Pediatric Drug Development: From “Wild West” to Model-Informed Confidence

Despite decades of progress in drug development, pediatrics remains one of the most complex and underserved areas of therapeutics. More than half of approved medicines still lack pediatric labeling, and especially in neonatal intensive care units off-label prescribing is often the rule rather than the exception. Clinicians routinely face high-stakes decisions with limited data, a reality many experts describe as the pediatric dosing “wild west.” 

Pediatric model-informed drug development (MIDD) is transforming this landscape by replacing empirical guesswork with quantitative, biology-driven decision making that improves safety, efficacy, and regulatory confidence. 

Why Pediatric Development Requires Model-Informed Approaches

Children are not small adults. Rapid, nonlinear physiological changes fundamentally alter drug absorption, distribution, metabolism, and elimination across development. Organ maturation, enzyme ontogeny, body composition, and renal function evolve dramatically from birth through adolescence, directly shaping drug exposure and response. 

Weight-based dosing alone rarely reflects this complexity. Combined with sparse sampling, small populations, and ethical constraints, traditional empirical approaches without modelling and simulation often fail to deliver reliable pediatric dosing strategies, making quantitative modeling essential rather than optional. 

Regulatory Expectations Are Shifting Earlier in Development

Global regulators now expect pediatric strategies to be embedded early in development rather than added as late-stage obligations. Stepwise Pediatric Investigation Plans (PIPs) in Europe, initial Pediatric Study Plans (iPSPs) in the U.S.,and modern extrapolation guidance such as ICH E11A all emphasize proactive justification of dosing, study design, and evidence generation using quantitative, model-based approaches. 

MIDD is no longer viewed as a supporting analysis. Emerging regulatory frameworks such as ICH M15 are already being incorporated into submission templates, signaling that regulators increasingly expect transparent, quantitative decision-making to underpin pediatric programs. 

Regulatory innovation is also extending beyond clinical trial design. The FDA’s introduction of New Approach Methodologies (NAMs) for safety and toxicity assessment reflects a broader shift away from traditional animal models toward mechanistic, data-driven evidence generation aligned with modern modeling strategies. 

Together, these shifts reflect a broader move away from trial-and-error development toward biologically plausible, quantitatively justified dose selection, often before first-in-child studies even begin. 

The MIDD Toolbox for Pediatric Drug Development

Pediatric model-informed drug development integrates multiple quantitative methods to translate developmental biology into actionable decisions: 

• Population Pharmacokinetics/Pharmacodynamics (popPK, PK/PD) modeling to characterize variability with sparse data
• Physiologically Based Pharmacokinetics (PBPK) to incorporate age-specific physiology and enzyme maturation
• Model-based meta-analysis (MBMA) to leverage adult or older pediatric data
• Quantitative Systems Pharmacology (QSP) to represent pediatric disease mechanisms 

Together, these tools enable simulation-based dosing strategies, safer trial designs, and stronger regulatory packages grounded in science rather than assumption. 

AI and Machine Learning as a Force Multiplier

Artificial intelligence (AI) and machine learning (ML) are increasingly enhancing pediatric model-informed drug development by accelerating model building, exploring complex parameter spaces, and identifying meaningful covariates across limited pediatric datasets. 

ML-driven approaches can rapidly test alternative model structures, support external validation, and uncover exposure–response patterns that may be difficult to detect with traditional workflows alone. 

Importantly, AI augments, not replaces, but enhance pharmacometric and clinical expertise and accelerate drug development timeline. Biological plausibility, transparency, and interpretability remain essential, particularly when decisions affect vulnerable populations. The most effective pediatric programs combine algorithmic efficiency with mechanistic understanding. 

From Modeling to Pediatric Precision Dosing in Practice

The impact of MIDD now extends directly to clinical care through Model-Informed Precision Dosing (MIPD). By combining population models with Bayesian updating, clinicians can individualize therapy using real-time patient data, making pediatric precision dosing achievable in everyday practice. 

Rather than fixed dosing tables, MIPD continuously adapts treatment to developmental stage, exposure levels, and patient response. 

Real-world applications such as maturation-informed sirolimus dosing in infants with vascular anomalies show how quantitative approaches improve outcomes when traditional trials are impractical or unethical. 

What This Means for Drug Developers

By embedding model-informed strategies early, sponsors can: 

• Reduce uncertainty before first-in-child studies
• Minimize invasive sampling and unnecessary trials
• Deliver transparent, defensible dosing rationale to regulators
• Accelerate translation from development to clinical use 

The most successful pediatric programs now integrate modeling by the end of Phase 1 or early Phase 2, not as a rescue strategy, but as a core development pillar. 

A Child-Centered Future Built on Data

Closing the pediatric evidence gap requires replacing empirical extrapolation with quantitative, biology-based decision making. 

Pediatric model-informed drug development is transforming the field from a “wild west” into a disciplined, predictive science, enabling safer therapies, faster development, and better outcomes for children.