Implementation of a Data Reliability Program: Implementation of a Data Reliability Program

Dedication

For Lizette, Mikhail, István, Christian, and Mikhail Jr

Disclaimer

The regulatory requirements and guidelines detailed in this book primarily focus on the lifecycle data of medicinal products for human use. However, it is important to recognize that the topics discussed are equally pertinent across various industries that handle critical data.

The recommendations for implementing a data reliability program presented in this book are based on the author's extensive experience in data management, spanning thirty-seven years.

While these suggestions aim to provide valuable insights, they should not be interpreted as formal processes mandated by relevant global regulatory authorities, standards, or guidelines. Furthermore, this book is not intended to establish a definitive paradigm or model.

Table of Contents

Dedication

Disclaimer

Preface

Chapter 1 Introduction

Chapter 2 Implementation of a Data Reliability Program

Chapter 3 Define Objectives and Goals

Chapter 4 Establish Governance and Ownership

Chapter 5 Data Inventory and Classification

Chapter 6 Data Quality Assessment

Chapter 7 Data Quality Standards

Chapter 8 Data Cleansing and Transformation

Chapter 9 Data Documentation

Chapter 10 Data Governance Framework

Chapter 11 Data Access Control Systems

Chapter 12 Data Monitoring and Auditing

Chapter 13 Data Lifecycle Management

Chapter 14 Continuous Improvement

Chapter 15 External Data Sources

Chapter 16 Reporting and Communication

Chapter 17 Compliance and Regulation

Chapter 18 Data Recovery and Disaster Planning

Chapter 19 Performance Metrics and Key Performance Indicators

Chapter 20 Feedback and Iteration

Chapter 21 Scalability and Future Readiness

Chapter 22 Data Reliability Implementation Program Applicable to Outsourced Activities

Chapter 23 Summary

Appendix I Glossary of Terms

Appendix II Abbreviations and/or Acronyms

Appendix III References

Appendix IV  Data reliability through various standards and frameworks

Appendix V Data Reliability in Medicinal Products Manufacturing

Appendix VI Are Data Quality and ALCOA attributes equivalent?

About the Author

Preface

Understanding and prioritizing data integrity, reliability, and quality are not just crucial for effective data management, but they also form the backbone of our profession. As data managers, data governance officers, and data quality assurance analysts, we bear the responsibility of ensuring that the data we work with is accurate, consistent, and trustworthy.

Data integrity [1]

Data integrity is not just a technical term; it is the bedrock of data management. It is the key to the accuracy and consistency of data throughout its lifecycle, ensuring that data is not corrupted, altered, or tampered with in unauthorized or unintended ways. For instance, a data integrity violation could occur if a hacker gains unauthorized access to a database and alters the data, leading to inaccurate results and potentially damaging decisions. Understanding and prioritizing data integrity is crucial for effective data management.

Data integrity is a technical aspect of data management and a fundamental principle that ensures the trustworthiness of data. It protects data from unauthorized modifications or errors, ensuring reliability and security.

Data integrity mechanisms, such as validation rules that check data accuracy and consistency, access controls that limit who can view or modify data, and encryption that protects data from unauthorized access, help safeguard data against integrity violations.

Data reliability [2]

Data reliability is not just about trust; it is about practicality. It is the assurance that data can be counted on to be accurate and consistent over time, making it a valuable asset in any data-driven organization. This practicality is what makes data reliability so important in our daily activities. It is not just a concept but a tool we can rely on in our work.

Data integrity is a prerequisite for data reliability. Data with high integrity, meaning it is accurate and consistent, can be relied upon to remain accurate and consistent over time, which is a key aspect of data reliability. In simpler terms, data integrity ensures that the data is accurate and consistent at all times. In contrast, data reliability ensures that it can be trusted to remain accurate and consistent over time, even as it is used in various transactions and activities.

Data reliability encompasses factors such as accuracy, consistency, completeness, and timeliness, collectively contributing to the overall quality of the data.

Data quality [3]

Data quality encompasses not only accuracy and completeness, but also other aspects. It is a comprehensive concept that encompasses data integrity and reliability. It focuses on the overall excellence of data, encompassing its accuracy, completeness, consistency, relevance, and timeliness, ensuring no aspect of data is left untouched. This all-encompassing nature is what makes data quality so valuable. It is not just about one aspect but the overall condition of data and its ability to meet users' needs and expectations.

Data quality efforts aim to ensure data fits its intended purpose, making it valuable and actionable.

Data integrity and data reliability are critical components of data quality. Without data integrity, data cannot be trusted, and without reliable data, the quality of data is compromised.

In summary, data integrity is a foundational element that ensures data is protected from unauthorized changes. In contrast, data reliability is built upon this integrity, ensuring that data is trustworthy and dependable. On the other hand, data quality is a comprehensive goal encompassing integrity and reliability, focusing on producing secure, consistent data that is valuable and fit for use in decision-making and analysis.

Organizations typically implement a combination of data governance, data management practices, data validation processes, and security measures to achieve high data quality, maintain data integrity, and ensure data reliability. Ultimately, the synergy of these three concepts is essential for organizations to make informed decisions, conduct meaningful analyses, and maintain the credibility and usefulness of their data assets.

This book intends to guide readers in effectively managing data and e-records reliability vulnerabilities and raise essential compliance.

Enjoy the reading. If you have suggestions for improvement or questions, send them to [email protected].

Orlando López

SME - E-Records Quality

References.

[1] Data integrity is the property of data that has not been retrieved or altered without authorization since creation and until disposal (NIST SP 800- 57P1, IEEE, ISO-17025, INFOSEC, 44 USC 3542, 36 CFR Part 1236, and other standards).

[2] Data reliability is not just about trust; it is about practicality. It is about being able to depend on data for subsequent transactions and activities.

[3] Data quality encompasses not only accuracy and completeness, but also other aspects. It is about the overall condition of data and its ability to meet users' needs and expectations. It is about being reliable, trustworthy, and suitable for analysis, decision-making, and other purposes.

Additional Readings.

HMA, "Reliability" in Data Quality Framework for EU Medicines Regulation, October 2023.

Chapter 1

Introduction

Data [1] is considered reliable if it provides a complete and accurate representation of subsequent transactions or activities and if it can be relied upon during these transactions or activities [2]. In the context of medicine manufacturing, where the quality and safety of the products are of utmost importance, data reliability plays a crucial role in ensuring the efficacy and compliance of the manufactured medicines.

Data reliability is not just about consistency and repeatability; it is about trust. It denotes the consistency and repeatability of data over time, regardless of the methods or tools used to collect it. A dataset is considered reliable if it consistently produces the same results, regardless of whether it is measured multiple times or obtained by different researchers or methods. For instance, if an experiment yields the same results every time it is performed, it is considered reliable.

Data reliability is fundamentally compromised when there is a failure to record or maintain complete and accurate records of test results or conditions associated with all tests. Furthermore, the lack of reliable data compromises the quality unit’s (QU) ability to ensure compliance with applicable standards.

Amman Pharmaceutical Industries US FDA Warning Letter, MARCS-CMS 668867, February 2024.

Data reliability is fundamentally compromised when there is a failure to record or maintain complete and accurate records of test results or conditions associated with all tests. Furthermore, the lack of reliable data compromises the quality unit’s (QU) ability to ensure compliance with applicable standards.

Amman Pharmaceutical Industries US FDA Warning Letter, MARCS-CMS 668867, February 2024.

Reliable data is one of the fundamental principles of data quality. Data quality measures how well data meets users' requirements and expectations for its intended purpose. Data reliability refers to the accuracy and truthfulness of the data.

In addition to reliability, several attributes govern data quality, including accuracy [3], completeness [4], conformity [5], consistency [6], objectivity [7], suitability [8], integrity, validity [9], source authentication [10], and timeliness [11] with the intended use. These attributes ensure that the data is not only reliable but also trustworthy [12], instilling a sense of security and confidence in the work being done.

Manufacturers operate under a quality system as critical players in the data reliability ecosystem. Their responsibility is not just to develop and document control procedures but to complete, secure, protect, and archive records, including data that provides evidence of their operational and quality system activities [13]. This empowerment to contribute significantly to data reliability in their operations is a testament to their integral role in ensuring data accuracy and integrity. Manufacturers play a key role in ensuring data reliability by implementing robust data collection and management processes, contributing to the quality and safety of the medicines they produce. Their actions directly influence the reliability of the data used in subsequent transactions and activities, making them key stakeholders in the data reliability process.

The reliability program ensures that a product or system functions consistently without failure over a specified period under normal operating conditions. The data obtained from the products or systems is reliable and trustworthy, providing evidence of the operational and quality system activities. This data can be used confidently in subsequent transactions and activities.

Reliable data is accurate, authentic [14], and has integrity; it is also highly usable [15]. This usability enables it to be used confidently for decision-making, analysis, and other critical purposes. For instance, in manufacturing medicine, reliable data can help predict equipment failures, optimize production schedules, and ensure compliance with regulatory standards. Data reliability in these processes cannot be overstated, as it forms the bedrock of trustworthy and effective decision-making, providing confidence and security in your actions. For example, reliable data can be used to predict the failure of critical manufacturing equipment, enabling timely maintenance and preventing production delays. Similarly, it can be used to optimize production schedules, ensuring efficient use of resources and meeting market demands. Lastly, it can ensure compliance with regulatory standards, avoid penalties, and maintain the manufacturer's reputation.

For features that ensure reliable data maintenance on your computer systems, refer to Trustworthy Computer Systems [16].

Book Chapters and Appendices.

This book is divided into twenty-two chapters and three appendices. It discusses the key elements to be addressed in a Data Reliability Implementation Program (DRIP), with a specific focus on production and quality control systems in medical manufacturing environments. Each chapter and appendix is designed to provide comprehensive guidance on different aspects of data reliability, from setting objectives to managing external data sources.

A reliability program ensures that a product or system functions consistently without failure over a specified period under normal operating conditions. Chapter 2 provides a detailed account of the DRIP.

Chapter 3, Define Objectives and Goals, emphasizes the significance of setting clear objectives and goals for a DRIP. These objectives and goals are pivotal in guiding the efforts and providing a framework for implementing and sustaining data reliability initiatives within an organization. This chapter is crucial to achieving data reliability.

Chapter 4 outlines the identification of individuals or teams responsible for overseeing the data reliability program and guiding efforts and initiatives to ensure data reliability within an organization.

Data inventory and classification processes enable organizations to understand better and manage their data assets. They ensure that data is adequately treated based on its significance, sensitivity, and importance to business operations. Chapter 5 discusses data inventory and classification.

Data Quality Assessments are an essential component of a Data Reliability Program. They are crucial in improving decision-making, building trust, ensuring compliance, enhancing operational efficiency, reducing costs, promoting data integration, increasing customer satisfaction, and driving continuous improvement. Chapter 6 provides the framework for the Data Quality Assessments.

Chapter 7, Data Quality Standards, defines the data quality standards and metrics that align with the organization's goals. These standards include accuracy, completeness, consistency, timeliness, and integrity.

Chapter 8 outlines the activities related to data cleansing and transformation. These activities focus on enhancing the quality and usability of the data, which is crucial for ensuring that the data is accurate, consistent, and meets the required standards for reliable analysis and informed decision-making.

Documentation is a critical aspect of data engineering. It refers to creating, maintaining, and organizing records for various aspects of the workflow and ensuring that these processes are well-understood, scalable, and maintainable. Documentation provides detailed information about an organization’s structure, content, and data usage. Chapter 9 discusses the purpose of documentation and the benefits of the data reliability program.

Chapter 10 outlines the Data Governance Framework, a structured methodology designed to manage an organization's data quality, reliability, and security throughout its entire data lifecycle.

A robust data access control system is crucial for maintaining the confidentiality, integrity, and availability of sensitive information. It is an essential component of overall information security strategies. Chapter 11 deals with data access control.

Chapter 12 discusses Data Monitoring and Auditing, which are crucial aspects of a reliability program. This is especially true in industries such as pharmaceutical manufacturing, where the dependability and performance of systems, processes, or products are crucial.

Chapter 13 discusses Data Lifecycle Management (DLM), a comprehensive approach to managing data from its creation or acquisition to its archival and disposal. DLM involves defining policies and procedures for data retention that comply with legal requirements and the typical stages in the data lifecycle.

Chapter 14 focuses on Continuous Improvements within the context of a data reliability program. Continuous improvements involve an ongoing commitment to refining processes, addressing issues, and optimizing the overall data ecosystem to ensure that data remains reliable, accurate, and valuable for decision-making. This emphasis on continuous improvement should reassure you of the forward-thinking nature of the data reliability program.

In the Data Reliability Program, External Data Sources refer to any data obtained from outside the organization or system where the data reliability program is practiced. While the benefits of External Data Sources are significant, organizations must implement robust processes within their data reliability programs to ensure the accuracy, reliability, and ethical use of external data. Proper data validation, documentation (Chapter 9), and ongoing monitoring (Chapter 12) are crucial for effectively managing external data within a data reliability framework. Chapter 15 explains robust processes applicable to external data sources within the context of a data reliability program.

In data management and business operations, compliance entails ensuring that an organization adheres to the applicable laws, regulations, and ethical frameworks relevant to its activities. Ensuring compliance is necessary for maintaining data reliability, security, and privacy. Chapter 16, Reporting and Communication, discusses this topic.

Chapter 17 covers Compliance and Regulations during the implementation of a data reliability program. Integrating compliance and regulation into the Data Reliability Program enables organizations to meet legal requirements, enhances data reliability and security, and fosters a trustworthy and responsible data environment.

Chapter 18 discusses data recovery and disaster planning, critical components of a comprehensive data reliability program. These elements ensure that an organization can effectively recover its data in the event of unexpected incidents or disasters.

Chapter 19 outlines the KPIs. In the context of a Data Reliability Program, defining key performance indicators (KPIs) and measurable metrics is a quantifiable indicator used to assess the effectiveness, performance, and quality of data-related processes and outcomes. The metrics discussed in this chapter provide a basis for measuring progress, identifying areas for improvement, and ensuring that the objectives of the data reliability initiative are met.

Chapter 20 discusses feedback and iteration, which achieve continuous improvement and adaptability in a Data Reliability Program.

Scalability and future readiness, as discussed in Chapter 21, are key features of a Data Reliability Program, particularly in the rapidly evolving landscape of data management and analytics.

A DRIP applicable to outsourced activities, as outlined in Chapter 22, is essential for organizations that rely on outsourced manufacturing, testing, or data processing, where data quality can have a direct impact on decision-making, regulatory compliance, and overall business performance.

This book applies to any data-driven industry. The author selected the medicinal products manufacturing industry as an example. In this environment, properly recorded information serves as the basis for manufacturers to ensure the identity, strength, purity, and safety of their products. This book highlights the implementation of data suitability, associated risk-assessed controls, and data handling.

The author refers the reader to relevant medicinal manufacturing product regulations and guidance for additional information. Some descriptions are based on listed regulations and guidance, and judicious editing is necessary to fit the context of this book.

This book aims to ensure that accurate, complete, and legible copies of data are available for review, guide readers in effectively managing data and e-record reliability vulnerabilities, and promote essential compliance in this area.

This book aligns with the definitions of data reliability in data engineering standards.

The recommendations to implement data-record controls, as described in this book, are based solely on the authors' standpoint and opinion and should be considered a suggestion only. They are not supposed to serve as the regulators’ official implementation process.

References.

[1]      The principles outlined in this book apply to data generated by both electronic and paper-based systems.

[2] NARA, "Universal Electronic Records Management (ERM) Requirements," Version 3.0, June 2023.

[3]      This is a dimension of data quality. Data accuracy refers to the extent to which data accurately represents real-life entities.

[4] It is one dimension of data quality. Data completeness refers to the extent to which all required data elements or values are present in a dataset without missing or null values.

[5] Data conformity refers to data adhering to specific definitions, such as its type, size, and format.

[6] Data consistency is one dimension of data quality. Data consistency refers to the accuracy, reliability, and coherence of data within a system or database.

[7] Objectivity is based on factual information, with minimal influence from personal opinions or feelings.

[8]      Data suitability refers to the appropriateness and fitness of a dataset for a particular purpose or analysis.

[9]      Data validity refers to the accuracy, correctness, and suitability of the data.

[10] Source authentication refers to data that has a verifiable source, accompanied by well-documented evidence or reliable witnesses.

[11] Data timeliness refers to the degree to which data is current and up-to-date at the point in time when it is needed for analysis, decision-making, or other business processes.

[12] Trustworthy data - Reliability, authenticity, integrity, and usability are the main characteristics of trustworthy data from a record management perspective. (NARA)

[13]US FDA, "Guidance