摘要
目的:探讨基于改进相对距离的生物相似性评价方法。方法:在仅考虑均数差异的传统相对距离基础上加入变异指标构造改进相对距离,通过模拟试验验证利用Bootstrap方法统计分析的一类错误和检验效能。结果:基于传统相对距离利用Bootstrap方法与常规方法表现相当,基于改进相对距离利用Bootstrap方法能有效控制一类错误,并保证较高的检验效能。结论:基于改进相对距离利用Bootstrap方法统计分析可更全面准确地评价生物相似性。
AIM: To introduce a statistical a- nalysis method based on modified relative distance. METHODS: Join the standard deviation in the rel- ative distance which to construct the modified rela-tive distance and test the type-I error and power using Bootstrap method by Monte Carlo simulation. RESULTS: When simulate based on relative distance, Bootstrap method perform fairly with the conventional method. When simulate based on modified relative distance, Bootstrap method could control the type-I error and maintain the power. CONCLUSION: The Bootstrap method would be a reliable method based on modified relative distance to assess the biosimilarity.
出处
《中国临床药理学与治疗学》
CAS
CSCD
2016年第3期293-297,共5页
Chinese Journal of Clinical Pharmacology and Therapeutics
基金
国家自然科学基金青年基金(81302512)