Data is the great enabler when it comes to manufacturing processes. But not all data is good. Industry pundits estimate that about 85 percent of the data generated in manufacturing has absolutely no value at all. Which makes one wonder if some of that useless information can cause problems.
What to do with all that data—the good and the bad— is not a new area of discussion, but it is still on the forefront of the minds of pharmaceutical executives, especially as they respond to Food & Drug Administration (FDA) enforcement actions. In recent years, there has been an increase in warning letters to companies around deviations from current good manufacturing practices (CGMP) in manufacturing facilities.
The number of drug GMP warning letters went from 42 in 2015 to 102 in 2016 to 114 in 2017. This puts more emphasis on data integrity in the industry.
The FDA uses the acronym ALCOA to define data integrity, meaning, it is attributable (who did it/source data), legible (recorded in a permanent medium and it is readable), contemporaneous (data recorded in real time), original (an original or certified true copy of data), and accurate (no errors or editing performed without documented amendments).
Addressing ALCOA took center stage during the recentInterphexconference in New York. A keynote presentation by Els Poff, executive director of the data integrity center of excellence at Merck, discussed details of data integrity from log-ins to definitions to validation and verification—as it relates to ALCOA. In addition, Poff, who is on theParenteral Drug Association (PDA)data integrity task force, said the organization is working on a comprehensive set of tools for industry that includes a technical report, which will be available at the end of the year, that will address new requirement challenges related to Industry 4.0 and Big Data.
The PDA has already released the Elements of a Code of Conduct for Data Integrity, which addresses the culture of data integrity. It touches on actions that can contaminate data, like employee errors or system configuration problem with electronic data handling.
This code of conduct is importantbecause, like cybersecurity, data integrity is everyone’s business. The document provides an in-depth guidance for life science companies with a focus around a handful of actions:
Applicability - The company should establish requirements and implement programs to provide employees with training on the fundamental principles of data integrity.
Data collection, analysis, reporting and retention - The company shall establish procedures, documentation and control systems to maintain data integrity by collecting, analyzing and reporting data on paper and in computer systems with the appropriate security, audit trails, validation and oversight of electronic documents.
Electronic access security measures - Any computer data acquisition system should have secure access to prevent unauthorized changes to electronic data.
Investigating wrongful acts – Procedures for investigating an alleged falsification shall include documented in-depth review, conducted in a fair and balanced manner by independent personnel.
Data integrity of outsourced services and purchased raw materials – as part of a vendor/supplier qualification program, the company will confirm that contractors, vendors or other third-party providers have established policies, standards, procedures or other documents that adhere to data integrity requirements.
The PDA code of conduct boils down to having a common understanding of the expectations for employee and management behaviors. Moreover, the path to data integrity should be “complete, consistent and enduring,” Merck’s Poff noted in her keynote. “Do you have all of the data and meta data over the lifecycle? Do you have processes in place with data and time stamps? Is the data on media that is still retrievable?”
This is not rocket science. But we are living in the era of organization. And, many of these things can get overlooked. To get your pharma house in order,downloadthe free PDA code of conduct document.
