QbD Success and Challenges: ONDQA Implementation
Dr. Christine Moore shared her perspectives on the current status of QbD. The talk was titled: ONDQA Implementation, QBD: the success and remaining challenges.
I am sharing my raw notes from her talk. Again, I will edit the article further after I return from my trip.
I encourage you to read the entire transcript below. Dr. Moore of ONDQA, FDA addressed many specific Quality-by-Design questions, including topics such as, regulatory relief, design space, global harmonization, criticality, QbD adoption, etc.
Dr. Christine Moore is the Deputy Director for Science and Policy and currently Acting Director of FDA’s Office of New Drug Quality Assessment. She started at the agency in 2004 as the Branch Chief of the newly created Manufacturing Science group. Christine has been actively involved in FDA’s Quality by Design initiatives and was a member of the expert working group for ICH Q8(R). Prior to joining the FDA, she worked for 10 years in API process development and scale-up at Pfizer and Searle/Pharmacia. Her background is in chemical and biochemical engineering, with degrees from Northwestern University and Massachusetts Institute of Technology.
Here is a short summary of (paraphrased) points:
1. Innovator companies have found that by using quality by design, … provides them a business benefit. They improve product quality and process robustness. They decrease operational cost, and that pays for itself.
2. We see it becoming increasingly adopted by generics and biotech companies….there is a limited use of regulatory flexibility that we’re actually seeing in applications in terms of design space, real time release testing..
3. Definition of “criticality” from Q8(R2) … should be based upon the potential impact of the parameter and not the residual risk after the implementation of the control strategy.
4. FDA would like to see the ICH terminology only.
5. We really think that quality by design, both science and risk based approaches has made great progress in industry. But, we also recognize that the regulatory post approval flexibility has not fully matured.
6. We have had very good success in our harmonization efforts between US and Europe and, surprisingly, including Japan as well. We have a high degree of similarity of regulatory expectations. We’re still working on aligning those even further.
Mentioned Links:
1. EMA-FDA pilot program for parallel assessment of Quality-by-Design applications: lessons learnt and Q&A
resulting from the first parallel assessment [Click]
2. Questions and Answers on Design Space Verification [Click]
Talk Begins
“Dr. Moore: Okay. First of all, I’d like to thank Previous Speaker for a critical look
at the industry perspective on QBD for small molecules. I
think the good news there is that none of what he presented
was a surprise to me and others in the agency.
Also, through some of the items that I’ll be presenting –
and I’ve seen Steve’s slides and talked to him about this –
I know that we’ll be addressing some of the points. But, we
do recognize that we will need to continue to work together
to fully advance and coalesce, really capture the full
potential of QBD.
Today I will be talking about, from a small molecule new
drug side, our perspective on the implementation of QBD in
terms of successes and remaining challenges. I’ll start
with just some quick background on quality by design and
then go into the current status.
I’ll spend quite a bit of time talking about our
collaborative efforts with our European regulatory
counterparts through our pilot for QBD and then some ideas
on next steps.
I’d like to start with a reminder of what the definition of
quality by design is. This comes straight out of ICH
Q8(R2). There are four points that I have listed here.
The first is that it’s a systematic approach to
pharmaceutical development and manufacturing. You start
with predefined objectives, and there’s an emphasis on
product and process understanding and on process control.
It’s based on sound science and quality risk management.
I thought it was important to start with this definition to
recognize that the, I’ll call it, textbook definition but
it’s really guidance definition, of QBD is about a science
and risk based approach.
It’s not about regulatory flexibility, however, by doing
QBD the presumption has been, and we’ve been trying to put
in place, systems that by having that science and risk
based approach certain regulatory flexibility followed.
Obviously, from Previous Speaker’s talk, we are not all the way there
yet.
But, let me go on and talk about what QBD is not. QBD does
not mean design of experiment, although you often use
design of experiments when you’re doing a QBD approach. QBD
does not mean design space. The two are not synonymous,
although, again, you can use QBD approaches to define a
design space. And, the use of QBD in development does not
automatically grant regulatory flexibility.
Rather, the applicant has to take that knowledge that
they’ve learned, that science and risk based approach, and
present proposals into the regulators about flexibility
that they’re looking for. ICH Q8(R2) defines a pathway for
flexibility in terms of design space.
The other ICH documents, as well as Q8(R2), talk about
flexibility such as real time release testing, and now
we’re going to move onto that discussion for other aspects,
such as comparability protocols, at least in the US and in
the European regions.
I have two circles here.
What you have in the blue circle there is science and risk
based approaches. That’s what the textbook definition of
QBD is. Then, there is a dotted arrow over to that blob of
flexible regulatory approaches.
The point is that one does not equal the other and that
there is a connection between the two, but it’s a dotted
line connection because additional actions have to happen
on the part of the applicant in terms of what you put into
your submission, what you claim, and how you justify those
flexible regulatory approaches.
Also, it’s a dotted line because there are not as many
pathways or not a complete pathway, and Previous Speaker touched upon
that right now to take that science and risk based
knowledge, those science and risk based approaches through
QBD, and turn it into full flexible approaches.
So, the state of quality by design. I think you’ll find
this is consistent with what the last speaker said. What
we’ve seen from our perspective, as well as what I’ve heard
in the many, many conferences I attend, is that the science
and risk based approaches of QBD are being embraced by most
innovative pharma companies for development.
Often, that dotted line’s not completed. So, they have the
blue circle. They did that work in development. They may or
may not have been fully transparent with that in their
application. That’s their choice. But, they’re not claiming
that regulatory flexibility such as design spaces.
On the other hand, as Previous Speaker mentioned this, the innovator
companies have found that by using quality by design, by
using those science and risk based approaches, it provides
them a business benefit. They improve product quality and
process robustness. They decrease operational cost, and
that pays for itself.
We see it becoming increasingly adopted by generics and
biotech companies. I don’t think that it’s as fully
embraced in those companies as it is in the innovator
companies, but it’s definitely growing. As also mentioned,
there is a limited use of regulatory flexibility that we’re
actually seeing in applications in terms of design space,
real time release testing, and other approaches.
I like to reference this paper. This came out just a little
bit over a year ago. I know at least one of the authors is
in the audience here. It was a survey of 12 companies, both
small molecule and biotech, regarding their experience with
quality by design. It came out over a year ago, so I would
imagine the survey was probably at least two years old.
(I’ll write an article regarding this paper soon)
Essentially, it was very positive in terms of the business
benefits of using QBD, and all of these different benefits
were listed in the paper in terms of improving your
product, improving your process, decreasing cost,
increasing speed to market, and increasing your
manufacturing capability. These are all positive aspects.
Also, some quotes from that paper talk about industry’s
perspective on the QBD.
Again, we’re talking about the blue box part here. From the
blue box part, again, it’s very positive where the people
who were surveyed in this work thought that QBD would
become more of the norm, and that it’s worthwhile doing,
and it’s already expanding its scope and will continue to
grow. If you haven’t read that paper, I suggest that you
do. That’s where we’re perceiving the industry is at.
We know at least for small molecules, the groups that my
office works with, they like the blue circle. They think
that it’s worthwhile to do that science and risk based
approach. That regulatory flexibility will get more to
that.
But, let’s talk a bit about how we’re responding to it,
what we’re doing, and what our status is, in terms of how
we review applications that contain quality by design
elements. First of all, we look at quality by design using
the standard application review process. We don’t have a
special process for these applications.
All of our review staff have been trained on elements of
quality by design in review through in-house training
opportunities and many, many external training
opportunities including some hands on training
opportunities for advanced methods.
We also have specially trained staff. We call them our QBD
liaisons that are trained very extensively. This was a year-
long training on and off on advanced quality by design
concepts. These folks serve as consultants and mentors, and
they’re available in every branch and division of ONDQA.
Another point of what we started when we started looking
intensely at quality by design is we started looking at
enhancing our team approaches of starting out with required
dedicated meetings where we get the whole review team
together. That includes the folks from compliance.
It includes investigators if they’re known at the time. It
includes other offices such as if we have microbiologists
reviews, et cetera, and defining the quality review team,
and looking at that interaction and enhancing and
continuing throughout the cycle of the review.
That approach has been very successful. We’ve now
standardized it for all of our new drug applications for
the last year, and we’re looking at extending that for all
of our applications and continuing that approach as we move
into the proposed office of pharmaceutical quality.
Recognize that the vision for office of pharmaceutical
quality is even more enhanced in terms of the team review
approach and focusing reviewers on areas of specialization.
The enhanced communication internally has been very
successful.
We’ve also transformed what we called at once our QBD
steering committee into a standing precedence meeting where
we consider new novel aspects and applications that might
be quality by design or other aspects, too. We do have
these discussions at the senior management level within
ONDQA so that we’re making unified decisions for the
office.
In general, what we see for our QBD review aspects is we’ve
moved from something we considered an extraordinary process
to a rather ordinary process. I’m not saying that your
application that you’ve worked very hard on is ordinary. I
am saying, however, that we do not have a special process
for reviewing quality by design.
Moving on beyond just FDA and ONDQA, I do have some
observations regarding some of the things Previous Speaker talked
about, too, and the global regulatory environment. I
recognize we can’t control other global regulators.
However, we can communicate with them and share our
knowledge with them.
In general, as I reflect upon those two aspects of QBD, the
blue circle should not be a problem. I would hope it’s not
a problem for different health authorities. They should be
able to understand review and appreciate that enhanced
science and risk based approaches.
But, we recognize, too, that the opportunities for those
flexible approaches might not be available in all regions.
That’s the pink blob over there, the regulatory
flexibility. There could be nonacceptance, or different
views of design spaces and the approaches to them,
agreements on using PAT or real time release testing.
Another area that we’ve been talking about and still are
working on is establishing clinically relevant
specifications. So, obviously, there could be different
expectations from different health authorities for QBD
containing applications and the amount and detail of
experience that they have in their ability to look at or
accept regulatory flexibility.
One of the things that we are doing to work on that is we
have a standing pilot between the European Medicines Agency
and FDA that’s called the EMA-FDA QBD pilot or the EMA-FDA
parallel assessment pilot. This has been going on for close
to three years now. You can find that information through
either the FDA web page or through the EMA web page.
The objectives of the program include to ensure consistent
implementation of ICH guidelines and to establish better
communication between our agencies to move towards more
consistency to have that collaboration.
Although this platform was meant for the review aspects, it
builds on previous platforms for joint inspections as well
as a program for scientific advice. The idea, though, is to
open communication pathways between the agencies, have the
discussions, and see what areas we can harmonize in.
The scope of the products that we have been looking for in
the pilot, we’ve accepted into the pilot, are those that
include quality by design or process analytical technology,
real time release testing sort of aspects. In the pilot we
explicitly excluded biotech products and priority review
applications although you’ll see in a minute we kind of
fudged on those a little bit.
There are multiple pathways in this program. It could be
for a simultaneous submission of either a NDA MAA, or a
supplemental NDA, or Type 2 variation in both regions.
That might qualify for the full parallel assessment. In
this pathway it’s requested by the applicant or the
sponsor, and, again, you’re submitting the application, the
supplement or the original application, to both agencies at
the same time or close to the same time.
The second pathway, which we found has become more popular
than the parallel assessment, is the consultative advice
pathway. In this case the application is only submitted to
one agency, and entry into the pilot can be initiated by
industry or by one of the regulatory agencies.
We’ve also thought of using this pathway for things that
were excluded by the parallel assessment, namely the
biotech and priority review applications. There’s also the
option of using scientific advice for CMC meetings, and
there is an existing program for that which we refer to.
What does it look like when something actually comes in as
a parallel review process? Separate reviews are conducted
by each agency, and then we have set points where we
exchange information, discuss the information, and see what
areas we agree on or do not agree upon. And, those are
based upon pre-established timelines.
This third bullet that is bolded is very important. There
is no change in the statutory review timeline. I know that
initially many companies were afraid if they went into the
pilot, the parallel assessment, that their applications
would be delayed. We’ve committed that that will not
happen.
Even though we’re looking to harmonize our questions or our
information request list of questions wherever possible,
we’re still using the current pathways by each of the
agencies. So, communications have been from the FDA to the
applicant and from the EMA to the applicant.
We just try and make those communications look the same as
much as possible. We strive, again, as much as possible for
a common list of questions or information requests. Then,
it’s also part of the pilot as outlined to have a lessons
learned meeting between the agencies at the conclusion of
the parallel assessment.
So, where are we today on this EMA-FDA pilot? We’ve done
one full parallel assessment and we’ve conducted four
consultative advice. Some of those are still ongoing. There
was also one biotech product in Steve and Jeff’s shop that
followed the same pathway for the consultative advice.
In addition to the specific applications, we’ve used the
relationships that we’ve established, the knowledge that
we’ve established, to have multiple discussions on quality
by design type elements in our review processes. We’ve had
multiple teleconferences both on applications and on
general topics. We’ve also had three face to face meetings
between EMA and FDA.
Regarding results from the pilot, we’re publishing now a
series of question and answers related to our outcomes to
the pilot. One of those questions and answers just
published on Monday, so I’ll talk about that in a minute.
Obviously, we’ve had many conference presentations.
An unexpected benefit of the pilot was that we had Japan
participate in many of our discussions, and they were
observers for the parallel assessment application that we
had. From our interactions they’re very keen on continuing
if not increasing that participation.
We are currently considering an extension of the pilot
beyond the current date that it expires which is March
2014. We have several companies that are interested in
doing the parallel assessment program.
So, regarding the communications we’ve had from this EMA-
FDA pilot, as I said, we’ve published two questions and
answers. ( Q&A 1 , Q&A 2) They can be accessed from either the EMA or the
FDA websites, or just Google EMA FDA QBD pilot and you
should be able to get the hit for it rather than typing in
these long numbers.
The topics that we’ve included in the Q and A are quite a
few. We’ve talked about the agencies’ expectations for
QTPPs, CQAs, criticality, and expectations for the level of
detail in the manufacturing process description.
We’ve also touched upon a growing trend we’ve seen in terms
of using quality by design approaches for analytical
methods, and there are certain terms that have arisen from
that sort of approach. The most recent one that was
published on Monday discusses design space verification.
I’m going to go quickly through these different topics just
to kind of give you an overview of where we’ve gotten in
terms of the questions and answers and the areas of
alignment. Again, I would refer you to the original
documents for more detail.
One of the first areas of agreement we came up with is QTPP
is quality target product profile, and CQA is critical
quality attributes. This one was pretty easy because we
want them in the applications.
In fact, if you look at ICH Q8(R2) it says they should be
in applications independent if you use an enhanced approach
or a more standard approach. That was a clear and easy area
of alignment between our agencies.
One area which was not as easy to come up with was that of
criticality. The FDA somewhat shifted their position in
this discussion. Previously, we were saying to ourselves we
don’t really care what you call things as long as you
clearly explain it. Then, after more discussions we
realized from a terminology perspective we get in trouble
by having too many terms and having companies use different
terms in different applications, because we have our own
knowledge management and communication problems.
I have listed here the definition of criticality from
Q8(R2) and that talks about the process parameter whose
variability has an impact on a critical quality attribute
and therefore should be monitored or controlled to ensure
the process produces the desired quality.
Criticality is also touched on through the points to
consider as well as the Q and A documents related to ICH
Q8, Q9, Q10. But, really the major point of agreement is
that criticality should be based upon the potential impact
of the parameter and not the residual risk after the
implementation of the control strategy.
Therefore, even if after you implement the control strategy
you can control that process well that doesn’t yield it as
gone critical. We’re looking off of the ICH definition.
We’re looking for consistency there.
We’ve also recommended that companies, at least in
discussion of their manufacturing process description, not
use terms like key process parameter – again, that leads to
confusion within the agency – and that only ICH terminology
be used.
At the same time we recognize that things such as key
process parameters are built into many organizations’
development approaches, and we do not object to the idea of
you using that internally or even putting that terminology
in your process development section.
But, please clearly define what you mean by your
terminology. We’ve seen other terms, too. Again, we want to
see that clearly defined. But, as far as in the
manufacturing processes description we’d like to see the
ICH terminology only.
We also have agreement on the level of detail in the
manufacturing process description. Previous Speaker touched upon this.
We expect that all relevant parameters both critical and
talked about here noncritically, though that’s not an ICH
term, are included in the process description and obviously
justify regarding their target values or ranges.
There is a difference between the agencies in terms of our
regulations here, because at FDA this expectation for the
complete description of the manufacturing process can be
satisfied in part through inclusion of the master
production record, more commonly know as the master batch
record or MBR, at commercial scale. There’s also a note
here that the requirement of the inclusion of the MBR is
required for both generic J products as well as V2
products. That’s by the regulations.
An area that was not anticipated when ICH Q8(R2) was put
together was the idea that you could take the same science
and risk based approaches that were being advocated for
evaluating processes and apply them to analytical methods.
We’ve now seen several companies do this. They tend to use
the terminology that I’m talking about.
There are some analogies between the quality target product
profile and what they’re calling the analytical target
profile as well as analogies between the design space and
what is being called the method operable design region.
So, this was part of our discussions in terms of
reharmonization with the EMA and, in general, we said we
were okay with that approach, with some caveats. We think
that using the idea of describing the analytical method,
the development, the robustness of the analytical method in
terms of ATP and MODR is an acceptable approach.
But, there are some limitations on what we think it should
be there. We do need some more experience to better qualify
those limitations.
One of the expectations, the second bullet on the ATP, is
we both agreed that this approach really would be
applicable to a single method of analysis. And, it would
not be applicable to, say, switch from a HPLC to a GC and
say well that’s under my filing because it’s defined in
that ATP. We believe you have to go through the standard
change management approach that’s laid out in other
guidances.
With the MODR that’s the analogy to the design space, we
again agree with that approach. But, in general, before we
put out more guidance in that area we would want to see
more examples from industry. The main point of the QBD for
analytical methods is that we’re amenable to looking at and
considering it, and we’d be willing to look at more.
This is the most recently published Q and A that just came
out on Monday. It talks about design space verification in
areas where we’re in alignment between EMA and FDA.
First of all, let me step back and talk about what is
design space verification. I do have the definition here.
That shows up in the Q and A. What we’re really looking at
is showing that the design space, which you may have
developed at laboratory or pilot scale, is suitable at
commercial scale and that it’s capable of manufacturing
quality product.
Now, what we came to agreement for initial design space
verification is we recognize the design space is typically
developed at laboratory or pilot scale. And, this is
specific to small molecules, so Steve or Jeff may have
something different about large molecules.
There are certain circumstances, limited ones, where we’ve
seen a design space developed and demonstrated at
commercial scale. And, some technologies like continuous
manufacturing allow that more readily.
But, in general we recognize that it’s developed at
laboratory or pilot scale and that often the initial
commercial scale demonstration is limited. It’s often at or
near where you would typically intend to operate that
product either within your normal operating ranges or
within a specific target ranges.
We also believe that it’s not necessary to repeat all of
those laboratory pilot experiments at commercial scale.
But, the initial design space verification really isn’t
enough to give us confidence of the change management and
the knowledge management throughout the product life cycle.
This is one of those gap areas that Previous Speaker spoke frequently
about. The way that we’re looking to help bridge this gap
is potentially through a design space verification
protocol.
This protocol could include things such as a definition of
potential scale-up risks, a recognition of which scale
dependent parameters you have that are unverified at scale,
and what is your control strategy related to those scale-up
risks and descriptions of any additional controls.
There are some important differences regarding the design
space verification part of, between the EMA and FDA, mainly
where we would like to see this information. The EMA
recommends that the design space verification be submitted
in the application in the regional information section. The
FDA, on the other hand, is recommending that if you choose
to have a design space verification protocol that it be
maintained at the manufacturing site.
We would also like to see a high level description being
provided in the application in some way saying that, yes,
we have this design space verification protocol, it covers
these topics, and you can find it available at our
manufacturing site.
Those are the areas where we’ve come to alignment with EMA
and that we’ve published on, but we do have some ongoing
discussions. We hope to be publishing additional Q and As
in the future.
Some areas that we’re still discussing include online
monitoring, namely MIR spectroscopy because that’s the most
common one that we’re seeing. And, some of the points that
we’re having the most discussion on include aspects of
calibration of the model, what samples do you have in that
calibration and at what scale, and how much of the
validation of the model do you need particular related to
from material manufactured at commercial scale.
We’re still discussing the level of detail in the
application regarding how much information the agencies
would like to see regarding the development of the design
space, your DOE, statistical data supporting it, et cetera,
and the level of detail regarding risk assessment.
Knowing that continuous manufacturing is on the horizon,
we’re also talking about regulatory considerations for that
technology that’s new to the pharmaceutical industry. Other
aspects are the approaches for large sample sizes, criteria
for content uniformity or the so called large M sampling,
and then utilization of protocols for post approval
changes.
I’m circling the last one although it kind of falls off the
page, because I’m going to talk about that a bit further.
Previous Speaker addressed that, again, we have the body of knowledge,
the science and risk based approaches, from QBD, but that
the regulatory systems are not yet in place to allow for as
much flexibility as industry would like to see.
I would say admittedly as regulators we would like to see,
too, regarding how to implement that regulatory, that how
to fully utilize the science and the risk based knowledge
that was gained during development. I don’t have a slide on
this.
But, if you heard Janet Woodcock speak on Monday, or if
you’ve heard her speak recently, you recognize that one of
the approaches we want to move towards in the future is by
using the combination of good science, good understanding,
as well as good manufacturing capability, good quality
systems to allow for improved regulatory flexibility.
I know Steve in his slides is going to touch upon that more
about how the two can work together. One of the limitations
we have is, first of all, we don’t have that quantitative
metrics on quality systems yet, but we’re working on
establishing that. The second is that we don’t necessarily
have the regulatory pathways to put it into place. But, one
regulatory pathway that we do have now are protocols.
That’s where I’m going to go to next.
For FDA these are the regulations under 314.70(e). EMA has
currently passed regulations for protocols, too. The point
of these protocols is that you tell us how you’re going to
make a change, how you’re going to evaluate that change,
what tests you’re going to do, and what the acceptance
criteria for that change is. You file that with us. We
agree to it.
\
It can allow reduced reporting categories, perhaps all the
way down to an annual report. By doing so you have higher
confidence that you will not have any regulatory hurdles in
terms of making that change, and also ease in terms of the
timing of the change and reducing the regulatory burden of
the reporting categories and waiting for approval, et
cetera. That, we think, certainly can help you in making
changes in a global environment.
The change protocols have the potential to enhance
regulatory flexibility and ease how you’re making those
changes. We really see this as one pathway, not necessarily
the only pathway, but one pathway to facilitate continual
improvement and process optimization. I really believe that
there’s a synergy with that risk based and science approach
that you learned about your product by doing a quality by
design approach.
The idea here is that you really want to understand the
relationships and the risks of your product all the way
back between the patient and the product and the
manufacturing process. One way to do that, you’ve probably
done if you’re doing quality by design, is to have risk
assessments. Those risk assessments can be supported by
laboratory or pilot scale data similarly to how you
would’ve supported your design space development.
There really is a potential to use change protocols to
support a wide breadth of changes in terms of optimizing
process, supporting continual improvement, and providing
flexibility in terms of scale and throughput. This is a
slight twist on how we had seen protocols used in the past.
It’s also somewhat in contradiction with our current draft
guidance on comparability protocols which seem to indicate
protocols are meant to be for narrow changes.
We’re trying to take that regulatory approach and broaden
it so that really we can provide the manufacturer with a
wide range of post approval changes that are supported by
their science risk based approaches and then implemented
under their quality system.
So, what are some considerations for these protocols? You
want to definitely understand the potential risks of any
changes you might be doing on the product attributes and
the quality of the product.
That not only goes for the unit operation where you may be
making that change, but downstream effects as well. It
could be possible to group together these changes in very
broad ways perhaps based upon unit operations or the
potential effect to a product critical quality attribute.
Really, an important consideration is how well can your
control strategy detect the effect of that change. If your
current control strategy, your standard approach, does not
detect a potential effect of risk of that change then you
would want to consider things such as enhanced sampling,
non-routine tests, maybe using advanced analytical methods.
But, what do you really need to do to verify product
quality? You also want to consider if by making this change
and if I have something like a design space approved not
only what’s the effect downstream but what’s the effect
with the rest of my approved areas of operation. That’s
what I have on protocols.
I’m going to talk a little bit about the next steps
regarding your work with EMA. As I mentioned, we currently
have several consultative applications ongoing. They
include newer technologies, continuous manufacturing, and
we also have several applications that were not part of the
pilot. They were pre-pilot. But, we got agreements from
the companies to include them in our discussions with EMA.
So, instead of just the five applications that I spoke
about, I think we have nine or ten in total that we’re
talking about based upon our experience with quality by
design. We’re looking at review approaches. That’s
especially useful for answering the question of what level
of detail we’re looking for in applications related to
design space and risk assessment.
As I mentioned before, we’re considering extending the
pilot beyond March. What we would really like to do is
advance the concepts I just finished talking about looking
at enhanced flexibility of post approval changes using
protocols as a pathway.
So, in conclusion, we really think that quality by design,
that blue circle, both science and risk based approaches
has made great progress in industry. But, we also recognize
that the regulatory post approval flexibility has not fully
matured. I think we’re in alignment with industry on that.
We have had, I think, very good success in our
harmonization efforts between US and Europe and,
surprisingly, including Japan as well. We have a high
degree of similarity of regulatory expectations. We’re
still working on aligning those even further.
We think that moving forward change protocols can be a
pathway to do that. It may not be the only pathway but it’s
one that we currently have available and we’re ready and
willing to implement.
That’s the end of my presentation. I want to thank you, and
I think we have some time for a few questions.
So, anyone have any questions?
Male : Alright, Christine. First of all, thanks, that was a fantastic
presentation. Regarding the change protocol, you had
mentioned that there are, let me look at my thing here,
regulations for the protocols. Were those regulations, is
that just the following slides right after that, or is
there somewhere where we can actually find…
Dr. Moore: No. See, you want to look in your 21CFR 314.70 (e)…
Male 1: Okay. Say that again.
Dr. Moore: 314.70(e).
Male 1: Okay.
Dr. Moore: And it’s just a paragraph there. Essentially what it says is
that you can propose a protcol. I have slides on this, but
not in this talk. The protocol can be part of an original
application. Or, it could be proposed as a pre-approval
supplement.
What the protocol needs to include is a description of the
change or potential changes, the method that you’re going
to use to evaluate those changes, so the test that you’re
going to do to evaluate the changes and the acceptance
criteria for the changes. You’ll also want to put in there
how you file it. There is comparable regulations in the
biologics. I don’t know if you know the number off hand.
Male 2: 601.12(e).
Dr. Moore: 601.12(e)?
Male 2: E.
Dr. Moore: B?
Male 2: No, E.
Dr. Moore: E, okay. 601.12(e) is on the biotech side. The language is
almost identical, with the exception that it doesn’t say in
that language that it can be included in the original
application, but we take them in the original application,
too. Yes?
Male 3: This question is maybe for both of you to touch upon. I was
really happy to see in your slides, Christine, the opening
talking about what quality by design is not. I think that’s
very important. In your bullet around quality by design is
not a design space that really hit home with me.
I think it would be very helpful for us to really engage in
what design space really means. I think we spent a lot of
time discussing design spaces and not enough time talking
about control strategies, and how those two are related,
and what those really mean.
From my perspective as a scientist, if I go off and do this
amazing work, and I generate multi variable analysis of
some step, and I understand my functional relationships
between parameters and control attributes, is there an
expectation now that I need to decide whether I’m going to
file design space? Or, in my past experience that would be
just an assumption that I would ask for a design space.
We’re getting questions now globally around why would you
file design space for a process that’s robust. So, there’s
a lot of confusion around whether we need a design space
for robust processes and how to use the design space. Is it
for compliance? Does it help with knowledge management?
And, is it really in fact a control strategy element in the
first place?
So, I really think your slide is important around quality
by design is not design space. But, I think we spent an
awful lot of time emphasizing the importance of design
space as apparent with the design space verification
protocol, the importance of that. I’d just like to get your
perspectives.
I think it’s really fascinating to me to see those two
different perspectives as to how and when a company might
choose to use the design space or propose a design space
when they have that full multivariate analysis in hand.
Dr. Moore: Okay. I was actually going to address a different part of your
question first. I might come back to that.
I think the point you said about getting questions from
regulators on why do you need a design space is something
that we need to work on from the regulatory thought process
side. Because I’ve heard, not necessarily with design
space, with other say post approval changes why is the
company proposing post approval changes.
I think that the regulators have to recognize that our
success is interdependent, that we have to allow you to
succeed in order to fulfill our public health mission. It’s
not just why do they have to do this, what was the problem
with it, but that your business reasons are legitimate
reasons for making changes.
That’s how I look at design space changes. If this provides
a business reason for you in terms of having flexibility of
continual improvement or optimizing your process, then
that’s a legitimate reason for us to do the work to
evaluate the design space. It shouldn’t necessarily be
viewed as a negative in terms of maybe you didn’t have a
robust process.
I forgot the actual question you asked.
Tim: It was a thoughtful question.
Dr. Moore: It was.
Male 4: Tim, I think what you want to know is what’s the advantage of
doing it, design space, and how would you characterize that
advantage, right. So, you actually, Christine, answered
what I would’ve answered. That if a company chooses for
business reasons to recognize a design space they would
recognize it in terms of what control strategy they
actually have.
Because at least at Pfizer, and I know at other companies,
we have decided that design space is one element of control
strategy. It could provide advantages if we characterize it
appropriately and can get it approved for post approval
change.
But, it may not always. We have had instances where we
don’t have truly multivariate interactions between
variables. We make the decision that it’s really not a
design space, so there’s no point in registering it. But,
our process is still robust and we have other controls that
can manage that.
Dr. Moore: Just to throw a teaser out there, I think that if we do a good
job putting additional flexibility in, however that pathway
is, through protocols, through qualifying or through
something related to the status of your quality system,
that there may not be a need to file and get a design space
approved and still have as much regulatory flexibility.
We’re not there yet, but I think that could be possible.
Tim: Any other questions?
Female 1: Thanks, Christine, especially for the protocols. This question
they asked also yesterday in the PAT session, because it’s
really burning now for all the companies that are doing
real time release testing. Recently we had the case where
we filed for real time release testing. We said it’s going
to be this spectrophotometer or equivalent.
The reviewers were adamant that we should take the
equivalent part off and that we should explicitly state what
kind of spectrophotometer we were going to use. From a
scientific point of view we know that we can actually get a
different spectrophotometer, and do the calibration, and
make sure that the calibration works, right, the same way
we calibrate methods.
So, the question is, if you have a filing, can you say that
this is my NIR, this is what I’m testing, this is what I
measure, but in the event that this spectrophotometer
breaks this is how I’m going to do my calibration to
another instrument, and this is my acceptance criteria,
therefore I don’t need to file for a variation?
Dr. Moore: Okay. There are two or three parts of that question. First,
regarding the interchangeability of spectrophotometers, it
sounds like from a scientific perspective that there’s a
difference of opinion between what the FDA reviewer thought
in terms of interchangeability versus what you’re thinking.
I believe that additional information – if you have
publications, et cetera supporting that interchangeability
– would be useful for us to learn from.
Regarding the alternative approaches if you have a PAT
method in place it comes down quite a bit to do you have
the same assurance of quality. That might touch upon
aspects such as validation. It’s really something that you
would want to bring in to have discussions with us for your
specific product, about the purchase that you’re proposing,
and then we can give you better advice. It’s potentially
possible, but it raises a lot of questions that we would be
better off discussing for specifics.
Female 1: You said this is what we brought you and we asked…
Dr. Moore: Yeah, we definitely are willing to talk about looking at online
monitoring or control in meetings and alternatives to your
standard purchase. Not a problem.
Male 5: Thanks, Christine. It sounds like lots of great dialogue going
on in the pilot between the different regions. That’s
really encouraging. But, some of the questions you had up
there around things like level of detail, I’m curious
whether you consider taking the dialogue in the pilot
beyond the initial approval.
Say into the first year or two post launch you can have
great discussions about the level of detail. But, if in the
year post launch you have to do variations. One region says
it’s no problem because [Inaudible 0:48:01] application.
But, another region says you’ve got a variation that takes
12 months to get cleared. By the time you go through such
questions, that may be insightful for the actual discussion
around the original approval beyond just getting it to the
goal line of approval. Because the theme that some of
Previous Speakers thinks raised around continuous improvement you
really don’t appreciate what the application brings in
until you actually manufacture it.
Dr. Moore: Okay.
Male 5: And are we losing an opportunity to put some things in the pile
by not carrying on the dialogue past launch?
Dr. Moore: We had not considered that, but we can. Alright. I know that
there’s going to be a Q and A session later, so shall we
cut it off?
Tim: Yeah.
Dr. Moore: Alright. Thank you very much.
Tim: Just a coda to what Jeff said. We had discussed extending the
original pilot but never got around to doing that. I think,
Jeff, that’s something we should probably talk about
considering in the future.”
What are your thoughts? Please share!