By Agnes Shanley
Anyone looking for hopeful signs of change in the ongoing industry/FDA impasse on smart manufacturing–specifically, PAT and QbD–was bound to be disappointed by this week’s FDA “Partnering with Industry” meeting in Bethesda.
The only real sign of change was CDER chief Janet Woodcock’s new look and hairstyle. I don’t mean that facetiously either. She looks great–light years away from the harried years of the previous Administration.
But many of the presentations at this meeting, hers included, were exactly what the industry has heard from everyone in PAT/QbD circles for the past seven years: “The early adopters get it, and eventually the rest of the world will come around, too.” There were the obligatory seven-year-old slides, some of which dated from the draft stages of the PAT guidance. Even the dreaded pyramid showed up a few times.
What was significant, though, was the emphasis on PAT, along with QbD. For the past few years, the industry has heard a lot about QbD but relatively little about PAT, the toolkit for achieving consistent quality. The result, for many, appears to be confusion.
The aim of this conference was clear: to remove that confusion and reinforce PAT’s importance as a way of achieving consistent right-first-time product quality.
Chris Watts, the last Beatle of the FDA PAT Team was there, and it was interesting to see and hear from some of the people behind the scenes at FDA–such as Jon Clark, Associate Director of OPS at CDER, Rick Friedman, Director of Manufacturing and Product Quality at CDER’s Office of Compliance, Rebeca Rodriguez, a member of FDA’s Pharmaceutical Inspectorate, who leads Puerto Rico’s inspection team, and Keith Webber, Deputy Director of CDER.
Unfortunately, Helen Winkle was ill and couldn’t attend. Former PAT Team leader Ajaz Hussain made some very important points, taking some former, and absent, colleagues at FDA to task for contributing to misunderstanding and creating an unnecessary rift between QbD and PAT.
He also paid homage to Aventis’ success with PAT, but, in the end, for anyone working in small molecules, portions of his speech might have been deeply unsatisfying. He almost seemed to be saying: “Look at what I’ve done in biology at Sandoz and Philip Morris, and learn.”
Before talking of applying a systems biology approach and a comparability protocol to manufacturing small molecule drugs, how about using old-fashioned chemical and control engineering approaches, then taking it from there? Girish Malhotra has much to say that’s spot on, in this regard.
“Experts don’t know everything. You need creativity and courage.”
There were some brilliant presentations on the program. Most notable was Jean-Marie Geoffrey’s talk. Apparently, at the last minute, Jean-Marie, who works as Senior Research Fellow for Takeda, replaced Ferdinando Aspesi, who had led Aventis’ efforts and is now with Pfizer-Wyeth, but was unable to attend.
Mr. Geoffrey gave several detailed examples of PAT applications in various unit operations, including fluid bed drying and granulation, and lessons he has learned at Takeda and Abbott.
However, he started his talk by alluding to personal experience: understanding his son’s illness in order to optimize the medical treatment and outcome, drawing comparisons and connections with the approach needed to gain process knowledge. “Experts don’t know everything. You need creativity and courage,” he said.
One important point he stressed was operator behavior and its potential impact on variability, which can be induced by vaguely worded SOPs: If you say “around 10,” in an instruction or SOP, for instance, an operator can work at 10 one day, 10.5, the next, and 11 the next day, introducing variability.
Inconsistency is a Killer
At the conclusion of his presentation, he addressed cultural issues. “Inconsistency in the support of the message is a killer. You need to sustain the sense of urgency,” he said. “Were we naïve to think we’d be done by now? The urgency is gone, compared with 2004,” he said, urging colleagues to build teams and a vision, and not to let up. It was an outstanding and memorable talk.
Gawayne Mahboubian-Jones, formerly with Optimal Industrial Automation, and now with Philip Morris International, gave a comprehensive talk that clearly explained, using numerous examples, what PAT is and how PAT and QbD interconnect.
He showed a hypothetical “progressive application” of PAT, sketching an implementation progressing in a series of stages. Many people doing pharma PAT today are only going part of the way, he said, and few get to the continuous improvement stage.
He also touched on data management and the importance of connecting manufacturing and development. ICH Q8(R2), he said, focuses on development only. “QbD is meaningless without integrating manufacturing, which is what PAT is primarily focused on,” he said. “In all other industries, one doesn’t differentiate between development and manufacturing, but talks, simply, about the product.”
Mahboubian-Jones recalled the situation that the electronics industry was in back in the 1970’s, when quality issues drove manufacturers to a Six Sigma approach. It took between 6 to 7 years to get there, he said, but now they’ve moved on to 8 sigma, so it can be done. Hmmmmm. The PAT guidance appeared six years ago . . . so why is pharmaceutical manufacturing still at 2.5 Sigma levels of quality?
PAT vs QbD?
The final part of his presentation addressed the critical topic of how PAT and QbD work together. “Sure, you can do QbD without PAT,” he said, “but once you see some unexpected variability, it will be as if you’ve built a house on sand. When the sand shifts, the house will fall.”
Impeding progress is the fact that there is no consistent understanding of what either term really means, he said. Each company has its own definition and understanding, and the actual industry understanding of QbD is very different from the real definition. “The skill sets needed for manufacturing and advanced control are absent in both pharma companies and regulatory agencies,” Mahboubian-Jones said, so teams cannot properly construct control strategies.
His conclusion: “Despite all the talk [and six years ago there was a lot of talk], the entire industry remains highly resistant to change, or being taken outside its comfort zone. They must understand that the future is built on change.” There are examples to follow for PAT, he said, mentioning Aventis.
Could Pharma Be Blindsided by Lack of Manufacturing Expertise?Mahboubian-Jones also raised the possibility of companies with strong manufacturing bases, food and beverage companies, for instance, moving into pharma as generic or even primary manufacturers. Their excellence in manufacturing would allow this to happen and the result could throw the whole pharma industry structure on its head.
For peer education and transfer of best practices, Martin Warman, Scientific Fellow at Vertex Pharmaceuticals, Inc., gave the conference’s best presentation on applying an integrated systems approach to quality management. Such work must focus on risk management, he said, and a roadmap must be built, starting with product definition, using target product profiles. Product definition leads to process definition. In addition, he clarified measurement system and process capability. We will post portions of his presentation, pending corporate approval.
Get ‘Er Done
Powerhouse speakers included Melissa Herkt, COO of Emerson Process Management, who used industry data convincingly to drive home the need for excellence, and transformational leaders, in drug development and manufacturing. The Alabama native injected humor and included a few down-to-earth references (even Larry the Cable Guy) in her presentation. (I’m not a big fan of Larry’s, but after Dr. Hussain’s Popperian traps, it was a welcome relief.) For a video excerpt from her presentation, click here.
Another standout was Bonnie Norman, head of quality and regulatory affairs for Intel’s Digital Health group, who spoke convincingly of the need for better risk management and control. It all starts, she says, by focusing not on speed to market, but on the end users of drug products. For a brief excerpt from her presentation, click here.
Also on the program was consultant Ian Bradbury and University of Delaware’s Chemical Engineering Professor Babatunde Ogunnaike, who has trained FDA staff and who spoke of his team’s research into online control of glycosylation in monoclonal antibody production.
Inspiration From Chemical Manufacturing
To inspire the small-molecule drug manufacturing crowd, Neil Everall discussed how ICI applied PAT was applied to TiO2 production, while Emerson’s Terry Blevins discussed use of simulation and data analytics, providing a pharma-relevant case study involving the specialty chemical company, Lubrizol. We will be posting a video clip excerpt next week.
Life with Deming
There was one interesting side note to the conference. J Edwards Deming’s grandson, Kevin Cahill, shared, albeit in a rather detached way, anecdotes about his grandfather.
Perhaps JE was too busy with his control charts to play much with little Kevin. But Kevin opened up a bit more at lunch, sharing memories of a little-known part of Deming’s identity: as a composer of music. He remembered sitting through one of his grandpa’s interminable choral compositions at a two-hour high mass, and attending, with his family and his proud grandfather, a special performance of one of Deming’s compositions by the Washington Symphony.
FDA speakers also provided insights into Agency plans and projects. Here, the best presentations were by Rick Friedman and by Keith Webber. We will be posting video excerpts from Mr. Webber’s presentation next week.
Mr. Friedman discussed the PAT compliance process. Industry infrequently explores its use, he said, but it’s all about good risk management and evaluation of potential vulnerabilities. As early as 1978, FDA regulations addressed variability (211.110a). The question is: ”How good are our measurements,” he asked. Mistakes often aren’t caught at the point of error.
For dissolution issues, he said, root cause analysis is often needed. Out-of-spec results can be caused by aberration of measurement or production processes, product related variability or measurement system variability.
Data Integrity a Focus in Inspections
He stressed the importance of data integrity, and the need to fix flawed processes.
Inspectors will pay particular attention to computer systems issues, and the possibility of staffers at a company overwriting results. Computer systems validation is becoming increasingly important, especially with lab data, and senior management will increasingly be held accountable. For a copy of his presentation, click here.
Some interesting questions came up during Q&A. . . . FDA hedged on when the new process validation guidance would become official.
But one audience member asked about how restrictive the design space had to be. “If product quality is improved by shifting slightly outside of the design space, would FDA accept the change?” Answer from FDA was that a supplement would be needed or the design space would need to be revisited, given the current nascent state of understanding among reviewers.
Professor Ogannaike wondered whether manufacturers will ultimately focus on product quality, as the chemical industry does in its continuous improvement programs, rather than the design space. “Otherwise, the design space could become a straightjacket,” he said.
Post-Approval Changes and Continuous Improvement
Consultant Ali Afnan, formerly with the PAT team, who developed the program for this conference while at OPS, brought up the question of post-approval changes, and cited 314.70, which asks for specific contact with the regulator. It needs to be rewritten, he says, and his upcoming June column, Step Change, will address this topic in greater depth.
Another audience member asked whether FDA would still consider mandating PAT, or whether fast-track approvals for applications using PAT could be used as an incentive to get more senior executives to support and fund programs. Keith Webber gave a clear, if balanced, “no” on both counts. For more, and video clips, click here.
Even if PAT’s progress, and that of QbD, have been frustratingly slow to those advocating smarter drug manufacturing, the conference shed important light on what’s going on within the FDA, and best practices within this, and other, industries, to advance the idea of manufacturing science.
The conference is expected to run again next year, this time with NIPTE as the partner. Perhaps more will have changed by that time, but even so, it was interesting to hear more from those within FDA, and industry, who are trying to drive that change.
- Pedro Hernandez’s Letter to Santa
- What does pharma need for Christmas? Pharmaceutical consultant Pedro Hernandez, PhD, takes up our challenge and offers his wish list...
- Many Gurus, Few Disciples at Second PAT, QbD Conference in Bethesda
- A very brief report links to most of the key presentations at an April conference, many of which could serve...
- April in Bethesda: FDA, Thought Leaders to Discuss QbD and What Lies Ahead
- This accredited, NIPTE-sponsored event features QbD thought leaders representing FDA, industry, academia, consultancies, and solution providers....
- Watts Leaving FDA: The End of an Era?
- The last of PAT's "three tenors" leaves the Agency....
- Emerson’s Herkt: Where Is the “Burning Ambition” for PAT/QbD?
- Pharma can do better, says Emerson Process Management COO Melissa Herkt. We need a “burning ambition” for PAT to take...