Methodology for Case-Control Studies with Prevalent Cases
针对现患病例的病例对照研究，提出一种调整生存相关偏倚的方法，仅需现患病例的诊断时间，利用加速寿命模型和拟似然估计，无需生存分布假设。
A method is developed for the analysis of case-control studies with prevalent cases. The method permits adjustment for potential bias induced by association of the exposure of interest and/or covariates with survival. The only additional information required is the time since diagnosis of each prevalent case. The method employs an accelerated life model, the parameters of which are estimated using quasi-likelihood techniques. In this way, no assumptions are required about the random component of the survival distribution. The method is shown to be typically less efficient than a study of similar size using incident cases. However, the ease of obtaining prevalent subjects offers considerable potential for assembling much larger samples of prevalent cases, leading to much more efficient studies overall.