An thần vừa phải thay thế các loại trước đây gọi là thuốc an thần có ý thức. Trong trạng thái này, bệnh nhân vẫn đáp ứng mục đích là mô tả cho tối thiểu an thần, mặc dù ý thức bị suy giảm nhiều hơn. Đường thở vẫn còn tình trạng ổn định mà không cần can thiệp, và huyết động học là "thông thường" không bị ảnh hưởng. | 286 Othman Fig. 4. Example of a run chart. Fig. 5. Example of a control chart bar graph with threshold. by the data itself or by benchmarking. Once initial quality improvement data collection has been completed the data can be used to generate a threshold performance standard. The mean or average occurrence can serve as a threshold. It is important to calculate the weighted average because each measurement is the average of the cases for that time period. Since the volume of activity varies the effect of each measurement should be in accordance with its proportionate volume 18 . An important concept in threshold development is variation specifically the variation around the mean. If the goal is to consistently meet performance standards the weighted mean average can also serve as the threshold of acceptable performance if a defined level of standard deviation SD of the mean is added 18 . Fig. 5 demonstrates the resultant threshold obtained by adding one standard deviation to the calculated mean average for the rate QA and CQI in Sedation Analgesia 287 Fig. 6. Example of a control chart bar graph with rolling mean. of oxygen desaturation over a 1-year period. If this threshold number is continually recalculated as more data points are collected over time a rolling mean is obtained that continually drives for improved performance Fig. 6 . This dynamic measurement approach is consistent with the philosophy of continuous quality improvement 2 . The second approach benchmarking is a comparison of the published or non-published experience or the results of other similar programs. These could be within the same institution with other similar institutions or to a national database. Benchmark data when compared to institutional data can identify obvious initial areas for improvement and often allow for identification of practices which might be used to improve performance. Over time if used alone benchmark data tends to result in acceptance of the status quo once the .
