2.1.1—Animal models

An animal model uses a non-human organism to study a biological process or disorder with the aim of learning something relevant to humans.

Models can allow controlled manipulation, repeated measurement and access to mechanisms that are difficult to study directly. Their value depends on construct validity: the model must share the relevant process, not merely show a similar outward behaviour.

A stress manipulation in rodents may test a pathway involved in depression, but species-specific behaviour, dose and context limit the inference. Human clinical evidence is still needed for treatment claims.

An animal result is not automatically generalisable or unethical. Evaluate welfare, model fit and the exact claim being transferred.

Evaluate animal ethics with replacement, reduction and refinement: replace animals with non-animal methods where the question can be answered adequately; reduce numbers while preserving a valid design; and refine housing, procedures and endpoints to minimize pain, distress and lasting harm. A harm-benefit review should consider species needs, scientific necessity, model validity, expected knowledge, trained care and whether less harmful alternatives exist. Weak transfer to humans reduces the ethical justification because animals would bear costs for evidence that cannot support the intended claim.