Meet Inna Ben-Anat, Global QbD Director of Teva Pharmaceuticals. Inna is a key thought leader in Quality by Design for generics. I met Inna in person at the 2013 IFPAC conference and we’ve been comrades in the QbD journey ever since.
I enjoy listening to folks who are actually working in the trenches of Quality by Design. After all, most of us are scientists working in the context of a large organization who we have to persuade. Inna is just that person.
Inna is charismatic in her presentations but personable in conversations. I hope you can experience both of her charms through today’s interview.
Inna’s journey begins with a story of an Analytical Lab Scientist transforming her career into a Quality by Design pioneer.
Listen and you will learn the topics below, covered in Inna’s interview:
- How to create a perfect QbD role for yourself
- What is unique about the generics pharmaceutical industry?
- Where are we with QbD in generics today?
- Best Quality by Design publications
- Balance between QbD effort and R&D cycle time
- Should we write SOP’s for Quality by Design?
- Using Risk Assessment routinely for prioritizing Design Space studies
- Importance of “Relevant” data – not just DOE
- Why QbD Risk Assessment is so important
- What are Regulators’ perspectives: GDUFA, QbR, Culture of Quality, ICH Q12, etc
- Success metrics for QbD
- How QbD metrics can align with Quality Metrics
- Statistical tools and examples
I highly recommend you download the slides before listening to the interview audio.
Download the Slides Here:
Click Here to Download
Interview with Inna’s Slides
(For best experience, listen to the audio while viewing slides.)
Sun: “ Hi Inna, welcome to QbDWorks. Inna Ben-Anat is the QbD director of Teva Pharmaceuticals and she’s a key thought leader in the QbD for generic space. I met her for the first time in person at the IFPAC Conference and we’ve been friends ever since and I really learned a lot from Inna. So welcome!”
Inna: “Thanks Sun.”
Sun: “I’m really excited especially about how you started the QbD journey. Could you take us back about 10 years ago? When you were working as a lab scientist and how you got the QbD lead role? Could you take us back to that time?”
Inna: “ Well, I graduated college with a Masters degree in chemical engineering, reliability engineering, and quality assurance. Soon after graduating I started my career over in Teva as an Analytical Lab Scientist located in Jerusalem.
After approximately one year, I moved to a transdermal start-up company where I worked for another 5 or 6 years. However I had to relocate from my Teva branch because my husband was also working in Teva at the time.
I think it was in 2006 when we moved to the United States. The relocation brought us to the Teva manufacturing plant in Seirraville, Pennsylvania. They were looking for chemical Engineers with both a Statistical and analytical background.
It was then that I started my journey with Teva U.S in 2007 as a process engineer position. Working here, my job consisted of optimizing a product and processes, post R&D and before it was put into commercial use. We didn’t call it QbD back then but we basically use the same tools, experiment designs, and risk assessments and other things.
Sometime around 2009, 2010 the whole QbD story came along. And then basically you know, we saw that there is a need to move this forward in Teva and what I did was kind of throw out a job description for myself because It is what i was doing already.
And this is my true passion so you know from here to there I kind of send it up to our upper management and now 5 years forward to the present this offer came. I’m leading the quality by design implementation globally for Teva and also the second part of my role is also to make sure that this is paying off commercial products robustness.
So I was really happy to be part of the journey and you know, to see how it was expanded from really very small local efforts to having a worldwide institution.”
Sun: “What a fantastic story. What was the response when you proposed that new job description for yourself? What were the responses of the managers?”
Inna: “Well I can’t elaborate on it too much but I really have to give a lot of credit to my upper management. They really had the vision back then. You know, it was a very unique position but I have to say that I had full support.
And you know I can’t say that it was really smooth from the beginning, it kind of was ok let’s start but because really the upper management had the great vision in this direction it kind of was very smooth sale from there.”
Sun: “Ok so it shows that the leadership support is necessary. So what is your day to day activity like as a director of global QbD strategy? What kind of tasks, projects, or programs do you do?”
Inna: “Basically right now we kind of already established the platform, so it’s mostly directing our QbD champions to make sure that the activities we do harmonize with each other and stay within our QbD guideline.
I also provide statistical support if needed to make sure we are in line with the regulatory requirement. This role also includes making sure that we have a robust product on the market and that the product sustains this stage with some proactive approaches.
My job is about ensuring the approach to quality designing products for robustness. Basically I watch the product’s design stage from the very beginning of R&D to the robustness and quality stage making sure that it sustains throughout the entire product life cycle.
So, again it’s many little things that combine our base but it covers all. I really have the privilege to say that I’m really passionate about what I do and this is really really you know, the place that I want to be at. A place where I coming to work every day.”
Sun: “ Wow I think that is the ideal kind of career path for QbD professionals. OK, great so now that you’re on your dream job as a QbD leader.
Sun: Let’s switch gears and go to your IFPAC presentation. Every year you present the status quo of what a generics industry is in the QbD implementation. And this year in the January IFPAC conference you presented QbD implementation in generic industry, are we there yet?
So let’s go back to that presentation. What is unique about generics regarding QbD? I mean how it is different from other pharmaceuticals and biotechnology as a whole?”
Inna: Personally, I think the whole business of modern generics would be the biggest difference. For starters, the market is super competitive right now.
It differs from the more complicated forms of generic development but it’s really becoming much more creative; but the timeline is still super-fast so they really need to find this unique balance.
It is required for generics to really maintain the efficiency using the resources from the business model but on the other hand they really don’t compromise on excellence and quality by design and product robustness.
So think all this really can’t…on one hand but I think it also makes ,it’s very interesting because you know in generic environment you can cover so many projects in pretty fast timeline.
So you really gain a lot of experience from many angles and also think, you know, applying quality by design principle in such a routine, it’s really the true test of really know how to apply QbD properly.
Because usually you only have one shot; you need to do it right the first time so plan ahead because if you don’t you’re going to be left behind.”
Sun: “Yes I think you’re right that there is a very unique proposition for generics and maybe that is one of the reasons why QbD has adopted so well in the generics industry compared to other industries like biologics and so forth.
You mentioned the balance between R&D and the timeline, how much work you put in versus how much, which marketing you attempt to market and also in your presentation you were referring to Michael Kimball’s presentation on GphA Fall Technical Conference from 2014. How do you apply that sweet spot of R&D cycle time versus QbD efforts?”
Inna: I think it’s really all about, like I said before, preparing your foundation by investing your time and efforts in the training, guidance documents, templates, and the system that is already in place. From there you can start your developmental tasks.
You have the tools already available for you so for. All that’s left is to develop your internal QbD guidance document, by defining you QbD strategy from the very beginning. You really need to assure that you have all of the functions involved on board with this.
Once you have your priorities down, you can move on to your risk assessment. It is here that you really want to get it right the first time. Because you did put in so much time and effort into planning and preparing all your tools ahead of time, it would be a huge waste to mess up here.
You know how scientist know the design of experiment; in reference to the statistics and how everybody is familiar with the regulatory requirement of the task; you really want that here as well for it will make your job all the much easier. Just remember that it’s always about efficiency”
Sun: “So I mean, that makes sense that you need the right platform systems to make it happen; risk assessments, DOE, and templates and so forth.
I’m sure it took a lot of time and effort to do that upfront work, how long did it take and what were the critical ones that you think were essential in the systems that you built upfront?”
Inna: “ The overall investment took us more than a year definitely, and again I can’t elaborate but I think it’s the same feedback that I get from all the industries. Pretty much everybody started with the same steps.
It’s all about defining a company’s internal quality by their design guidance document first. I think you can also see on one of my first slides that the FDA case studies were really really helpful.
It should be easier for generic industry to start implementing QbD because the FDA was so supportive. They published many documents that could really prepare the industry to get all those internal guidance’s ready.
Just remember that you can’t start implementing such a program before you invest in the training for it. And it’s really about training each one of the scientist about what QbD is and what QbD means to your company.
Like ‘What is the main milestone?’ or ‘ What is the practical implementation; statistical and training?’. So all this need to be done and build ahead of time.
And to think one of those documents were also in my presentation; this is the “Lawrence used…and Understanding from a Technical Quality by Design.” The biggest version was then published in March and I would really recommend it to anybody who still doesn’t have it.
I’m pretty sure everybody has already read it but I think this paper is really really simple and is a great way to explain quality by design and practical implementation.
I really like, it really explains the understanding of products and process, all the input, all the output you know and how you link between those. So I think this paper also can really bring in all the practical great in-ground on how to implement QbD in efficient way.”
Sun: “Right, very nice and I’ll have a link on the website on your post regarding all these papers you sited on your paper and it will have the slides up for reference as well.”
Sun: “So you mention guidance documents training and, so do you also have like SOPs? I know I get requests for SOPs for QbD and so forth. Do you go down to that level?”
Inna: “ I think for quality by design implementation I would maybe, stay away a little bit from SOPs and follow the guidance document. Because we don’t want to put everything into the boxes and templates because we still need some kind of freedom to adjust the milestones and logic when it comes to selling products and services.
And I know that in many companies SOPs could be really strict and restricting so I like to think that the guidance documents should serve this purpose really well. Maybe at some point there will be some type of limit on SOP, but I recommend starting from the guidance and protocol.”
Sun: “OK, that’s good to know. Now I want to jump to where we are today in generics. You have a slide specifically titled QbD implementation and generics, where are we today. Talk about QbR format, its almost 100% applicants and so forth. Could you talk about where we are today in generics?”
Inna: “As I mentioned before, generics is now in a very good place. During the IFPAC Conference, I could see that many of the presentations started to get more practical than theoretical.
Hearing those presentations made me really happy. Based on these feedbacks we can see that the applications submitted based on QbD format are almost at 100%. It’s also about the content, it’s not just about putting those colorful table and templates and pretty… you know justifying if it’s the proper idea, justifying your risk assessment and so on.
And that’s why when they put the QbD published guidance from FDA, they were immediately responding to that release. I know that initial from FDA was great, the industry is implementing basic samples but you know please don’t think too many… because sometime we can see, you know copy paste from the guidance’s.
It’s really much related but I think the industry is already way ahead of this and we’re doing much better. However if you want to put it there, it’s important to understand that those aren’t the kind of guidance documents and examples that we want.
This is the way we need to do it and that’s why I think its just as though he was our client, it doesn’t matter for which… you need to follow the same logic, the same milestone which will be pretty self-explanatory.”
Why Risk Assessment is the Core Element in QbD…and Defines Your Experimental Studies
Inna: “Which kind of brings us to risk assessment and I think you Sun, has had tons of experience with risk assessment tools and have had great success with them.
It all comes down to the principle really. In order to form the risk assessment proactively, we should first go do any of the QbD and DOE studies to find the risks and prove that it works for you.
I would say it’s one of the core elements of QbD implementation because it really defines your experimental strategy, and showing what is critical. It’s linking all the output to the inputs for a better understanding of your products purpose.
A lot of effort that is placed in the company should be put on which risk assessment tool to use to train the people proactively.
Especially in the generic environment. Like you said we need to effect some decisions that are related to just the timeline and how the equipment is utilized in the set timeframe. The amount of API and all of this come to the point of where we really need to assure a first time execution of DOE. We can only do this if we conduct our risk assessment properly.”
(Disclaimer: Teva Pharmaceuticals is a user of Lean QbD software.)
Sun: “Right, OK.”
Inna: “And of course, you know, statistics is our best friend. Again statistics can go to many directions which is I request training with statistics because it’s very easy to be, ‘garbage in garbage out’ so to say. It’s my true passion and right now l more or less put my hands on the data because we have such a great team.”
Inna: “But I will say that the statistics, when used properly, can really bring much insight and understanding to our products and their purpose.”
Sun: “Sure and when people think about QbD they usually think that it is equivalent to almost DOE or statistics.”
Inna: “There is some samples that I’m really also trying to stress. I’d like to point out that QbD is really not equal to DOE.
QbD is a framework. It’s about all those steps. And DOE is a tool. A tool to gain better understanding of our product’s purposes. However, it’s still a tool.
So QbD is much more than DOE and if there still are some misunderstanding, I think that once we start implementing this tool in the entire methodology, it will become much more clear.”
Sun: “OK, great. Let’s go to the regulator perspective that you’ve mentioned in your slides. There are quite a few initiatives including GDUFA and updated QbR, ICH-Q12 and response letters. Could you talk a little bit about that please?”
Inna: “ There are many changes in the regulatory landscape and I think many of those changes kind of make QbD a little bit more formal. It also helps to understand it a little bit better.
For example if we talk about GDUFA (slide 20) we can say many many things about GDUFA and just joke for hours about it. In terms of how its related to quality by designs however, it’s really about increasing quality.
First time, high quality submissions be exact. You want to focus on proper development in order to make sure that all the functions are aligned and ready to execute a high quality file for submission and eventually bring high quality products to the market.
GDUFA really encourages this with all the metrics and making sure that the highest quality submissions will be on top of the pile.
When we talk about QbR though, I think it was really great thing for the industry because we finally know all those QbD milestones that were in question.
They’re formalized in the final document and now all the industries just have one standard. You know they’re examples of Teva’s and it’s kind of making the QbD more formal and harmonized. A good example of a development report would be one that was published in the case studies.
And again regarding the system complete response letter that I put on slide 23.
Sun: Complete Response Letter from FDA.. So what was the response of the team who submitted the dossier?
I know you’re going to ask me if they’re ours or not ours, I can’t elaborate on that. My main point is that we can see that we’re really switching from one factor at a time.
Our approach is more interactive when you look at the parameters assessment and purposes when you design the experiment.
A multivariate approach that utilizes interactions and to really gain an understanding of products and purposes and you know from our perspective utilizing design of experiment and submissions is one of our KPI’s but I would really recommend to the company to start utilizing this tool. I don’t think it’s really a mandatory but I think if you really want to gain true understanding of your product and process DOE’s is the only way to go.”
Sun: Alright, OK. Now how does ICH Q12 play into the QbD?
Inna: “ We don’t have much clearance on Q12, but I think it’s a great guidance to that is coming from QbD’s . At previous conferences we were talking about guaranteed efforts.
As of now it’s really about lifecycles. We heard about lifecycle so much during IFPAC but there still some uncertainties in term of submission changes and how its QbD type should do.
Life cycle approach is something Q12 established and it will provide many answers in this area. It also puts in the stand that QbD is really about the life cycle approach. ”
Sun: “Right, excellent. So let’s go to one of the main questions. How do we or industries measure success of the milestone.
Since you’ve talked, already about couple milestones; so how you set up for QbD milestone and success. Could you talk about the short term KPI’s and the long term metric cause this is one of the biggest challenges for the QbD practitioners.”
Inna: “Right, and again KPI is something that we really want to define for ourselves once you measure a problem. It is then that the KPI is going to present it here. They’re not necessarily Teva’s but I think they’re pretty much similar in form.
Again I would define QbD as KPI’s short term and long term because it’s really about first implementation. Just look at it short term.
We really need to be at the point where our submission is really based on quality by design and methodology so we can put, for example , the percentage of submission based on QbD.
I would definitely recommend handing percent KPI there. Of course you also want to use the enhanced approach in QbD format while elaborating all the tools that are available, for instance PAT tools.
Again I don’t know if each company based is on their resources or on technical capabilities. They can define their own target but I would recommend it to start from year to year of implementation.
In terms of how of the submission contains DOE, its about how much you utilize PAT-tools and so on. In the short term, its more about really high quality submission and monitor implementation process.
But of course the long term magic is really the true benefit, the true test, you know if you really want to make sure when those product are going to market, it really has a more success rate of scale-up, we want a more success rate of validation, we want on time launch and it’s really about reaching.
And also generics from, you know, first you file to your market, then you bring high quality products to the market.
We want to see all product quality and robustness improvement. And I think many companies right now are tracking robustness of commercial product.
So once you have those tools such as process capability and statistical process control, it’s really easily. You can track those progresses from regulatory.
And we want to see our submissions to be of a high quality with fewer deficiency letters. We can see this already happening so I think it also should be some kind of metric.”
Sun: “This is next one guideline, for those who want to implement some kind of KPI’s and metrics for the QbD initiatives and it fits right in with the quality metrics that they’re trying to implement now as well. It’s very helpful.”
Inna: “Yes, I think they need to know, is the quality metrics I would, like I think you know, the older I speak PDA initiative is going on, It’s really about you know not just metrics that are kind of lagging metrics and already reactive metrics but also about right first time and proactive metrics and quality design, process culpability right first time kind of taking place.”
Sun: “Right, so one of the short terms KPI metric system is how many tools in the QbD framework you use?
And I think you point out the PDA – the table of Statistical tools to support product life cycle on the slide somewhere and I see it and it has like descriptive statistics, DOE, Multivari Chart, regression and so forth.
So do you use that as a guideline, like a check box to see what kind of tools are used in their submissions?”
Inna: “ We do have SOPs in place where we point to a specific set of tools and how and where to use it.
I guess my main message here was that everybody is talking about the design of the experiment which we think is a great tool but there are many other tools we can use throughout an entire product life cycle.
This can really bring additional angle on how we look at the deed and how we build in the story from the deed. My next few slides kind of go in through those tools and for example, slide 21 shows that it’s really about the design of the experiment which kind of self-explanatory.
You know to balance the spending of what’s critical and driving the relationship from there for example slide 32 is about the platform and regression to assess stability data.
This is very important also because the recent change for generic stability requirement says we need to submit 3 documents at once with six months of data instead of usual 3 months.
So the bulk of the data is much larger and we can use it our advantage. We can use statistical tools to demonstrate where we have a stable product.
We can also predict shelf life and by using proper statistical tools. We can really make stat analysis very easy and very fast for example on slide 33 which is Variance Components Analysis.
Basically, when FDA withdraws the guidance for stratified sampling and blending infirmity, the company needs to find a way to demonstrate how this certain blend or mixture is homogeneous, and how the finished product is uniform so variance component analysis is a really great tool to show viability.”
Inna: “ On the graph we can already see for example how active distribution is between different samples from left and right. I think it’s really useful to extract as much information from your data as you can.”
Sun: This is a great feature in JMP. For Minitab folks there are the Gage R&R and MSA as well. They can use those modules to emulate variance of components.”
Inna: “ Let me say its very important to make sure our measurement system is of the right integrity and sensitivity before we go into more complex analysis such as the design of the experiment which you know.
Slide 34 for example, shows even a more interested approach which is the Monte Carlo simulation. Which we can joke so much about but the bottom line is if we build in a good model base on the design of the experiment we can really leverage this model to estimate our future project robustness.
And this could be done very easily by introducing potential fluctuation equipment and critical process parameters. You should know what to expect in an analytical methods variability.
Say we simulate 5000 Runs within this to introduce variability. In this case, instead of one prediction to value with confidence, we actually get a distribution.
This way if you have a specification defined, you know you can see what percent, if any, of your distribution is going to be out of step. And it’s kind of very early simulation of your process culpability which of course later on if you move through your products understanding in life cycle you can consume this but use an actual data, an actual process culpability.
Which I think on the next slide there’s an example of course, you know, this is probably self-explanatory statistical process control I think all the companies are using today such as control chart. I think those are very simple but great tools again using correctly this can really bring a lot of insight from data understanding.”
Sun: “Excellent, one of the common mistake using statistics especially DOE, is wanting to go straight into DOE without realizing that all they could see is probably noise and one of the conditions before DOE is having to process stability.”
Inna: “You know that’s a very good point because there is so much homework you need to do before you go through a design experiment, even if it’s a bad design experiment…
but I think it’s like you say, it’s really important to do the training, to get the homework and you know today with so many statistical softwares available it’s very easy for the scientist to go straight to the software and start clicking buttons and doing something that is you know can be very dangerous, so you really need to do your homework and really gain this understanding.
And it’s just when the signs and purpose understanding and statistics come together only then when it’s becoming successful.”
Sun: “Excellent, yes I want to be mindful of your time and our listeners’ time. So I want you to give us a prediction of where the QbD is going in generics.”
Inna: “Well I think if you look at the last slide it’s kind of where this turtle is going, it’s pretty much very close to the finish line. I don’t think we are quite there yet, but I think we’re getting there.
It’s kind of a moving target because I think standards are constantly raising and you know a requirement is constantly raising So I think very steady but slowly but industry is progressing to the right direction.
I think you know in a few years we going to see much more sensitivity implementation, much more case studies. It probably going to be 2 steps forward one step backward.
I think you know most importantly you should really believe in it and make sure that there is a real culture of change into the company into this direction to be proactive to make sure we design the quality and we will eventually see the return on investments.
I think the main message should be business case aside we are making drugs, we are making medicine for people so we really need to make sure especially on generic side, we bring in the highest quality of affordable medication to the public because this is what it was made for.”
Sun: “Well said. Thank you so much for your time today.”
Inna: “Thank you so much for this opportunity Sun and you know I think it was really great as always talking to you.”
Sun; “Same here, thank you.”
I hope you enjoyed the interview with Inna Ben-Anat of Teva Pharmaceuticals.
If you enjoyed this article, please subscribe for more and comment on how QbD for Generics will evolve in 2015 and beyond!