Reduction requires limiting animals used in the study to the absolute minimum required for scientific validity. This minimum may be based on a statistical analysis (power analysis), or an estimate of the minimum quantity of antibody or tissue required by a study. Another statistical method currently being employed to reduce the numbers of animals is sequential analysis. For those studies where this method is appropriate a notable decrease in animal numbers can be realized. Changes in experimental design can also reduce the number of animals required. For example, use of an unbalanced design (fewer control animals than experimental animals) may be possible for some studies. Other studies may permit use of the animals as their own control thus eliminating the need for a separate control group. In some cases a control group may not be necessary at all. For example, in a study of a new immunosuppressive drug an investigator may want to include a control for the vehicle in which the drug was dissolved. However, previous studies of immunosuppressive drugs already had used this vehicle as a control. It was not necessary to repeat the control. The use of animals to determine standard baseline values is also not necessary in many cases since many such values may be obtainable from the literature.
Less obvious methods include the use of genetically homogeneous animals from known stocks to reduce the variability due to the organism itself. Controlled environments and standardized husbandry practices also contribute significantly to the reduction of organismic variability and thus to the reduction of the number of animals required for a study. Animal sharing, if done appropriately, can reduce the number of animals used in research. The key is appropriate use. Animal sharing does not mean the use of an animal in multiple painful or stressful procedures. A clear example is the use of an animal under anesthesia in a feasibility study just before it is to be euthanized for tissue harvest by use of anesthesia overdose in a second study.