Storm-Versloot M N, Ubbink D T, Chin a Choi V, Luitse J S K
Department of Emergency Medicine, Academic Medical Center, Amsterdam, The Netherlands.
Emerg Med J. 2009 Aug;26(8):556-60. doi: 10.1136/emj.2008.059378.
To compare inter and intra-observer agreement of the Manchester Triage System (MTS) and the Emergency Severity Index (ESI).
50 representative emergency department (ED) scenarios derived from actual cases were presented to 18 ED nurses from three different hospitals. Eight of them were familiar with MTS, six with ESI and four were not familiar but trained in both systems. They independently assigned triage scores to each scenario according to the triage system(s) they were familiar with. After 4-6 weeks the same nurses again judged the scenarios in a different order. Unanimity in judgement and unweighted and quadratic-weighted kappas were calculated.
Unanimity in judgement for MTS was 90% and for ESI 73%. One-level disagreement was found in 8% and 23% of the cases, respectively. Interobserver unweighted kappas were 0.76 (95% CI 0.68 to 0.83) for MTS and 0.46 (95% CI 0.37 to 0.55) for ESI. Quadratic-weighted kappas were 0.82 (95% CI 0.74 to 0.89) and 0.73 (95% CI 0.64 to 0.83), respectively. At 4-6 weeks, one-level intra-observer disagreements were 10% and 22% and 2-level disagreement 1% and 2%, respectively. Intra-observer unweighted kappas were 0.84 (95% CI 0.73 to 0.94) for MTS and 0.65 (95% CI 0.59 to 0.72) for ESI.
Using paper-based clinical scenarios, MTS was found to have a greater inter and intra-observer agreement than ESI.
比较曼彻斯特分诊系统（MTS）和急诊严重程度指数（ESI）在不同观察者之间以及同一观察者内部的一致性。
从实际病例中选取50个具有代表性的急诊科场景，展示给来自三家不同医院的18名急诊科护士。其中8人熟悉MTS，6人熟悉ESI，4人不熟悉但接受了两种系统的培训。他们根据自己熟悉的分诊系统，独立为每个场景分配分诊分数。4至6周后，相同的护士再次以不同顺序对这些场景进行判断。计算判断的一致性以及未加权和二次加权的卡帕值。
MTS判断的一致性为90%，ESI为73%。分别在8%和23%的病例中发现了一级分歧。MTS观察者间未加权卡帕值为0.76（95%可信区间0.68至0.83），ESI为0.46（95%可信区间0.37至0.55）。二次加权卡帕值分别为0.82（95%可信区间0.74至0.89）和0.73（95%可信区间0.64至0.83）。在4至6周时，同一观察者内部的一级分歧分别为10%和22%，二级分歧分别为1%和2%。MTS同一观察者内部未加权卡帕值为0.84（95%可信区间0.73至0.94），ESI为0.65（95%可信区间0.59至0.72）。
使用纸质临床场景时，发现MTS在不同观察者之间以及同一观察者内部的一致性高于ESI。
