在現狀設限數據的研究中,感興趣的毀壞事件是一次檢查設限的,所得的資訊僅為檢查時間,和在此時間毀壞事件是否發生的狀態指標。對於現狀數據分析,Willamson, Lin and Kim (2009)在基線分布為Weibull的比例風險模型下,提供了一個樣本數計算方法,然而一個現狀數據不必然服從比例風險模型和Weibull基線分布假設。 本論文考慮包含Weibull, Log-Logistic, Log-Normal等等,較具彈性的基線分布(Sparling et al, 2006),在比例勝算模型,提出一個現狀數據的樣本數公式。模擬試驗驗證了公式的正確性,使用了兩個真實研究案例,作為樣本數公式的應用例證。
In the study of current status data, when the failure of interest examined once, the available information are the examination time and the indicator of the failure has occurred or not by the examination time. Willamson, Lin and Kim (2009) proposed a sample size formula for current status data under the proportional hazards model with Weibull baseline distribution. However, a current status sample is not necessary to follow the proportional hazards model or satisfy the Weibull baseline assumption. In this thesis we consider another popular survival model, the proportional odds model, with a flexible baseline distribution (Sparling et al. 2006) includes Weibull, Log-Logistic. Log-Normal distributions. Under this model, we proposed a sample size formula for current status data. Simulations certify the validity of the formula, two real examples illustrate the applications of the method.