Thoughts on the Use of Weight of Evidence Approach in Nonclinical Safety Evaluation — PMDA Early Consideration Summary

In Plain Language

The document addresses situations where conventional animal studies may have limits in predicting how a medical product will behave in humans, including when humans and animals respond differently or an appropriate animal species is difficult to select. PMDA describes using a weight of evidence (WOE) approach—a comprehensive evaluation of multiple available data and information—as a useful way to draw conclusions that one piece of data alone may not establish. It says all available information should be fully used to gain insights relevant to human safety and pharmacokinetics, and that data from New Approach Methodologies (NAMs) may be used alongside animal-study data as part of this approach. PMDA expects evaluation methods independent of animal experiments to support efficient, logical, and scientifically valid nonclinical evaluation and improve safety prediction in humans. The document principally sets out PMDA’s context and planned activities; it does not provide a step-by-step assessment procedure or a list of criteria for deciding whether a particular WOE package is acceptable.

What This Guidance Covers

This Early Consideration presents PMDA’s thoughts on nonclinical safety evaluation using the WOE approach, including the use of data obtained through NAMs. Its subject is the nonclinical evaluation of medical products, the term used in the document. It gives context for why a broader evidence base may be useful and describes actions PMDA intends to take to encourage development and consideration of this approach.

The document discusses new modalities as part of that context. It defines a New Modality as a classification of medical products using new molecules and technologies different from those used in conventional small molecule drugs and antibody drugs. Examples named are oligonucleotide therapeutics, cell therapy products, gene therapy products, multispecific antibodies, and antibody-drug conjugates. The document does not limit the WOE concept to only those examples.

This is a short statement of considerations and planned agency activities, rather than a detailed technical protocol. It does not set out product-specific study requirements, a fixed method for combining evidence, numerical decision thresholds, or criteria for accepting individual NAMs. It also does not present specific case examples in the text provided. Those limits matter when interpreting the document: the reader should not treat it as a ready-made algorithm for reaching a safety conclusion.

Why It Matters

The document identifies several developments behind interest in WOE. First, safety evaluation has been conducted with consideration of the 3Rs of animal experiments: reduce, refine, and replace. PMDA notes increasing international demand for ethical consideration in animal experiments as social concern for animal welfare grows.

Second, the development of new modalities has made some challenges more evident. Species-specific differences—differences in how humans and animals respond—can complicate interpretation of animal results. Choosing an appropriate animal species for nonclinical safety evaluation can also be challenging. In some cases, these issues make appropriate risk assessment difficult using conventional animal experiments.

Third, the document points to advances in scientific technology and the growing attention paid to NAMs as scientific evaluation methods that can replace or complement animal experiments. It anticipates that data from these methods may contribute to improving predictions of efficacy, safety, and pharmacokinetics in humans. These points form the rationale for making fuller use of available information rather than relying on a single kind of evidence where it may not provide a clear answer.

PMDA also describes a practical aim: prompt and more accurate risk evaluation based on scientific evidence, support for the 3Rs, and efficient development of medical products. The document frames WOE as a way to bring relevant evidence together in support of human-focused conclusions, while acknowledging that knowledge and international trends continue to develop.

Key Recommendations

The document’s main direction is that all available information should be fully used to gain insights relevant to safety, pharmacokinetics, and other human considerations in medical product development. PMDA describes proactive use of WOE as useful, with NAM data potentially combined with animal experiments. The detailed actions in the document are actions PMDA says it will take, rather than a list of sponsor deliverables.

1. Use a comprehensive evidence perspective. WOE is defined as the comprehensive evaluation of multiple available data and information to draw conclusions that are not clear from a single piece of data. The document presents this as a way to support nonclinical safety evaluation in view of the limitations of animal experiments and the importance of information relevant to humans.

2. Consider NAM data as part of WOE. The document says NAMs may be used in combination with animal experiments as a basis for WOE. It describes methods independent of animal experiments as expected to promote efficient, logical, and scientifically valid evaluation and contribute to improved prediction of safety in humans. This is a statement of the approach’s expected value, not a claim that every NAM or every dataset is suitable for every assessment.

3. PMDA will organize its thinking using accumulated experience. PMDA says it will organize its thoughts on WOE-based nonclinical safety evaluation by using accumulated consultation cases and discussions at AMED research groups, academic conferences, and other settings. The document does not specify a timetable or a formal output for this activity.

4. PMDA will discuss NAM data and consider acceptance. PMDA says it will discuss the use of NAM data in nonclinical safety evaluation and related challenges with parties inside and outside PMDA, and consider acceptance of data obtained through NAMs. This describes future consideration; it does not state that all NAM data will be accepted.

5. PMDA will share thoughts on specific cases. PMDA says it will present its thoughts on specific cases of WOE-based nonclinical safety evaluation at academic conferences, in research papers, on the PMDA website, and through other channels. The document does not itself supply those case examples.

PMDA says these activities are intended to promote the 3Rs, improve safety prediction in humans, and support the development of safer and more effective medical products in collaboration with related parties.

Practical Implications

Interpretation for development teams: The document signals the value of planning to consider evidence as a whole, especially when animal models have limitations relevant to the product. In practice, a team could make its reasoning clearer by identifying which information comes from animal studies and which comes from other methods, explaining what each source contributes, and describing how the combined information informs the human safety assessment. This is an interpretation of the WOE concept, not a process specified by PMDA in this document.

For nonclinical scientists and statisticians, the emphasis is on the context and relevance of multiple sources of information, rather than treating any single result as necessarily conclusive. If sources differ, a clear account of what each source can and cannot show can help readers understand the overall evaluation. The document does not prescribe a scoring system, formal evidence hierarchy, statistical model, or rule for resolving conflicting findings, so these should not be attributed to PMDA on the basis of this text.

For regulatory affairs teams, the document indicates that PMDA is working to organize its thinking and consider the use of NAM data, including its challenges and acceptance. Teams may therefore find it useful to describe why selected information is relevant to the product and the human safety question. The document does not establish a submission format or promise a particular regulatory outcome; it describes PMDA’s intended activities and general perspective.

For clinical operations and investigators, the document’s focus is nonclinical safety evaluation, not clinical study conduct. Its relevance is indirect: the stated goal is better prediction of safety in humans, but the document does not set out clinical monitoring or trial-design recommendations.

Technical Details

The document provides definitions rather than equations or a stepwise statistical method. Its key concept is that WOE combines multiple available data and information to reach conclusions that are not clear from a single piece of data. The emphasis is on comprehensive evaluation: evidence is brought together to inform conclusions, rather than relying only on one result. The document does not explain how evidence should be weighted, how uncertainty should be quantified, or how to make a decision when different evidence sources point in different directions.

PMDA defines NAMs as tools for evaluating the efficacy, safety, and pharmacokinetics of medical products without using conventional animal models. The definition encompasses technologies and methodologies including in silico (computer-based), in chemico (chemical-based), in vitro (outside a living organism, such as in cells or tissues), and ex vivo (using material taken from a living organism) approaches. These plain-language explanations clarify the terms; the categories and scope are those given in the document.

The document’s use of the 3Rs refers to reducing animal use, refining animal procedures, and replacing animal experiments with other approaches. It situates WOE and NAMs within that context, but does not say that animal experiments must always be removed: it explicitly notes that NAMs may be combined with animal experiments.

Finally, the defined term New Modality refers to medical products using new molecules and technologies different from those used in conventional small molecule drugs and antibody drugs. The examples in the document are illustrative examples of this category, not a stated exhaustive list.

Who Should Read It

This document is relevant to nonclinical safety scientists, toxicologists, pharmacologists, and other specialists who evaluate evidence for medical products, particularly where the suitability of conventional animal models is uncertain. It may also interest statisticians and data scientists involved in integrating results from different methods, although it does not prescribe a statistical integration technique.

Regulatory affairs professionals, medical writers, and project teams can use it to understand PMDA’s stated direction on WOE and NAM data. Researchers developing or applying NAMs may find the planned discussion of data use and acceptance relevant. The document is also useful for readers seeking context on why animal welfare, species differences, and prediction of human safety are discussed together. It should be read as an Early Consideration that reflects PMDA’s views at the time of publication, not as a detailed product-specific assessment framework.

Official PMDA Source

• Official title: Thoughts on the Use of Weight of Evidence Approach in Nonclinical Safety Evaluation (Early Consideration)
• Issuing body: Pharmaceuticals and Medical Devices Agency, Center for Product Evaluation
• Date on cover: October 24, 2025
• Official PMDA PDF: Thoughts on the Use of Weight of Evidence Approach in Nonclinical Safety Evaluation (Early Consideration) (PDF)
• PMDA list page: PMDA Early Consideration documents

Source: Pharmaceuticals and Medical Devices Agency (PMDA) website, PMDA Early Consideration documents. This page was created by editing and summarising the PMDA document.

Consult the official document for the full text; this summary does not replace it. PMDA provides the English version for reference, and the Japanese original prevails if the two versions differ. PMDA describes Early Consideration documents as reference information reflecting its views at the time of publication; those views may change as scientific knowledge develops.

Related Clinical Biostats Resources

• All Clinical Biostats tutorials — Browse other tutorials, including material on nonclinical studies and PMDA Early Considerations.
• Statistical Calculators — Access statistical calculators that may be useful for quantitative work in research and development.