臨床試驗的統計方法學研究

The Development of Statistical Methodologies for Clinical Trials

劉容慈(Liu, Jung Tzu)

指導教授 ： 鄒小蕙 ； 曾晴賢

國立清華大學/生命科學院/生物資訊與結構生物研究所/博士(2015年)

https://doi.org/10.6843/NTHU.2015.00254

摘要

藥物的開發是一個耗時，非常昂貴，且具有較大風險的過程。在臨床藥物的開發中，大多數經費用於臨床試驗的研究和開發。因此，急需統計策略來提升臨床研究設計、減少所需的樣本數、縮短藥物上市所需的時間，以及增加新藥研發的成功率。在本論文中，我們提出三個新的臨床試驗統計方法學，其中包括：在一個包含安慰劑的三臂試驗 (three-arm trial) 中建立二元反應變數 (binary outcome) 之非劣性 (non-inferiority) 新藥藥效試驗方法學研究；在銜接性試驗中，我們建立一個利用加權組合 (weighted combination) 的統計方法，結合原始區域和銜接性區域的資訊來評估新藥在銜接性區域的藥效；和在多國多區域臨床試驗 (multiregional clinical trials) 中，利用離散型分布的隨機效應模型來評估區域間已知或未知藥效的一致性研究。根據我們提出的方法，推導出所對應需要的樣本數。我們也提出型一錯誤率和檢定力的模擬研究。根據模擬研究的結果，發現我們的方法有穩健的真陽性率，以及適當控制的偽陽性率。我們使用真實的數據和模擬的數據來說明我們提出的方法，並且呈現了從統計觀點和經濟學觀點的應用。此外，所有方法的程式都是在R環境中執行，而且可以透過向作者要求而獲得。

關鍵字

非劣性試驗 ； 銜接性試驗 ； 多國多區域臨床試驗 ； 隨機效應模型 ； 一致性

並列摘要

Pharmaceutical development is a time-consuming, very expensive, and highly risky process. Most of budget for research and development goes to clinical trials in clinical development. Therefore, there are urgent needs of statistical strategies to enhance clinical research design, reduce a required sample size, shorten a duration of drug development, and increase a success rate of new drug development. Three new statistical methodologies of clinical trials are proposed in this dissertation, including establishing non-inferiority efficacy of a new treatment with binary outcomes in a three-arm trial; a weighted combination approach combining information from original region and local bridging region in a bridging study; and assessing a consistency of a known or a unknown treatment effect under a discrete random effects model in multiregional clinical trials (MRCTs). Sample size requirements for the proposed methodologies are derived. Simulation studies of type I error rate and power based on the proposed methods are given. Simulation results show that the methodologies are robust and well-controlled in terms of true-positive and false positive rates. We illustrate the methods using real data sets and simulated data for examples, and then we present applications from both a statistical and an economic point of view. Furthermore, all programs of proposed methods are executed from the R environment and the programs are available by request to the authors.

並列關鍵字

non-inferiority trial ； bridging study ； multiregional clinical trial ； random effects model ； consistency

參考文獻

2. Hsiao CF, Hsu YY, Tsou HH, Liu JP. Use of prior information for Bayesian evaluation of bridging studies. Journal of Biopharmaceutical Statistics 2007; 17:109–121.

3. Lan KKG, Pinheiro J. Combined estimation of treatment effects under a discrete random effects model. Statistics in Biosciences 2012; 4:235–244.

6. Committee for Medicinal Products for Human Use (CHMP). Guideline on the choice of the non-inferiority margin. EMEA/CPMP/EWP/2158/99, 2005. http://www.ema.europa.eu/docs/en_GB/document_library/Scientific_guideline/2009/09/WC500003636.pdf (accessed 1 July 2011).

7. Pigeot I, Schäfer J, Röhmel J, Hauschke D. Assessing non-inferiority in a new treatment in a three-arm clinical trial including a placebo. Statistics in Medicine 2003; 22:883–899.

8. Koch A, Röhmel J. Hypothesis testing in the “Gold standard” design for proving the efficacy of an experimental treatment relative to placebo and a reference. Journal of Biopharmaceutical Statistics 2004; 14:315–325.