Tang Yongqiang
a GlaxoSmithKline , Research Triangle Park , North Carolina , USA.
J Biopharm Stat. 2015;25(5):1100-13. doi: 10.1080/10543406.2014.971167. Epub 2014 Oct 16.
A sample size formula is derived for negative binomial regression for the analysis of recurrent events, in which subjects can have unequal follow-up time. We obtain sharp lower and upper bounds on the required size, which is easy to compute. The upper bound is generally only slightly larger than the required size, and hence can be used to approximate the sample size. The lower and upper size bounds can be decomposed into two terms. The first term relies on the mean number of events in each group, and the second term depends on two factors that measure, respectively, the extent of between-subject variability in event rates, and follow-up time. Simulation studies are conducted to assess the performance of the proposed method. An application of our formulae to a multiple sclerosis trial is provided.
推导了用于分析复发事件的负二项回归的样本量公式，其中受试者的随访时间可以不相等。我们得到了所需样本量的精确下限和上限，计算起来很容易。上限通常仅比所需样本量略大，因此可用于近似样本量。下限和上限可以分解为两项。第一项依赖于每组事件的平均数量，第二项取决于分别衡量事件发生率在受试者间变异程度和随访时间的两个因素。进行了模拟研究以评估所提出方法的性能。还提供了我们的公式在一项多发性硬化症试验中的应用。
