The role of validation in establishing the scientific credibility of predictive toxicology approaches intended for regulatory application
Open accessAbstract
New approaches in toxicology based on in vitro methods and computational modelling offer considerable potential to improve the efficiency and effectiveness of chemical hazard and risk assessment in a variety of regulatory contexts. However, this presents challenges both for developers and regulatory assessors because often these two communities do not share the same level of confidence in a new approach. To address this challenge, various assessment frameworks have been developed over the past 20 years with the aim of creating harmonised and systematic approaches for evaluating new methods. These frameworks typically focus on specific methodologies and technologies, which has proven useful for establishing the validity and credibility of individual methods. However, given the increasing need to compare methods and combine their use in integrated assessment strategies, the multiplicity of frameworks is arguably becoming a barrier to their acceptance. In this commentary, we explore the concepts of model validity and credibility, and we illustrate how a set of seven credibility factors provides a method-agnostic means of comparing different kinds of predictive toxicology approaches. It is hoped that this will facilitate communication and cross-disciplinarity among method developers and users, with the ultimate aim of increasing the acceptance and use of predictive approaches in toxicology.
The claims and statistics of this paper haven’t been checked yet.
Run the check — free, no account needed. It reads the open-access text and checks the sources it cites.
Integrity signals
Registry
Cites
Statistics
Claims the paper makes
The claims and statistics of this paper haven’t been checked yet.
Run the check — free, no account needed. It reads the open-access text and checks the sources it cites.
Claims other papers make about this one
Each is checked against this paper’s own text: does it say what the citing paper says it does?
Mechanistic validation may provide a measure of test method performance even when a good reference dataset is not available.
“Such an approach may provide a measure of test method performance, even when a good reference dataset is not available (Leist et al., 2012; Patlewicz et al., 2013; Patterson et al., 2021; Tigges et al., 2021).”
From Recommendations on fit-for-purpose criteria to establish quality management for microphysiological systems and for monitoring their reproducibility (2024) · cites this paper as Patterson et al., 2021
In “Validation principles and frameworks > Similarities and differences between the frameworks” (page unavailable), the paper says that when relevant in vivo data are scarce, an in vitro test’s credibility can rest mainly on mechanistic relevance and thorough test-method characterisation. It does not say this provides a measure of test-method performance; it says predictive capacity cannot really be demonstrated in that situation. (LLM, conf=0.84)
Experimental — accuracy not yet measuredRegulatory bodies, validation authorities, method developers, and industry toxicologists realize the need to increase confidence in the scientific validity of novel in vitro methods, especially those proposed for regulatory application.
“Regulatory bodies, validation authorities, method developers, and industry toxicologists realize the need to increase confidence in the scientific validity of novel in vitro methods -especially those being proposed for regulatory application (Patterson et al., 2021).”
From How a GIVIMP certification program can increase confidence in in vitro methods (2021) · cites this paper as Patterson et al., 2021
Self-check
Percentages, sums and reported differences re-computed from the paper’s own numbers.
The self-check needs the paper’s full text.
The citation lists haven’t been fetched yet
References 52 · Cited by 47
Research works printed but not linked: 4
- [23]ECHA Guidance on Information Requirements and Chemical Safety Assessment. Chapter R.4: Evaluation of available information 2017 ECHA Helsinki, Finland
- [26]Balci O. Verification, validation, and testing, in Handbook of Simulation: Principles, Methodology Advances, Application and Practice 1998 John Wiley & Sons (Engineering & Management Press) New York 335 393
- [47]OECD, 2015, OECD Style Guide, 3rd Edition, Paris: OECD.
- [60]Fuller S. Social epistemology 2nd Edition 2002 Indiana University Press Bloomington, IN
- Authors
- Patterson, E. A., Whelan, M. P., & Worth, A. P.
- Affiliations
- University of Liverpool (GB)
- Joint Research Centre (IT)
- Corresponding author
- Worth, Andrew P.
- Venue
- Computational Toxicology
- Publisher
- Elsevier BV
- Published
- 2021-02
- Volume
- 17
- Pages
- 100144
- ISSN
- 2468-1113 (ISSN-L)
- PMC
- PMC7903516
- OpenAlex
- W3098314637
- Cited by
- 47
- References
- 52
- Citations by year
- 2026: 12 · 2025: 5 · 2024: 10 · 2023: 6 · 2022: 8 · 2021: 6
- FWCI
- 0.92
- Type
- journal article
- Language
- en
- Open access
- hybrid
- doi.org
- Licence
- Locations
- Computational Toxicology (publishedVersion, cc-by)
- PubMed Central (publishedVersion, cc-by)
- PubMed (publishedVersion)
- Joint Research Centre (European Commission) (submittedVersion)
- Funders
- Joint Research Centre
- Topics
- Health Sciences › Veterinary › Small Animals › Animal testing and alternatives
- Physical Sciences › Computer Science › Computational Theory and Mathematics › Computational Drug Discovery Methods
- Life Sciences › Biochemistry, Genetics and Molecular Biology › Cancer Research › Carcinogens and Genotoxicity Assessment
- Keywords
- predictive toxicology, method validation, scientific credibility, model validity, credibility factors, regulatory toxicology, in vitro methods, computational modelling, chemical hazard assessment, risk assessment, regulatory acceptance
About these checks
Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.
Statistics and arithmetic are recomputed, not judged.
Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.
What counts as a claim ›How we measure confidence ›How we check this ›
Are you an author, or do you see a problem with this page? Request a review
