How is animal research regulated?
Through licensing systems built around a framework known as the 3Rs — replacement, reduction and refinement — with oversight that is considerably stricter than public debate usually acknowledges, and which critics argue still permits too much.
The 3Rs, proposed in 1959 and now embedded in law in many countries:
Replacement — using non-animal methods wherever possible: cell cultures, organ-on-chip systems, computer modelling, human volunteer studies and existing data.
Reduction — using the minimum number of animals, including through better statistical design, which is also better science.
Refinement — minimising suffering through housing, handling, analgesia, humane endpoints and enrichment.
How licensing typically works. Three layers: a licence for the establishment; a licence for the project, assessed by a regulator weighing likely benefit against likely harm; and a licence for each individual carrying out procedures, requiring training and competence. Local ethical review bodies assess proposals before submission, and inspectors visit, in many systems unannounced.
The harm-benefit assessment is the core of it: a project is refused if the expected harm is not justified by the expected benefit, or if a non-animal alternative could answer the question.
What the numbers actually show. Published statistics in several countries report that the large majority of procedures are on mice, fish and rats, with dogs, cats and primates a very small fraction; that a substantial proportion are for breeding and maintaining genetically altered animals rather than experiments; and that most procedures are classified as mild or sub-threshold, with severe procedures a small minority.
The genuine criticisms: that severity classification is self-assessed; that poor reproducibility in animal studies means some suffering produces no usable knowledge; that translation from animal models to human outcomes frequently fails, particularly in neurological and psychiatric research; and that regulatory testing requirements sometimes mandate animal use where alternatives now exist.
Where change is happening: validated non-animal methods are being accepted for some regulatory testing, and several jurisdictions have set out roadmaps to reduce reliance on animal models.
General information; frameworks vary by country.