Insights from Peter Baker at the 46th International GMP Conference

When it comes to data integrity, Peter Baker has seen it all. He is well-known in the industry as a former FDA investigator, and now heads his own firm, Live Oak Quality Assurance. And, as he told the audience in March at the International GMP Conference in Athens, Ga., head-in-the-sand avoidance is common, while the right approach to data integrity issues—openness, honest self-reflection, and objectivity—can be hard for some organizations to follow.

[Related: For more on data integrity, view a recording of our May 2022 webinar that addressed data integrity issues in the laboratory.

And when they cannot follow that path, they may limit an investigation’s scope, avoid systemic causes during root-cause evaluation, and taint the scientific justification.

This article is Part I of a two-part series on insights from Peter Baker’s presentation. This first half focuses on the systemic dynamics that lead to data integrity problems, while Part II will present Baker’s insights on managing data integrity issues with quality risk management (QRM).

All those ALCOA+ principles continuing to be implicated in about 80% of all the warning letters that we see—that is pretty significant

Companies often avoid that head-on approach despite the fact that data integrity issues are going to happen. After all, Baker’s talk is titled, “Handling the Inevitable: Investigations Where Data Integrity Has Been Implicated.” And he estimates that every site will probably face a data integrity issue within a year or two.

“What we see is 80% of the warning letters coming out of CDER cite these data integrity keywords,” Baker explains. “All those ALCOA+ principles continuing (sic) to be implicated in about 80% of all the warning letters that we see—that is pretty significant.”

Did you know that the top three data integrity observations last year were system controls, backup and archival processes, and data manipulations?

With in-depth Redica Systems data for Inspection Preparation, including inspection trends and investigator tendencies, you will be able to identify where data integrity deserves more attention—before you get a visit from the next Peter Baker.

Denying the Inevitable, Limiting Investigations

Denying the inevitable is trouble. But managers often think that if data integrity is implicated, an investigation must be concluded as quickly as possible, says Baker.

This pressure from the top “forces the QA department to limit scope, to try their best to limit their deep-diving to the root cause, [and] they are forced to conclude ‘no product impact,’” he points out.

How do managers limit investigations? “Oh, ‘our people would not do that’—you hear that a lot,” says Baker. “Or, ‘this could never happen because of X, Y, and Z.’” But without evidence, those are assumptions—uncertainties that color an investigation.

A hasty investigation is probably not following a systematic and risk-based approach, and an investigator is “easily going to find the limitation of the scope, the product impact statement not being scientifically justified,” Baker says.

Our people would not do that’—you hear that a lot

Instead, the effort companies put into a data integrity-related investigation should match the risk to patient safety, as it says in ICH Q9: Quality Risk Management.

Management groups who allow their staff—especially QA—to perform the investigation as required are the ones who will really succeed, Baker says.

Self-Reflection: Killer of Data Integrity Root Cause Evaluations

Root cause evaluations are where investigations can really start to go bad, says Baker. “Because this is where the company is forced to do some serious self-reflection.”

Focusing on its own procedures and processes during the root cause phase “can be very uncomfortable, especially in some companies,” he says.

The “blame game” gets going when an organization cannot face that discomfort. “You are going to see it being blamed on frontline employees,” says Baker. “You are going to see power dynamics and structural implications; you are going to get closed-loop thinking; you are going to see ‘scratching the surface’—and most importantly, the introduction of bias.”

Facing Shortcomings with a Strong Quality Culture

An organization with a strong quality culture, especially at the top, can tangle with data integrity properly. A strong quality culture aims to prevent future occurrences.

In a data integrity-related root cause evaluation, investigators who are part of a strong quality culture ask, “what about the quality culture may have led to this failure?” Baker says. “What about the management strategy that may have led to this failure?”

When a data integrity issue revolves around an operator, a strong quality culture should prompt questions like, “what about the SOPs that led to this failure?” says Baker. “Before we can retrain our operators to this SOP, what needs to be changed? How can we make the job of the operator easier tomorrow than it is today?”

Part II will examine how to address data integrity issues using quality risk management (QRM).

Warning Signs for Quality Culture

It is hard to gauge the strength of quality culture, either in a single company or across the industry. Inspectors do not name it specifically in 483s and warning letters.

But a weak quality culture leaves traces. Using Redica Systems to find the phrase “quality unit inadequate” in 483 observations offers a rare window into the state of organizational quality culture.

For example, from 2017–2021, Redica Systems data shows that these observations have grown every year for human drug GMP operations across the globe, reaching a peak in 2019 (the drop in the subsequent two years is likely due to the reduction in inspections due to the COVID-19 pandemic).

With Redica Systems, you can see the warning signs of weak quality culture across the industry by region, time period, and facility type. You can view the most current “quality culture” information from investigations at the facilities of peers, competitors, and vendors.