上肢肌肉骨骼不適之危險因子的評估方法中，直接量測法有較精確量測的優點，而觀察法與自我評估問卷於工作現場評估之執行面上較為容易進行。本研究比較直接量測與觀察法對於相同作業內容的危險因子評估結果之相關性，比較觀察法中較具可行性與精確性之評估指標，作為評估勞工現場手部活動情形之依據。本研究訪視10家公司，對19位受試者進行工作現場不同作業別之手部活動量測，所有受試者之工作皆以上肢作業為主。直接量測儀器包括表面肌電訊號感測元件(surface EMG)，安裝於受試者前臂，以及雙軸電子量角器安裝於手腕部，對受試者工作時之手部活動情形作監測與紀錄。並以錄影機拍攝至少15分鐘或5個工作週期，之後再由查核人員使用Strain Index(SI)及ACGIH-HAL檢核表評估動作負荷暴露，進行手部動作分析評估之觀察比較。本研究使用之直接量測與觀察法皆對施力強度及施力頻率等變項作評估，並使用無母數統計方法檢定不同方法對相同變項評估結果之相關性。直接量測法與觀察法量測所得之施力頻率達統計上顯著相關(r=0.62, p<0.01)。SI檢核表之每分鐘施作次數與雙軸電子量角器測得之動作重複頻率(mean power frequency)相關性最高(r=0.66, p<0.001)。不同評估工具於施力強度部分之量測值並沒有顯著相關，但使用觀察法對施力頻率作評估具有良好可信度與有效性，可作為未來大樣本工作現場手部作業暴露評估的量測指標之一。

To evaluate the risk factors of upper extremity musculoskeletal disorders, direct measurement has the advantage of better validity while observation method and self-report have the strength of convenience in operation. The aims of this study were to evaluate the correlation of exposure variables monitored by direct measurement and observation method, and to find out the most feasible and valid parameters by observation method to assess hand activity at workplace. A total of 35 different tasks by 19 subjects of 10 companies were included in this study, which were primarily involved in upper extremity activity. All subjects filled in a modified Nordic Musculoskeletal Questionnaire to report their health outcome prior to the beginning of this study. Then, the study subjects were asked to work as usual for the designated task for 15 minutes or more than 5 cycles per task. Meanwhile, direct measurements, including surface electromyography and biaxial electrogoniometers for wrist, were applied to record the bioelectronic signals for hand activities. All the subjects' hand activities were videotaped and, later, rated by a reviewer with Strain Index (SI) checklist and hand activity level (HAL) checklist issued by the American Conference of Governmental Industrial Hygienists. Both direct measurement and observation method provided exposure variables for exertion force and frequency. Exertion frequency obtained by direct measurement and observation method was significantly correlated (r=0.62, p<0.01). The effort per minute estimated by SI checklist was correlated with the mean power frequency of exertion (r=0.66, p<0.001) and the root mean square of wrist deviation speed measured by biaxial electrogoniometer (r=0.51, p<0.01). However, such consistence was not found for exposure variable of exertion force. The findings of the present study suggested that the exertion frequency of upper extremity motion estimated by observation method could be used as a feasible indicator for physical work loading assessment at workplace if a direct measurement is not feasible for all study subjects.