【作者】 李娜；

【导师】 谭德讲； 许卉；

【作者基本信息】 烟台大学 ， 药学硕士（专业学位）， 2021， 硕士

【摘要】 药品分析方法替代研究是方法生命周期中重要的一环,是分析方法更替变化的内动力。进行替代研究可以加速新方法的接收转化,丰富药品检测方法,促进药品检测技术的提升,从而确保药品质量和用药安全。替代研究涉及的法规文献众多,随着对替代研究理解的深入,有关替代的法规指南也在不断扩充、发展和完善。但目前仍存在以下问题有待解决(1)药品种类和分析方法众多导致方法替代分类的不明确;(2)方法替代的比较类型及其所用统计分析方法存在科学性问题,导致结果失真;(3)有关分析方法替代的实验设计不严谨;(4)测定方法样本量不足及效能过低导致结果误判;(5)替代研究涉及大量统计学知识,少有明确的可操作性的替代统计操作规范;(6)在替代结果的统计分析方面缺乏简单可靠的工具以供使用。基于以上问题并结合前期的研究进展,开展了以下论文研究工作:1、明确了方法替代的分类。根据产品检测属性分类,主要分为相同属性的替代研究和不同属性的替代研究两种。根据检测结果的数据类型分类,主要分为定量数据的替代研究和定性数据的替代研究两种类型。其中定量数据可细分为3种类型:a)可接受方式、b)性能等效方式、c)结果等效方式。定性数据的比较方式又称为d)结论等效。其中a.可接受方式是指新方法经过严格的验证后即可替代原方法的方式。这种方式会大大增加方法的误判风险;b.性能等效是指两种方法在经过验证且确保专属性的基础上,在使用范围内对准确度和精密度分别进行等效性和非劣效的区间估计方式,以便掌握新方法是否在性能指标上等效或优于原方法;c.结果等效是使用两种方法对某些样品检测的连续性结果进行定量数据的比较;d.结论等效是使用两种方法对某些样品检测指标的结论进行定性数据的比较分析。以上替代类型几乎适用于目前替代研究统计比较的所有情形,进一步促进替代研究科学化发展。2、引入了更科学有效的检验方式。以往对两方法的比较采用差异性检验的方式,本论文创新性的将等效性检验引入到方法的替代研究领域中,特别是在考察方法的性能指标时,进行了准确度的等效性检验(同质均匀样本等效性检验公式为:(?)、非同质均匀样本的等效性检验公式为:(?))。同时也考察了精密度的非劣效检验(同质均匀样本的非劣效检验公式为:(?)、非同效检验质均匀样本的非劣性公式为:(?))。两种检范围验有效地证明新方法与原方法在区间范围内的一致性。进一步规范了替代研究的比较标准,即认定与原方法比较新方法只有等效或部分非劣效才能说明两方法可以替代。3、明确了方法替代的实验设计方案、样本量和效能问题。进行替代研究应事先制定合理可行的设计方案,根据替代类型的差异对应不同的设计方案,在性能等效中提供了均质和非均质供试品两类实验方案。在结果等效中对非独立样本进行重复测定的实验设计方案。样本量的考察也是实验设计中的关键要素,样本量小或效能低均无法充分说明所得结论的可靠性,本论文根据实验设计不同提供了样本量和效能的计算公式,从而确保替代实验的严谨性。4、构建了首款专门针对药品领域方法替代研究的统计比较软件。基于前期理论研究,结合计算机科学开发了替代研究的统计软件(ALTER),解决了目前方法替代统计计算困难和统计评价指标不足的问题,同时,通过与法定示例和专业统计软件进行替代结果比对验证,证实所开发软件结果和结论的可靠性和权威性,为今后方法替代研究提供了方便实用的统计工具。总之,本论文是将统计分析与药品领域中的方法替代理论交叉的跨学科研究,为药品领域中替代研究的发展应用提供新思路,为替代研究统计规范化提供新模板,为今后实验室进行替代研究提供新工具。更多还原

【Abstract】 The study of analytical method alternative is an important part of the life cycle of analytical methods and the internal force of the change of analytical method alternative.Alternative research can accelerate the acceptance and transformation of new procedures,enrich drug detection methods and promote the improvement of drug detection technology,so as to ensure drug quality and drug use safety.Alternative research involves many regulatory documents.With the deepening of understanding of alternative research,the relevant regulatory guidelines are constantly expanding,developing and improving.However,there are still the following problems to be solved at present.(1)There are many types of drugs and analysis methods,which lead to unclear alternative classification of methods;(2)There are scientific problems in the comparison options and the statistical analysis methods used,which lead to the distortion of the results;(3)The experimental design of alternative analytical methods is not rigorous;(4)Insufficient sample size and low efficiency of the determination method lead to misjudgment of results;(5)Alternative research involves a lot of statistical knowledge,and there are few clear and operable alternative statistical operation norms;(6)There is a lack of simple and reliable tools for statistical analysis of alternative results.Based on the above problems and combined with the previous research progress,the following research works have been carried out1.1.Defining the classification of alternative method.According to the classification of product detection attributes,it is mainly divided into two kinds:the alternative research of the same attribute and the different attribute.According to the data types of detection results,it is mainly divided into two types:quantitative data alternative research and qualitative data alternative research.Quantitative data can be subdivided into three types:a)acceptable approach,b)performance equivalent approach,and c)result equivalent approach.The comparison of qualitative data is also called d)conclusion equivalence.a.Acceptable approach refers to the approach in which the new method can alternative the old method after strict validation.This approach greatly increases the risk of miscalculation of the method;b.Performance equivalence refers to the interval estimation of the equivalence and non-inferiority of accuracy and precision respectively within the range of use on the basis of the validation and specificity of the two methods,so as to know whether the new method is equivalent or better than the old method in terms of performance indicators;c.Results equivalence refers to the alternative of quantitative data on the continuous results of some samples detected by two methods;d.Conclusion equivalence refers to the alternative analysis of qualitative data on the conclusions of some sample test indicators using two methods.The above types of alternative are almost applicable to all the situations of the current statistical comparison of alternative research and further promote the scientific development of alternative research.2.Introducing a more scientific and effective way of test.In the past,difference test was used to compare the two methods.In this paper,equivalence test was innovatively introduced into the field of alternative research of the method.Equivalence test of accuracy was carried out when examining the performance indicators of the method.(The equivalence test formula of homogeneous materials is:(?)The equivalence test formula of non-homogeneous materials is:(?).At the same time,the non-inferiority test of precision was also investigated(The non-inferiority test formula for homogenous materials is:(?)The non-inferiority test formula for non-homogeneous materials is:(?)The two tests effectively prove the consistency between the new method and the old method in the range of interval.The comparison standard of alternative research is further standardized,that is,the new method can be alternatived only if the new method is equivalent or partially non-inferiority compared with the old method.3.Defining the experimental design protocol,sample size and power of the alternative research.Reasonable and feasible design schemes should be developed in advance for the study of alternative,and different design protocols should be developed according to the difference of alternative options.In the performance equivalence,two kinds of experimental protocols of homogeneous and non-homogeneous materials are provided.An experimental design for repeated determination of independent samples in result equivalence.The investigation of sample size is also a key element in experimental design.Neither small sample size nor low efficiency can fully explain the reliability of the conclusions obtained.This paper provides the calculation formulas of sample size and power according to different experimental designs,so as to ensure the rigor of alternative experiments5.Building the first statistical comparison software specifically aimed at the research of alternative methods in pharmaceutical field.Based on previous theoretical research,the combination of computer science to develop the study of alternative statistical software(ALTER),the solution to the current alternative statistical computing difficulty and the problem of insufficient statistical evaluation index,at the same time,with the legal sample and the professional statistical software for alternative results compare,verify the reliability of software is developed on the results and conclusions and authority,for the future of alternative research provides a convenient and practical method statistical toolsIn conclusion,this paper is an interdisciplinary study that intersects the statistical analysis with the method alternative theory in the field of drugs,and provides new ideas for the development and application of alternative research in the field of drugs,a new template for the statistical standardization of alternative research,and a new tool for future alternative research in laboratories.更多还原