在藥物的第二期臨床實驗中，為了降低受試者用藥後的風險，常採用Simon二階段設計來進行實驗。在這個設計中，實驗進行方式為，在第一階段中受試者用藥後的療效反應之平均達到預先設定值時，才會進行第二階段的實驗。就統計分析師的立場而言，當實驗順利進行完兩個階段，此時第一階段反應之平均會服從某種截斷分布，例如截斷常態分布。若直接採用兩個階段實驗中全部受試者的平均反應，來估計藥物的療效，則由於未考慮到截斷的影響，很可能會產生偏誤。所以本論文將分別在參數方法及非參數方法下修正截斷所造成的影響，並利用模擬方式探討這些方法的準確性。

Generally, Simon two-stage designs are usually employed in Phase II clinical trails to avoid giving patients an ineffective drug. In the design, if the average of patients’ responses at the first stage reaches to a pre-specified value, then the second stage is allowed to continue. From statistician’s viewpoint, the average of responses at the first stage follows some truncated distribution, such as truncated normal distributions. Because of ignoring the truncation effect, the average of all responses in two stage experiments is an overestimated estimator. Therefore, parametric and nonparametric estimation methods are applied to take into account for the truncation effect. Furthermore, a simulation study is conducted to investigate the accuracy of these methods.

Dudweicz, E. J. and Mishra, S. N. (1988). Modern Mathematical Statistics. Wiley. 556.
Efron, B. (1979). “Bootstrap methods: another look at the jackknife.” The Annals of Statistics, 7(1), 1-26.
Lehmann, E. L. (1986). Theory of Point Estimate. Wiley, 1st Ed. 406.
Li, G., Shih, W. J., Xie, T. and Lu, J. (2002). “A sample size adjustment procedure for clinical trails based on conditional power.” Biostatistics, 3, 277-287.
Simon, R. (1989). “Optimal Two-stage Designs for Phase II Clinical Trials.” Controlled Clinical Trials, 10, 1-10.
Tsai, W., Chi, Y. and Chen, C. (2006). “Interval Estimation of Binomial Proportion in Clinical Trials with a Two-Stage Design.” submitted to Statistics in Medicine.