Goodman William M, Lamers Angela
University of Ontario Institute of Technology, Ontario, Canada.
Nurs Forum. 2010 Oct-Dec;45(4):246-52. doi: 10.1111/j.1744-6198.2010.00199.x.
Applying simulations in healthcare practice and education is increasingly accepted, yet a number of recent authors have questioned the effectiveness of these technologies. The contention is that while high-fidelity simulators may contribute to educational gains, their gains compared to low-tech alternatives are often "not significant." That assessment, however, and the evidence it is based on, may be a consequence of asking the wrong questions. Typical studies often compare a measure for "average success" for one group's members versus another's on some criteria, but this can mask important information about the "tails" of the distribution for how trainees are performing. An alternative approach, adapted from quality control, compares error rates for each group in the experiment, in aggregate. The statistical results of evaluations can change if this method is used, as illustrated by a recent study showing that simulation training can significantly reduce the frequency of medication administration errors among student nurses on placement. The paper includes a case study to tangibly demonstrate how the way we frame our evaluation test question can reverse the apparent statistical finding of the significance test.
在医疗保健实践和教育中应用模拟技术越来越被认可，但最近一些作者对这些技术的有效性提出了质疑。争议在于，虽然高保真模拟器可能有助于教育成果，但与低技术替代方案相比，其成果往往“不显著”。然而，这种评估及其所基于的证据，可能是提出错误问题的结果。典型的研究通常会根据某些标准比较一组成员与另一组成员的“平均成功率”指标，但这可能会掩盖有关学员表现分布“尾部”的重要信息。一种从质量控制中改编而来的替代方法是汇总比较实验中每组的错误率。如果使用这种方法，评估的统计结果可能会改变，正如最近一项研究所示，模拟培训可以显著降低实习护士用药错误的频率。该论文包括一个案例研究，切实展示了我们构建评估测试问题的方式如何扭转显著性检验的明显统计结果。
