DataIntegrity: 2020 FDA Data Integrity Observations in Review

数据完整性：2020年FDA数据完整性缺陷分析

In 2020 thusfar, numerous warning letters have been issued from the FDA citing dataintegrity violations. Common findings violate principles of ALCOA+, 21 CFR Part11, and the FDA’s data integrity guidance document. Some of these findingsinclude:

2020年至今，FDA以数据完整性违规为由发出了许多警告信。常见发现违反了 ALCOA+、21 CFR  Part 11 和 FDA 数据完整性指南。其中一些缺陷包括：

• Deletion or manipulation of data

• 数据删除或操纵

• Aborted sample analysis withoutjustification

• 无正当理由中止的样品分析

• Invalidated OOS results without justification

• 判定 OOS 结果无效而未经论证

• Destruction or loss of data、

• 数据缺失或丢失

• Failure to document workcontemporaneously

• 未能在工作时同时记录

• Uncontrolled documentation

• 不受控的文件

Data integrity is of utmost importance inany CGMP manufacturing setting now more than ever. While data integrityguidance is not a new concept, there is increased scrutiny of data integrityand 21 CFR Part 11 compliance as electronic records have become the industrystandard. For manufacturers trying to attain a high degree of data integrity,the FDA’s Data Integrity and Compliance with Drug CGMP Guidance for Industrydocument is a great tool to frame an approach to vigorous internal policies.This document is repeatedly referenced in FDA warning letters when companiesare in violation of principles stated in the guidance. This document encouragesmanufacturers to implement effective and robust strategies to ensure thataccurate and secure data management systems are in place and routinelymonitored by the quality unit.

数据完整性在任何 CGMP 制造环境中都比以往更加重要。虽然数据完整性指南并不是一个新的概念，但随着电子记录已成为行业标准，对数据完整性和 21 CFR Part 11符合性的监管也日益加强。对于试图获得高度数据完整性的制造商来说，FDA 的数据完整性和符合药物 CGMP 行业指南是制定实施有力内部政策的方法的重要工具。当公司违反指南中规定的原则时，FDA 警告信中反复引用这些文件。本文鼓励制造商实施有效和稳健的战略，以确保准确和安全的数据管理系统到位，并由质量部门定期监控。

Anotherstandardized tool referenced in numerous data integrity guidance documents formanaging data integrity risks is ALCOA+. ALCOA+ provides a set of principles tobe applied to all CGMP records; Attributable, Legible, Contemporaneous,Original, Accurate, Complete, Consistent, Enduring, and Available. AlthoughALCOA+ is not the final word for data integrity compliance, using theseprinciples to challenge data generation and management strategies provides astrong framework for internal data management policies. Managing data integrityrisks is essential to the safety, purity, quality, and efficacy of drugproducts.

许多数据完整性指南中引用的另一个用于管理数据完整性风险的标准化工具是ALCOA +。ALCOA +提供了一套适用于所有CGMP记录的原则：可归属，清晰，同步，原始，准确，完整，一致，持久和可用。尽管ALCOA +并不是保证数据完整性合规性的最终词汇，但使用这些原则来对标数据生成和管理策略可为内部数据管理策略提供强大的框架。管理数据完整性风险对于药品的安全性，纯度，质量和有效性至关重要。

21 CFR Part11 is also exceptionally important for data integrity of electronic records andelectronic signatures. 21 CFR Part 11 is a legally binding statute thatcompanies must comply with when generating electronic records. There are manyways that validated instruments, software, and procedures can manage andmitigate data integrity risks and demonstrate compliance to 21 CFR Part 11.Software should be chosen that does not allow for data modification without apermanent record and that creates records attributable to who did the work,when, and why. Once an appropriate software is chosen, it should be validatedto the rigors of 21 CFR Part 11 and data integrity guidance to activelydemonstrate complete compliance and secure data. This year alone, numerouswarning letters have been issued for data integrity violations. Many are obviousviolations while some are discreet and could go undetected. The following areexamples of data integrity observations issued from the FDA thus far in 2020and considerations for avoiding such observations.

21 CFR Part 11对于电子记录和电子签名的数据完整性也非常重要。21 CFR Part11是具有法律约束力的法规，公司在生成电子记录时必须遵守该法规。经过验证的仪器，软件和程序可以通过多种方式来管理和减轻数据完整性风险，并证明其符合21 CFR Part11的规定。应选择在没有永久记录并创建可归因于谁？何时？什么原因？的记录情况下不会允许进行数据修改的软件。一旦选择了合适的软件，就应该通过21 CFR Part11以及数据完整性指南的严格验证，以证明完全符合和数据安全。仅今年，就违反数据完整性的问题发出了许多警告信。许多是明显的违规行为，而有些则是谨慎不易发现的。以下是FDA到目前为止在2020年发布的数据完整性缺陷项的示例，以及避免此类缺陷的注意事项。

In March ofthis year a manufacturing site was found to be intentionally shredding anddestroying records. Additionally, the site failed to integrate certainchromatography peaks due to disabled peak detection, which was against itsprocedure. Furthermore, investigation into unknown chromatography peaks was notconducted per procedure. This same company produced OOS results that were notproperly investigated and replaced the OOS with a resample repeatedly,resulting in three failed results and an insufficient laboratory investigation.This occurred before accepting the fourth passing result. Based on guidance,data should never be destroyed, replaced, or overridden. Data should beprotected and enduring. When data result in an unknown or OOS, the event shouldbe documented and thoroughly investigated. Corrective and preventative actionsshould also be taken to protect the data lifecycle.

在今年3月，FDA发现一个制造工厂故意粉碎并破坏了记录。另外，由于禁用了峰检测，该工厂未能积分某些色谱峰，这与该公司的程序不符。此外，没有按照程序对未知色谱峰进行研究。该公司产生的OOS结果未经适当调查，并反复用重新取样代替OOS，导致三次测试结果不合格，实验室调查不足，然后在第四次得到合格结果并接受该结果。根据指南，绝不能销毁，替换或覆盖数据。数据应受到保护并持久。当数据导致未知或OOS时，应记录该事件并进行彻底调查。还应采取纠正和预防措施来保护数据生命周期。

Another FDAobservation this year found employees signing logbooks and batch records thatshould have been reviewed in weeks past. This clearly violates the“contemporaneous” requirement of ALCOA+. The gathering, review, and release ofdata should be documented contemporaneously, that is, at the time the work isperformed. Procedures should dictate this requirement and analysts should betrained to document work at the time of completion to avoid discrepancies. Manysoftware platforms take the work out of this for users by permanently recordingwho performed the work, when, and why in a secure audit trail – all in realtime, with the inability to edit. These are all examples of overt dataintegrity violations. It is critical for any CGMP site to review, maintain,secure, and retain CGMP documentation. While many of these observations areovert violations, and some even appearing intentional, many violations areinadvertent and difficult to detect.

FDA今年的另一项缺陷项，员工签署了本已在过去几周内完成审查的日志和批记录。这显然违反了ALCOA +的“同步”要求。数据的收集，审查和发布应同时（即在执行工作时）记录。程序应规定此要求，分析人员应进行培训，以便在完成工作时记录其工作，以免出现差异。许多软件平台通过在安全的审计追踪上永久记录谁执行这项工作，何时以及为什么，为用户解决了这一问题，而这些工作都是实时进行的，并且无法进行编辑。这些都是公开数据完整性违规的示例。对于任何CGMP工厂而言，查看，维护，保护和保存CGMP文件记录都是至关重要的。尽管这些缺陷中有许多是明显违规，甚至有些是故意的，但也有不少是无意的，很难发现的。

The FDAstated a recent finding from another company for aborted chromatography sampleset runs. Although aborting a run is not necessarily forbidden, in the rareinstances where a run must be aborted, there should be clear procedures andaborted runs should be investigated thoroughly. These actions were notperformed by this company for the aborted runs. The same warning letter alsostated deleted data as a finding. Data should never be deleted and everythingpossible should be done to secure and retain all CGMP data. Again, manysoftware platforms eliminate the risk of deleting data all together by makingthis action prohibited. A warning letter issued to another company found thatthe company was printing uncontrolled copies of batch records and maintenancesheets. These copies were pre-printed and kept in an unsecured area. The samecompany could not locate raw data from a standard curve from an analyticalassay. Analysts and operators should be properly trained to retrieve all datapertaining to CGMP processes for inspectors or to make important qualitydecisions regarding batch disposition. They were also found to be usingscrappieces of paper to record data of an analytical assay. Another company wasfound unable to produce laboratory data on product batches, violating theavailable principle of ALCOA+. A laboratory technician reported that raw datais routinely discarded. These observations all pose many data integrityconcerns — data on uncontrolled scraps of paper and loss of data violate almostevery principle of ALCOA+. Documents should be controlled and in original formto avoid loss or obstruction of original data.

FDA对另一家公司的缺陷项是色谱样品组运行中止。尽管不一定非要中止运行，但在极少数情况下必须中止运行的情况下，应有明确的程序，并且应彻底调查所中止的运行。该公司未针对已中止的运行执行这些操作。同一封警告信还指出数据删除的缺陷。永远不要删除数据，应该尽一切可能保护和保存所有CGMP数据。同样，许多软件平台通过禁止执行此操作来消除删除数据的风险。向另一家公司发出的警告信发现该公司正在打印不受控制的批记录和维修表格的副本。这些副本已预先打印并保存在不安全的区域。该公司还无法从分析测定的标准曲线中找到原始数据。应当对分析人员和操作人员进行适当的培训，以为检查员检索与CGMP流程有关的所有数据，或就批处理做出重要的质量决策。还发现他们使用废纸来记录分析测定的数据。另一家公司被发现无法提供有关产品批次的实验室数据，这违反了ALCOA +原则。实验室技术人员报告说，原始数据通常被丢弃。这些缺陷项都带来了许多数据完整性问题-不受控的废纸数据和数据丢失几乎违反了ALCOA +的每一项原则。文件应受控并保持原始格式，以避免丢失或阻塞原始数据。

A securequality system must be in place to prevent any intentional or unintentionalloss of CGMP data. Many software programs used with analytical instrumentsprevent loss of data under any circumstance, mitigating risks for such dataintegrity violations mentioned above. Software programs offer audit trails,various user permissions, inability to delete data, and many other features makingit simple to automate compliance to data integrity guidance and 21 CFR Part 11.The bottom line is that data, paper or electronic, must be retained to makeappropriate quality decisions and demonstrate proper control. All the abovefindings were referenced from warning letters issued thus far in 2020 alone.

必须建立安全的质量体系，以防止CGMP数据有意或无意的丢失。分析仪器使用的许多软件程序在任何情况下都可以防止数据丢失，从而减轻了上述违反此类数据完整性的风险。软件程序提供了审计追踪，多种用户权限，数据无法删除以及许多其他功能，使自动遵守数据完整性指南和21 CFR Part11变得很简单。最重要的是，必须保存数据，纸质或电子数据，做出适当的质量决策并证明适当的控制。以上所有发现均参考仅在2020年发布的警告信。