Note: In this article, all references to USEtox refer exclusively to the far-field model. Near-field exposure (consumer product use, indoor air, dermal exposure, direct human contact) is NOT covered here and will be addressed in a separate article.

This article represents the view of the author.

Why Different Models Give Different Answers — and Why USEtox far-field Is Not a Universal Consensus

The result is simple but often overlooked: the conclusions of an LCA can change depending on the model chosen. In some cases, two products may even switch ranking entirely. This article clarifies why this happens, how consensus efforts like USEtox far-field emerged, and why significant limitations remain — especially for toxicity-related impact categories.

One method, many models: why results differ

In LCA, the inventory (energy, materials, emissions) must be translated into impacts through characterisation models. But unlike climate change — where the IPCC provides a globally accepted scientific foundation — many LCA impact categories have been shaped by independent academic teams working in parallel.

As a consequence, impact categories such as human toxicity, ecotoxicity, water use, resource depletion, land use, photochemical ozone formation, and freshwater eutrophication have historically been calculated using different models, each based on distinct system boundaries, fate/transport equations, exposure assumptions, effect factors and datasets.

Not surprisingly, these differences lead to impact scores that may vary by factors of 10, 100 or even 10,000. Two LCAs of the same product can give different answers, and two alternative products can switch ranking depending on the model used.

The toxicity  models disagree by orders of magnitude

Toxicity impact categories are where disagreements were historically the strongest. Different research groups created competing models — CalTOX, USES-LCA, IMPACT 2002+, TRACI, EDIP, MEEuP, etc. — each with its own algorithm and assumptions.

Results often differed by several orders of magnitude. A chemical classified as highly toxic in one model could appear almost harmless in another. Product toxicity scores could be altered simply by switching model. Early toxicity models also aligned poorly with regulatory systems such as REACH/CLP.

Risk assessment experts criticised these early approaches for being hazard-based only, with minimal treatment of exposure, fate, degradation, mixture effects or long-term accumulation.

The emergence of USEtox far-field: a pragmatic attempt at harmonisation

In response to these inconsistencies, UNEP and SETAC launched an international task force to build a consensus model for human toxicity and ecotoxicity: USEtox far-field. Today, USEtox far-field is incorporated into ILCD recommendations and the EU Product Environmental Footprint (PEF).

USEtox far-field harmonised many modelling assumptions, reduced variation, introduced multimedia fate and exposure modelling, and created a shared reference point for toxicity assessment in LCA.

But the USEtox far-field consensus is not equivalent to global scientific consensus

Despite its value, USEtox far-field is often misinterpreted as a universal scientific consensus. It is not. The USEtox far-field consensus was built within a limited circle of LCA practitioners, not by the toxicology, regulatory or environmental chemistry communities.

Unlike climate change modelling, USEtox far-field does not benefit from a global scientific consensus. Many toxicologists do not endorse USEtox far-field as a valid toxicity assessment tool.

Use acceptance also remains low because:

• USEtox far-field input data do not align with REACH/CLP regulatory datasets

• PBT/vPvB substances are underrepresented relative to their importance  

• Only inhalation and ingestion exposure routes are included  

• Toxicity scores are often dominated by metals from energy production  

• Up to 50% of emitted substances still lack characterisation factors  

In practice, despite having a fate factor (or exposure factor), USEtox far-field remains a hazard-based indicator rather than a risk assessment tool. USEtox outputs should never be interpreted as a safety indication. Products with a lower USEtox are not safer, it just means that potential pressure on the human or environment should be lower.

No LCA toxicity model can be experimentally validated

A fundamental challenge of LCA impact assessment is that results cannot be validated experimentally. There is no measurable “true” toxicity footprint that can be compared with model predictions.

Environmental fate, degradation, mixture toxicity, long-term exposure and multi-compartment interactions cannot be monitored in a way that allows validation. All toxicity models in LCA remain theoretical constructs whose outputs must be interpreted as relative indicators, not absolute measurements.

What this means for sustainability assessments

LCA remains a powerful tool, but its toxicity modules — even with USEtox far-field — must be interpreted with caution. For sustainability and Safe & Sustainable by Design (SSbD), LCA should be complemented with:

• REACH/CLP hazard classifications  

• identification of SVHC, PBT/vPvB and endocrine disruptors  

• exposure and risk assessment  

• material/substance substitution strategies  

• product-specific chemical safety analyses  

Regulatory-aligned approaches such as ProScale, which uses REACH/CLP data directly, may provide clearer decision-making support.

Conclusion

LCA impact assessment models are essential, but they are not neutral or interchangeable. Different academic models can lead to drastically different results, even opposite conclusions when comparing products.

USEtox far-field was created to reduce inconsistencies in toxicity modelling, and it represents a major improvement over earlier models. However, it is not a universal scientific consensus, is not endorsed by toxicologists, and cannot be validated experimentally.

The good news is that LCA practitioners increasingly recognise these limitations. Over the last decade, the LCA community has been working toward greater harmonisation, improving data quality, and refining model structures to reduce methodological variability.

The European Commission has also addressed this issue through the Product Environmental Footprint (PEF), which recommends specific models for each impact category to ensure that LCAs performed across Europe are comparable, reproducible and based on a shared scientific foundation.

Still, no single model fully captures the complexity of chemical impacts or environmental dynamics. For meaningful sustainability assessments, LCA must be used as part of a broader decision-support toolbox. Recognising the strengths and limits of each model — and using them transparently — is the first step toward designing products that are truly safe and sustainable by design.

#LCA #LifeCycleAssessment #Sustainability #Toxicity #USEtox #PEF #SSbD #ChemicalSafety #EcoDesign #EnvironmentalFootprint #GreenDeal