Categorized | QbD News

Process Variability? FDA, Heal Thyself

    Process Variability? FDA, Heal Thyself

    By Agnes Shanley and Michele Vaccarello Wagner

    At a recent webcast, participants noted, off-line and off the record, that today, delays in Quality by Design (QbD) application review were coming, not from “old timers” within FDA, or traditionalists wedded to old-fashioned quality control methods, but from newly recruited reviewers.

    There can be variability in the quality, relevance and quantity of questions raised, they said, and some of the questions that applicants must address can appear to come from left field, increasing the cost and time required for applications based on QbD principles, without necessarily advancing the science involved.

    Part of the problem, they suggested, is the fact that FDA has hired over 1,300 new scientists since 2007, and some of them have not had any experience in the industry.

    Criticism of FDA is nothing new, and the Agency has been working for some time to train staff and get its reviewers, inspectors and regulatory affairs experts to communicate and collaborate more closely. Increasingly, the Agency is calling on people from outside the Agency to help in this process.

    However, where the old science board meetings on drug manufacturing and quality  involved guest lecturers from academia and industry, FDA is increasingly involving paid management consultants, of the type who advise Big Pharma in its continuous improvement efforts.

    Last month, Janet Woodcock, director of the Center for Drug Evaluation and Research (CDER), announced that the Agency had hired McKinsey & Co. to help it better improve its review process, to determine the root causes of a huge backlog of generic drug applications, estimated at 2,000.

    FDA’s increasing use of management consultants has led to criticism, according to a recent article in the Wall Street Journal. According to, the WSJ wrote, McKinsey has performed $17 million worth of consulting work for FDA since 2008, while Leadership Performance Solutions has earned nearly $8 million in fees.

    The Agency has also been chastised for working with some of the companies it reviews. During the heparin crisis, FDA worked on analytical method development with Momenta Pharmaceuticals, which, with Sandoz, manufactures a generic form of the name-brand blood thinner, Lovenox, approved by FDA last year.  The company’s founder, an MIT professor, shared expertise in a specialized area of bioanalytics, at a time when answers were needed, quickly.

    However, the free consulting and appearance of a relationship between Momenta and FDA may have helped motivate two lawsuits from Momenta’s competitor Amphastar, whose product has failed to gain FDA approval, even though it was submitted for review before Momenta’s.

    Although Dr. Woodcock and Momenta management were cleared of any allegations of financial conflicts of interest, a November GAO report publicized in the Wall Street Journal criticized FDA for giving the appearance of conflicts of interest, in utilizing Momenta’s expertise.

    Given all the controversy over FDA training and culture, it may come as little surprise that both were prominent topics of discussion at ISPE’s annual meeting on November 8.  A regulatory town hall forum and a track on quality systems and ISPE’s Product Quality Lifecycle Implementation (PQLI) brought FDA officials and industry professionals together to discuss FDA’s expectations for Quality by Design (QbD) and changes in the way the Agency is handling review, inspection and training.

    Moderating one of the regulatory sessions and a PQLI session was Joe Famulare, former Deputy Director, Office of Compliance, who is now global head of Quality, External Relations and Collaboration with Genentech.

    Among the topics discussed at the regulatory forum were joint inspections, such as have taken place for API’s in Ireland, as well as issues of change management and Quality by Design. Speakers on the Quality Systems/PQLI track focused on the development of new guidance documents, which ISPE expects to release next year, which are expected to clarify such contentious concepts as “critical process parameters,” and what Bruce Davies referred to as “the threshold” and “the continuum” of criticality. Line Lundsberg Nielson also discussed plans for clarifying control strategies.

    The two sessions overlapped on the subject of FDA.

    At the afternoon PQLI session, one member of the audience commented on what he described as the variability of FDA reviews, especially those based on QbD, after a question was raised about staff training and readiness to handle QbD product review, given the Agency’s recent massive recruiting efforts.

    Christine Moore, Deputy Director for FDA’s Office of New Drug Quality Assessment (ONDQA), responded to the question briefly from the audience, recapping the Agency’s Inspectorate and other training initiatives, as well as its attempts to bring more reviewers into GMP inspections and to create opportunities for holistic training.

    She had addressed these issues the previous day in a session on the regulatory track. Overall, she says, at ONDQA the goal is for inspectors, reviewers and those in the compliance office to work more closely together, and to share information in an integrated, rather than linear fashion. This will be essential, she noted, for the more complex approaches demanded by the QbD framework, which assumes cross-functional problem solving and knowledge transfer.

    As Moore noted, reviewers have always been able to take part in cGMP inspections. In the past, however, few reviewers seemed to want to, and there was little reviewer participation, for example, in pre-approval inspections.

    However, walls between reviewers and inspectors may be starting to come down. In the past, Moore said, reviewers from the Office of Biological Products and the Center for Biologics Evaluation and Research were more likely to take part in inspections than their other FDA counterparts, she said. But now, in addition, more reviewers at ONDQA are participating in inspections, with 16 reviewers taking part in inspections this year, compared with four in 2008. This year, reviewers have been able to participate in PAIs for NDAs, PAIs for sNDAs, Pre-Operational Reviews, and For-Cause Inspections.

    Having reviewers on inspections adds value, she said, giving them an increased understanding of the product and process. It also allows them to better understand the rationale for scale-up and process control, and the implementation of online monitoring and the use of models.

    But the knowledge flows both ways, Moore said, since reviewers also give compliance and regulatory affairs staff a much better understanding of product, process and other risks.

    ONDQA prepared a memo to help improve the way FDA reviewers can communicate application-related information to the Offices of Compliance and Regulatory Affairs, Moore said. It was prepared early in the review cycle to help plan inspections, and features a product description, process summary, critical steps and control, as well as a summary of product-and process-related risks.

    The memo was designed for complex products and or processes, or situations with complex regulatory approaches, such as real-time release testing, and applications where there are questions about data integrity or manufacturing capabilities, she said.

    Ms. Moore then discussed how things would be different under Q8, Q9 or Q10. Although the assessment methodology and scope would be the same, inspections would now be more focused, as FDA tackles such questions as:

    • What about implementing process parameters (both critical and non-critical)?
    • How to perform change control in the design space?
    • How will an organization know whether it is inside/outside the Design Space?
    • How to manage even when “out of design space”?
    • Is the manufacturing site capable of implementing the control strategy (e.g., for real-time release)?
    • Is the manufacturing site capable of developing and implementing an appropriate batch release strategy based on GMP and control strategy?

    As Moore noted, drug product development predictions will increasingly be made using predictive models, so there will be a need to address flexible change management, as well as protocols for change control, monitoring, managing of deviations, including out of spec results.

    Predictive models will be verified and validated at the commercial site and throughout the product’s lifecycle, Moore said, and subsequent adaptation under PQS will be monitored by inspection oversight.

    Previously on the program, Tara Gooen of CDER discussed the pre-approval inspection compliance program guidance manual, and changes designed to bring it more in line with the “Desired State,” as articulated in FDA’s Pharmaceutical Quality for the 21st Century.

    In revising the manual, she said, FDA staff collaborated more closely, implementing a knowledge-transfer program to facilitate the sharing of product and process knowledge between CDER and ORA for high risk or complex products or processes.

    Priority inspections are typically triggered, she says, when:

    1. the establishment is named in an application to FDA for first time
    2. the first application is filed by applicant
    3. the first ANDA is filed for an approved drug
    4. the finished product contains a New Molecular Entity (NME)
    5. the finished product or API is manufactured by a substantially different manufacturing process, previously covered by the establishment
    6. the API process involves high risk or its intended use has significantly changed
    7. there have been numerous application submissions or certain site/process/product changes that could pose significant challenge to the state of control of facility
    8. the profile class status of application product or API is unacceptable or has not been updated via a site inspection within the past two years.

    The manual calls for a data integrity audit, Gooen said, with FDA taking action against companies that commit data fraud or provide false information, citing the Ranbaxy “application integrity” policy set in September 2008.

    Between 2003 and 2009, she said PAI withheld recommendation most often for one of the following reasons:

    • enforcement action was pending, or previous deviations persisted
    • firm not ready/drug not made here/facility withdrawn
    • insufficient development data
    • inadequate lab controls
    • inadequate QA functions

    Attention to detail is important, she noted, in particular the listing of all manufacturers in one place including such basic information as Name, address, email, fax, contact, registration numbers, clear description of responsibility.

    Later, Ms. Gooen outlined PAT and QbD, and presented some studies involving products that failed to pass review, where a more thorough QbD approach could have led to very different outcomes.

    Baltimore District investigator Ernie Bizjak discussed the new preapproval inspection compliance program based on his experience.  As he explained, the new program focuses on evaluating:

    1. readiness of commercial manufacturing
    2. conformance
    3. data integrity

    Manufacturing readiness, he says, is determined by:

    • Assessing manufacturing and lab investigations related to proposed commercial processes
    • Sound and appropriate sampling and testing program for components (for instance, heparin); asking about supply chain, questioning sampling programs
    • Sufficient facility and equipment controls to prevent cross-contamination
    • Adequate procedures for controlling changes, investigating failures, adverse events, conducting recalls, reporting information to FDA: Quality systems
    • Evaluating the feasibility of the proposed commercial process and manufacturing batch record. This will include a review of scale-up batches, with an eye to how well the firm understands its manufacturing processes, and a review of process validation activities, if they’ve been completed.

    Conformance to application will be noted by observing, processing and/or testing operations and the establishment’s knowledge transfer program, he said, while an audit hardcopy and supporting raw electronic data will be used to authenticate data submitted in the CMC section of an application. More details, he said, can be found on the FDA website,, by 7346.832 into search bar.

    Related posts:

    1. It’s Official: FDA, EMA Will Pilot Parallel QbD Review
      FDA and EMA announced a new pilot program that will allow parallel evaluation of QbD-based drug submissions....
    2. Should Fast-Track Process Serve as a Model for All FDA Approvals?
      New research suggests that an accelerated review process results in “no material differences” in post-approval activity for the manufacturer and...
    3. Snippets from AIChE: Navigating QbD’s Shades of Gray
      A Q&A session focused on subtleties of QbD and QbD-based regulatory filings, suggesting that not everything always needs to be...
    4. FDA Expected to Join PIC/S This Year
      The event would mark another significant step toward global regulatory harmonization and information sharing....
    5. 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....

    3 Responses to “Process Variability? FDA, Heal Thyself”

    1. epcotint says:

      Having worked for a government agency, new recruits are going to ask questions as if they are the “king of the hill” and this is disaster. More delays.

      Question has to be asked did McKinsey improve the application review performance by a half a day. McKinsey is good but do the MBA’s know what chemistry or chemical engineering questions to ask. They do not.

      Until we move to “Science and common sense” based review, regulation based process is going to drive us six feet or more under.

    2. malcolm ross says:

      Many thanks for this report for those of us who could not attend. Anyone who asks the question “why isn’t the industry rushing at QbD” should simply read this article. There are more and more agency recruits with little or no industrial experience and sometimes not even relevant academic experience and this is not an FDA problem-it applies to all regulatory agencies.

      I run training courses such as “Analytical Chemistry for Regulatory and QA staff” and I have had some quite surealistic experiences including a QA person responsible for QC in a very large pharma company. I could see that she wasn;t with me during the very opening training session so I asked what her background was—-a degree in consumer affairs-taken on to deal with consumer complaints which then became part of QA and a work experience program moves everyone round every year!!!!

      In the agencies themselves the situation is very similar-recruits are from outside the industry-the salaries are low-they are overworked and don’t get to go to training courses or interact openly with industrial people until they are at the top level. And of course, on the way up many leave and go into industry because of the supposed advantage that they have regulatory experience.

      No wonder we have a problem.

    3. Pedro Hernandez says:

      Very nice summary article!

      This is an area where a more frequent and focused use of Pre-Operational visits (POV) could benefit both sides. The agency could benefit from the technical/scientific knowledge and practical experience of the industry and the industry could get a feel for agency expectations. And more important, by rotating inspectors, it would expand the inspector’s expertise and help avoid silos in which certain teams may set different precedent/requirements for similar operations from different manufacturers.


    Leave a Reply

    You must be logged in to post a comment.