Integrated Approaches to Testing and Assessment

Integrated approaches to testing and assessment (IATAs) provide a means by which a diversity of relevant and reliable information about a chemical is used to answer a defined hazard characterization question. The information may include toxicity data, computational model predictions, exposure routes, use cases, and production volumes. This information is used to characterize outcomes that can inform regulatory decision-making.

There is increasing interest in replacing tests that use laboratory animals with human cell-based, biochemical, and/or computational methods to predict chemical toxicity. However, because of the complexity of toxicity mechanisms, data from several alternative methods usually need to be considered together to adequately predict toxicity. IATAs provide a means for considering these data in combination. IATAs can also guide the generation of new data, preferably using non-animal approaches, to inform regulatory decision-making.

A defined approach to testing and assessment can be used as part of an IATA or on its own as a source of chemical safety information. While IATAs may rely on expert judgment, defined approaches are structured and reproducible, and therefore provide objective outputs. NICEATM, in collaboration with other scientists, has developed defined approaches for skin sensitization, endocrine disruption, and identification of eye irritants.

A defined approach to testing and assessment relies on:

Input data generated from identified methods.
A data interpretation procedure that is used to translate data produced from a defined set of information sources into a prediction.

Examples of defined approaches:

Sequential testing strategies use a fixed, stepwise approach to obtain and assess test data. They include interim decision points at which the user may either proceed to additional testing steps or stop testing and make a hazard prediction.
Integrated testing strategies use an approach in which multiple sources of data or information are assessed at the same time to arrive at either a hazard prediction or a decision that more testing is needed.

More information about IATAs and defined approaches is available on the OECD website.

Defined approaches for eye irritation

NICEATM is collaborating with PSCI, EPA, and member companies of the trade association CropLife America to identify promising methods to use in an in vitro defined approach for classification of eye irritation potential of agrochemical formulations. Development of these methods is described in published papers (Clippinger et al. 2021; van der Zalm et al. 2024; Daniel et al. 2025) and a NICEATM report. Use of defined approaches for identifying chemicals that could cause serious eye damage and irritation was internationally accepted in 2022 with adoption of OECD Test Guideline 467 (updated in 2025).

Defined approaches for skin sensitization

In 2018, EPA released a draft Science Policy to reduce animal use by using defined approaches to identify potential skin sensitizers. The draft policy is the result of national and international collaboration among ICCVAM, NICEATM, Cosmetics Europe, EURL ECVAM, and Health Canada’s Pest Management Regulatory Agency. These collaborations included the following projects:

NICEATM and international collaborators convened a 2016 workshop on skin sensitization (Casati et al. 2018) that produced a proposal for an OECD project to develop an international harmonized guideline for defined approaches for skin sensitization (Guideline 497, adopted 2021, last updated 2025).
NICEATM and industry scientists from Cosmetics Europe wrote open-source code to reproduce defined approaches submitted as case studies to OECD for skin sensitization testing and assessment. Data on 128 substances to evaluate these defined approaches were also developed as part of this project (Hoffman et al. 2018; Kleinstreuer et al. 2018).
NICEATM and ICCVAM developed a defined approach that uses non-animal data to predict three skin sensitization endpoints: murine local lymph node assay hazard (Strickland et al. 2016), human skin sensitization hazard (Strickland et al. 2017), and human or animal skin sensitization potency (enables classification of skin sensitizers as "weak" or "strong" without animal tests; Zang et al. 2017).
NICEATM and collaborators with Procter & Gamble developed open-source software to implement an integrated testing strategy (a type of defined approach) for skin sensitization potency based on a Bayesian network. The model provides a numerical probability that a substance belongs in each of four potency classes (Pirone et al. 2014).

NICEATM and the consumer products company Unilever collaborated to develop the Skin Sensitization Risk Assessment – Integrated Chemical Environment (SARA-ICE) defined approach (Reinke et al. 2025). SARA-ICE is the first defined approach accepted in an OECD Test Guideline (OECD Guideline 497) to provide a human-relevant numerical estimate of skin sensitizer potency, which is required in some regulatory contexts.

Defined approaches for endocrine disruption

NICEATM and EPA developed a defined approach that combines data from 11 high-throughput screening tests with a computational model to identify substances with the potential to interact with androgen receptors in a cell (Kleinstreuer et al. 2016). Androgens are a class of hormones, produced largely by the testes, that serve as the primary male hormones.
NICEATM validated a defined approach developed by EPA that combines data from 18 high-throughput screening tests with a computational model to identify chemicals with the potential to interact with the estrogen receptors in a cell (Browne et al. 2015). Estrogens are a class of hormones, produced largely by the ovaries, that serve as the primary female hormones.
NICEATM and EPA analyzed the high-throughput screening based models described above to identify a subset of assays that would provide similarly accurate predictions of endocrine disruption while minimizing the resources required to run defined approaches. For detecting estrogenic substances as few as four assays could be used (Judson et al. 2017). For substances interacting with the androgen pathway as few as five assays provided similar performance as the full model (Judson et al. 2020).