收藏 分享(赏)

PDA TR59中英文对照(利用统计学进行产品监控).pdf

上传人:精品资料 文档编号:9739086 上传时间:2019-08-30 格式:PDF 页数:89 大小:2.74MB
下载 相关 举报
PDA TR59中英文对照(利用统计学进行产品监控).pdf_第1页
第1页 / 共89页
PDA TR59中英文对照(利用统计学进行产品监控).pdf_第2页
第2页 / 共89页
PDA TR59中英文对照(利用统计学进行产品监控).pdf_第3页
第3页 / 共89页
PDA TR59中英文对照(利用统计学进行产品监控).pdf_第4页
第4页 / 共89页
PDA TR59中英文对照(利用统计学进行产品监控).pdf_第5页
第5页 / 共89页
点击查看更多>>
资源描述

1、 Utilization of Statistical Methods For Production Monitoring Technical Report No.59 http:/ 制药技术的传播者 GMP理论的践行者 2 目 录 1.0 INTRODUCTION 简介 4 1.1 PURPOSE AND SCOPE 目的和范围 . 4 2.0 GLOSSARY OF TERMS 术语表 6 3.0 STATISTICAL PROCESS CONTROL TOOLS 统计学过程控制工具 15 3.0.1 Prerequisites for Data Analysis 数据分析前提 . 15

2、3.1 RUN CHARTS 运行图 22 3.2 CONTROL CHARTS: INDIVIDUALS 控制图:单值 22 3.3 MOVING RANGE CONTROL CHARTS 移动极差控制图 . 24 3.4 AVERAGE AND VARIABILITY CHARTS 均值图和变异性图 25 3.5 HISTOGRAMS 柱状图 . 26 3.6 PROCESS CAPABILITY (CPK, PPK) 工艺能力( CPK, PPK) . 30 3.7 EXPONENTIALLY WEIGHTED MOVING AVERAGE CHARTS 指数加权移动平均控制图34 3.

3、8 CUSUM CHARTS 累积和图 . 36 3.9 EXAMPLES OF EFFICIENT MIXTURE OF THE STATISTICAL TOOLBOX 统计工具箱有效混合使用的实例 . 38 4.0 ACCEPTANCE SAMPLING 验收抽样 . 40 4.1 TYPICAL APPLICATIONS 典型应用 . 40 4.2 KEY TERMS 关键术语 40 4.3 TYPES OF SAMPLING 抽样类型 . 41 4.4 TYPES OF ACCEPTANCE PLANS 接收方案类型 . 41 4.5 PROS AND CONS 优缺点 42 5.0

4、APPENDICES: TECHNICAL DETAILS AND EXAMPLES 附件:技术细节和实例 44 5.1 RUN CHARTS 运行图 . 44 http:/ 制药技术的传播者 GMP理论的践行者 3 5.2 CONTROL CHARTS: INDIVIDUALS 控制图表:单值 46 5.3 MOVING RANGE CONTROL CHARTS 移动极差控制图 . 52 5.4 AVERAGE AND VARIABILITY CHARTS 均值图和变异性图 57 5.5 HISTOGRAMS 直方图 70 5.6 CPK , PPK FOR PROCESS CAPABILI

5、TY 过程能力 CPK,PPK 值 . 71 5.7 EXPONENTIALLY WEIGHTED MOVING AREA CHARTS (EWMA) 指数加权移动区域图( EWMA) 76 5.8 CUSUM CHARTS 累积和图 . 80 5.9 SAMPLING 抽样 . 86 http:/ 制药技术的传播者 GMP理论的践行者 4 1.0 Introduction 简介 As manufacturers seek to improve the quality of their goods, statistical methods have been rediscovered as v

6、ital tools for successful development and manufacturing. Industries like automotive, electronics, and consumer products grow and change partly as a result of adopting statistical methods. 随着生产商期望提高产品质量,统计方法逐渐成为开发和生产成功的重要工具。汽车、电子、消费品行业也部分因为采用统计方法发展和改变。 The pharmaceutical and biopharmaceutical industr

7、y increasingly recognizes the importance of statistical methods to consistently create products that conform to predetermined quality characteristics. Statistical methods provide objective evidence in meeting this goal and are fundamental for understanding the process, which enables further improvem

8、ent and development. 制药和生物制药工业越来越认识到统计方法的重要性,以持续生产符合预定质量标准的产品。统计方法为符合这一目的提供了客观证据,也是理解工艺的基础,使持续改进和开发成为可能。 Industry and regulatory bodies are working together to provide guidance and frameworks on the use of statistical methods. The International Conference on Harmonization, International Standards Or

9、ganization and European Union have provided guidance on the use of statistical methods. 工业和监管部门正在一起努力建立统计方法使用指南和框架。 ICH、 ISO、 EU 已经有了统计方法的指南。 In light of the increased focus on this topic, this PDA Task Force recognized the need to provide guidance to help companies identify and use statistical meth

10、ods. The primary objective of this Task Force was to convey the appropriate use of statistical methods at a level most can understand. 鉴于人们对这一主题 越来越重视, PDA 工作组认识到应该建立指南帮助公司识别和使用统计方法。该工作组的基本目的是让大多数人能够适当应用统计方法。 1.1 Purpose and Scope 目的和范围 The purpose of this document is to present relevant and easy-to

11、-use statistical process control (SPC) methods that are applicable to the pharmaceutical/biopharmaceutical industry. Advanced statistical methods, such as multivariate models and Design of Experiment (DoE) will not be considered. An overview of acceptance sampling is also included in Section 4.0. 本文

12、件是为了提供的相关、易用的统计过程控制方法,以应用于制药 /生物制药工业。 高级统计方法,如多变量模型和试验设计( DoE)不包 含在内。 4.0 节对接收取样进行了概述。 1.2 Implementation to Support Decision Making 利用统计方法为决策提供支持 Statistical methods are intended to improve the quality of decision-making. They are simply a means to a result. If the manufacturer does not first under

13、stand why it is utilizing a statistical method, problems such as failing to detect important signals or over-detecting unimportant normal variation can occur. Caution should be exercised to first establish the question to be answered and then statistical method to aid in answering the question. 统计方法

14、可用于提高决策质量。它们只是达到结果的一种手段。如果生产商没有理解为何采用统计http:/ 制药技术的传播者 GMP理论的践行者 5 方法,就可能出现无法检测重要信号或过度放大微小、正常的变动。应注意首先确定需要回答的问题,再采用统计方法帮助回答问题。 The statistical methods may be used in an ongoing program to analyze collected data. Timely evaluation of data allows the prompt detection of undesired process variation, wh

15、ich facilitates process understanding and may support responses to control variability. 统计方法也可用于一个持续计划中对收集的数据进行分析。及时进行数据分析能够迅速发现 非预期的工艺变动,增加对工艺的理解,并及时采取措施对变动进行控制。 To best aid the end-user, each statistical method is described in the following format: 为了更好地帮助用户,每一统计方法按以下形式论述 Description 描述 Pros and C

16、ons 优缺点 Typical Applications 典型应用 Technical Details and Examples (see appendices)技术细节和举例(见附件) The guidance contained in this document is not intended to establish mandatory standards for using statistical methods across a products lifecycle. 本指南不用于建立产品生命周期内统计方法应用的强制标准。 http:/ 制药技术的传播者 GMP理论的践行者 6 2.

17、0 Glossary of Terms 术语表 Some of the key concepts are illustrated below; additional explanation will be in subsequent sections 以下介绍了部分主要概念,进一步 解释见后续章节。 Limits 限度 There are different types of limits to be considered. Some are binding (specification limits); others are for orientation (process control

18、limits). 有不同类型的限度,有的具有约束力(质量标准);其他的用于指导目的(过程控制限度)。 Specification limits 质量标准 Specification limits are set by the manufacturer based on therapeutic product and regulatory requirements. ICH Q6A defines specification as: A list of tests, references to analytical procedures, and appropriate acceptance c

19、riteria which are numerical limits, ranges, or other criteria for the tests described. 1t establishes the set of criteria to which a drug substance or drug product should conform to be considered acceptable for its intended use. Conformance to specifications means that the drug substance and / or dr

20、ug product, when tested according to the listed analytical procedures, will meet the listed acceptance criteria. Specifications are critical quality standards that are proposed and justified by the manufacturer and approved by regulatory authorities. (1, 2). 质量标准由生产商基于药品和监管部门要求制订。 ICHQ6A 将质量标准定义为:一个

21、测试清单,参考对应的分析步骤,以及以数值、范围或其他测试标准表示的适当的可接受标准。它建立了一系列标准,为了符合预定用途药物必须符合这些标准。符合标准意味着药物按照所批准的分析步骤检测时,可以符合可接受标准的要求。质量标准是生产商提出,证明其合理性,并得到监管部门批准的关键质量要求 (1, 2)。 Specification limits 质量标准 : Denote the boundary between acceptable and unacceptable, the quality threshold. 指示合格和不合格界限,质量阈值 Describe what the process

22、must achieve. When specifications are exceeded, there is a loss of value, time or cost. 描述了工艺必须达到的要求,当超出限度要求时,就会有价值、时间的损失或成本的增加。 Should be driven by therapeutic effect and toxicological impacts as relevant to the patient. Specifications control risk to the patient 应根据治疗效果和毒理作用制订。质量标准应能控制对患者造成的风险。 Pr

23、ocess Capability 过程能力 Process capability descries how the process performs in relation to the specifications. 过程能力描述了过程如何符 合质量标准相关要求 High Capability Processes 高过程能力 http:/ 制药技术的传播者 GMP理论的践行者 7 High capability processes have low inherent variation relative to the specification or goal. When a number

24、of results are plotted on a histogram, it is unlikely that there will be many result occurrences near the specification limits. 高能力过程与质量标准有关的固有变动较小。当采用一定数量结果作直方图时,不会有许多结果出现在限度附近。 Figure 2.0-1 Example of a High Capability (Low Variability) Process Capability Histogram 高能力(低变动)过程能力图 Process Capability

25、 (cont.) 过程能力(续) Low Capability Processes 低能力过程 Low capability processes have greater inherent variation relative to the specification or goal. When a number of results are plotted on a histogram, it is likely that results will occasionally occur beyond the specification limits. 低能力过程与质量标准相关内在变动较大。当

26、用一系列结果作直方图时 ,结果有可能超出限度范围。 Figure 2.0-2 Example of a Low Capability (High Variability) Process Capability Histogram 低能力(高变动)过程能力图 Statistical Process Control Limits 统计过程控制限度 Statistical process control limits are statistically derived measures that are used to define the typical operating range for t

27、he process. Process control is the focus of this document. Unlike specification http:/ 制药技术的传播者 GMP理论的践行者 8 boundaries which are related to product impact, control limits are boundaries that annunciate when process performance may have shifted. 统 计过程控制限度是用于确定常规工艺操作范围的统计措施。过程控制是本指南的重点。质量标准限度与产品功能有关,而

28、控制限度能提示工艺性能的漂移。 Process control limits 过程控制限度 : Denote the boundary between typical and unusual operational performance ranges for the process. 指示典型工艺操作范围与异常工艺操作范围的界限 Are calculated from prior performance data and used to detect when a process is unstable or “out of statistical control“. 通过历史操作数据计算而

29、得,用于监测工艺是否不稳定,或超出统计控制范围 The prompt recognition of deviations from typical performance enables review. The review can lead to an understanding of how a process may be improved. It is essential to understand that process control boundaries only relate to the ability to discern numerical differences. T

30、hey are not equivalent to specification limits that describe the conformance / non-conformance boundaries. Statistical significance is also not the same as practical significance, which is a difference that has a meaningful impact on the process. Depending on the circumstances of the process, when t

31、here are sufficient samples present, it is possible to detect statistical differences that have no practical importance. It is also possible in processes with tight specification limits and noisy measurement systems for it to be challenging to detect important performance changes. The users of proce

32、ss control tools must apply their understanding of the process when evaluating the relevance of process control detections. 迅速发现偏离典型操作范围有助于及时进行回顾。回顾能够加深对工艺的理解,并改进工艺。关键要理解过程控制界限仅与区分数值差异的能力有关,而质量标准则不同,质量标准规定了符合 /不符合界限。统计学意义与实际意义也不同,这种不同将对工艺造成影响。根据工艺条件不同,当有足够样品时,也可能检测到一些无关紧要的统计学差异。严格质量标准限度和噪声测量系统也可能检测出

33、重大工艺性能变化。使用过程控制工具时,应根据对工艺理解,进行过程控制结果的相关性评估。 Statistical Process Control Limits (cont.) 统计过程控制限度(续) Stable Process (Statistically In Control) 稳定工艺(统计学意义上受控) Stable Process (statistically in control) is a process that is consistent and predictable. It does not exhibit special (or assignable) cause var

34、iation as identified through the use of statistical techniques. The variation present is due to only common cause variation 稳定工艺(统计学意义上受控)是指稳定一致且可预测的工艺。它不会发生特殊(或可指明)原因的变动,这些变动可通过统计技术的应用而发现。现有的变动仅由于正常原因变动构成。 Figure 2.0-3 below shows a stable process (statistically in control). The distribution of the

35、 process is consistent over time (as shown on the left). Each successive outcome is random, but aligns with the expectation for the process. http:/ 制药技术的传播者 GMP理论的践行者 9 下图 2.0-3 显示了一个稳定工艺(统计学意义上受控)。在整个时间范围内结果分布是一致的(如左图所示)。每一个连续结果是随机的,但都与对工艺的预期保持一致。 Figure 2.0-3 Examples of a Stable Process (Statisti

36、cally In Control) 稳定工艺(统计学意义上受控) Unstable Process (Statistically Out of Control) 不稳定工艺(统计意义上不受控) Unstable process (statistically out of control) is a process that is not consistent or predictable. It exhibits special cause variation as identified through the use of statistical techniques. This is al

37、so sometimes called “Out of Trend“, where the current performance is detected to be not part of the prior (level, uniform) trend. Figure 2.0-4 below illustrates an unstable process. The distribution of the process is not consistent over time (as shown on the left). Each successive outcome is random,

38、 but does not align with the expectation for the process. 不稳定工艺(统计意义上不受控)是一种一致或无法预计的工艺。通过应用统计技术,可以发现特殊原因变动。有时也称作“趋势异常”,这时可发现现有结果与以前趋势(水平的, 均一的)不一致。下图 2.0-4 显示的是一个不稳定工艺。在整个时间范围内结果分布不一致(如左图所示)。每一连续结果都是随机的,但与工艺期望值不一致。 Figure 2.0-4 Examples of an Unstable Process (Statistically Out Of Control) 不稳定工艺(统计意

39、义上不受控) http:/ 制药技术的传播者 GMP理论的践行者 10 Statistical Process Control Limits (cont.) 统计过程控制限度(续) Warning or Alert Limits 警戒限 These types of limits are determined and used typically during development, environmental or manufacturing process monitoring for the purpose of detecting trends or to get a better

40、understanding of the manufacturing process and its consistency. These limits are usually not intended to require a formal investigation, a report or an involvement of Quality Assurance in case of detection. In some instances (e.g., microbiological monitoring) a Standard Operating Procedure (sop) can

41、 define follow-up steps that may include actions, such as increased frequency of testing or possibly proactively increased actions to mitigate risk of future impact. 这类限度一般在开发阶段、环境控制或生产工艺控制中制订并使用,以监测趋势或获得对生产工艺及其一致性的进一步理解。超出这些限度一般不需要进行正式的调查,报告或 QA 的参与。某些情况下(例如,微生物监控),可在一个标准操作程序确定后续措施,包括采取的 行动,如增加测试频次

42、或主动降低风险措施 Variability 变动 Variability is categorized here by two types; i.e., common cause and special cause. It is important to recognize the distinctions when understanding a process and taking the appropriate action to improve the process. 变 动分为两类:即普通原因和特殊原因。加强对工艺的理解,并采取措施改进工艺时,重要一点是能够区分这两类变动。 Com

43、mon Cause Variability 普通原因变动 Common cause variation is the result of the combination of all of the typical variability in the materials, process and measurement system. These many small components of variation are expected to be present in the manufacturing process. 普通原因变动是物料、工艺、测量系统所有典型变动共同作用的结果。在生

44、产工艺中可以预期这些微小变动。 Special Cause Variability 特殊原因变动 Special cause variation is a change caused by special circumstances not expected from the process. These occurrences are not predictable, and may come and go sporadically. Special cause events are detectable when compared to statistical control limit

45、techniques, hence the term used to denote this condition is described as “out of statistical control“ (limits). The prompt detection of special causes allows the proper investigation and evaluation of impact to subsequent production or downstream processes. 特殊原因变动是正常生产过程不会发生的特殊情况导致的变更。这些事件是无法预期的,只是偶

46、然发生。与统计控制限度比较,就可发现特殊原因事件,因此这种情况被称为“统计控制超标。迅速发现特殊事件后,可及时进行适当调查,评估对后 续生产或下游工艺的影响。 Trends 趋势 Trends are changes in the average or variability of the result. An analysis of data often exhibits an ongoing upward or downward pattern that is not due to random noise. Analyzing trends is useful in http:/ 制药技

47、术的传播者 GMP理论的践行者 11 detecting patterns that could lead to future quality problems, and in anticipating upcoming performance. Statistical process performance monitoring tools as defined in this document can be used to detect trends using objective numeric tools. 趋势是结果平均值或变动的一种改变。数据分析通常呈现为持续向上或向下的图形,这并

48、不是由于随机噪声 导致的。对趋势进行分析有助于发现可能导致后续质量问题,估计后来的工艺性能。利用客观数据,本文中的统计过程控制工具可用来发现工艺趋势。 In the interpretation of GMP requirements, regulatory authorities increasingly request an assessment of trends in inspections. The prompt detection and evaluation of trends (3) supports the implementation of corrective and p

49、reventive actions (CAPA) as suggested by ICH QI0 (4). There is always an uncertainty if a trend is relevant to product quality. When establishing monitoring practices, the potential impact of process shifts should be considered when balancing risks of failure to detect potential hazards versus the risk of annunciating unimportant changes. “The level of effort, formality and documentation of the quali

展开阅读全文
相关资源
猜你喜欢
相关搜索

当前位置:首页 > 企业管理 > 管理学资料

本站链接:文库   一言   我酷   合作


客服QQ:2549714901微博号:道客多多官方知乎号:道客多多

经营许可证编号: 粤ICP备2021046453号世界地图

道客多多©版权所有2020-2025营业执照举报