Demystifying physiologically based pharmacokinetic modelling among non-modelers towards model-informed medicine use in under-served populations.

Physiologically based pharmacokinetic (PBPK) models represent computational technology to characterize drug behavior within the context of detailed human physiology. Today, PBPK is routinely used in drug development and regulatory approval to support decisions on how a medicine can be used under certain clinical conditions. As such, PBPK has the potential to enhance medicine use for populations that are often under-served globally in drug development and clinical care, namely pediatric patients, pregnant and lactating women. To facilitate broader applications of PBPK for these populations, we joined force and organized five hands-on workshops primarily to non-modelers on the principles of PBPK and its potential applications in pediatric and obstetric pharmacology in 2021 and 2022. In this open letter, we report learning objectives and content of such workshops and to highlight the significant value of these educational efforts.


Disclaimer
The views expressed in this article are those of the author(s).Publication in Gates Open Research does not imply endorsement by the Bill & Melinda Gates Foundation.The views may not be understood or quoted as being made on behalf of or reflecting the position of the Medicines and Healthcare products Regulatory Agency, UK.

Ethics statement
As it is not a medical/health research, no otherwise sensitive information was collected, participation was voluntarily and data were processed anonymously, explicit written informed consent was not sought.

Value of physiologically based pharmacokinetic modeling for under-served populations
There is a general lack of on-label drug dosing information for pediatric patients, pregnant and lactating women 1,2 .Historically, these populations have hardly been evaluated in clinical programs during the development of drug products.Even when clinical studies are conducted, the generalizability of such studies can be limited because the effect of relevant age and/or pregnancy-related physiological and anatomical variations can significantly complicate a drug's pharmacology in a specific subgroup of these populations, calling for different dosage adjustments in different subgroups.To this end, computational tools such as physiologically based pharmacokinetic (PBPK) modeling which characterize the physiological changes can be used to support dosing decisions [3][4][5] .
A PBPK model consists of several compartments representing organs and tissues.Maturational effects of physiology can be described using (gestational) age-dependent equations.This allows the creation of a set of virtual subjects, each representing anatomical and physiological characteristics of pediatric subjects or pregnant and lactating women at a specific developmental stage (the system model).In parallel, one can compile drug-dependent parameters to construct a drug model, and couple it with the system model through relevant physiological processes concerning the absorption and disposition of the drug.A user-defined virtual trial then can be conducted to simulate pharmacokinetic (PK) characteristics in a specific virtual subject or sub-population.Such PK information can then form the basis for dosing decisions 6 .
In the last two decades, advancements in computation and pharmacological sciences fueled the industrialization of PBPK.Today, PBPK is a platform technology used routinely by drug developers and regulatory agencies to inform dosing recommendations for a wide range of applications 7 .Interest in PBPK modeling to predict PK in pediatric and pregnant subjects has increased dramatically over the past few years 6,8 .Of all regulatory submissions involving PBPK modeling received by the US Food and Drug Administration (FDA) in 2018 -2019, 9% were related to pediatric population 7 .Application of pediatric PBPK models also increased in clinical and research settings, for both biologics and small molecules 9 .While in silico PBPK modeling is generally conducted to inform clinical trial design in pediatric and pregnant populations, it can also be applied in the post marketing stage to inform drug dosing 10 .

Demystifying PBPK among non-modelers towards broader application in under-served populations
PBPK is ideal to investigate and predict effects of dynamic physiological changes on a drug's pharmacology.Although the FDA reported emerging use of PBPK in pediatrics, no application was recorded to inform drug dosing in pregnant and lactating women 7 .This gap can be attributed to the low confidence in PBPK for these intended uses, and/or the lack of understanding of today's PBPK as a platform technology by broader medical communities.While the former can be continuously improved through increased PBPK research in the space of pediatric and obstetric pharmacology 6,8 , the latter requires introduction of PBPK to stakeholders who have little or no experience in pharmacology modeling.As such, we joined force and conducted five workshops in 2021 and 2022 primarily targeting these non-modeler audiences (Table 1).
These workshops target clinicians, pharmacists, academic researchers and scientists from non-governmental organizations, ranging from the (PhD) student to senior fellows, as well as clinical assessors from regulatory agencies, with an interest in pediatrics and/or obstetrics pharmacology.Recruitment of participants were by targeted advertising and selection for MHRA workshop, and through announcement by Radboud Summer School, European Society for Developmental Perinatal and Paediatric Pharmacology (ESDPPP), and American Society for Clinical Pharmacology & Therapeutics (ASCPT).No inclusion/exclusion criterion was specified.No prior knowledge of PBPK modeling and accessibility to PBPK modeling software was required for Radboud Summer School, ESDPPP, and ASCPT events.
All workshops shared similar features by including theoretical and hands-on sessions.The theoretical sessions were designed to Presentations of theoretical sessions were tailored according to the specific audience in a particular workshop and hands-on sessions were optimized with clinical cases that are generally familiar to the audience.The hands-on sessions were led by tutors in small groups each including 5 to 8 participants.Attendees accessed the Simcyp software via Amazon WorkSpaces.A mix of tutors from academia, industry or regulatory agencies enabled diverse discussions within each group.The technological requirement of a proprietary software such as Simcyp and the need to have tutors who are experienced users of the software for our workshops can be replicated using other PBPK software platforms, such as GastroPlus ® (https://www.simulations-plus.com/software/gastroplus/) and PKSim ® /Mobi ® (https://www.open-systems-pharmacology.org/).A hands-on exercise was followed by a plenary session, in which results and perspectives by each group were shared with other groups to stimulate interactive discussions.In one case study of a typical workshop (e.g., ESDPPP), delegates first simulated PK of anti-HIV drug efavirenz after a typical dose in the adult population, followed by simulating different doses in the pediatric population according to different pediatric dosing guidelines (e.g., those from US Food and Drug Administration and World Health Organization).The groups then explored the impact of either ethnicity, a weekends-off treatment regimen, or varying assumptions on CYP2B6 ontogeny on efavirenz pharmacokinetics.Results were presented in the plenary, and delegates reported their findings and related real life clinical experiences to the simulation results.
Final agendas including speakers, presentations, and handson case studies for these five workshops can be found in Supplementary File 1.

Effectiveness of our workshops
To understand the impact of our workshops, we conducted online surveys before and after the workshops (Supplementary file 2).Because the surveys are voluntary, not all participants responded (Table 1).It was more challenging to obtain complete feedback after the workshop (for example, 36-75% of original responders before each workshop responded to the same question after the workshop at Radboud Summer School, ESDPPP and ASCPT events, Table 1), which limits the assessment of the impact of the workshop on changing attendees' perspectives.As such, caution should be made when interpreting the survey results, especially when using percentage of responders to evaluate change in perspectives before and after the workshops.Some general findings from the surveys were: -The majority of responders across all five workshops were considered 'non-modelers' because they were not familiar with PBPK (Figures S4-S6, S23, Supplementary In-person

London (UK)
One day + two (1hr) online pre-event sessions 32 delegates b (mostly UK)

Pregnant population
Not applicable as all feedback was collected after the workshop 20 (63% file 2).For Radboud, ESDPPP, and ASCPT workshops, many of the respondents were PhD students, followed by postdoctoral researchers and pharmacists, mostly working in the field of pediatrics.Most delegates had only heard and read about PBPK modeling prior to the workshop, did not read PBPK modeling papers frequently, and had the impression that PBPK studies are difficult to interpret.Participants at the MHRA workshop were mostly pharmacists or clinicians working in obstetrics, and none of the responders were confident in PBPK before the workshop.
-The workshops appeared to be effective in leveling the ground for attendees to (i) understand state of science and technology and (ii) confidently apply PBPK.When asked 'do you have sufficient knowledge and skills at this moment to conduct PBPK model simulations?',most of the responders answered no before the workshops, yet they felt more capable of doing so after the workshop (Figure S7 of supplemental file 2).There was an increased percentage of responders after the workshop considering that 'possessing mathematical and programming skills' (Figure S10, supplemental file 2) is not necessarily essential.Similarly, responders at MHRA workshop tended to feel more confident in PBPK after the event (Figure S23-24, Supplemental file 2).The increased confidence and updated perspectives in attendees can be attributed to easy access to PBPK as a technology platform during the workshop, instead of modeling in the form of differential equations.With technology at hand, attendees can explore various modules constructed within the PBPK platform to visualize relevant physiology changes and pharmacological processes, execute simulations, and discuss simulation results within 30-45 minutes of a hands-on session.The experience of efficiently applying PBPK to answer clinical questions helped attendees to have updated views of PBPK.
-The attendees appeared to appreciate the value of PBPK in optimizing medicine use in under-served populations.For example, responders' answer to the question 'are we at the point of using PBPK model output to inform drug labelling' tended to transition from being split to being confident that the field is ready to apply PBPK output to inform drug labels (Figure S15, Supplemental file 2).Responders also acknowledged the capability of PBPK to simulate drug exposure in fetuses upon maternal administration, as well as drug exposure in preterm and term neonates after the workshops (Figure S15, Supplemental file 2).
-All participants indicated the practical examples on dose adjustments with hands-on modeling as the most useful aspects.Many were interested in simulating dosing scenarios for specific drugs in tailored virtual patient populations, fostering follow-up discussions on research collaborations.
-Clinically oriented users asked for a simpler interface for the PBPK modeling software and expressed an interest in more real-world examples, particularly related to the top 10 frequently used medicines in clinical practice.Whilst some expressed the desire to have access to more user-friendly dosing tools based on contemporary PBPK technology, others however indicated that they still wanted to understand the science behind the software.

Final thoughts
An adequate understanding of the concept and the state-of-thetechnology of PBPK by healthcare providers is a critical step towards successful application of the tool to optimize pharmacotherapy in specific patient populations.Through these workshops, we effectively raised awareness of the relevance of PBPK in the context of clinical care, which reduces the barrier to implement PBPK modeling in informing drug dosing in children and pregnant women.We continue to collaborate in this space.
Freriksen et al. report on their experiences with providing workshops focused on educating nonmodelers on the principles and potential applications of PBPK, with a more specific focus on pediatric, pregnant, and lactating populations.These educational initiatives are important for helping demonstrate the utility of this modeling approach to clinicians and researchers.General feedback from attendees appeared to confirm that these workshops did help "demystify" attendees' understanding and ability to use/apply PBPK, and general findings from these trainings are helpful as a starting point for figuring out how to educate people on a larger scale.
Please consider the following suggestions: Pg. 4, Table 1: consider providing a breakdown of the amount of time spent listening to didactic sessions vs. hands-on learning activities alongside the total duration of the workshops 1.
Pg. 4: The statement on "The technological requirement of a proprietary software…" felt a bit out of place and it did feel like some brief commentary on this aspect could also be helpful for readers.I would consider pulling this out into a separate paragraph/subheading on the ability to apply this to other software, but I do think explicitly noting open-source options and some of the pros/cons with this approach would be helpful if the authors are familiar with these other programs.Cost and ability to access software seem like important points to consider in addition to educating non-modelers in this space to ensure that these programs are used after the training.

2.
Pg. 5: There is a statement on "We continue to collaborate in this space."However, I do think some high-level commentary on future directions and ways to potentially increase access would be helpful to give readers a sense of how this educational space will continue to evolve.

Ananth Kumar Kammala
The University of Texas Medical Branch at Galveston, Texas, USA The manuscript, "Demystifying physiologically based pharmacokinetic modelling among nonmodelers towards model-informed medicine use in underserved populations," highlights the success of workshops designed to demystify physiologically based pharmacokinetic (PBPK) modeling for non-modelers.These workshops have effectively raised awareness and understanding among clinicians, pharmacists, and academic researchers about using PBPK to optimize drug dosing in underserved populations.
The workshops engaged diverse participants, including PhD students, postdoctoral researchers, pharmacists, and clinical assessors from regulatory agencies.This wide range of attendees indicates the broad relevance and applicability of PBPK modeling in different fields.
Although it has several advantages but there are few limitations of the manuscript to be addressed.
The workshops relied on the proprietary software Simcyp®, which may not be accessible to all scientists.Although the manuscript suggests similar workshops could be conducted using other PBPK software platforms, such as GastroPlus® and PKSim®, the lack of direct experience with these alternatives is a limitation.

1.
The manuscript notes that interpreting survey results, particularly when comparing 2.
percentages to absolute numbers, can be challenging.This complexity may introduce bias in understanding the true impact of the workshops.By addressing these strengths and limitations, the manuscript provides a comprehensive overview of the value and challenges of using PBPK modeling to inform drug dosing in underserved populations.

Where applicable, are recommendations and next steps explained clearly for others to follow? Yes
Competing Interests: No competing interests were disclosed.

Anil Maharaj University of British Columbia, Vancouver, Canada
The submitted open letter summarizes 5 educational workshops on fostering an understanding of PBPK modeling to non-modelers and provides an analysis of feedback received from attendees.As PBPK modeling is commonly inferred to be an inherently complex process, the submitted letter, which details how educational workshops can be used to bridge the knowledge gap among nonmodelers, is of contemporary interest.A key strength of this letter is the collection and analysis of quantitative survey data from workshop attendees as it serves to substantiate assertions/positions discussed by the authors.The following comments are focused on improving the clarity of the submitted letter: 1. Abstract: "As such, PBPK has the potential to enhance medicine use for populations that are often under-served globally in drug development and clinical care, namely pediatric patients, pregnant and lactating women." Logical flaw -The opening sentences of the abstract do not provide an appropriate context/rationale as to why PBPK modeling is suitable for examining PK in underserved populations.
2. "To this end, computational tools such as physiologically based pharmacokinetic (PBPK) modeling which characterize the physiological changes can be used to support dosing decisions (3-5)" References (particularly ref 4 and 5) do not discuss the use of PBPK modeling to support drug dosing decisions.
3. "In parallel, one can compile drug-dependent parameters to construct a drug model, and couple it with the system model through relevant physiological processes concerning the absorption and dis-position of the drug." Confusing sentence -unclear as to what is the output from combining drug-and system-specific parameters.Consider revising to indicate that combining drug-and system-specific parameters allows models to generate PK estimates that are developmentally specific.
4. Suggest defining the acronym MHRA in the body of the letter.
5. Suggest indicating how many (of the 5) workshops were online and how many were in-person within the body of the letter.6.Why were MHRA survey results not integrated with the rest of the workshops in supplementary document 2? Were the survey questions for this workshop unique or different?7. Consider including dialogue on how future workshops could be improved (or separating the 'Effectiveness of our workshops' section between strengths and areas for improvement).

"
The increased confidence and updated perspectives in attendees can be attributed to easy access to PBPK as a technology platform during the workshop, instead of modeling in the form of differential equations."Do survey results directly support this statement?If so, please include dialogue that links this proposition to quantitative results from the survey.

Does the article adequately reference differing views and opinions? Yes
Are all factual statements correct, and are statements and arguments made adequately supported by citations?Yes

Where applicable, are recommendations and next steps explained clearly for others to follow? Yes
Competing Interests: No competing interests were disclosed.
I confirm that I have read this submission and believe that I have an appropriate level of expertise to confirm that it is of an acceptable scientific standard, however I have significant reservations, as outlined above.Ogochukwu Ukamaka Amaeze 1 University of Washington, Washington, USA 2 Clinical Pharmacy and Biopharmacy, University of Lagos, Yaba, Federal Capital Territory, Nigeria The paper "Demystifying physiologically based pharmacokinetic modelling among non-modelers towards model-informed medicine use in under-served populations" outlines the details and outcomes of five workshops designed for non-modelers.These workshops aimed to impart knowledge of the science and technology behind PBPK modeling and foster confidence in its application to enhance drug therapy in underserved populations, including children and pregnant women.The paper highlights the benefits of the workshops in equipping participants with the necessary understanding and confidence to potentially harness PBPK modeling effectively in clinical practice.Please find my comments below for the author's consideration.
While the abstract effectively defines under-served populations, the paper omits a definition of the term "non-modelers".It would be beneficial for the authors to incorporate a clear explanation of what constitutes a non-modeler in the second section, titled "Demystifying PBPK among non-modelers for wider use in under-served populations."This addition will 1.
provide valuable context for the workshop target audience -non-modelers.Page 3 -In the statement, "The workshops appeared to be effective in leveling the ground for attendees to (i) understand the state of science and technology and (ii) confidently apply PBPK," consider adding "PBPK" before "science" in (i) to specify the domain.Additionally, for (ii), consider expanding on the applications, such as applying PBPK in clinical practice, research, and related contexts.

2.
It would be helpful for the authors to include a discussion on the challenges encountered during the workshops and any inherent limitations.Drawing from their experiences and insights from organizing these workshops, offering recommendations for effectively organizing similar events worldwide would be invaluable.These recommendations could encompass considerations such as the mode of delivery (online versus in-person), optimal duration, software technologies, and other pertinent factors.This guidance would assist individuals interested in organizing similar workshops in the future.

3.
A minor point -the text in the final row and column in the table is missing a closing bracket.4.

Eric Chan
National University of Singapore, Singapore, Singapore The reviewer concurs that an understanding of the concepts of PBPK by healthcare providers is crucial for the successful application of PBPK modeling to optimize pharmacotherapy in specific patient populations.Through the workshops, the authors effectively raised awareness of the relevance of PBPK in optimizing clinical pharmacotherapy.The workshops also help to reduce the barrier clinical practitioners in implementing PBPK modeling to informing drug dosing in children and pregnant women.Similar workshops should be conducted in other parts of the world.
Below are comments for the authors to consider.
Abstract.It is stated that 'As such, PBPK has the potential to enhance medicine use for populations that are often under-served globally in drug development and clinical care, namely pediatric patients, pregnant and lactating women'.Perhaps the authors should consider including geriatric patients in this statement even though their focus is on pediatric and obstetric pharmacology subsequently.As geriatric patients are often excluded from clinical trials just like the pediatric patients, pregnant and lactating women, citing them may create the awareness of urgent work needed in the geriatric areas as well. 1.
Page 4. It is mentioned that 'The technological requirement of a proprietary software such as Simcyp and the need to have tutors who are experienced users of the software for our workshops can be replicated using other PBPK software platforms, such as GastroPlus ® ( https://www.simulations-plus.com/software/gastroplus/) and PKSim®/Mobi® ( https://www.open-systems-pharmacology.org/)'.The reviewer is wondering if such alternative software platforms have been tested before in similar workshops conducted by the authors?If yes, sharing their experiences in this paper briefly may be important as not all scientists have access to the Simcyp software.

2.
Using Figure S12 as an example, it is relatively difficult to interpret the bar graph.For example, under 'no' as a response, the percentage before workshop was about 10% while the percentage after the workshop was about 20% (2-fold higher).But if we examine the absolute number of responders, they are comparable.The reviewer understands the nuances between percentage versus absolute number, but the interpretation of the results can be biased depending on which read-outs the authors select to present and discuss.Please clarify.

3.
The paper discusses many benefits of the workshops.Are there any limitations, challenges and considerations?Please discuss.

Is the rationale for the
(a) Disseminate the basics of PK and PBPK modeling.(b) Demonstrate the utility and impact of PBPK modeling in drug development and clinical practice.(c) Raise awareness of the impact of pregnancy and age-related changes on drug disposition.(d) Discuss key regulatory and clinical considerations when applying these models to inform drug dosing.The hands-on sessions employed a PBPK platform technology (Simcyp ® v21, Certara UK Limited, Simcyp Division, Sheffield, UK) to help attendees to (a) Become familiar with basics of PK and PBPK modeling.(b) Visualize the impact of anatomical and physiological changes on drug disposition.(c) Understand the implications of dose modification in children and pregnant women.

Reviewer
Report 19 July 2024 https://doi.org/10.21956/gatesopenres.16225.r36766© 2024 Maharaj A. This is an open access peer review report distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

Reviewer Report 19
July 2024 https://doi.org/10.21956/gatesopenres.16225.r36760© 2024 Guo-Fu L. This is an open access peer review report distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.Li Guo-Fu 1 China Pharmaceutical University,, Nanjing, China 2 Subei People's Hospital, Yangzhou, China I find this open letter written by renowned and experienced experts in the field of PBPK modeling to be both interesting and informative.I have one suggestion: should the authors emphasize the best practice strategy in the use of PBPK modeling to address specific medicine use questions in underserved populations?Highlighting the best practice, particularly regarding the differing levels of quantification or verification of PBPK, could enhance the confidence of non-modeler attendees in using their PBPK models to address their intended questions.Is the rationale for the Open Letter provided in sufficient detail?Yes Does the article adequately reference differing views and opinions?Yes Are all factual statements correct, and are statements and arguments made adequately supported by citations?Yes Is the Open Letter written in accessible language?Yes Where applicable, are recommendations and next steps explained clearly for others to follow?Yes Competing Interests: No competing interests were disclosed.Reviewer Expertise: Physiologically-based pharmacokinetic modeling; clinical pharmacology; clinical pharmacy; biostatistics I confirm that I have read this submission and believe that I have an appropriate level of expertise to confirm that it is of an acceptable scientific standard.Reviewer Report 19 July 2024 https://doi.org/10.21956/gatesopenres.16225.r36318© 2024 Amaeze O.This is an open access peer review report distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Are all factual statements correct, and are statements and arguments made adequately supported by citations?YesIs the Open Letter written in accessible language?YesWhere applicable, are recommendations and next steps explained clearly for others to follow?Partly Competing Interests: No competing interests were disclosed.Reviewer Expertise: Clinical Pharmacology, Maternal and Pediatric Pharmacology, PBPK Modeling.I confirm that I have read this submission and believe that I have an appropriate level of expertise to confirm that it is of an acceptable scientific standard, however I have significant reservations, as outlined above.Reviewer Report 14 November 2023 https://doi.org/10.21956/gatesopenres.16225.r35428© 2023 Chan E. This is an open access peer review report distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

Table 1 . Five PBPK modeling workshops held in 2021 -2022
. a Radboud Summer School of 2021 also included exercises related to applying PBPK modeling to evaluate in vitro-to-in vivo extrapolation and drug-drug interactions in healthy subjects.b MHRA workshop aimed at those primarily interested in PK changes in pregnancy, rather than PBPK modeling specifically.(Supplemental file 1)

Letter provided in sufficient detail? Yes Does the article adequately reference differing views and opinions? Yes Are all factual statements correct, and are statements and arguments made adequately supported by citations? Yes Is the Open Letter written in accessible language? Yes Where applicable, are recommendations and next steps explained clearly for others to follow? Yes
Competing Interests: No competing interests were disclosed.Reviewer Expertise: Clinical pharmacology, infectious diseases, pediatrics, pregnancy I confirm that I

have read this submission and believe that I have an appropriate level of expertise to confirm that it is of an acceptable scientific standard.
Reviewer Report 19 July 2024 https://doi.org/10.21956/gatesopenres.16225.r36765© 2024 Kammala A. This is an open access peer review report distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

Open Letter provided in sufficient detail? Yes Does the article adequately reference differing views and opinions? Yes Are all factual statements correct, and are statements and arguments made adequately supported by citations? Yes Is the Open Letter written in accessible language? Yes Where applicable, are recommendations and next steps explained clearly for others to follow? Yes Competing Interests:
No competing interests were disclosed.