The Digital CMC Sandbox is a core enabler of the Digital CMC CERSI, providing a controlled “safe space” for regulators, industry and academia to explore, evaluate and build confidence in computational models. It directly supports the CERSI mission to advance regulatory science by enabling innovation in a structured, risk-informed environment while maintaining product quality, patient safety and regulatory integrity.
In this context, the sandbox allows organisations to test, validate and understand digital tools without regulatory risk, supporting their future use in submissions or inspected environments and helping to align global expectations.
Two Complementary Elements
1. Static Environment – Structured Evidence Generation
The static component of the sandbox captures outputs generated through the DCRL workflow. As users progress through each stage, it builds a comprehensive model credibility assessment, including context of use, risk evaluation, data strategy and validation evidence.
• Provides a structured, traceable record aligned to regulatory expectations
• Supports documentation for regulatory submissions and inspections
• Enables consistent interpretation of evidence across organisations and regions
This element ensures that digital tools are supported by robust, transparent and regulator-ready evidence.
2. Interactive Environment – Dynamic Model Evaluation
The interactive component allows users to actively engage with models in a controlled setting. Users can adjust inputs, rerun models and explore outcomes under different scenarios.
• Visualises model performance, uncertainty and robustness
• Enables exploration of factors such as data quality, quantity and model assumptions
• Supports scenario-based analysis aligned to the defined context of use
This dynamic environment enhances transparency, explainability and trust, helping users understand both the capabilities and limitations of digital tools.
Exemplar Case Studies in the Sandbox
The Digital CMC Sandbox includes four illustrative case studies that demonstrate how the DCRL framework can be applied across different types and regulatory risk levels of computational models. Together, they support the CERSI aim of enabling innovation through practical, regulator-ready examples that build confidence across industry and global regulators.
• Case Study 1 – Hybrid (knowledge- and data-driven ML models) Model for Predicting the Solubility of an Organic Compound in a Solvent
• Case Study 2 – Data-driven / AI Model: Digital Polymorph Risk Assessment using solid form informatics tools and automated testing
Focuses on AI/ML approaches, illustrating key considerations such as data quality, training/test splits, model uncertainty and explainability. It shows how risk-based evaluation is critical where models influence quality or control decisions.
• Case Study 3 – Empirical model: Stability Related Dissolution Changes in Oral Solid Dosage Forms to Support Shelf-Life Assignment: An Empirical Modelling Approach
• Case Study 4 – Hybrid Digital Modelling Framework for Continuous Direct Compression: Predictive Loss-in-Weight Feeder Models for Soft Sensor Applications
[IH10.1]
These case studies provide complementary perspectives across model types and use cases, helping users understand how the DCRL framework can be applied consistently. They act as practical reference points to support training, regulatory dialogue, and the safe adoption of digital tools in CMC.
Figure 6: Digital CMC Sandbox Exemplar Case Studies
Training and Capability Building
The sandbox also serves as a practical training platform, supporting a wide range of users—from regulators and regulatory professionals to scientists and model developers. Through guided interaction with models and structured credibility assessments, users develop a deeper understanding of:
• Model behaviour and uncertainty
• Data dependencies and risk considerations
• Evidence requirements for regulatory acceptance
More advanced scientific and mathematical training can be layered on where required.
________________________________________
Why It Matters
The Digital CMC Sandbox exemplifies the CERSI approach: creating shared environments where innovation and regulation evolve together. By combining structured evidence generation with hands-on evaluation, it helps:
• Reduce uncertainty and perceived risk
• Build confidence in digital and AI-enabled approaches
• Enable consistent, globally aligned regulatory decision-making
Ultimately, the sandbox supports the safe, transparent and effective integration of digital technologies into medicines development and manufacturing.