The global pharmaceutical industry has been grappling with the regulatory agencies’ focus on data integrity for the past several years. Data integrity guidance documents issued by regulatory agencies have resulted in the examination of the lifecycle of data coming from the analytical and microbiological laboratories. For a contract laboratory, reliability and trust in the data delivered to the sponsor are paramount. A contract laboratory cannot afford to be implicated in a breach of data integrity, no matter how small. A breach in data integrity not only impacts the data and standing with regulatory agencies, but also compromises relationships with sponsors, and in turn, the health of the business. The data integrity guidance(s) do not need to be considered an obstacle but a tool to improve the trust in the data. Data integrity is not a new concept: it has been part of science for centuries. However, the technology of today has added complexity to proving that the data is beyond reproach.

One of the foundational tenets of science that young scientists are introduced to in university studies with respect to data integrity is the consequence of “dry labbing” or falsifying results. Instructors emphasize the importance of accurately recording data by issuing the ultimatum that falsifying laboratory data would be met with immediate failure of the course. As scientists enter the pharmaceutical laboratory work force, the effects of falsifying results become clearer as the impact to patient safety is directly impacted by the laboratory data. In today’s environment, acts that breach data integrity guidance by a laboratory analyst should result in disciplinary action. Each discovered case of data falsification or of an uncontrolled process that could easily allow data falsification, adversely impacts the integrity of the data and degree of trustworthiness in the results from a laboratory.

The principles that guided the documentation of scientific data in the past, when data was recorded primarily in hardcopy laboratory notebooks, are the same as the ALCOA+ principles used now, there just was no acronym used. These principles and definitions require data to be:¹

• Attributable – Who performed the activity, and when?
• Legible – Can it be read? Must be human-readable and a permanent record
• Contemporaneous – Recorded at the time the activity was performed
• Original – Original record or certified true copy
• Accurate – Error free
• Complete – Everything is included, and nothing is missing
• Consistent – The data should be acquired in a chronological order, with time stamps included for any addition to original data
• Enduring – Data is on a media that is maintained and protected with an emphasis on ensuring data is readable and available long after it is recorded
• Available – Records are accessible when needed

For decades, the documented hardcopy laboratory data in bound notebooks was reviewed, and any discrepancy in the data was easily detectable and could be questioned. However, even with hardcopy data, the validity of the data was directly impacted by the integrity of the bench analyst and laboratory management’s commitment to ensuring the recording of complete and accurate data.

With the introduction of computerized chromatographic data acquisition systems in 1980s and 1990s, a conundrum was created. Was the printout of the data or the electronic file the raw data? If the raw data was the electronic file, was the electronic file secure from being easily altered? In some systems, the files could be written over, renamed, etc., without any audit trail documenting the action. Subsequently, the integrity assurance of the laboratory data was diminished unless laboratories added configurational or procedural controls over the data acquisition systems.

Compliance to 21 CFR Part 11 – Electronic Records; Electronic Signatures, has been part of our industry for more than two decades, but data integrity guidances have relatively recently enhanced the control of electronic records. Warning letters that cited data integrity gaps  have emphasized the lack of control of computer systems and associated records. However, the data integrity principles also apply to paper-based records, such as printouts from balances, pH meters, documentation captured on loose forms, and observations written directly into a bound laboratory notebook. While a laboratory can achieve compliance with electronic data record integrity, the job is not complete. Laboratories must find and root out vulnerable data documentation practices beyond electronic records. A useful tool to find areas of vulnerability is data mapping. Data mapping is a type of mapping with an emphasis on the lifecycle of data, considering the following:²

• Generation and recording of data
• Processing data into usable information
• Checking the completeness and accuracy of reported data and processed information
• Data (or results) being used to make a decision
• Retaining and retrieval of data which protects it from loss or unauthorized amendment
• Retiring or disposal of data in a controlled manner at the end of its life

Figure 1 provides an example of a high-level data map of sample testing, starting with sampling and ending with reporting test results, with a breakdown of one step in the process: sample testing. The map provides the visual assessment of how data is generated, the required supporting records, where data is recorded, and how data is recorded. 


Many of the steps/boxes in the sample testing data map can be further detailed to eventually list every document that supports each process. Each of the underlying records must meet ALCOA+ principles. Figure 2 exemplifies how the preparation of the reference standard solution step can be further broken down to list each of the underlying records that support the process of the preparation of the reference standard solution.


Each of the supporting records should be assessed to ensure all the records meet the ALCOA+ requirements. For example, the weight data from the balance is commonly printed with the metadata of date, time, and balance ID. However, some older analytical balances allow relatively easy access to the date and time settings which are sent to the printer along with the weight data. In such a case, a procedural control is needed to verify that the time and date was correct when the weight was printed. Similar situations may also be present in other small instruments such as pH meters or Karl Fischer titrators. Integrity of date, time, user, and other metadata can be assured when a laboratory adapts a validated e-notebook system.

Changes in data lifecycle are inevitable, and when changes are proposed it is important to understand the lifecycle elements for each type of the affected data or record, and ensure controls which are proportionate to data criticality and risk at all stages.² Data mapping the changes can aid in determining if the change increases the risk on data integrity. For example, in the past the logging of the receipt and use of reference standard may have been recorded on a dedicated form or 3 x 5 card that was kept in a file.

However, was this document (form or card) tracked? Was the document secure at all times? What is the risk to the products if this document was accidently lost/destroyed? Would an electronic system be better? Not necessarily if the electronic records do not meet the ALCOA+ principles; however, even a bound logbook of the reference standard receipt and use would improve the integrity of the data over a loose form. 

A new data process may have conveniences over the traditional bound notebook, although if changes are not easily detectable through audit trails, controlled through access control, and accountable to individuals, the new process adversely impacts the integrity of the data. Data mapping would explore these potential shortcomings and allow mitigation to ensure a compliant process.

Data integrity can impact any stage in the data lifecycle, whether the data is electronic or a hardcopy. Therefore, it is important to understand the lifecycle elements for each type of data or record in terms of criticality and risk. Thoroughly mapping each data acquisition process can be a useful tool to find areas of data integrity risk and a potential risk to products. A map of data acquisition in a laboratory should break down the process so that each step in the map can list out every record supporting each process on the map. Then, each of the records can be assessed individually for the risk to data integrity. 

References

1. Pharmaceutical Inspection Convention/Pharmaceutical Inspection Co-operation Scheme. “PIC/S Guidance: Good Practices for Data Management and Integrity in Regulated GMP/GDP Environments”, PI 041-1 (Draft 3), November 30, 2018, available for download from https://picscheme.org/en/publications#selCategory_
2. European Medicines Agency, “Guidance on good manufacturing practice and good distribution practice: Questions and answers: Data integrity,” August 2016, https://www.ema.europa.eu/en/human-regulatory/research-development/compliance/good-manufacturing-practice/guidance-good-manufacturing-practice-good-distribution-practice-questions-answers#data-integrity-(new-august-2016)-section

---

Tim Rhines, Ph.D., Director,
Lachman Consultant Services, Inc.

Timothy Rhines is a director in the science and technology practice at Lachman Consultants. He is a seasoned CMC pharmaceutical/biopharmaceutical professional with more than 27 years directing CMC activities, leading analytical chemistry teams, addressing compliance gaps, compliance department leadership with P&L responsibility, implementing process excellence initiatives, developing pharmaceutical product stability operations, and CMC project management. Dr. Rhines has a strong technical background encompassing both small and large molecules. For more information visit www.lachmanconsultants.com.