Credibility assessment of model informed drug development approaches in paediatric and rare diseases: a review of EMA submissions and assessment reports from 2014 to 2024
Flora Musuamba Tshinanu1, Dr Ine Rusten3, Dr Hélène Haguet1, Lisa Hanquet1, Marco Dionisi1, Shalom Govere1, Happy Djokoto1, Camille Massaux1, Pr Jean-Michel Dogné1
1University of Namur, 2Federal Agency for Medicines and Health Products, 3Systems Resource Lab AS
"The currently used model for regulating drug development was established more than 40-years ago, and is compartmentalized based on the 3 traditional types of data: in vitro, animal in vivo, and clinical. While there is a continuous adoption of the regulatory system to meet the needs of the key stakeholders, such as updating existing and issuing new guidance documents, adapting regulatory procedures, providing advice on specific methodological innovations and drug developments, etc., this main structure with these three different domains with separate responsibilities has been maintained. The availability of new data sources and types (including Real World Data (RWD)) and innovative methods and the possibility for cross-domain data to integration offer new opportunities for drug development and assessment, in particular for pediatric and rare diseases. To address this issue, the ERAMET project (Horizon Europe grant 101137141) proposes the use of the question-centric approach for drug development. The question-centric model is implemented using the credibility assessment framework combined with an attribute framework and uncertainty quantification techniques relevant from the perspective of regulatory assessors. The question-centric model is iteratively developed and refined within the project and will be tested by being applied to several use-cases. One of the challenges being addressed is the assessment of evidence generated using alternative or more innovative methods by regulatory authorities given the lack of systematic and integrated approach to assess and establish their acceptability. The ICH M15 and the ERAMET project aim to define the framework for assessment of Model Informed Drug Development (MIDD) evidence to inform decision-making.[1] The ERAMET project aims to build an ecosystem including a repository of key regulatory questions linking the questions with the data and the methods used to answer them. The aim of the work is to describe an integrated repository of key regulatory questions that are consistently encountered during drug development and across regulatory procedures (MAA, scientific advices and PIPs) and evaluate the place of modeling and simulation methods for answering these questions. Regulatory submissions from January 2014 to December 2024 were explored for this purpose. Benchmarking of modeling and simulation methods versus traditional methods was performed by using the table for assessment of MIDD evidence, a communication tool between applicant and regulatory authorities that has been set up in the ICH M15 guideline.[1] Methods Five (therapeutic) areas were selected for a streamlined repository building: paediatric pain, severe asthma, haemoglobinopathies, CAR-T therapies and haematological cancers, and assessment of QT prolongation. The relevant regulatory procedures were screened to identify the questions of interest, as well as data and methods to answer them. We reviewed the paediatric investigation plans, the scientific advices, EPARs and eCTD summaries (quality, primary pharmacodynamic, toxicology, clinical efficacy, clinical safety and clinical pharmacology) for these treatments and/or applications and translated the information into essential regulatory questions. To benchmark modeling and simulation methods with traditional methods, credibility activities were performed by applying the credibility framework as described in the ICH M15 guideline . The credibility assessment was implemented in relation to their context of use, the patient risk and the regulatory impact. Results An integrated repository of Questions, data and methods was built are hierarchically organized around aspects related to modality of drug use, benefit/risk assessment, efficacy, safety and pharmacokinetics. These questions were classified and organized based on their level (i.e. cellular, molecular, tissular, individual or population) and broken in sub- questions when relevant for more granularity. A graphical and a tabular representation of the questions, data and methods used to answer them was built. Analytics were generated to characterize the importance of MIDD approaches. The most common models were Exposure-response and POPPK models. In addition, we identified a series of questions for which modeling and simulation methods could be used but have not been proposed by the applicant. Credibility matrices were filled for questions answered by modeling and simulation methods. To exemplify the implementation of the credibility assessment we selected a POPPK and an exposure-response models with medium to high model risk and regulatory impact and subsequently benchmarked them against alternative approaches used during the concerned drug development. For the models with medium or high regulatory impact, credibility assessment reveals various levels of maturity. Therefore, their regulatory acceptance was also variable. Overall, we found that the use of MIDD approaches has a great potential to address drug development and regulatory questions in the development of drugs for pediatric and rare diseases if informative data can be generated to develop and qualify them. Conclusion Our review of regulatory submissions over 10 years revealed that MIDD approaches have a great potential in the development and assessment of drugs for pediatric and rare diseases if informative data can be generated to develop and qualify them. We identified the applications of MIDD approaches and benchmarked them vs traditional approaches. The use of the question-centric approach and related credibility assessment led to detect various levels of maturity when MIDD approaches were used in paediatric and orphan drug development. Their acceptance by regulators was also variable over the past years. "