Sherlaw-Johnson Chris, Morton Anthony, Robinson Michael B, Hall Alistair
Clinical Operational Research Unit, University College London, Department of Mathematics, UK.
Int J Cardiol. 2005 Apr 20;100(2):301-7. doi: 10.1016/j.ijcard.2004.12.009.
Quality control in clinical care is becoming increasingly more prevalent, resulting in a need for tools that can be readily used by clinical teams for monitoring their own performance. The aim of this study was to devise a practical and effective scheme for monitoring coronary care mortality in real-time.
We obtained data for 2153 consecutive patients admitted after acute myocardial infarction between 1st September and 30th November 1995 to one of 20 acute hospitals in West Yorkshire participating in the NHS R and D funded EMMACE-1 study. We developed control charts for each centre to monitor 30-day mortality. These control charts used two complementary tools: the Risk-Adjusted Cumulative Sum chart (CUSUM) and a new 'Rocket Tail' chart based upon the Variable Life-Adjusted Display (VLAD). We also combined information from each of these to devise a further chart.
Control charts are shown for two centres in order to illustrate the important features of the different but complimentary monitoring tools. The Risk-Adjusted CUSUM is shown to be useful for detecting isolated runs of unsatisfactory outcome results but is not 'intuitive', and does not give any indication of the preceding history of outcomes. The Rocket Tail chart is shown to give a good summary of outcome history and also how overall performance compares with what would be expected for the case-mix. A chart that combines both approaches appeals to the advantages of each.
We propose a visual approach to health-care monitoring that beneficially combines and extends the different information of the previously used CUSUM and VLAD charts.
临床护理中的质量控制日益普遍，这就需要临床团队能够方便使用的工具来监测自身表现。本研究的目的是设计一种实用且有效的方案来实时监测冠心病护理死亡率。
我们获取了1995年9月1日至11月30日期间连续入住西约克郡20家急性医院之一的2153例急性心肌梗死后患者的数据，这些医院参与了由英国国家医疗服务体系研发基金资助的EMMACE - 1研究。我们为每个中心绘制控制图以监测30天死亡率。这些控制图使用了两种互补工具：风险调整累积和图（CUSUM）和基于可变寿命调整显示（VLAD）的新“火箭尾”图。我们还将这些信息结合起来设计了另一张图。
展示了两个中心的控制图，以说明不同但互补的监测工具的重要特征。风险调整CUSUM图被证明对检测孤立的不满意结果序列有用，但不“直观”，也没有给出任何结果的既往历史信息。火箭尾图被证明能很好地总结结果历史，以及整体表现与病例组合预期的比较情况。结合两种方法的图兼具了各自的优点。
我们提出一种可视化的医疗监测方法，它有益地结合并扩展了先前使用的CUSUM图和VLAD图的不同信息。
