Practical Tips on QbD from FDA, MHRA & EMA

Posted On

This post is packed with practical tips on QbD (Quality-by-Design) and other topics, straight from the global regulatory agencies. Officials from Regulatory Agencies (FDA, EMA, MHRA) opened up the floor to the industry for questions. Since the scope was beyond QbD, I bolded the QbD-relevant questions and responses from the regulatory agencies. Below is a list of key topics:

• How should Design Space be incorporated into the Product Lifecycle Management or Quality Management System?
• Have Pre-Approval inspections and biannual inspections been reduced due to Prescription Drug User Fee Act (PDUFA) and Food and Drug Administration Safety and Innovation Act (FDASIA)?
• Can FDA’s proposed Quality Metrics measure R&D’s health or Return on Investment?
• How are regulatory agencies (FDA, EMA, MRH, etc.) harmonizing requirements from the pharmaceutical industry?
• How do we shift industry’s paradigm from “just compliance” to “continuous improvement culture”?
• Why FDA monitors the health of the company’s business and senior management?
• How will FDA treat Traditional versus Quality-by-Design Submissions?
• What are the minimal expectations from the Quality-by-Design framework?
• What did FDA observe in industry’s Risk Assessment practices during inspections?
• What are the key questions FDA inspectors ask regarding Risk Assessment?
• What are the things that makes FDA reviewers and inspectors upset and you see it very often in submissions?
• What are some practical tips from the FDA regarding Pre-Approval inspections and submissions?
• How to bring FDA reviewers to your side.
• What are the updates on process validation guide, Annex 15, and
• the NIR (Near InfraRed) guidelines?
• Why we can’t use USP guidelines for Regulatory submissions

Panelists were:
Rick Friedman (FDA), Steve Kozlowski (FDA), Tara Gooen (FDA), Christine Moore (FDA), Ian Rees (MHRA), Sabine Haubenreisse (EMA)

To keep the questions and answers in context, I bolded them in the original conversation from my notes as a starter.

In the near future, I will update this with the summary of Q&A. (Subscribe to receive this)

 

Here is the conversation:

———————————
Which end shall we
start?

Rick: burning issue.

Moderator: Let’s start with Rick.

Rick: My name is Rick Friedman. My affiliation is CDER Office of
Compliance, and my burning issue is just general implementation of quality
systems, robust quality systems, in the industry.

Tara: Hi. I’m Tara Gooen. I think you all heard me talk earlier.
Actually, it was
pretty empty, then, so welcome.

Moderator: It wasn’t personal.

Tara: My burning topic I’d like to talk about is in terms of focusing
on quality systems,
but change management systems, how to beef those up, improve
those, and your thoughts.

Steve: I’m Steve Kozlowski. I’m Director of the Office of
Biotechnology Products at the FDA, and for burning topic,
I sort of want to continue what we
talked about earlier, which is how to formalize what are
regulatory commitments and the role of control strategy might
play in that.

Christine: I’m Christine Moore, Acting Director, Office of New Drug
Quality Assessment,
FDA CDER. Our office reviews INDs, NDAs, and supplemental NDAs
for new drugs, and I think you guys will have plenty of burning
topics.

Ian: Good afternoon. I’m Ian Rees from the U.K.’s MHRA. I’m a GMP
inspector, so
I’ll tackle GMP type issues but not assessment type issues. The
kinds of things that keep us awake in Europe are shortages, but
the other big topic of interest is support for innovation.

Sabine: I’m Sabine Haubenreisse, European Medicines Agency
Liaison Official at the FDA. In my role, I’m facilitating dialog between both
agencies, so my wish list is moving from confidence building to
mutual relations in the area of inspections.

Moderator: Thank you. Well, the way we have this set up, it’s very
important that you have
questions from the floor. I have some warm-ups here while you’re
getting your thoughts together and we get the conversation
going.

Oops. This microphone’s in a funny spot.

So feel free to hand question cards into the center. Folks will
be coming around handing them out and handing them in if you
want to remain anonymous with the cards. Otherwise, Roger here
has the microphone.

I think Steve Kozlowski has sort of queued up a question
that was talked about this morning, but I’ll just take one angle
of it, and that was the question about regulatory commitments,
and I’ll equate that not rather to control strategy at this
point, but quality systems.

I noted in the talk that Tara gave this morning that if you give
us some upfront information about your facility — let’s see if
I quoted right — we may be able to waive the PAI if we have some
good supplementary information in there about the facility and,
of course, you have a good track record. I would expect
also

And from what I heard in Lynne’s discussion this morning, there
was this PALM tree, I mean PALM plan, that was put in place that
sort of gave you some sort of a picture of life cycle management
and control, which was sort of the theme of Q10.

So in this QBD paradigm now and sort of successful approval of
design space, which puts some of the change management in the
hands of the company, how do we distinguish these very clear in
the applications, or what examples of success might have been
seen?

We can start, maybe, Steve, Tara, Tara, Steve, and let anybody
shoot at it.

Steve: So examples of success in terms of having a change
management theme?

Moderator: Well, even an understanding that if I take Lynne’s example, you
want to make
sure that CPPs are noted in the design space of control, but
then there’s a broad picture of CPPs. So it seems like the PALM
plan, in that one example, helped you gain an understanding of
that.

Steve: Right. Again, I think, as was discussed earlier, the issue
of criticality can
become….

Moderator: Can folks hear you?

Steve: The issue of criticality may become something that there’s
disagreement over,
negotiation over, but the question is what is the ultimate
strategy for controlling your product and process, and are there
changes that can be describe in some way to put clarity on what
the reporting requirements would be otherwise, like a protocol.

I think the PALM certainly dealt with changes within the design
space. Then there are things, though, that are not covered, and
those would be dealt with outside of the PALM and in terms of
the normal way of risking changes.

Moderator: So would you say the quality system would handle those changes
in some sense?

Steve: Well, I think if it defaults to what we’ve been doing
before, then it depends upon
whether it’s reportable change or not.

Moderator: Oh, yes. That’s a baseline.

Steve: So, yes. But I think the PALM, in having this information,
makes more clarity
about what things need to be reported and how to measure risk. I
think it’s an advance, but when you don’t have that, you default
to the classic ways of looking at risk.

Moderator: Well, thanks. Any comments from the PAI or GMP colleagues as to
how to make
that integration? Tara, you mentioned something about PAI, I
understood, and how certain things in the submission could help
that process.

Tara: I think fully understanding what the responsibilities are at
each site will help us
make a good decision about whether or not it’s valuable to do a
preapproval inspection versus a post-approval inspection versus
both or versus neither, just cover it on a future surveillance
inspection.

I think clarity on those responsibilities and oversight is
really good to have in the application from the decision making
perspective.

Moderator: Any questions from the audience so far?

Barry: I do, Joe.

Moderator: Could you state your name and affiliation, sir?

Barry: Sure….. As part of the
PDUFA legislation
was supposed to be a gradual changeover to biannual inspections
as opposed to individual and fewer PAIs, could you tell us where
that stands?

Tara: You want to tell them. I’ll start.

Rick: Okay.

Tara: Okay. For years, there was always the hope to do biannual
inspections at
manufacturing sites but the PAI program really ended up driving
a lot of the resources.

So the inspections were just being done at the sites that needed
a preapproval inspection and moving away from the two-year, kind
of forgetting about the firms that haven’t submitted a recent
application.

Since, I think, 2009 — and I think Alicia had this reflected in
one of her slides — we’ve been moving away from that and
focusing more on doing the routine surveillance inspections so
that we don’t have applications that are being held up for
surveillance preapproval inspection.

I think that there are efforts to move away from that. I think
we’ve made a lot of headway in that as well in terms of
different focused areas to be able to move away from a
preapproval inspection that’s only done because it’s coupled
with a surveillance inspection.

Rick: Inspections won’t necessarily be biannual for several years. As
you know, we’ve
had risk-based planning and continue to refine the model, but the
parity between domestic and international will, per the FDASIA,
for that amendment to the act, be a requirement in a risk-based
planning, and in fact, we have been working on exactly that, and
it is a consideration in how we are doing our annual planning
now.

Moderator: Is it correct to say then, Rick and/or Tara, that as a baseline
for generic drugs
under the user fees now, there is support and funding for post-
approval inspections, which is somewhat different than we’re
used do with the Prescription Drug User Fee Act, which was
primarily focused on preapproval activities?

Rick: There’s more funding because of the PDUFA provisions than
FDASIA, yes, and
that actually will allow us . . .

Moderator: It’s somewhat of a distinction.

Rick: . . . to make the increase in surveillance inspections that we
need to make. For all
inspections, we have to use appropriated money also with user
fee money because PDUFA obviously doesn’t allow us, for the
innovators, to also do the proper parity in domestic and
international, so we’re still going to have to use . . .

Moderator: It has a trigger in it.

Rick: . . . appropriated money for that. But it allows us to do the
full inspections
program the way it really should be done, and also it’s going to
help the generic industry with a backlog, because it wasn’t very
funded, as we know, the whole generic review and inspections
program.

Moderator: Question from the floor?

Sun: My questions relates to quality metrics and
Quality By Design and Innovation.

In quality metrics–the proposed four– are all
related to manufacturing, and that is related to response after
the fact, not so much on the upfront work where the Quality By
Design is trying to change.

So is the scope of quality metrics to measure indicators of yesterday?
Can it be tied? Can it have a larger scope where we can
encompass a new metric not just only related to manufacturing,
but also upfront product development as well?

Rick: It’s a tough metric, but we think that design programs could be
semi-quantitatively, probably, graded in terms of level of maturity.
But our first goal on the metrics workgroup was to create a few
easy metrics that are currently collected by the industry so
that we can actually ask the industry their experience with
those metrics.

We understand that they’re not sophisticated enough or
sufficient to actually gauge the state of control on their own
of a company, but when we find the right metrics and we start
mixing that and we augment what we have from inspection data,
we’re going to have a much better picture when we’re ultimately
done in this.

Those lagging indicators that you refer to, there will be others
also that are leading indicators. Again, one of them that we
have discussed in the workgroup was some sort of gauge of how
robust the company’s development program is.

That could be how well the program is designed in terms of its
procedures and standards and whether it uses DOE and things like
that.

But I think it would also have to have outcomes connected to it,
like does it scale up successfully? Does a company have a
history of successful scale-up?

Steve: It would almost seem having good quality metrics allows
you to better judge
development programs. So even if you don’t have specific metrics
for them yet, clearly, you’d be able to get a relationship
between what’s put into process development and measures. And
even though they’re lagging measures, they’re certainly not
lagging compared to shortages or quality problems or recalls.

Moderator: Well, that’s interesting. One of the findings in the ISPE drug
shortage report was
when a problem developed over the commercial life cycle, then
there was a sub-question in the survey: was it known at the time
of development or at the time of tech transfer? In most cases,
yes.

We got — I don’t remember the numbers — but we sort of got a
hit on that question. It was sort of known at the time of
development, and it just didn’t get resolved, and then, hence,
they had experienced a shortage.

That was an interesting bullet in Lynne’s slide, I thought,
about getting through this QBD process as proactively preventing
shortages, which is something I think we’re thinking more and
more about in the preapproval process, particularly with
breakthrough therapies.

It’s almost like the dividing line. If it’s approved, can you
now supply the patients it was intended for?

Christine, I thought you were about to get activated now.

Christine: Yes, I just wanted to add in there that I think those two
things are indirectly
Related, in that if a good job is done in development and
there’s a thorough understanding of the products and processes
and the risks and it’s communicated well to the manufacturing
sites, that should be reflective in the performance measurements
that are collected. It may be a few years down the line, but I
think they are already interconnected.

Moderator: Questions on the floor here? I think you used all your spots on
the metrics section,
Lynne.

Deb: I’d like to ask about more than just metrics, Jim. Deb. It’s nice
to see my former colleagues here. I had a question for Sabine
and maybe for Ian if you would comment.

Sabine, you mentioned on your wish list is moving towards mutual
reliance in inspections. I was wondering if you could talk more
about the progress that’s being made and how you see that going
forward.

Ian: Perhaps I could make a start on that. I supposed there is a
difference. It sounds
subtle, mutual reliance as opposed to mutual recognition. Mutual
recognition is a formal agreement between
the EU and, in this case, the U.S. FDA. Mutual reliance still
requires an agreement, but it’s not as formal, so it’s an easier
target I guess.

What we’ve had for quite a few years, there’s the API Pilot
Programme. There’s the API Pilot Programme and the finished
product join inspections.

What we had in place, which probably needs to be reinvigorated,
was a program that was an agreement in 2011 where we could —
and this is a reciprocal arrangement — where we could waive or
defer inspections.

That did happen, and it did start. I know MHRA did defer a few
inspections, but I think, probably, that program could be
reinvigorated, perhaps with the trade agreement that’s in
progress at the moment.

But there’s certainly quite a lot of experience with joint
inspections, good working between EU authorities and FDA
inspectors. And then you have to say, “Well, how much more do we
have to do before we start taking more active steps in mutual
reliance?”

I think I may have mentioned in another session that just the
U.K. alone does about 40 inspections per year in the U.S.,
somewhere between 40 and 50. Okay, we may have to do some of
those still. I don’t know if they’re new sites to us, but if we
can redeploy on some risk basis that inspection resource and
then rely on work that’s done here by FDA, then surely that’s
better for all.

Sabine: Just to complement that, obviously, you know that we have
the European
Falsified Medicines Directive, which came into force this year,
and the U.S. has been listed as well as our
authority. So there, from Europe, we were relying now on the
help from authorities to help us, as
well to oversee the quality of API manufactured outside of the
Europe, and the API that will come into Europe.

So there, I think, the Europeans have created a pathway to rely
more on each other, and we hope that this could be a good module
as well for the implementation of FDASIA, to do this in a timely
manner, to really help. Rather instead of going and inspecting
the same site, rather being able to have a wider range of
manufacturing sites.

Me being embedded in the FDA, I’m obviously talking regularly
with colleagues on the API inspections but also on the
inspection of finished product. So we’re hoping there as well
and on the ground to move this forward.

Deb: Thank you. That’s interesting. I will tell you. You probably
know this, but the
comments I hear a lot in the industry are, “Well, agencies are
going to rely on each other to take negative action against us,
but not to forgo additional inspections or approval.”

I think that if that causes some resistance, and I think the
more that industry can see regulators working together in a way
that alleviates regulatory burdens, I think it will really help
to get more support in the industry. So it’s good to hear you’re
taking steps forward.

Moderator: Deb, thanks for your question. We addressed the GMP, but in
terms of CMC
cooperation, would the panelists care to talk about how Europe
and the United States and other regions are cooperating?

Christine: Well, I’ve already spoken in the last session rather
extensively about our work on
the EMA/FDA pilot for QBD containing applications. At this time,
the way that pilot is set up is that the reviews are done
independently by both of the organizations, by both of the
agencies, and from our regulatory standards, they really need to
be done that way.

As a first step, though, we’re trying to harmonize our
approaches, trying to harmonize or use common lists of questions
or information requests to the applicants. There is not a plan
anywhere in the near future for mutual recognition of
applications, but I think we’re taking a first step.

Moderator: Thank you. We have a question from the floor.

Steve: No, just to follow up.

Moderator: Oh, I’m sorry.

Christine: I think Sabine has something.

Sabine: Just in this context, I also wanted to remind the audience
that we do have as well
the possibility of parallel scientific advice. That’s anything
that is open as well for industry. If you have a parallel
application and you want to have advice from both regulators,
that’s a very good opportunity.

More and more companies are coming to us, then, and do submit
applications in parallel. And we have the opportunity there
between agencies to look at the questions, discuss between
experts, and then relate that feedback to the company, and that
should be as well helping to move the global development forward
also on the CMC side.

Steve: For biologics, as Christine mentioned, we participated in
terms of an EMA pilot
that started by ONDQA, but I want to throw an idea out which has
been brought up at number of different venues.

If you really wanted to push a global approach to an
application, could you find a rare disease where the population
was spread in many different regions and basically make the
argument that the clinical trial population is so small that it
basically has to be a global trial, that it’s an orphan drug,
and that the resources to do this would be very, very limiting,
and if it’s children, have the parents of those children lobby
their governments?

Because, in some sense, the global application is both a
political issue as well as a scientific issue, right? So if you
found a group where there was tremendous political interest in a
global application, that might serve as an example.

I don’t know if you can find the disease and organize the
parents, but it would certainly be an interesting strategy.

Christine: That was a lot of ifs.

Steve: Yes.

Moderator: Over here, question on the floor.

Rocco: Yes, my name is… First, a
reflection. We’ve had
some really good presentations, technical presentations on
quality management systems, quality risk management, and to be
honest with you, personally, I attend these conferences almost
every year.

The quality of the presentations are great. There’s a lot of
energy. But there is still a gap between what happens at these
meetings and what happens when you go back to your firm.

At a minimum, these firms are looking to comply, all right?
Meaning that instead of focusing on quality management systems
and continuous improvement, we’re always looking to ask what the
regulators want, what the regulators are looking for, instead of
understanding our business.

The question I have is, in terms of unintended consequences, in
particular with the FDA proposed plan for KPIs and metrics, what
is the reflection from the regulators in terms of how we can
shift the paradigm and how we can help the quality organizations
in particular to address this lack of focus on a quality management systems
perspective?

And again, in regards to the surveillance program, beyond the
quality metrics and the other metrics that the agency typically
collects, are the regulators looking at signals that are related
to the economic health of the firm and some of the
organizational changes?

Because I would propose that the level of QMS integration and
importance of quality in quality management systems is directly
proportional to the health of the firm, meaning that these firms
that are losing patent protection on products are cutting
resources.

There’s no time to do tech transfer. There’s no time to do
transition plans, just your thoughts on that.

I mean, that’s the reality on the ground. I think sometimes we
come to these meetings and it’s kind of a Pollyanna kind of
atmosphere. What’s the response from the regulators on this?

Moderator: Maybe you could repeat that question? (Crowds Laugh) Folks may want to take
different segments
of it — metrics, quality system, and getting in with industry.
Maybe you want to take on from our colleagues here.

There was also something about some unique things about the
finances of the firm and all, so maybe, Ian, you could go with
that.

Ian: Perhaps I can make a start. Since 2009, in the U.K., we’ve put
in a risk-based
inspection approach. It was fairly . . . I don’t know. I
wouldn’t say we gained a lot, but we started to gain some
experience.

There were mostly factors on-site where we assessed the risk of
the company, and then that informed, then, the re-inspection
into the duration of time we’d spend on-site.

Just earlier this year, since it’s still early stages, what
we’ve put in, we’ve invested quite a bit of money to put
together a software package which has an algorithm in it which
looks at a very broad range of factors in a company, not just
compliance history, and it’s taken quite a bit of doing.

It’s still at the formative stage where we were looking at some
of the more complicated structures and dependencies that the
companies have and putting in factors like mergers and
acquisitions and broader factors which might inform the risk
either at a site or perhaps a related group of sites.

So it’s not there yet, but it’s a work in progress, and we hope
that will have a broad perspective. As you say, there are some
broader, perhaps more subtle, factors that we can look at which
might actually be quite informative or revealing about the state
of compliance or risk and how that might change in a company.

Rick: The thing about mergers and acquisitions is people say that
there’s the adverse
impact on the company and their quality system often a couple
years after an M&A, not right at the time of the M&A, because
they’re trying to integrate at that point, and that causes
confusion and problems and inertia in decision making, including
improving on areas that are causing failures, for example.

That was a good question, a tough one. It had a few parts to it.
I think one of the ways that we can start, that the FDA can do,
as you said, to counteract this business influence that may be a
financial-as-job-number-one type of mindset that certainly gets
most prominent when there are financial difficulties is, I
think, we can ask more sophisticated questions at inspections
about senior manager responsibility and their awareness of
issues and whether there is right escalation processes and what
they do when they see those issues.

I think, from my experiences at what we would call an unhealthy
quality system over the years from when I did inspections back
in the day and today see in compliance cases, is really nice
charts are sent up to senior management saying, “Everything’s
going fine.”

But when I would actually look at the data at the process level
or equipment facility level, I saw there were a lot of things
going on. I was like, “Is senior management aware of this?” I’d
say, “Well, show me the chart you gave to them,” and it would be
all averaged out and everything would look great.

I don’t think that type of approach is statistically honest.
It’s not showing whether there are problems or whether there are
signals that a burgeoning issue may get out of control and right
now is simmering but actually can get to the crisis level out of
the crisis [inaudible 00:28:28] was talking about. So that type
of escalation process in organizations and things like that, I
think we have to ask more questions about and not stay only in
the weeds in inspections.

Inspections do a very good job. I think a lot us realize you
don’t always agree with everything that an inspector might say
in one of your inspections, but they do a very good job of
sniffing out issues.

What we want to make sure they do is that they connect to the
systemic failure and the management failures a little more
often. I think what the FDA can do to improve is asking those
questions and then having senior management explain, come in to
the conference room or whatever, not just the discussion to explain what their thought process was
in neglecting a recurring issue, for example.

Steve: To me, the core of this question was if you produce
metrics or very specific
standards to me, then industry will basically meet that
regulatory standard and will not create the culture of quality
we’ve talked about.

So I think . . .

Christine: You stole my line.

Steve: Well, sorry. It’s a very useful line. Christine wants
credit. But I think the way to
deal with that is that we hold industry accountable to
standards, and we’re very dynamic in the metrics and tools we
use. These metrics may be a starting point.

We also do acknowledge management be a learning organization in
the sense that we put all this together. OPQ , one of the key
ideas of it, is that we begin to integrate all this different
data, and it may turn out we find patterns that predict problems
in industry in a multivariate way that we haven’t seen as well
because we’re not looking at it that way.

I think as long as standards and principles are what’s fixed,
but the way we assess them is dynamic, then you can’t just study
for the test, because we’re changing.

Rick: The last part of your question was economic health, are we
looking at it? And yes,
we’ve had meetings to discussing how we’ll look at it, which is
carefully figuring out how we look at the balance sheet of a
company, not necessarily is a blockbuster drug coming off.

Moderator: We’ve got a lot. Any other comments on this question?

Tara: If I could comment on that as well. In terms of follow-up to
what Steve was
saying, I think it’s really important that there are different
areas and different perspectives on coming up with what the
right metrics are.

I think it’s really important that a site’s quality systems be
very involved in that as well and have those different
perspectives, because, otherwise, we’re going to see the same
kind of silo approach that you’re thinking might be a result of
this effort.

But if we bring the right people in, I think that’s part of the
proposed reorganization within CDER as well, to bring those
different perspectives together to come to an even greater
outcome. I think it’s important to do that within companies as
well and bring those different perspectives together.

Moderator: Thank you. We have a question. Let’s get the next floor
question. Excuse me.

Diane: Thank you. I’m going to
apologize for taking
us back to the previous question, but I have full confidence
that we’ll end up back to the metrics question later on in the
discussions.

I really appreciated the discussion and the mutual or, I guess,
the parallel assessment that’s going on, that collaboration
that’s going on across the agencies as we look at the
breakthrough therapies.

The comment I wanted to add there is really to encourage you, in
that program, to stay connected together through the post-
approval changes that will inevitably come for those
breakthrough therapies.

I think that when we talk about significant challenges for the
industry — and I want to acknowledge Roger, who really
presented this, I think, also in the session earlier today with
the slide that showed one product with now 24 different varying
registration requirements across the world — this is really
significant for us. I also just want to share another example
and a number for you.

So the way we manage these different changes as an industry is
we typically do what’s called lot flagging, and we have lot
restrictions on our batches, and in the complexity of this
environment, I was speaking with a colleague over these last two
days who mentioned that for one of their products, they have now
48 batch restrictions on a single batch awaiting approvals in
different countries.

I just really want to offer that to help you understand the
magnitude of what we’re struggling with. It used to be that some
of these countries would accept approvals from major markets and
they would then approve, but that’s not happening now.

Again, I just really want to encourage the work you’re doing —
it’s fantastic; we need more of it — and really encourage this
concept of getting us to mutual reliance on applications.

Christine: Thank you for bringing that up, but I think it is going to
deserve some additional
discussion internally, I think this is an idea that we will
bring back and have those discussions, so thank you.

Moderator: Another question from the floor.

Ralph: Hi. I just want to
make a comment, if I
may, on Rocco’s commentary about economics. Although I think
it’s a very good question, I think it’s interesting because when
we look at this, the companies that are laying off thousands of
people, quite frankly, it’s not like they’re in the red.

They’re just not making the projections that we wanted to. I’d
be very interested to attend a future session within this
conference on reconciling the business model with the goals of
pharmaceutical quality systems.

I guess we could go back to that slide that I think it was, Joe,
that you had where we had the results of a survey where we asked
the top executives about, or someone within your session there,
about sacrificing long-term company health with short-term goals
or something like that.

Moderator: That was…

Ralph: Thank you. My question is . . .

Rick: No, it was me.

Moderator: Oh, it was Rick. All right.

Ralph: I didn’t think we needed a survey for it, either.

Moderator: Okay.

Ralph: Christine commented that generics have made some QBD
submissions. Could
you give some rough percentages of each of the total ANDAs and
NDAs that are going in as QBD submissions?

Then, in regards to the generic QBD submissions, what’s the
perception of the overall quality of those? Could you elaborate
if there are shortcoming and what those are in comparison?

Christine: Okay. I think Steve mentioned this in his talk. We made a
mutual decisions at the
CDER review divisions that we’re not going to be calling
applications QBD applications because they’re really isn’t a
clear dividing line between what should we consider Quality By
Design and what should we not.

As I discussed in my presentation, Quality By Design is a
science and risk based approach, so how much makes a QBD? If I
use some science and risk-based approach, is that . . . anyway.
So we were taking metrics in ONDQA.

In terms of counting, we weren’t counting the submission
contents but the regularity flexibility from that, and that
became confusing to people. I think it was self-defeating, too,
because it really was looking at the wrong metrics.

Now, regarding generic drugs, there has been quite a bit of
discussion from generic drugs on using QBD concepts, the
elements that they’re recommending. Well, frankly, some of them
are minimal expectations under ICH QA — identifying QTTP, CQAs,
critical process parameters, and clearly defining a control
strategy. Those are things that, again, in terms of ICH Q8 R2,
are expectations for the traditional approach.

What’s different is really an emphasis in the generic drugs on
better enhanced process understanding, product knowledge of
process understanding. That has been emphasized through
publication of various case studies, both for immediate release
and modified release applications.

Again, it would be difficult for me to quantify. I think at some
point someone was doing that what percent has QBD has elements
in it and what those QBD elements are. Again, it’s not an easy
black and white designation, so I can’t give you statistics on
that.

We know that QBD approaches, those science and risk-based
approaches, are being done more and more. I don’t know of any
examples of design spaces, for example, if you talk about the
regulatory flexibility being included in an ANDA, but that
doesn’t mean that they haven’t been, because I work on the new
drug side.

We know it’s increasing in terms of generic drugs as well as
biotech products. It’s difficult to measure.

Moderator: Does anybody have a . . . who has the hottest card on a new
topic they’d like to
pick up? Steve’s got it up.

Steve: Okay. I have a question about information in the Quality
Overall summary.
What is an example of a regulatory commitment to be included in
the Quality Overall Summary, and is change management
appropriate to discuss in the Quality Overall Summary?

I don’t think that either of those are necessarily covered in
the Quality Overall Summary the way that we utilize it, but they
could be in the application. So there’s a lot of discussion
around regulatory commitments and what sections, if it’s limited
to sections, would tend to contain regulatory commitments. They
can be pointed to by a post-approval management plan that we
heard about before.

Regulatory commitments can be in the application. For instance,
process description could be part of regulatory commitments. The
control strategy is clearly a regulatory commitment.

And change management, I think there’s a lot of discussion about
how much of QNF’s change procedures need to be in an
application, if any.

I know you’ll jump in.

All I’ll say is I know for biologics in Q11, it says that within
a design space, you can have a protocol describing how you would
manage the change, and the PALM that we heard about today is in
fact an example of managing changes.

So there are levels of change management. Clearly, the details
of change management are part of a company’s SOP. They’re looked
at in inspection. They’re not in the application, but certainly,
some aspects of change management in the context of the control
strategy, potentially in the context of protocols, whether about
a design space or other things, clearly would need to be
described somewhat in the application.

Christine: If I can follow up from Steve. In terms of looking at
regulatory commitments, I
think about, well, in terms of two aspects, things that
essentially are in the application as ways that you can operate.
If you look at the regulations, they give you the ability to
propose alternatives to your process.

For example, I look at design space as an alternative to your
process that you’re operating within. You can operate within
this particular range without any additional filings.

When you look at the first paragraph of the change regulations,
314.70, it talks about changes to the application beyond those
already approved. So I think we do need to clearly define what
is already approved.

Going beyond those changes, the protocols are very useful
mechanism to do that, and that’s where Steve came in with the
PALM.

I look at it as two parts. One is that you have approved
alternatives, approved variations to your applications. And then
you can also, through a protocol, have a description of how
you’re going to define changes that could then justify in
reduced reporting.

Regarding tying it into the QOS, the QOS, as Steve said, could
consolidate and point to the areas in the application where
these different things would be explained. For example,
alternate processes would most likely be explained in the P3
section with the process descriptions, etc.

So it could be a kind of place to tie it all together, but
that’s still an area being developed.

Moderator: Tara, go ahead.

Tara: Thank you, Joe. If I could talk more about the idea of change
management within
the application or being part of the QOS, I’d be very cautious
about submitting the procedures and very explicit details of
actually how you plan to manage changes within your facility.

That process needs to be managed within your quality system. It
needs to be improved over time. It’s not something that can be
reviewed on paper, if you will. It’s really a living, breathing
system at the facility.

So that being said, I did appreciate some of the language that
was in Roger’s presentation about the reviewers having
confidence in a pharma’s ability to manufacture or set up
systems properly.

I think that a description could be provided as long as it’s a
current description in terms of providing confidence, not
necessarily locking in that particular process or change
management system.

Moderator: Yes, sir. We have a question from the floor?

Steve: Good afternoon. I just wonder
if you would just actually comment a little bit about the
potential for risk-based reviews that’s been discussed a little
bit and how that might look in the future.

Christine: I like this topic. There are a lot of changes that we’re
looking at in the proposed
Office of Pharmaceutical Quality, and I didn’t show the slide.
Steve had some of the bullet points. I’m trying to see if I can
remember all of them.

Some of the aspects we’re looking to do is better integration
between review and inspection, more risk-based approaches
towards review, having that surveillance function that we’ve
been speaking about, and there’s a fourth one that I’ll probably
remember when I move on.

Let’s focus on the risk-based review perspective. I think one
thing that we did that was pretty smart of Dr. Woodcock when she
started to look at the reorganization is rather than spending
the first few months talking about organizational boxes and
what’s in what and who goes where, we talked about how we’re
going to work.

One of the subgroups that I lead, I think, was very exciting. It
talked about how we as review staff and inspection staff can
integrate risk review throughout the product life cycle.

We came up with some very good ideas that I’m hoping we can fold
into the OPQ work processes on how we can look to the risk to
the patient through the product design, through the process
design, and through the implementation and use that as a basis
for our decision making, our regulatory decision making, use
that as a basis to make better decisions on our workload
management, and use that as a basis of knowledge management.

Again, it’s in the very conceptual stages, but I think we have a
commitment to move that forward. I’m pretty excited by it
because I think it’s very powerful, and I think when you look at
what we’re tasked to do, especially with some of the things we
need to do delivering to generic drugs in the backlog, we have
to work more efficiently.

You in the industry, I think you’ve seen that risk-based
approaches can make you work more efficiently. You’ve shown
that. I think we’ve realized that, and we need to start doing
that more ourselves.

Moderator: Thank you. Anything on the floor here? We’ll go to the next
card. My colleagues,
Ian and Sabine, I know you had some cards.

Ian: Yes, okay. So I’ve got a question here about what are some of
the roadblocks to
harmonizing inspectional resources between different
authorities. In other words, why is no progress being made on
sharing inspectional information between countries?

Perhaps rather than thinking of it as a roadblock, perhaps we
could think of it more as roadworks. That may be better, because
it isn’t a complete blockage. There are successes. There are.
Certainly, I can speak from the European perspective. We’ve got
mutual recognition.

So for example, we rarely now have to go to Canada. We never got
to Switzerland. Sadly, we never go to Australia or New Zealand.
We rarely go to Japan.

Within the near future, when they join PIC/S, the mutual
recognition agreement will be extended, so we won’t go there for
biologics or sterile products either.

So there’s that very formal side, and that’s the mutual
recognition. Then there’s the mutual reliance side.

Oh, just before I leave that, there’s a European database which
was called the EUGMP. It’s now called EUGMDP because it’s got
distribution in it. And that’s the repository for inspection
information, GMP certificates, and manufacturing authorizations
as well.

And so now that’s something that we’ll, even in the States, have
to place all this information, but we’re beginning to see
countries from outside of the EU beginning to populate that as
well. So there’s a resource to share information.

Then, going on further, what I would really like to see being a
big success is PIC/S. It’s now got 43 participating authorities.
Half of those are EU authorities. More recently, we’ve got the
FDA, and there’s a whole stack of others, including China. Their
express intent is to apply to become PIC/S participating
authorities within the next 10 years.

So that’s a forum. There’s an accession procedure. You have to,
then, as a result of that, have confidence in the system that
each of them has. They may not be identical, but they’re
equivalent. And so in that way, then, you can take into account
the work that’s done by other authorities.

I think as well that there’s probably more sharing of
information than is probably recognized. Certainly, the U.K.,
for example, has a confidentiality agreement with FDA, so we
share information backwards and forwards. So I think there’s a
quite a lot that’s already there.

Hopefully, we can get on and perhaps do some further roadwork
and get a nice, smooth blacktop, we call it, or tarmac, we call
it.

Moderator: We have a question on the floor.

Mike: My question is
regarding risk assessments. Given that risk assessments are or,
I guess, should be revised as more information and more
knowledge is gained and after experimentation or whatnot, how
have you seen risk assessments be captured or recorded?

Are they part of control documents in an electronic system at
sponsors, or is it more of like technical reports adding with
supplement reports, or even just with meeting memos and formally
summarized in things like filings or supplements or follow-up
type things like that?

I wonder if you could comment on that.

Panelist: Tara, what do you see on inspections?

Tara: I see a variety of different approaches. I like to see the
output of the risk
assessment instead of looking at the entire document, so I ask
for or I look at the risk assessment in tangential ways. I look
at the change management system, and in that way, I look at
changes that have been evaluated and how the risk of making that
change fits into the entire risk platform for that particular
product as well as through deviations and CAPAs as well.

That’s at least what I see. I’m not seeing a standardized
approach from company to company. I think that it really
depends.

I am encouraged by seeing risk assessments span more widely
across manufacturing organization than maybe I did in the past.
It was very just one person writing the document, and then it
sort of sits on a shelf. But really integrating risk management
approaches throughout a lot of different aspects of
manufacturing throughout the whole facility, really having that
be a foundation of everyone’s job, I see that.

But am I looking for formal documentation or a formal report or
something like that? Generally not.

Rick: If risk assessment is truly supposed to be an interim process,
which all the
presentations that we’ve heard in these conferences over the
last few years since Q9 was published have said, then the risk
assessment that was done at the time prior to launch, which
still can’t see into the future enough to know what process
experience is really going to be, that risk assessment should
not be static and shouldn’t go on the shelf.

But I think it’s possible that industry might want to think
about — we’ve had this discussion kind of in the hallways and
at lunch with some of the industry folks at this conference —
that maybe at year two or batch 100, there is a risk review done
of the original risk assessment that was submitted or that was
done at the time of scale up to see if process experience tells
you that the control strategy that you chose or various other
things that decisions that were made based upon the original
risk assessment are still correct.

Obviously, at an earlier point, if problems are very apparent in
the first 10 batches, revisiting the risk assessment through
risk review and Q&I would need to be done.

But I think it might be a useful checkpoint at a certain number
of batches or a certain number of years of manufacturing to
really make risk assessments truly iterative.

I agree with Tara, also, that risk assessments are used in
different areas, including design of the equipment and
facilities, that we’re encouraged by.

Tara: If I could add as well, if I think about my history as an
investigator, from the very
beginning, I was trained to go into a facility and say, “What
are your critical unit operations?” before we even had a
definition for critical. “What are the important unit
operations? What are the important parameters when it comes to
this unit operation?”

And now, when I would go in and inspect, if there aren’t clear
answers or responses to those types of questions, there’s more
kind of formal language that’s been developed through Q9 and
through a lot of these discussions. There are different ways to
look at it.

Steve: I think in parts of the risk assessment that are reviewed,
reviewers will ask
questions in terms of are they valid risk assessments. So there
needs to be enough explanation of how they’re derived to
convince the reviewer that they’re valid to make judgments on
the control strategy front.

Some of that validity is based on the procedures that are used.
“Are you really using the right subject matter experts? Are you
facilitating them correctly?”

So those procedures would be important to really have worked out
because they’re part of defending the validity of those
assessments.

Moderator: Yes, Mr. Tyler.

Mr. Tyler: Hi, Joe. Yes, just to build on that, I think it was in
Christine’s presentation, but
you have the internal, the map of the
document, and I think in the context of that document, it says
that, typically, maybe in some submissions, you have risk
assessment just like what you were saying. It also states that
the reviewer, in the absence of a risk assessment, is encouraged
to conduct a risk assessment of their own.

I was just wondering what the agency has done to train
reviewers. I know that there’s training and stuff. Is that
really actually being done to take a step and to connect your
own risk assessments on the information that we supply as
sponsors?

Steve: I think that our reviewers have been exposed to a number
of workshops about risk
assessments and had some training on them. I think we’ve always
done risk assessment, but we haven’t necessarily done risk
assessments that are documented or through a formal process.

I mean, any review is really looking at, “What do you see as
risky in this manufacturing process? How could it affect the
product and, ultimately, the patient?”

So we do need to think, in the cases where we either disagree
with the risk assessment or there isn’t one, about maybe more
formal processes to think about doing that. I don’t think we’re
quite there yet. I know Christine wants to comment more on that.

Christine: We have had some of our reviewers in ONDQA, at their own
initiative, do risk
assessments on products and use that as part of their assessment
of evaluation. We’ve additionally done formal risk assessments
for certain drug shortages or crisis issues.

I know we did a major one when Heparin happened years ago, and
we’ve done extensive risk assessments on certain topics like
nanotechnology, but it’s been mostly on an ad hoc basis.

As I mentioned before, one of the things that I’m really hoping
we will put into the work processes of our proposed Office of
Pharmaceutical Quality is a structured risk assessment format as
part of the review process.

Again, the purposes are to both focus that review to use that as
part of our regulatory decision making, as part of our workload
management, and as part of our knowledge retentions.

Steve: Actually, to add to that, Patrick Swan, formally of
the FDA, has presented in
a number of conferences an FMEA that he designed for assessing
the risk of certain manufacturing process changes. Again, that’s
not a formal tool that everybody uses, but there are a number of
reviewers who, based on that, have continued to use that in
thinking about some of the manufacturing changes they see in
biotech products.

Tara: We have an add-on to this topic, risk assessments. Risk
tolerance tends to vary
from person to person and company to company. Is there a push to
have a standardized scale for severity and occurrence scores?

That’s open to anyone.

Christine: I think that’s a bit down the line. Perhaps we’ll find that to
be useful after we put
these processes in place and really move these concepts forward,
but we’re not ready to do that yet.

Rick: The stuff we just talked about, like formal risk assessment
training and showing
objectivity in the risk assessment process, people taking their
hats off in whatever organizational units or biases or pressures
they have going into the risk assessment, bringing the right
SMEs to the table so that you get the best decisions, all those
kind of skills can help take some of those issues away as well
as getting the right data.

As there’s more data, risk tolerance actually starts to go away
and value judgments they teach you in risk management, it starts
to go away because the data starts to tell the story. So to the
extent that you have a medical doctor to say whether this really
could hurt a patient, and you have data which says that this is
a weak or not weak point in the process, all that stuff can then
start clarifying where there’s ambiguity.

When there’s uncertainty in risk assessment, that’s when
people’s values start entering in and when risk tolerance
becomes prominent in the risk assessment.

Steve: One possible approach to move towards more shared ideas
about risk is for
conferences to be organized around a very specific issue, a unit
operation, a chromatography step, whatever it is, sort of as was
done a while ago when we had Conformia, to have three or four
companies each say how we would for the risk assessment and
compare what parameters they rank as high.

Make it a whitepaper. It’s not a regulatory document. Start
having a shared venue of science around what parameters are
really risky in a process, for instance, and then we learn both
about the risk assessment tools, but maybe even get a standard
idea of how to think about risk of certain operations.

Attendee: a very good idea and something I
especially take as a hint to
do, and I think that’s a fantastic idea actually. I just have one question.
I’m known to always ask weird questions, but here it is.

Christine: You promise.

Attendee: Obviously, you are reviewing a lot of publications — QBD, non-
QBD, one QBD,
half QBD, 0.3 QBD, whatever — and as you review them, sometimes
you feel like pulling your hair and saying, “Come on. Why are
they doing this?”

What is the thing that makes you more upset and you see it very
often in submissions? What is the one thing you would like
companies to change?

Christine: I don’t know if we see it often in submissions, but I get very
upset when a
company puts an inappropriate claim in 6-point font in the
footnote of a table, which we have seen.

Moderator: Any other comments on that topic?

Steve: Well, one issue is to not sort of invent a lot of new
terminology but to stick to
ICH terminology and to explain what you’re doing in a way that
allows a general standard to be applied.

Tara: One thing that we see is that the responsibilities of different
manufacturers that
are listed in the application is not a clear list or provide all
the level of detail that we need as well as sometimes not being
accurate either.

I know that changes happen, but it can be very frustrating to
schedule an inspection, get out on site, and then they say,
“Well, we’re not doing that anyway,” or “We didn’t even know we
were listed in that application,” which happens a lot.

I was out on a surveillance inspection and a preapproval
inspection. It took them a full week to figure out what they
were supposed to be doing for that application because they
didn’t even know that they were listed.

I can tell you that’s something that bothers me.

Moderator: That’s the classic case, when the firm doesn’t know they were
mentioned in
somebody else’s application.

Tara: Yes.

Moderator: Right, question from the floor.

David: We’ve got Q10. We’ve got big
revisions to chapter one
of the EU Guide to GMP. One of the really big changes is the
emphasis on the responsibility of senior management for quality
and the quality management systems.

Could we hear what the regulators expect, the evidence that they
would ask for, and a summary of what they’re currently finding?

Moderator: Why don’t we start with Europe on this one? Thank you, Ian.
We’re afraid to hear
Rick’s answer, actually.

Ian: Yes, the chapter one changes came into effect in January of
this year, so it’s
relatively new. I suppose in general there are the, I don’t
know, slightly hackneyed statements that senior management must
be involved.

I think in smaller organizations, they can’t help but be
involved, and that’s usually a good thing. It depends on the
quality of the senior management, I suppose. In larger
organizations, it’s a bit more developed, and there are a lot of
statements to say that they will be involved.

Typically, though, they are only involved when there’s an
escalation of issues and when something is going seriously pear-
shaped, what we have classed as a critical finding, then they’re
definitely and squarely involved.

But I think that’s after the event. We’d rather see it more
proactive, and there have been some very good discussions
earlier in the week about getting senior management involved.

Yes, there are a lot of issues on their plate, but some of these
more structural issues, they need to put their mind to because
they’re the ones that can cause issues long term. It’s a change
of culture. Again, this is something that I’ve heard a lot, and
it’s one of those difficult, intangible things. You can’t weigh
it. You can’t measure it. It’s a cultural change.

But, yes, operationally, we certainly see the QP involved and
all the heads of departments, but quite often, it doesn’t
trickle up. I don’t know if things can trickle up. That sounds
like it’s defying gravity, and maybe that’s the problem. That
would be nice to see, to have more genuine and proactive
interactions.

I think it’s a work in progress.

Moderator: Rick?

Rick: I think proactive versus reactive organizations get the benefit
in the long term. It’s
like the short-term earnings versus long-term earnings
discussion, the question we heard before.

When the first times a CEO or a president of a company finds out
that there is a recurring issue is when they receive a warning
letter from the Food and Drug Administration, I think we can say
that was not a well-functioning quality system.

I think that even delegation is not
abdication and that there have to be certain systems in place
that assure that significant issues are escalated, as Ian and I
have talked about for many years, these types of issues in
quality systems that make them work right.

The right issues, obviously, certain issues are addressed at a
lower level with a flatter process and are not too bureaucratic
and there are reasons not to raise issues at a higher level in
some cases, too, but periodic reviews, ongoing quality review
boards, and ongoing reports to upper management, all those
things are mechanisms that are not just in the pharmaceutical
industry but they’re in basic management texts that people read
in business school and in systems theory.

All these types of things, I think we’re trying to look for
inspections that are more sophisticated maybe then getting your
VP involved with every single issue. It’s more knowing when to
do it and having the right systems to do it.

I think Tara might have some stuff to mention also about other
things like internal audits, etc.

Tara: Sure. Certainly, the FDA doesn’t look at the details of
internal audits, but we may
look at the frequency of internal audits and the procedures
surrounding them. That’s something we could look at. I have seen
some more successful models on inspections where there is a
routine meeting set up with senior management on site as well as
maybe some off site for the larger companies. That’s worked out
fairly well in terms of making sure that the important issues
remain on the radar and come to resolution.

I would also like to just kind of bring this around to the m-
word as well — metrics. I think that combining appropriate
metrics to be able to . . . because I think it’s a communication
issue as well.

Senior management is dealing with so many different topics and
so many different issues that it’s about distilling down the
major issues to senior management, and one way to do that is
through setting up the right metrics in conjunction with having
a strong vision for a quality culture at a facility.

So those two things combined are so, so important, and I think
that we have a lot of opportunities with the metrics to look at
it from a FDA perspective. Yes, that’s one perspective, but I
would say that’s just one perspective. I think there is a lot of
value in defining the right metrics within your company to be
able to look at how your systems are working and how successful
they are.

Steve: Just to make a comment because Tara used the word culture,
I have heard that the
CEO of Coca-Cola visits a number of bottling plants each year
independent of problems or no problems. The question I would ask
is how often does a CEO visit a manufacturing plant at a large
pharma company without a problem? Because that sends a message,
too, right? Even if it’s one visit a year.

Christine: That’s something.

Moderator: Steve, do you ever watch that show “Undercover Boss?”

Steve: The point is send a message that the
CEO cares. It’s not
undercover, right?

Christine: Steve is going to come to work in disguise one day.

Panelist: Tara, you just mentioned an interesting distinction between
external audits
and internal audits when you combine that with the idea of
metrics. In this conference, we’ve talked a lot about OOS as
being a lagging indicator, OOT being a leading indicator,
reviewing process capability.

Would it make sense . . . because I think that part of the
reason we’re afraid of adopting some of this is the idea that
perhaps this would all become externally auditable.

But does it make sense that if OOS is already the flag for the
external audit that we might consider OOT or process capability
as a self-review, an internal audit, something that still falls
under that idea of the business taking an honest look at itself?

Tara: I’m sure Rick will jump in here as well. I think the idea is to
find the problems
before they happen, so focusing after the problems happen is a
little too far forward. I think it’s important to look at some
of those leading indicators as well.

Panelist: But for the agency to look at those indicators or for industry
to look at those
indicators? Part of the problem is that we’re trying not to see
the problems until they become so large that everybody sees the
problems.

Tara: I think that transparency is a vision. I hear your concerns. I
think those are valid
concerns. But I think that the agency needs to help prevent
problems as well instead of finding them after they happen.

Ian: If I could just add a comment as well, I think there’s an issue
of confidence
between the company and the inspection authority as well. If
you’re honest enough to want to do the right thing, to find
whatever the appropriate metric is, you find it and you start
dealing with it. Then you find the authority then looks at it
and then beats you up or cites you for it.

Whereas, hopefully, and it’s going to take a while, to have that
confidence, if you like, the degree of maturity as well in our
inspections staff — I can speak for the U.K. for sure — but
they see this as a positive thing rather than saying, “Hooray.
Here are some deficiencies I can cite.”

But actually, the company’s been honest enough. It’s found some
issues. It’s dealing with them. I think that can only be a
positive thing, and we need to make sure that where we come
across those sort of things that they are properly rewarded.

Otherwise, you’ll have them hidden away. You wouldn’t go to your
tax man saying, “I’m having a few problems with my offshore
account. Could you help me?” But we’d certainly like to. That’s
a level of engagement and maturity between industry and the
regulators.

Tara: It reminds me of one of my first preapproval inspections that I
went on. The
Director of Quality stood in the doorway on the first day, and
he said, “I know all the problems that you’re going to find in
this inspection.” I said, “Okay. Just give me a list, and I’ll
get out of here. Sounds great.”

But I can tell you when I do look at anything that is similar to
an internal audit, like senior management meetings and that sort
of thing, I’m only looking at it from the perspective of, “What
are they not following up on? What has been lingering for a
significant amount of time, and does this have a potential to
impact product that will be distributed?”

So it’s not so much seeing that program to find all the
problems. It’s seeing what hasn’t been followed up on and the
significance of those issues.

Lina: This is a question
to my
European colleagues. Is there anything you can share in regards
to the progress of the process validation guide, Annex 15, and
the NIR guidelines?

Ian: Sure. I don’t know if you wrote the question, but I have a
question that’s almost
identical to that. Perhaps just to give you an overview for
those that aren’t familiar, for the European review process for
any bit of GMP, first of all, the Inspectors Working Group is
the custodian of that, so they meet at Inspectors Working Group.
I attend that.

It’s a pretty big group, so we have all competent authorities
from around Europe. We usually have mutual recognition partners.
We have colleagues from FDA there as well, not always, but when
they can, and of course, they’re very welcome.

What we do is we find, “Are there any issues to be fixed? Is
there something that hasn’t been reviewed in a long time? Are
there new trends that aren’t covered appropriately in GMP?”

So we raise a problem statement. As you mentioned, Annex 15 is
on that track. Now, it’s moved on to the next stage, which is
then, “Yes, we’ve agreed that it needs something doing about
it,” and it goes out for public consultation as a concept paper.
That’s a document which is a kind of direction of travel, if you
like, a high-level document. So that’s been out.

The text from that then has been looked at, and then the
concepts from there have been worked up into a detailed revision
of the text. That text should be agreed at the next Inspectors
Working Group, which is in a few weeks’ time. It’s at the end of
this month.

By the time that’s finalized, I would guess, it’s probably
unlikely that it will come out at the end of the year. Probably
very early next year, you’ll have then the revised, detailed
text of Annex 15.

Generally, those are out for three months, but if there are
quite a lot of changes, then it might well be out for six
months. That’s the next stage. Hopefully, then, we get a good
response to that, and then the appropriate changes will be made,
and then they’ll be agreed.

So just to close the process off, then what will happen is the
drafting group, which, by the way, includes FDA, so we try to
get as much alignment in terminology and concepts as possible.
That then will come back the Inspectors Working Group and agreed
on. Then that has to go to the Commission for final legal review
to make sure we haven’t overstepped our legal, I think, is the proper term in Europe, we haven’t
overstepped the proper footprint.

Then, eventually, that will get adopted and published. Yes, in short, it should come out early
next year.

Lina: There was also the process validation Guide and the NIR.

Ian: Okay.

Lina: Sorry.

Ian: Yes, what’s happened is the Quality Working Party, which is the
reviewers, if
you like, the assessors, their document, there was a joint
meeting with Quality Working Party and the Inspectors Working
Group a couple of months ago, and I know that their document was
very near being finalized. I’m not directly involved in that,
but its publication is imminent.

The NIR guide, again, I’m not so involved with, but I think
that’s at a similar stage. It may well even be published by now,
but I’m not so sure. I know it’s very close, anyway.

Attendee: I have a question following up to the NIR guide, actually,
based on what we heard
this morning from Christine about the point where there is
agreement or non-agreement for the joint review processes. If
there is not agreement between the FDA and NEA about the NIR,
does NEA still plan to publish that guidance? I don’t know if
Christine knows anything about that.

Christine: I don’t know. We’re not influencing their schedule to publish.
In general, we’re
actually in pretty close agreement on NIR. We’re in agreement. I
wish I had those slides with me right now. We’re in agreement
with the general scientific principles of model building.

There are some differences that might seem slight to many in the
audience, but for people who work at NIR
put could be quite significant regarding some of what the
appropriate calibration sets are and validation sets.

But as far as the general principles of model building, we’re
aligned with that, and we have those other details to still work
out.

George: Hi. ISPE just recently put
out a process
performance and product volume monitoring system guide, and part
of the guide discusses proactive process monitoring, things like
PI systems.

These types of systems have a lot of advantages for a firm, such
as to keep your process in a constant state of control or your
critical parameters. It ensures reliability and makes your year-
end reviews real simple because you have constant readouts on
this.

I’m just wondering, from your perspective and from your view of
the industry, what percentage of the firms do you think take
advantage of these constant readouts in real time to maintain a
state of control?

Moderator: Okay. Don’t all go at once, please.

Christine: We’re giving it to the inspector here.

Ian: “I don’t know,” is the honest answer. But I think it’s a fairly
small number, a small
percentage of firms that do that. I couldn’t tell you the
number. I’d have to consult with colleagues who are doing
inspections more frontline. But I think it’s a still a
relatively small number.

Moderator: All right, Chuck.

Chuck: Quality metrics, obviously, Janet highlighted that and
sessions here have been
quite populated with that. Obviously, some of the challenges you
hear are that industry needs about 20, 30 batches to accrue
enough data. Obviously, with orphan products for certain that
takes some time.

So the question I have is how long does it take to really accrue
meaningful data to get feedback to get some regulatory
advantage? I know, obviously, you’re looking to quality. Then a
question in follow-up to the European colleagues — is this a
divergence in harmonization, or are you moving on the same
track?

Moderator: Well, the first question is statistically based. I guess
anybody could take that in
terms of getting comfortable with batch reliability, I think.

Chuck: And what the timing is.

Moderator: The timing in terms of developing metrics.

Christine: Our process of developing metrics is very early and definitely
ongoing, so I just
don’t think we have enough information to answer that question.

Moderator: You don’t have enough metrics on the metrics.

Christine: Rick, you don’t want to venture on that one, do you? I just
don’t think we know.

Rick: The question is a little vague about regulatory advantage. I’m
not sure, either.

Chuck: Well, there has been some discussion that you have your
inspections, that the
advantage, if you can demonstrate that you have excellent
quality metrics, is that there will be fewer inspections and so
forth.

So the question is that’s the carrot in some respects. The point
is that you have to wait with an orphan product 50 years to
accrue 20 batches. So what? But the other question is are the
European regions moving in the same way, or is this going to be
a sort of a divergence on expectations?

Panelist: it’s up to you.

Ian: I would hope that there is no real divergence. I think that
would be difficult to
know. Again, orphan products, there aren’t that many. Very
often, they’re in facilities that are making other products, so
you can assess rather better from those other products the
general state of control.

What we do see are some very, very small, ultra orphan type
products. They’re not even licensed. In Europe, we’ve got a
system of unlicensed products. It may be tracheal replacements
or something like that, laryngeal replacements, which can be
classified as medicines. These facilities are only making one or
two or three a year.

Partly out of professional curiosity, we’re there. Not,
hopefully, in a burdensome way, but maybe, shockingly, even
helpfully. We’re learning with the organization and trying to
assess their degree of competence. By definition, almost, these
are very new organizations. There’s a learning process there.

I don’t know if that helps, but that’s if you’d like an example
of something that’s ultra orphan, which is tailored specifically
to that.

Moderator: We have a question on the floor.

Attendee: What is your general opinion about the new validation approach,
the CQV
validation approach of ASTM E2500, versus the traditional
validation? Has it helped you guys in your inspection process?

Moderator: Well, that’s a small question in the last 10 minutes. Five
minutes now.

Rick: We looked at that standard when we wrote the Process Validation
Guidance, and
the continued process verification advanced technology
discussion in the guidance incorporates that mindset, so we’re
very flexible to the type of approaches that could be used to
verify that a process continues to be in a state of control
throughout the life cycle.

In fact, a CQV type of approach is used. It’s very similar to
verification concepts that we have in the guidance. It’s a good
supportive standard. If you want to know what FDA is looking for
specifically, you should look at the Process Validation
Guidance.

Moderator: We’ve asked our European colleagues about validation. Anything
else to add?
Okay, another question to answer.

Barry: I forgot. This morning, in the session
on stratified sampling, it was mentioned that even if a company
has a very good robust process validation study, it would still
not be acceptable to use the USP content uniformity criteria for
batch release, that during an inspection, the field
investigators would assess all of the company’s data on
uniformity to decide on whether a company released a batch
acceptably.

I was wondering if you could comment on that. What criteria
would field investigators use?

Tara: I believe that the issue with the 905 is the number of samples
that are taken in
terms of content uniformity. So the number of samples that are
described in USP is not necessarily a statistically valid number
of samples for any particular batch.

So the idea is, then, the numbers of samples collected to cover
content uniformity needs to be statistically representative, and
I know that investigators have gone out and had questions as
well as when we’re looking at CMC sections in terms of prepping
for a preapproval inspection, we have had questions like that as
well, so we make sure to loop in the stats guys, which you’ve
heard from in that section and make sure that we’re seeing a
sampling plan that’s statistically valid.

Barry: Is there any guidance for industry because industry needs
to know what to expect?
It can’t be on an ad hoc basis.

Rick: First of all, as a general concept, which you’re aware of
Barry, the sample sizes in
USP are meant as a referee test, but in general notices, it
states that the sample sizes specified are not intended to meet
the regulatory requirement. A company should always look at
their sampling plans to determine whether they’re statistically
significant when they’re establishing their batch release
protocols.

In terms of being more specific about expectations for sampling,
we have been having discussions on statistically significant
sampling plans — not easy for me to say — and we are
considering the possibility, it’s not certain yet, of putting
out a guidance.

We do have some written matter in-house that we’ve developed as
a concept paper for ourselves, and it’s possible that we will be
communicating more, not only on CU 905, but on general
expectations for a company to establish GMP compliant, most
important, statistically significant sampling plans.

There will be more interaction with reviewers, also, I think, in
the future. I think Christine addressed that.

Moderator: You back there and the other
members. Do you
want to just comment on your initiative and what you’re engaged
in?

Attendee: Sure. We actually had a session this morning where we talked
about some of the
sampling plans and how you go about it. We hope to take the
information that’s been said today and some of the feedback
we’ve gotten from the audience.

Turn that into a whitepaper that we’re going to target the end
of January to finish, get it published with some preliminary
ideas, get people thinking about it in industry, and then that
paper will hopefully be published around April or May. It will
set the stage for the June 2014 meeting where we’ll have a
session in it.

In that session, we’d like to have an open mike, a big part of
it where people come in, give you 10 minutes, something like
that, forward, justify it. We’ll take that
information back and then write another paper that may or may
not be the basis for a future guidance guide.

Panelist: Let me just put a plug in that ISPE’s June conference will have
this workshop, and
it’s looking for a lot of substrate from everybody. Currently,
the membership in that group is FDA, academia, industry, and
USP, so it’s really a cross-representation. . . .

Attendee: just one other thing is the presentation
that most impacts us
was Jim Bergum’s and I believe ISPE will have these all on
their website or something like that. Jim Bergum and Alex
Viehmann presentations addressed this.

Moderator: There is a team, as Chuck referred to, under the Regulatory
Compliance
Committee that he chairs, and PQLI will be assisting in that.

I think we’re near the end of time according to my timekeeper. I
want to say just one thing in response to the question we had
earlier before I thank the panel.

There was a question in terms of how do we keep these things
going in between the conferences? You did hear one indication
about that.

We have communities of practice within ISPE on various topics.
We have the product quality life cycle initiative. These are
active volunteers on many topics including metrics, process
validation, RCC stratified sampling, breakthrough therapy, etc.

We just don’t come and make presentations and go home. We look
for any opportunities to interact amongst industry, between
regulators, etc., to keep these topics going.

For those who are interested in that in between and want a
little bit more volunteer work here, just leave your card right
here in front of Chuck, put down on the back of it what your
interest is, and we will call you. Don’t worry about calling us.

Obviously, we’d like to thank the regulators who did an
outstanding job in fielding all the question and ably handling
them. So let’s give a round of applause.

We thank you as well for attending. Safe travels.

———————————

What are your thoughts? Feel free to share by leaving comments.