Sieve maximum likelihood regression analysis of dependent current status data

• 成果类型: Article
• 署名作者: Ma, Ling; Hu, Tao; Sun, Jianguo
• 署名单位: University of Missouri System; University of Missouri Columbia; Capital Normal University
• 刊物名称: BIOMETRIKA
• ISSN/ISSBN: 0006-3444
• DOI: 10.1093/biomet/asv020
• 发表日期: 2015
• 页码: 731738
• 关键词: PROPORTIONAL HAZARDS MODEL; nonparametric-estimation; efficient estimation; survival

摘要

Current status data occur in contexts including demographic studies and tumorigenicity experiments. In such cases, each subject is observed only once and the failure time of interest is either left- or right-censored (Kalbfleisch & Prentice, 2002). Many methods have been developed for the analysis of such data (Huang, 1996; Sun, 2006), most of which assume that the failure time and the observation time are independent completely or given covariates. In this paper, we present a sieve maximum likelihood approach for current status data when independence does not hold. A copula model and monotone I-splines are used and the asymptotic properties of the resulting estimators are established. In particular, the estimated regression parameters are shown to be semiparametrically efficient. An illustrative example is provided.