Misclassified current status data occur when each subject under study is observed only once and the failure status at the observation time is determined by a diagnostic test with imperfect sensitivity and specificity. In this article, we provide a methodology for the analysis of such data under a wide class of flexible semiparametric transformation models. For inference, a nonparametric maximum likelihood estimation procedure is proposed along with the development of an EM algorithm. Furthermore, we show that the resulting estimators of regression parameters are consistent, asymptotically normal and semiparametrically efficient. A simulation study and a real data application demonstrate that the proposed approach performs well in practice and has substantial superiority over the naive method that ignores the misclassification.

中文翻译：

---

错误分类当前状态数据的回归分析

当每个研究对象仅被观察一次并且观察时的失败状态由具有不完美敏感性和特异性的诊断测试确定时，错误分类的当前状态数据就会发生。在本文中，我们提供了一种在多种灵活的半参数转换模型下分析此类数据的方法。对于推理，随着 EM 算法的发展，提出了一种非参数最大似然估计程序。此外，我们表明回归参数的结果估计量是一致的，渐近正态和半参数有效。模拟研究和实际数据应用表明，所提出的方法在实践中表现良好，并且比忽略错误分类的朴素方法具有实质性优势。