Refinement of rodent research through environmental enrichment and systematic randomization (2007)

Wurbel, H., Garner, J. P.

Abstract

Environmental enrichment may prevent abnormal behaviours and improve animal well-being, but concerns have been raised that it might also disrupt standardisation, thereby reducing the precision and replicability of animal experiments. In this article, we review the logic and evidence surrounding this debate. We show that animal welfare can be improved by beneficial enrichments without disrupting standardization. However, we also argue that standardization is a flawed concept, which entails the risk of obtaining results of poor external validity and therefore needs to be profoundly revised. Environmental enrichment may or may not improve animal welfare depending on whether or not the enrichments are biologically relevant and beneficial to the animals. In mice (especially males), for example, nesting material may currently be the only enrichment to conventional barren "shoe-box" type cages that can be recommended unreservedly, although more may be needed to guarantee acceptable well-being. Further research is therefore needed to develop practicable housing systems and enrichments.Contrary to common claims, there is no evidence that enrichment increases variation in the data of animal experiments, or that it increases the risk of obtaining conflicting results in replicate studies. However, both theoretical considerations and empirical evidence indicate that the concept of environmental standardization (intended to minimize both variation in the data and the risk of obtaining conflicting results in replicate studies) seriously limits external validity of many animal experiments and therefore actually decreases replicability of results. Systematic environmental randomization provides a means to increase the external validity (and hence replicability) of experimental findings without inflating the numbers of animals used. Together, environmental enrichment and systematic environmental randomization therefore contribute to the refinement of animal experiments in the best of meanings of the 3Rs concept.