Robert Seltzer, compliance manager at GSK, and highly active in American Society for Quality (ASQ) circles, gave a compelling presentation recently at a local chapter meeting on root cause analysis and corrective and preventive action (CAPA), and why pharmaceutical manufacturers may not be doing it all that well. This post summarizes (not in order of presentation) some key points from his talk.
It’s no secret that insufficient root cause analysis and CAPA have figured in many 483’s and regulatory failures of late. Richard Friedman of FDA has tied the fact to outmoded manufacturing and quality practices, suggesting that industry management may not always see the cost of compliance and the business value of proactive, redundant control.
Seltzer, who believes that pharmaceutical manufacturing and quality professionals can be ranged in a bell curve based on their openness to cGMP’s, believes that pharma takes too linear and sequential a view of root cause analysis. In addition, he suggests, some QC professionals may be too limited by fishbone diagrams, one of many tools they might use to analyze quality problems and their sources.
In an ideal system, workers will be in a state of self control, Seltzer says, using Joseph Juran’s definition of control, manifesting itself in:
• Knowing what the target is
• Having a means to measure performance
• Being able to take corrective action.
When workers are not in a state of self-control, Seltzer suggests, corrective actions must be management’s responsibility. One assumes this is part of the reason for some of the industry's recent problems. Seltzer peppered his presentation with quotes that haven’t yet been overused. Two examples?
“Walter Shewhart conceptualized the “Plan, Do, Check, Act” cycle, however many companies only practice the second step, which results, instead , in a “Do Do” organization
"We can't solve problems by using the same kind of thinking we used when we created them." --Albert Einstein
Seltzer suggests that industry professionals consider Reality Charting, a method espoused by Dean Gano in his book, Apollo Root Cause Analysis. Gano views causes and effects as part of a “causation system.” Lack of knowledge is what limits understanding of that system.
In reality charting, Seltzer summarizes, a team brainstorms at least causes of a trigger event, in the form of an action and condition along with evidence supporting those conditions. The team then asks why of each stated cause until there are no more answers, and produces a cause and effect chart that shows all known causes and their inter-relationships. All causes are then examined and a solution proposed that is within the company’s control, prevents recurrence, and meets the company’s goals and objectives.
Once a corrective action has been pinpointed, the team then identifies driving forces supporting it, as well as restraining forces to be overcome or minimized. Drivers, Seltzer explained, include justifications, resources and people, business needs, financial benefits and risks reduced.
Seltzer cautions against relying too much on one risk assessment tool. He also says that the “five why’s” lead to oversimplification.
He also observed that investigations can generate the following reactions:
• “It will never work here.”
• “No one will buy it.”
• “We already tried that once.”
• “It’s not in the budget.”
• “Good thought, but impractical.”
• “No one else is doing it that way.”
Seltzer judges the completeness of any corrective action based on the following:
• Necessary and sufficient training (or re-training) has been administered
(does rule exist of not more than 20% absentees from training?);
• Necessary and sufficient procedures, instructions, and/or physical changes are “in place” (i.e., written and issued, or installed and commissioned);
• Necessary and sufficient qualification, validation, and/or other implementation (i.e., “in use”) has/have been confirmed by the site quality and then stated or recorded as such.
Nothing in a key GMP procedure or expectation should be superfluous, he says, and there must a minimum threshold set for adequate corrective action measures.
However, he notes, completeness is not the same thing as effectiveness, which is defined by most or all of the following:
• The causation system of the trigger event’s outcome has been identified is better understood.
• Defensible, risk-based solution(s)/corrective action(s) has(have) been determined and adopted controllably (e.g., via change control, validation).
• Proper monitoring has been put in place around the implemented corrective action(s).
• A determination has been made that corrective action(s) has/ have not resulted in any other undesirable event outcomes as a by-product.
• From the determined causation system, the original trigger event’s outcome doesn’t reappear. This criterion will likely require data trending.
He shared his mantra (some of which follows):
• Don’t reinvent the wheel: use other people’s wheels and axles.
• Use technical terms, acronyms, and abbreviations; but first define them.
• Learn, use, and hone (i) rules of logic, (ii) statistics, (iii) quality tools.
• Question assumptions, particularly long-held and cherished ones; and be willing to tell the emperor/empress (s)he has no clothes.
• Ask anybody any question (except rhetorical ones—they insult!).
• Be a maverick/iconoclast instead: it can save one’s company and career.
• Kill two or more birds with one stone when possible. It’s called leveraging!
• Work more via professional, duty-bound procedures, job descriptions and less via “good old boys’ clubs” or the currying of favors or mutual back-scratching. The latter leads to bias. Also conflicts of interest must be avoided at all cost.
• Be ready to warn people of what may risk bias or conflict of interest. Ethics encompass and mark a person for life. Jobs come and go!
Here’s hoping that more people take these guidelines to heart, whether they’re managers or rank and file employees. We’re trying to convince Seltzer to write an article for us this year. Stay tuned. In the meantime, where would you put yourself, your managers and your colleagues on the "GMP bell curve?"