Tambour M, Zethraeus N
Centre for Health Economics, Stockholm School of Economics, Sweden.
Health Econ. 1998 Mar;7(2):143-7. doi: 10.1002/(sici)1099-1050(199803)7:2<143::aid-hec322>3.0.co;2-q.
Recently, a number of papers have brought up the issue of how to make cost-effectiveness (CE) studies stochastic, i.e. how to obtain confidence intervals for CE ratios. In this note we present a bootstrap procedure for estimating bias-corrected confidence intervals for CE ratios. The bootstrap procedure is tested in a simulation study based on the assumptions made in a recent paper by Wakker and Klaassen in this journal. We test two variants of CE ratio bootstrap confidence intervals. The first is a bootstrap analogue of the parametric method proposed by Wakker and Klaassen which gives results similar to those obtained with the parametric method. However, computing bootstrap confidence intervals directly for the CE ratio produce results closer to the predetermined significance level.
最近，一些论文提出了如何使成本效益（CE）研究具有随机性的问题，即如何获得CE比率的置信区间。在本注释中，我们提出了一种用于估计CE比率的偏差校正置信区间的自助法程序。该自助法程序在基于Wakker和Klaassen最近在本期刊发表的一篇论文中的假设进行的模拟研究中进行了测试。我们测试了CE比率自助置信区间的两种变体。第一种是Wakker和Klaassen提出的参数方法的自助类似物，其给出的结果与使用参数方法获得的结果相似。然而，直接为CE比率计算自助置信区间会产生更接近预定显著性水平的结果。
