 from around the world. Thank you for joining us today. My name is Jan. I am co-moderator together with my colleague Christian, Saina, and we would like to welcome our speaker for today, Dr. Amanda Giraldelli, PhD, for our Introduction to Analytical Quality by Design, AQBD Principles. And yes, we are going to start. This will be a one-hour webinar. Everyone is automatically on mute. And if you have any questions, please be so kind to type them into the Q&A box. You should be able to see. If you're working from a laptop, it should be in the lower right corner. And we'll be feeling those questions as we go along. Please allow me to just briefly introduce Amanda. She is a graduate in Pharmacy and Biochemistry and holds a PhD in Analytical Chemistry from the University of São Paulo, Brazil. Her focus on her thesis was Plant Metabolomilics and by UHPCL, HRMS, GCMS, and Application of QBD Principles in Analytical Procedure Development. Currently, Amanda is a scientific affairs manager with us at USP and visiting professor at the Federal University of Campinas, Unicomp at the Institute of Chemistry. She's mentoring students in research project on AQBD. She's a specialist in chromatography and mass spectrometry and has more than 12 years of experience in R&D areas with strongest expertise in analytical procedure development and validation. AQBD, Stability Studies of Drug Products, Analysis of Pharmaceutical Impurities and Degradation Products, Chemometrics, Omics Science and Characterization of Compendial Standards. Previously, she worked as Senior Scientist at USP Reference Standard Laboratory for eight years with characterization of compendial standards. She also worked as an R&D Scientist in the Brazilian Pharmaceutical Industry and visiting Scientist at Technician University in Berlin, Germany, working on Proteins Characterization by LCHRMS and at Leiden University in Netherlands at the Center for Proteomix and Metabolonix, working on method development for characterization of biological samples. She's also member of North Jersey Chromatography Group of the American Chemical Society. So, yes, thank you again for joining us. And Amanda, I'll give the floor to you. We will have at the very end, once you leave and close our webinar window, there will be a short survey. So please take your time if you can to fill out our survey. It's very important to us for our quality purposes. We really appreciate your feedback. Please be forthright with us. It really helps as well to improve on the experience that you have with us. And yeah, thank you again. Yeah, so hi. Good afternoon, everyone. Thank you again for the kind introduction. It's really a great pleasure for me to have the opportunity to provide you a brief introduction to the analytical quality by design principles. Hopefully you can see all my screen right here, Christian. Yes, we can. Amanda. I hope everyone. Okay, great, great. Well, and the agenda for this talk today is the following. So you're going to start our discussion discussing about the AQBD concept, and then I'll delight to share with you a brief overview of the USP general chapter 1220, which talks about analytical procedure lifecycle. And then I'm going to share a case study where we're going to develop together a method by using UHPLC UV, using the AQBD principle. So the idea here is to apply the AQBD elements to the method development. And then the last topic would be to discuss about the benefits of applying AQBD to method development. Well, the first question I have here is, what is AQBD? What is the AQBD concept? So if you have a look at the ICH guideline Q-weight for pharmaceutical development, you're going to find the QBD concept, which is the following. It is a systematic approach to development, which emphasizes product and process understanding and process control, based on some science and quality risk management. And in the analytical procedure environment, we can use the term AQBD, which stands for analytical quality by design, referring to the application of quality by design principles to the method development. And USP has several public standards addressing the different stages of the procedure lifecycle. For instance, the chapter 1225 on validation of procedures, 1226 on verification of procedures and so on. And here I also delight to highlight the new proposed chapter 1220 on analytical procedure lifecycle, which presents a holistic approach for management of the knowledge and procedure lifecycle. You're also finding in Fermacopeo Forum, several stimuli articles published by the USP, addressing different key aspects of the procedure lifecycle. And you're going to see that all these key elements are almost the same key elements of the AQBD, right? Because the procedure lifecycle is driven by the quality by design principles. But with this slide here, I would like to show you the evolution of the quality by design concept and how this has been applied over the last years. So we know very well that Dr. Joseph Tudan was the person who first introduced the quality by design concept. He was an engineer and management consultant who are advocated for the quality management, right? So he has read in several books around quality management and also QBD. And one of them that I would like to highlight today, it is drawn on quality by design. So where you are going to find the quality trilogy, which is the following. So we initiate with quality planning, which is the sense of quality by design. Then we do the quality control and improvement, right? And he was used to say that quality should be designed into a product and most of the quality problems are related to the way in which a product was designed in the first place. And after the year 2000, we started seeing FDA encouraging the industry to apply and adopt risk-based approaches in QBD principles in drug product development and manufacturing. And they were used to say at that time already that increase the testing doesn't necessarily improve product quality. The quality must be built into the drug product, right? In the drug product design. And this is one of the most famous concept of quality by design or description for quality by design, that the quality must be built into the product design. And in our case for procedures, the AQBD, during the AQBD, we need to build the quality into the procedure design, right? Well, between 2004 and 2012, we have seen several ICH guidelines being published outlining QBD concept, such as the ICH guideline Q8, Q9, Q10, Q11. We can also include here in this list the Q12 on product lifecycle, which is also driven by QBD concepts. And in 2011, DMA and FDA launched a pilot program to assess implementation of QBD in drug product development and manufacturing. And one of the main objective of this pilot program was to ensure a consistent implementation of the ICH guidelines outlining QBD concepts. And it was also seen as a way to facilitate for regulators to share the regulatory decisions on a very new regulatory concept, which is quality by design. And then in 2017, the AQBD working group of the International Consortium for Innovation and Quality, they published a very interesting survey to assess implementation of AQBD. So you're going to see in the next slides that the AQBD workflow is consisted of different elements, right? So we see that we can apply a design of experiments, but it's also we also need to use quality risk management tools, right? And we see that at that time, 2017, 70% of the respondents of this survey said that they were already implementing some of those elements, right? And with this survey, it's also interesting for us to see which were the challenges flagged by the industry, right? And one of them was exactly that we still have, we don't have actually guidelines to show us how to implement the AQBD concepts. In 2018, the ICH shared that they will be creating a new ICH guideline, the Q14, to discuss about analytical procedure development. So they will be discussing about different approaches for method development, traditional and enhanced approaches. And we are expecting to have access to this text still this year. So they are about to publish this for public consultation. So let's see if the ICH will publish this year, this guideline for revision. And a very interesting survey was also published by the regulatory agency of the UK in partnership with British Pharmacopeia. So they published in 2019, a hypothetical monograph, which was developed using AQBD principles. And basically, they published this consultation including some aspects of AQBD in this monograph, such as the analytical target profile, which is basically the objective of the method, and also including the operating range, or the method of parable design region. And I'm going to explain you about the concepts for these key elements of AQBD later. And last year, in 2020, they published the consultation response, showing how to apply AQBD to Pharmacopeia standards development. And we saw with this consultation response, a very positive reaction of the stakeholders, showing their interest in seeing in the monographs, the ATP and the MODR, right. And last year, the USP also published the chapter 1220 on procedural life cycle, also discussing about key some key elements of the AQBD. And at the beginning of this year, USP and BP organized a workshop on AQBD and life cycle, procedure life cycle, and where we have really a nice engagement from the participants, especially with great questions around how to implement AQBD in the industry and to develop Pharmacopeia standards. And in April of this year, the BP published in their website, a supplementary chapter for public consultation around the application of AQBD to Pharmacopeia methods. So we have seen that regulators in the industry, they are increasingly embracing the quality by design concept, right. And we have seen over these last years, that some quality paradigm shifts, moving from using a quality by testing approach, moving to a quality by design approach, and also moving from using a compliance-driven approach to use based on approach to improve the quality of medicines, right. Well, we know that guidelines to show us how to implement the AQBD approach, they are still lacking, right. However, we can adapt the quality by design workflow applied to drug manufacturing and drug formulation. We can use and adapt this workflow, we can use in the analytical procedure context, right. And here on the slide, I'm going to show you exactly this, this workflow that we can use and apply to develop methods using quality by design principles. So the first step is to establish the analytical target profile. And this will be the objective of the method, take into consideration the performance requirements for that method. And then we go to the stage one of the procedure lifecycle, where we start with deciding or selecting the critical procedure attributes, those will be the analytical responses that we will be monitoring during the procedure design, those responses should reflect the method of quality and good performance of this method. Then you're going to select the critical procedure parameters. And those are the analytical conditions or factors that will be studied and seeing how the changes in these analytical conditions will impact on analytical responses. And for this reason, it's important that we use our prior knowledge, we can also use here some risk assessment tools. And I'm going to show you an example in the case study to help us to select those critical parameters, which may have a significant impact on method performance. And then we are going to do some experimentation, right? So we can use a systematic way of doing experiments, we can use design of experiment, run on some screen tasks optimization designs to optimize performance and robustness. And here just to give an idea how we can conduct those studies. So let's assume that you have a process we would like to to optimize this would be a method to quantify some impurities using HPLC UV. So the process will be basically the separation of the target impurities from adjacent picks, for example. So first, we are going to select the input factors, they could be columns, age of mobile phase, organic solvent, you're going to select a suitable type of design of experiment. And then we are going to acquire data, and we are going to monitor your responses, they could be resolution between critical pair, number of picks with resolution above 1.5, taking factor, retention factor of a specific impurity. And then that we have the response. And you know, with our input factors, we are able to do regression analysis and generate a prediction model, which will predict the responses based on the analytical conditions, right? Right now, we are able to identify conditions for optimized responses and optimize the the method itself. And after validating those prediction models by conducting statistical analysis, we have access to the knowledge space. And the knowledge space will be a space where we know how the analytical responses will behave based on the changes of the analytical conditions. And within this knowledge space, after conducting this assessment, so here, you're going to also estimate the uncertainty associated with the results, with the responses, right? Now we are able to to select or delimitate the method operable design region, which is the M-O-D-R, right? And within this operating range, all study factors combination will have supplement performance and robustness. And with this, after validating the M-O-D-R, we are able to identify risks and propose control for those risks, right? To avoid this risk or minimize possible sources of variability. And we can then establish some analytical comfort strategies to help us to assure that the method is fit for its intended purpose along the whole procedure life cycle. And then we go to state two of the procedure life cycle, which is method validation followed by the state three with, which is performance verification. And we see that all those stages are interconnected. It's much easier now to conduct the knowledge management and the life cycle management, right? So we see that applying this approach, this will allow a feedback and feed forward interaction among all stages. And this will give us the chance to maintain or to assure that the performance of the method is great along the complete life cycle of the procedure. Well, the evaluation of the risks posed by the variables or analytical conditions and also how the variables will impact on performance should be based on scientific knowledge, prior experience, and also experimentation. And design of experiment, it is a very powerful data analysis in collection and data collection tool that we can use in several experimental situations. And we can also apply this to develop methods, right? This will help us especially to understand how the variables will impact on performance through risk assessment, experimentation and modeling as needed, right? It's very common that the beginning of the method development, we don't have so much information about how the factors, possible factors may impact on performance. We know a little bit about this based on pattern knowledge, right? However, sometimes hard for us to estimate or understand which will be the effect of interaction between factors on performance. And in order to gather knowledge, we can start the method development by conducting some screening tests. So here we see some different types of design of experiment that you can use for this proposal. Basically, the main goal will be to understand the effect of parameters on performance. And then we are able to apply some quality risk management tools here. We are going to select best performance conditions and also workable regions in order to further refine them using an optimization design. Basically, during the screening design, we will be focusing on selecting and identifying variables which have a great, a significant impact on selectivity. And then we are going to select those conditions, run the optimization design. So here we're going to find some suitable types for optimization design. And the greatest goal here is to optimize performance, right? So we'll also be applying here and using quality risk management tools to identify risks and propose how to manage those risks. Here, during the optimization design, it's pretty important that we can use some approach and strategies to estimate the uncertainty associated with the responses, right? With the results. And then we can delimitate the operating range, assuring that we have an operating range where we can assure the good performance and good robustness of the method. And here, during the optimization design, we'll be also focusing on accuracy and precision, not just only on selectivity. And then based on a great understanding about the method that we acquired during this method design, we are able to establish suitable contrast strategies for the method. Well, and what about the current practice for validation? So we know that the pharmaceutical industry, they are developing and validating the methods in alignment with ICH guidelines, with USB guidances, right? However, the method development, method transfer, have become an exercise more focused on satisfying regulatory recommendations rather than trying to understand, identify the sources of our ability and trying to minimize those sources, right? Sometimes those guidances are treated as in a check box manner, right? Well, and those guidances in general will recommend some performance characteristics, right, to be assessed during validation. They will be showing us some some default criteria that we should meet with the procedure, right? And sometimes there is not a transparent and ratio male behind the establishment of those criteria. Those criteria should be selected and established during method design based on the identified risks that we are evaluating during method design. And all of those performance characteristics here, they are a good indicator of a good performance, right? However, sometimes this indicator here cannot provide a direct measure of quality or error associated with results. And this current practice will give less consideration to understand the total measurement and how this can impact the decisions to be made, especially if a product is out of specification or not, right? So we see that by using an approach which give us the chance to estimate the variability and uncertainty associated with the measurement, this is pretty important since this variability may pose a significant challenge, especially when we have a result which is out of specification. And here, just a brief introduction to the Chapter 1220 on Procedure Life Cycle. This chapter will present a framework which will show the procedure life cycle approaches consisted of three stages. The Stage 1 procedure design, which is the method development, Stage 2, which is validation and then verification of performance. And we see that all of those stages there are interconnected, right? And this will give us the chance to promote continuous improvement. And you're going to see that this general chapter and this approach will emphasize importance of using scientific knowledge and quality risk management to manage the procedure life cycle and showing the importance of considering the estimation of measurement uncertainty and also establishing tolerances for the measurement uncertainty in order to give us an idea how to make the correct decision about the product and about the performance of the method. And I think at this point you may have in your mind some benefits of applying AQBT to method development, but before we summarize those benefits, I would like to share a case study where we developed a stability-indicating method using UHPLCUV to quantify some impurities in the drug substance. Well, the first step, as we saw, it is the analytical target profile establishment, right? So here for this particular case study, the ATP will be a procedure to quantify under-the-tron-related compounds. We have seven impurities in the drug substance under-the-tron hydrochloride, and in parallel to the ATP establishment, we need to start doing quality planning, which is the initial risk assessment for method design. We are going to start here getting prior knowledge about which kind of processing impurities, the CPMs we may find in drug product, which kind of degradation products we may see arising during the stability of that product, right? Then we need to access your chemical structures, evaluate their physical chemical properties, because this will guide us to select a suitable technique to quantify those impurities with the required sensitivity with good selectivity. In this example, we have under-the-tron, which is the API, which is a non-polar compound with basic moiety. We see here that we may explore reverse phase from autography, but we may face some challenge, especially because we have the impurity F and E, which are very polar compounds. We may face some poor retention of those impurities in reverse phase from autography. We also have ionizable compounds, right? And neutral compounds. And we may also explore the addition of ion pair reagents to improve the retention factor of impurity F and E, and also change the selectivity for the ionizable compounds. And on the next slide, I show you the risk assessment tool we used to help us to select the critical parameters, right? And here, the idea is not to go into details about these parameters, but just to show you the tool that I used, which is a causal effect diagram, where we can see the parameters, which are the critical parameters, we selected for the screening one. Basically, those factors selected here, we selected them based on their potential impact on selectivity. Because we know that in chromatography, a primary goal is to have good resolution between the target compounds and other impurities or adjacent pigs. And we know that the resolution in chromatography is described based on three terms, right? The retention factor, the efficiency, and also the selectivity. And we know that selectivity will have the greatest impact on resolution. For this reason, we are selecting for the screening one conditions and parameters, which may have a significant impact on selectivity. Well, here we have the overall strategy for this method development. So the first thing, we did the risk assessment, we purchased the non-impurities and we prepared a solution with a mixture of impurities and the API. And then we've run the screening one using design of experiments. So we use here the optimal design. We screened four different columns, four different ion pair reagents, different organic solvents. We also blended the organic solvents to explore different selectivities of organic solvents. And we also explored gradients low. And after running the screening one, doing data interpretation, we could select the best performance conditions which were further explored using Ice Screening 2. So basically here, we added a new level of column. We added Disurbax bonus column. Since we saw, based on the screening one, the BEH should column give the best performance. And we would like to explore here a similar technology. Since BEH should has a polar group embedded in the alkyl chain, so we selected Disurbax bonus column which also has a polar group embedded in the alkyl chain but now trying to reduce the percentage of carbon load and trying to explore different interactions between the stationary phase and the analytes. And we also added a new column, a new factor which is column temperature. Since we know that this factor may have a significant impact on selectivity of ionizable compounds. And we are working with ionizable compounds. And then the best conditions, we selected to be photorefined using an optimization design. So here we use a geoptimal design fitting the data into a codec model. And we run also the in silica robustness test and we could delimitate the MODR. And finally, after defining the final condition, we could conduct this stress testing, selectivity studies and validation. Well, for the screening tool, let's see how we did the data interpretation. So basically we run this screening too using design of experiment. We generated the prediction model and after validating those prediction models, we could then project all the responses that we are monitoring the quality that they are using to monitor the quality of the method that you can see here and here are the acceptable criteria for those responses. We are projecting those predicted responses on the graphics. So those are here, the acceptable performance region graphics. And the golden rule to interpret those data is to find white area. Because where we have white area, it means that all acceptance criteria for all responses were mapped. So we have good performance. So we see here that by using the bonus column, with NFPA and HFPA, we have white area with the different column temperatures. You see here that we have white area. We have good performance using this column. We can also further extract knowledge by evaluating here that those graphics from the BH shield column. We see that we also have white area, right? But we need to limit the use of this column to the use of organic solvent, acetyl nitro, because the corner of those triangles are related to the use of different organic solvents. So we see that using the BH shield column and 100% of acetyl nitro, we can have great performance and we can increase this when we increase the column temperature to 45 Celsius degree. And because we have a great flexibility to work with bonus column with the different conditions, we selected this column for refinement using unoptimization design. While we can further extract knowledge from the data, we see here the counter plots obtained during the screening too for the response resolution between a critical pair, which is under the throne and an inferiority. And we clearly see here that by increasing the temperature of the column, right, by setting the temperature at 45 Celsius degree, and using the organic solvent acetyl nitro, which is the A-corner here, we are increasing the resolution and for sure this will be the conditions we will be selecting to further refine using optimization design. However, we still see some issues, right? So we still see here some pre-retention of the very polar compounds impurity E and F and we also have an additional critical pair which is between the impurity B, which is an ionizable compound and impurity C, which is a neutral compound. And during the screening too, we evaluated the different IMPERI agents with a concentration of 5 millimolar. So the idea was before we go to the optimization design, let's just try to optimize the IMPERI agent concentration. So we increased this concentration and we use here, we run some analysis using HFBA at 10 millimolar. And for sure we were expecting that the ionizable compounds would have a longer retention in the altitude chain. So we could solve the very poor retention we have before for the impurity E and F. And because impurity B is ionizable compound, it was retained longer in the column, impurity C is a neutral compound, it didn't move, right? And we could solve these issues here and select the best concentration of HFBA, which is 10 millimolar, right? Well, now we understand how we selected the best performance conditions from screening two to be further optimized using an optimization design. So here we explored four different factors, right? So we explored the concentration of the IMPERI agent, which is the HFBA. So we saw that we selected 10 millimolar as set point and we're going to explore the around set point, the same for column temperature, flow rate and gradient slope. And we use here the optimal design, as I said, and here are the results. So after writing the optimization design, we can build some graphics to extract knowledge from the data, right? So here we have the knowledge space. Remember that the golden rule to interpret the data here is to find right area because it means that all conditions will give us on a supplement performance in good robustness. However, at this stage, we still didn't conduct risk assessment. We still didn't identify or estimate the potential uncertainty and variability that might be associated with the results, right? So we don't have information on robustness. We cannot just select a final condition based on this study. And basically, when we are running the optimization design, we have a single injection for each experiment, right? So we don't have access to the uncertainty associated with that measurement of that response. And what we are going to do here, we are going to simulate multiple injections in order to assess the uncertainty associated with that measurement. So here we use Monte Carlo simulation to do this and also running the in silica robustness study where we are going to simulate the variation around the set point for each organ, for each factor, right? And see how this will impact on your response. Now we are going to have new prediction models taken into consideration the variability associated with the measurement. And now we are able to identify risks and select final conditions with great robustness. And the second stage of the quality risk management here during optimization design will be the establishment of a specification for performance characteristics for the critical attributes. And then we are able to calculate the process capability index which will give us an idea about the ability of the process to meet the specifications. And then we are able to select those conditions which will give us a great performance and also robustness. And after conducting an identify of risks we are able to delimitate the method of parable design region and we need to validate this MODR. And as I said we don't have any protocol to help us to validate properly this MODR. Basically what we are proposing in this case is we will conduct the validation of prediction models by doing statistical analysis and then select some working points within the MODR and do some verification runs. Then we are able to compare the experimental and also predict the responses right and validate the MODR. And then we are going to optimize the responses and I'm going to show you how we did this for this particular case study. Well let's go back again to the knowledge space. Now we are going to project the responses predicted by the new prediction models now considering the variability associated with the responses right and when we do this we see we get a surprise because we don't have any white area anymore it means that we have good performance but we don't have good robustness so we cannot select a final condition here and this is the beauty of using equality by design because during the method design we are able to identify the risks and design the method in a way to prevent it to fail before it fails during method transfer during the analysis the routing analysis using this procedure and in this case we are designing the method in a way changing the settings of some factors to find a robust condition which will give us a good robustness and good performance. So if you will see if you have a look at this graphic here we are applying some constraints for the organic solvent the percentage of this organic solvent so we had before arranged between 38 to 45 percentage of organic solvent and now we are applying some constraints to find a robust area where we can assure that the MODR will give us conditions with a supplement performance and robustness and here just illustrate how we can conduct the MODR validation after during the statistical analysis of prediction models and evaluating the ability of prediction of those models we can run some verification runs so here we selected the edges of the MODR in the target condition we run the verification runs and we could monitor the performance in all of those points right so we see here that we have great performance we had great selectivity with nice nice resolution between all the target impurities we have great retention factor for the vericolor compounds and we have a final condition where we have a single method that we can quantify all the impurities and here just illustrate how we can further optimize the response after delimiting the MODR so we can use response for phases counter plots to help to optimize the response and here's an example so we see the counter plot and response surface for the response resolution between the impurity B and C which is one of the critical pairs of this study and we see that which we set the column temperature at 45 Celsius degree and we have the concentration of IMPERI agent set at around 10 millimolar 10.5 we are increasing the resolution for impurity B and this is the final condition well and here you see the final condition we have a visual graphic with the MODR we can also have the MODR in a table where we see the operating range right which was validated based on quality risk management and also based on quality risk management tools we were able to acquire a lot of understanding about the method and identify risks and establish suitable counter strategies in this case suitable system suitability criteria for us to verify the performance of the method along the whole procedure life cycle well and with this some concluding remarks will be exactly the benefits of applying quality by design principles to method development so we see that by applying AQBD to method development we can first acquire a high degree of understanding about the method right we know that one of the key neighbors of AQBD it is to do the management of knowledge and apply quality risk management tools and we can only identify risks and manage risks if we have a high degree of understanding about the method and with this we can also select optimized and robust conditions we can then access the operating range which is the MODR by identifying the risks proposing some controls for those risks and we can also understand and control sources of our ability and manage the risks using QBD we can also identify the measurement uncertainty and establish tolerances for this measurement uncertainty and with this we can establish suitable control strategies the system suitability criteria we can also establish more suitable control requirements for method transfer right and we can also increase reliability of deciding if a product is out of specification so we see a clear benefit here of applying QBD it is to improve the quality of medicines right and this will help us to this will facilitate knowledge and analytical procedure life cycle management and also help us to promote continuous improvement for the procedure and quality of medicines oh and this brings me to the end of my presentation I just would like to share with you a very nice and simple sentence that you are going to find in the USP stimuli article on analytical comfort strategy around the concept of QBD well fundamental to the concept of quality by design is to start with the end in mind and the start will be the method design and end will be the quality so we need to take care and consider build the quality of the procedure and product in deep method design right well and with this I thank you so much for for the attention and I'll be more than happy to answer any form of question that you may have and if you're free to retell to me in this address in case you have some some questions in the future and would like to have some discussions around a QBD thank you all right thank you Christian would you like to feel any questions answers yes of course thank you Jan hello also from my side I am Christian looking off over the questions today and we have a couple of questions number one question as always or as very often could we have the slides after the presentation so Amanda would we be allowed to make a PDF of the PD of the PowerPoint and and send it to participants for excellent excellent thank you very much so that was question number one we have also one question what does the MODR stand for and I think it became later than clear it was at the very beginning I think it became later clear but that's the method operable design region and this is more or less the equivalent to the design space when we talk about normal quality by design so yeah I sorry Amanda I know I was answering your question but but maybe you have something to add to that yeah sure you are totally correct Christian the design space outlined in the ISH guidelines where we apply QBD concept to drug manufacturing is exactly the same MODR that we are considering for the development of procedures right so the method operable design region will be as I said a region where we can assure that all conditions will provide us a good performance and good robustness right and we can only access this through using a systematic approach to our data to identify risks and we also need to access the robustness of conditions otherwise we cannot delimitate the MODR and something very nice to point out here is that by using the MODR in the future we may have some regulatory flexibility why because as soon as we are assuring the good performance and robustness of the method we can in the future make some movements within the MODR and they still assure a good performance and robustness right and sometimes we know that during the product lifecycle we need to to to have some changes in the synthetic growth of the API or maybe in the formulation of the drug product a classical case would be the nitrosamine case right where some manufacturers need to change the synthetic growth of some API to reduce the level of nitrosamine in the product or they still eliminate this impurity and this may change the organic impurity profile and can impact on the performance of the of the procedure the procedure may not be any more suitable for that purpose and for these reasons it's interesting to have a MODR where we can do some movements within the MODR and we don't need them to go back and redesign the method again and validate everything again when we have those changes in the product lifecycle yeah thank you very much we also have very often the question coming up for the plots that were shown especially on slide 19 but all that goes about the design of experiments which software were used there and and can we do the same that was also a question yeah sure well for this particular case we used a fusion software which is a product by design software from us matrix however we have many other software that we can use I can just share some of them they would be the ACD Alpha Chrome we also have Chrome Sword we also have dry lab design expert expert and we can also use very simple statistical software such as MediaTab we can also use an open source software which is the AIR program right we can even do design of experiment using Excel very well another question thank you another question we have is is AQBD applicable to both isocratic and gradient methods yes what but in case we would like to have a gradient or we'd like to explore the gradient we can also include this as as variable or we can maintain this as a constant factor but it's possible to actually we can explore all parameters that we would like to explore even using an isocratic method or using gradient yes perfect and another question was apply the design of experiment if you don't have a linear relationship of variables or variables to responses yeah yeah so we can also model this and then we need to know that we might need to use some suitable design of experiment types to allow us for example to estimate quadratic terms or higher order terms right and and normally we are doing this when we are doing optimization designs or using optimization designs because we are using design types that can give us this this opportunity to to estimate the quadratic terms and then we can see some inflection points in case the your response is not linear yes very well there are a number of questions let me see what what is a good one here to crush I think we also need we will not be able to to answer all questions here today but Amanda gave her email address already awg at usp.org you can also send to myself christian.ziner at usp.org I gave my email address in the chat already but here is another question for you Amanda I realized that automation equipment takes an important role in AQBD but if this automation is not available could you give us an estimation of the time that's required to establish the MODR of a method yes for sure yeah so for sure we can accelerate the method development when we can have an automatic communication between the the software and the the instruments that we are using for that specific technique right and sometimes we don't have this this this opportunity what we need to do then we need to we are going to do the data processing right and after doing the data processing we are going to have the responses and we need to enter manually those responses into the software or using excel or other kind of statistical statistical software and for sure this will add some delay in the method development but it's totally I think it's totally fine I think one of the hardest step that we need to spend a bit of time is exactly data processing because if you are doing the data processing in a wrong way or or integrating the pigs incorrectly you can add bias and have and mess up the the prediction model right yeah another question here have you experienced any cases that even after all these AQBD measures the method developed becomes inadequate after some time has elapsed so no I still haven't experienced this since but this might happen and this might be because we didn't select the critical parameters to be started or we didn't select the critical attributes to be started as I said this is a very important stop of AQBD of the AQBD approach because if we are not selecting the critical attributes we will not be monitoring correctly or we might not have the quality we would like to have at the end right so this is a very important step to pay attention at the very beginning of the method design yeah there's another question what is the practical risk if we wouldn't apply Q analytical QBD for analytical methods as the MODR is not of much a lower as the MODR is not much a lower thing I think this question is no this question is slightly can you read it as well on your end it's at 4047 find this one so okay maybe I didn't find this one Christian yeah you can send me let's let's take another let's take another one you were saying that validation is not a direct measure of quality but doesn't validation contribute to ensuring consistency consistency which is then in itself an element of quality yeah for sure actually what I said is that some performance characteristics such as a curiously precision etc they are good indicators of performance but during method design those those indicators may not provide a direct measure of of of the quality and ever associated with the measurement right so we need to conduct or estimate the uncertainty associated with with the result and with the measurement and and also variability for us to be able to identify where we may face some lack of robustness do you expect Amanda there's another question do you expect the analytical quality by design concept to be included in submission does yes at some point this might be possible we still see that are some knowledge gaps for the industry to use and implement this and is the same for for the regulatory agencies right so we we are in a time that we need to exchange experiences and exchange knowledge on this and I guess that after the ICA guidelines published on Q4 the Q14 is published we might have a clear a clear overview or view of how this might happen again okay we are coming to the end of today's webinar but we might have time still for maybe one or two questions I have one more here in case we have a manufacturing process already developed using quality by design and already in GMP manufacturing so if we want to move within design space to manufacture some product will there be any regulatory consequences on the analytical quality by design then this might have some consequences since in case it changes the organic imperial profile for example of my particular case study today we might need to evaluate this yes and I think the last question that we should take well yeah let's use that one here because it's a very general one I would think do we have to validate all the parameters that are covered in my workspace for example if my method is workable throughout 30 to 55 degree Celsius so yeah well the idea is that by validating the MODR then you will just validate the final condition right so for this reason it's so important that we work and start changing exchanging knowledge around protocols for MODR validation because we'd not be necessary in the future to validate all working points within the MODR this is the idea that we don't need to validate them but we need to access and evaluate the performance requirements throughout the MODR okay thank you very much Amanda for for all the answers to that questions and also for this wonderful presentation we got a lot of comments also that that the participants liked this presentation very much and found it really it said perfect here as well so so very good results and we could not answer all the questions here for today but we we have the chat protocols and we'll try to answer as much of those offline as well if you have further questions or if you think your question has not been answered then please reach out either to to amander under awg at usp.org or to me christian.ziner at usp.org my email address was also in the in the chat already mentioned and with that I give back the word to for final words yes thank you very much christian thank you so much Amanda for this very interesting webinar and this great information you've provided I hope for all our participants this has been enjoyable for you as well and informative again as christian said please reach out to us directly by email if you have any further questions and comments we'll just make some quick announcements for the next rounds of webinars that we have scheduled for the Europe Middle East and Africa region so on 29th of April on Tuesday from 10 to 11 we have qualification of the solution testers usp performance verification test PVT on 11th May Thursday from 10 to 1115 what you always wanted to know about usp reference standards but we're maybe afraid to ask and on 20th May Tuesday we'll have general chapters 1220 analytical procedural life cycle with Dr. Rossio Papa for those who might be already familiar with him Amanda will be back with us on Tuesday 1st of June from two to three usp proposed general chapters 1469 nitrosamine impurities and on 15th June Tuesday from 10 to 11 we'll have the value from a repeal reference standards which christian will be presenting all right well thank you so much Amanda is just something you just briefly want to mention for the nitro webinar you have lined up for the 1st of June or is it too early still oh no no I can I can share basically we will be discussing about testing methods for for nitrosamine analysis in pharmaceuticals and and discussing about analytical challenges that we may face when developing methods and validating methods and I also delight to share the four procedures that we have in the in the chapter 1469 excellent thank you and christian do you just want to just share something really quickly about for 15 June the value of pharmacopoeial reference standards some thoughts yeah that that that presentation is based on an article that we recently published in in pharmaceutical technology we will talk there about the the risks that that are that that you are taking if you don't set up secondary standards in the in the right way for example and we we explain a little bit what yeah what the risks are there we explain the guardband principle and and look a little bit into uncertainty it's just for the sake of of showing the risk there so it's it's it's a presentation that was not shown so far to to that regard yeah excellent all right if you would like to stay informed with further the webinars that we have information please visit our home page at usb.org and select the tab that says products and services to subscribe to our newsletter where you can receive regular updates so we can stay in touch thank you so much again when you close the window there should be a short server that should pop up so again please we welcome your feedback please be forthright and honest with us it really will help us as well and thank you so much everyone for participating please stay safe please stay healthy and we hope to welcome you back soon all right thank you very much everybody bye bye thank you bye