Computational modeling and simulation (M&S) has quickly moved from promise to practice in early-phase drug development. By reducing reliance on animal testing and accelerating timelines, in silico methods have become an integral part of the discovery process, backed by recent policy shifts like the FDA Modernization Act of 2022.
At the heart of this transition is model-informed drug development (MIDD), a framework that leverages M&S tools to streamline development, reduce uncertainty, and enable more confident regulatory decision-making. A recent FDA report by the Modeling and Simulation Working Group details tangible progress and case studies that highlight how MIDD approaches are already influencing regulatory decision making, from evaluating drug impurities to predicting toxicity and optimizing clinical trial design.
This momentum aligns with broader changes within the FDA, including the agency’s commitment to phasing out animal testing requirements and the integration of Elsa, an artificial intelligence (AI) tool designed to assist with reading, writing, and summarizing. As M&S is increasingly being used as a primary tool for evaluating safety, efficacy, and manufacturing robustness, it’s important to understand how these approaches are being integrated into regulatory review practices and codified within global frameworks.
How the FDA is applying these tools
Across the FDA, different centers are now using MIDD tools to support clinical trial design, regulatory decision-making, and policy development. Common approaches include physiologically based pharmacokinetic (PBPK) models, quantitative structure–activity relationship (QSAR) models, quantitative systems pharmacology (QSP) models, and digital twins. But what exactly are these approaches, and how do they work in practice?
- PBPK models simulate how a drug moves through the body, including absorption, distribution, metabolism, and excretion, using virtual populations. This helps predict responses in children, the elderly, or those with organ impairments before clinical testing.
- QSAR models use the chemical structure of compounds to predict specific outcomes, such as toxicity or mutagenicity. They flag high-risk molecules early in discovery, helping researchers prioritize compounds and reduce unnecessary lab or animal tests.
- QSP models combine drug-specific data, including PBPK and QSAR outputs, with detailed biological pathway information to simulate a drug’s effects on a disease system over time. They guide dosing, trial design, and patient selection, predicting both efficacy and safety throughout preclinical and clinical development.
- Digital twins are virtual replicas of physical manufacturing systems. They enable engineers and regulators to test process changes, assess risks, and optimize quality control without disrupting actual production lines.
For example, the FDA’s Center for Drug Evaluation and Research (CDER) uses PBPK modeling to assess drug interactions and optimize dosing. CDER has also developed QSAR models to predict genetic toxicity when standard test data are limited, and has created digital twins of continuous manufacturing lines for several solid oral drug submissions since 2019.
This growing regulatory use of M&S reflects a parallel trend in industry. As Piet van der Graaf, Senior Vice President and Head of QSP at Certara, told DDN, “A 2024 analysis of FDA submissions found that the number of QSP models provided to the agency has more than doubled since 2021.” These submissions supported small molecules and biologics in multiple therapeutic areas and were applied not only to drug efficacy (>66 percent of cases), but also to safety assessments such as liver toxicity, cytokine release syndrome, and bone density changes. QSP was also applied to dose optimization, the simulation of therapies for rare diseases, and the creation of virtual populations to capture variability in pharmacodynamic responses.
From guidance to global standards
The FDA is not the only agency integrating MIDD approaches; international regulators such as the European Medicines Agency and Japan’s Pharmaceuticals and Medical Devices Agency are also advancing parallel frameworks. To avoid fragmentation, the International Council for Harmonization (ICH) has initiated efforts to standardize how M&S outputs are planned, evaluated, and documented for cross-border submissions. Rather than treating models as standalone evidence, this framework emphasizes embedding M&S within broader, evidence-based assessments.
As Mark Davies, the Head of Quantitative Pharmacology and Data Science at Physiomics, told DDN, “we are close to establishing a universal practice guideline through the ICH’s M15 guidance, which was published as a draft in 2024. This work builds on a long-standing interdisciplinary community exploring best practice, and the guidance is written to apply to both current and emerging MIDD approaches, while providing recommendations for related regulatory interactions, reporting, and submissions.”
In parallel, the FDA has launched pilot programs, such as the Innovative Science and Technology Approaches for New Drugs (ISTAND) initiative. This is designed to qualify novel M&S tools as drug development methodologies under regulatory review. The program is focused explicitly on nonanimal-based methodologies and technologies including AI that “use human biology to predict human outcomes in order to help reduce and replace animal testing as part of drug development.” If successful, these efforts could accelerate the shift from case-by-case evaluation toward codified standards that apply across therapeutic areas and modalities.
The growing role of AI
Beyond traditional M&S, the FDA is also beginning to integrate AI tools into its workflow. Elsa, the agency’s large language model-powered assistant, is being tested to accelerate review tasks such as summarizing adverse event reports and evaluating trial protocols. While this use is relatively limited today, it signals a broader shift in how regulators may use AI to support both operational efficiency and scientific evaluation in the future.
By harnessing these tools, regulatory agencies can create faster, data-driven pathways for innovation while ensuring that safety and efficacy remain at the core of regulatory science.
This piece is part of a series examining the growing role of modeling and simulation in both drug development and regulatory decision-making. For more on the industry side, read our other article on how modeling and simulation is transforming early-phase drug development and trial design.
Frequently asked questions (FAQ):
How do modeling and simulation (M&S) benefit drug development?
M&S approaches can predict safety and efficacy outcomes before clinical testing, reduce reliance on animal studies, identify potential risks earlier, and shorten development timelines. They also allow regulators to explore scenarios that would be impractical or unethical to test in real patients.
What are some common M&S approaches used by the FDA?
- PBPK models simulate how drugs are absorbed, distributed, metabolized, and excreted in the body.
- QSAR models predict outcomes such as toxicity based on chemical structure.
- QSP models integrate biological pathways and drug data to predict efficacy and safety.
- Digital twins replicate manufacturing systems to test process changes without disrupting production.
Are international regulators adopting M&S approaches too?
Yes. Agencies such as the European Medicines Agency (EMA) and Japan’s Pharmaceuticals and Medical Devices Agency (PMDA) are advancing parallel frameworks. The International Council for Harmonisation (ICH) is developing the M15 guideline to establish global best practices for planning, evaluating, and documenting models in regulatory submissions.
What role does AI play in regulatory science?
AI is beginning to complement traditional M&S. At the FDA, tools like Elsa — a large language model–powered assistant — are being tested for tasks such as summarizing trial protocols.
What is the ISTAND initiative?
The Innovative Science and Technology Approaches for New Drugs (ISTAND) program is an FDA pilot designed to qualify novel drug development tools — including M&S and AI-based methods — as regulatory methodologies. A major focus is on non-animal-based approaches that use human biology to predict human outcomes.
