摘要
目的探讨设计以率作为终点评价指标的单组目标值试验时,不同样本量计算方法间的区别。方法通过样本量计算原理与计算结果的比较,分析不同样本量计算方法间的差异,进一步通过Monte Carlo随机模拟方法,探讨使用不同方法所得样本量估计实际的检验把握度,验证所得结果的正确性。结果当预计事件发生率p和目标值p0相差10%时,近似正态法和一般精确概率法所得样本量和把握度较相近,但是当p接近1.0时(p>0.85),精确概率法所得样本量略低于近似正态法,且把握度明显低于近似正态法。小概率事件的精确概率法所需样本量始终低于近似正态法和一般精确概率法。随机模拟显示,在相同的参数设置下,近似正态法给出的样本量能够提供足够的研究把握度,而小概率事件的精确概率法所得样本量能提供的把握度过低。结论如果要检验某医疗器械的使用成功率是否不低于某个临床认可的标准,按照近似正态法计算的样本量,更能提供足够的检验把握度,尤其对于小规模的临床试验。
Objective To explore the differences of sample size calculation methods used for single arm OPC trials in which primary endpoint is proportion. Methods Sample size calculation principles and results were directly compared. Monte Carlo random simulation was used to validate the results. Results When expected event rate p and target value P0 has a 10% difference, sample size and power obtained from normal approximation method and binomial exact method were similar. However, when p was close to 1.0 （ P 〉 0. 85 ） , sample size and power obtained from binomial exact method were significantly lower than normal approximation method. Sample size obtained from rare event exact method was constantly smaller compared to normal approximation method and binomial exact method. Results from simulation showed that normal approximation method could offer sufficient power, but power obtained from rare event exact method was too low. Conclu- sions If the hypothesis of study was to demonstrate the success rate of certain medical device was larger than a clinical ap- proved target value, the normal approximation method provides strong statistical power, especially for small clinical trials.
出处
《中华疾病控制杂志》
CAS
北大核心
2013年第11期993-996,共4页
Chinese Journal of Disease Control & Prevention
关键词
单组目标值试验
样本量
近似正态法
精确概率法
ingle aim OPC clinical trial
Sample size
Normal approximation method
Binomial exact method
