How Legal Liability Concern is Slowing the Future of NAMs

By Mackensey Alexander| August 6th, 2026

I. Introduction

New Approach Methodologies (NAMs) have gained prominence over the last decade, with international policymakers increasingly endorsing alternatives to animal models. In the United States, landmark bipartisan legislation like the FDA Modernization Act 2.0 (passed as part of the Consolidated Appropriations Act) codifies this shift, by explicitly authorizing the use of data from non-animal methods – in lieu of or in addition to animal models – to demonstrate safety and efficacy prior to the start of human trials. Building on this momentum, the FDA modernization Act 3.0 urges the FDA to replace any references to "animal" tests with references to “non-clinical” models and research. This bill was passed in the house committee with an impressive 44-0 vote. Effective NAMs, such as Microphysiological Systems (MPSs), have demonstrated strong human-relevant results in areas including monoclonal antibodies (mAb). In certain toxicity applications, these systems may outperform animal models because they more accurately replicate human physiology. Despite legislative and scientific progress, otherwise effective NAMs like MPSs still face significant practical barriers to routine use in most jurisdictions, because institutions, regulators, and sponsors remain reluctant to move away from animal testing. Much of this hesitation stems from fear of potential legal liability when data is obtained from NAMs rather than animal models. Without a robust regulatory acceptance pathway, legal liability concerns will continue to delay the widespread development and adoption of NAMs.

II. Building a Legal Shield: The Gold Standard

Animal models have been cemented as the legal and cultural precedent for biomedicine since the 1960s, when regulatory agencies like the FDA deemed them the "gold standard." Animal models were incorporated into formal guidelines and legally mandated for safety and efficacy assessments of new drugs. Standardization of animal testing methods became harmonized around the same time, which bolstered their acceptance and use across the globe.

Were an institution to face legal backlash for a product approved using animal-derived data, the institution could argue that their compliance with FDA standards evidenced adherence to federally established safety requirements when defending against negligence claims. Review standards for negligence claims in non-clinical toxicology testing center on ICH guidelines, which interrogate in part whether the correct animal model was used to garner sufficient results. This “gold standard” label for preclinical tests led to pervasive reliance on animal models in all areas of research: from university training to textbooks and accreditation bodies. In practice, this rendered any sort of deviation from such doctrine legally risky.

This pattern has been referred to by scholars as path dependency; a phenomenon typically categorized by social and cultural resistance to change. Path dependency in the biomedical sciences is created by the historical lock-in of the animal model, where actors’ dedication to fulfilling the status quostatus quoactively works to prevent the consideration of alternative methods.

Over the decades, animal use has been largely noninformative in certain areas of human health research – specifically behavioral neuroscience. There are current deep shifts away from animal use as the “gold standard” for safety and efficacy testing, because the biological mismatch between animal models and humans creates translation problems when drugs that may or may not have been effective in animals must perform similarly in human clinical trials. With over 90% of behavioral neuroscience experiments failing to translate to humans, this problem demands scientific innovation. While animal models have certainly yielded successes, we may be bumping up against their limits.

III. Legal Defensibility: The Validation Gap as a Driver of Legal Uncertainty

Using animal models is, in part, a legally strategic decision. Their use is well-established in decades of published literature detailing the animal model validation triad: predictive, face, and construct validity. Predictive validity assesses whether the model’s response to treatment will predict human therapeutic outcomes; face validity assesses whether the model’s observable symptoms resemble those in humans; and construct validity assesses whether the model replicates the biological mechanisms underlying human disease. The validity triad is readily understandable to judges and juries and is cumulative in nature, as each published study contributes to a broader body of evidence that can be presented as scientific consensus. Institutions can use this scientific consensus to defend their research decisions.

NAMs have no comparable unified measure of validation. Validation for NAMs is governed by a patchwork of guidance from different regulatory bodies, including the FDA's fit-for-purpose standard, the ICCVAM  context-of-use framework, the  OECD's  Integrated Approaches to Testing and Assessment (IATA), and the EMA's Qualification of Novel Methodologies process. While these frameworks share general principles for validation, they differ in application. This gap presents a significant and compounding legal challenge. Every new use requires new validation, making it particularly difficult to accumulate the kind of legal defensibility that animal model validation provides. An institution that satisfies one framework has no guarantee it has satisfied any other, and no court has been asked to decide which one sets the legal standard of care. The result is a closed loop, where NAMs cannot generate the scientific consensus courts need to determine the appropriate standard of care without widespread use, and widespread use is unlikely to happen without assurance of NAMs’ legal defensibility.

In negligence actions, the standard of care is generally defined by what a reasonable person (or in this case, a reasonable institution) would have done under the same circumstances. Because animal models are supported by well-established precedent, institutions can easily demonstrate that their conduct conformed to scientific norms. Without similarly reliable precedent, it is difficult to determine what constitutes reasonable conduct when NAMs are used in place of animal models. While the FDA Modernization Act 2.0 has shifted regulatory language surrounding the permitted, and even encouraged, use of non-animal methods, there remains a lack of sufficient language to clarify the legal defensibility of developing and adopting NAMs.

IV. Conclusion

A critical step to establishing legal confidence in NAMs is devising a validation framework that is transferable across jurisdictions, institutions, and experiments. The creation of a robust regulatory acceptance pathway could offer actors similar legal certainty for NAMs as the animal validity triad does for animal models. It is not a lack of innovation that creates caution in the industry around NAMs, but rather a question mark surrounding the legal consequences of their integration. Explicit legislative permission to use NAMs is not the same as legal protection for using them. While it’s true that science must drive the transition to non-animal methods in biomedical research, investment is necessary for NAMs to evolve. Until institutions, regulators, and sponsors can identify a clear, transferable validation framework that courts recognize as establishing a reliable standard of care, legal uncertainty will continue to suppress the development and adoption of NAMs.  

The views expressed do not necessarily reflect the official policy or position of Johns Hopkins University or Johns Hopkins Bloomberg School of Public Health.