Continuous versus group sequential analysis for post-market drug and vaccine safety surveillance.
Continuous versus group sequential analysis for post-market drug and vaccine safety surveillance.
复制标题
DOI：
10.1111/biom.12324
复制
发表时间：
2015-09
期刊：
影响因子：
1.9
通讯作者：
Kulldorff M
中科院分区：
文献类型：
--
作者：
Silva IR;Kulldorff M
在上市后药品安全监测中使用顺序统计分析正在迅速兴起。连续分析和群体序列分析都已被使用，但对于何时使用哪种方法缺乏共识。我们比较了连续分析和组序贯分析的统计性能在I型误差概率方面；统计能力;零假设被拒绝时发出信号的预期时间；以及在不拒绝零值的情况下结束监视所需的样本量。我们提出了一个数学命题，证明对于任何群序列设计，总是存在一个一致更好的连续序列设计。结果表明，越频繁的测试总是越好。此外，对于基于泊松的概率模型和基于对数似然比的平坦拒绝边界，我们比较了各种连续和组顺序设计的性能。使用精确的计算，我们发现，对于所使用的参数设置，总是有一个连续的设计与较短的预期时间信号比最佳组设计。本文的两个关键结论是：(i)任何上市后安全监测系统都应尽可能频繁地获取数据，以及（ii）当新数据到达时，应始终进行顺序测试，而不是故意等待其他数据。
The use of sequential statistical analysis for post-market drug safety surveillance is quickly emerging. Both continuous and group sequential analysis have been used, but consensus is lacking as to when to use which approach. We compare the statistical performance of continuous and group sequential analysis in terms of type I error probability; statistical power; expected time to signal when the null hypothesis is rejected; and the sample size required to end surveillance without rejecting the null. We present a mathematical proposition to show that for any group sequential design there always exists a continuous sequential design that is uniformly better. As a consequence, it is shown that more frequent testing is always better. Additionally, for a Poisson based probability model and a flat rejection boundary in terms of the log likelihood ratio, we compare the performance of various continuous and group sequential designs. Using exact calculations, we found that, for the parameter settings used, there is always a continuous design with shorter expected time to signal than the best group design. The two key conclusions from this article are (i) that any post-market safety surveillance system should attempt to obtain data as frequently as possible, and (ii) that sequential testing should always be performed when new data arrives without deliberately waiting for additional data.