Yan Ying, Zhou Haibo, Cai Jianwen
Department of Mathematics and Statistics, University of Calgary, Calgary, Alberta, Canada, T2N 1N4.
Department of Biostatistics, University of North Carolina at Chapel Hill, Chapel Hill, North Carolina 27599, U.S.A.
Biometrics. 2017 Sep;73(3):1042-1052. doi: 10.1111/biom.12657. Epub 2017 Jan 23.
The case-cohort study design is an effective way to reduce cost of assembling and measuring expensive covariates in large cohort studies. Recently, several weighted estimators were proposed for the case-cohort design when multiple diseases are of interest. However, these existing weighted estimators do not make effective use of the covariate information available in the whole cohort. Furthermore, the auxiliary information for the expensive covariates, which may be available in the studies, cannot be incorporated directly. In this article, we propose a class of updated-estimators. We show that, by making effective use of the whole cohort information, the proposed updated-estimators are guaranteed to be more efficient than the existing weighted estimators asymptotically. Furthermore, they are flexible to incorporate the auxiliary information whenever available. The advantages of the proposed updated-estimators are demonstrated in simulation studies and a real data analysis.
病例队列研究设计是在大型队列研究中降低收集和测量昂贵协变量成本的有效方法。最近，当关注多种疾病时，针对病例队列设计提出了几种加权估计量。然而，这些现有的加权估计量没有有效利用整个队列中可用的协变量信息。此外，研究中可能存在的昂贵协变量的辅助信息不能直接纳入。在本文中，我们提出了一类更新估计量。我们表明，通过有效利用整个队列信息，所提出的更新估计量在渐近意义上保证比现有的加权估计量更有效。此外，它们可以灵活地纳入可用的辅助信息。所提出的更新估计量的优势在模拟研究和实际数据分析中得到了证明。
