE

摘要

Interrater reliability was assessed for a video-based, quantitative exposure assessment method for estimation of hand forces and wrist postures for a short duty-cycle job. Six raters analyzed a 25- second video of a worker packing a case and rated bilateral hand forces and wrist postures at ¼- second intervals. Consistency was analyzed using intraclass correlation coefficients (ICC). Interrater reliability was high for hand forces (left: ICC=0.78; right: ICC=0.77), moderate for wrist flexion/extension (left: ICC=0.61; right: ICC=0.59), and poor for wrist deviation (left: ICC=0.15; right: ICC=0.34). For one subtask repeated six times per cycle, force variation was inadequate for ICC calculation, but ICCs were similar for wrist postures (mean ICC=0.55 for flexion/extension, mean ICC=0.39 for deviation). Results indicate raters were consistent in estimating hand forces. Discrepancies in estimating wrist postures were consistent with existing literature. ICCs work well in situations with more variability (i.e. more activity), but not in static, unchanging conditions.