This procedure described below will help you to randomize animals into animal groups according to the needs of a G(x)P environment:
The algorithm is based on a document published by Sage. You can find it here. The title is Randomization of Animals by Computer Program for Toxicity Studies. The process will follow the exact process using 5 different weight range groups. The only deviation on the process is that for we can use variable amount of animals for each group. The document assumed that all animal groups had an equal number of animals assigned.
This process runs separate for males and females.
- First the animals are sorted on weight.
- Then they are divided into 5 different weight groups.
- Each animal is assigned a unique random number within the weight range group.
- Then the animal groups will start picking the animals from the weight range groups. The system calculates based on the number of animals that need to be allocated in the animal groups how many animals need to be taken from each weight range group. Then the first group starts taking the animals from the different weight range groups with the lowest random number. When the first group is finished, the second group starts taken animals from the weight range groups. It will start with the weight range group that follows the group where the first exposure group picked its last animal from and so on.
Here is a simple example:
40 animals are divided into 5 weight range groups, means 8 animals per group.
Animal group 1 needs 20 animals.
Animal group 2 needs 10 animals.
Animal group 3 needs 10 animals.
Animal group 1 will take 4 animals from each weight range group.
Animal group 2 will take 2 animals from each weight range group.
Animal group 3 will take 2 animals from each weight range group.
All animals are now randomly assigned to the different exposure groups and they are equally picked from the weight range groups.
- Then the animals are placed in cages so the animals in one cage have the best matching weight.
- Then the cages are randomized within the exposure groups.
- Then the animals are assigned animal number and animal group.
For each group the Mean value, Standard deviation and RSD is calculated. Then we calculate the maximum mean difference between the group. Good practise is to keep this number below 1%. The algoritm will help you on this. If the percetage is to high, then rerun the allocation and it will give a new result. Rerun untill you find a satisfying result. If the result is not satifying, this will mean that the weight of the animals is not suited for the allocation and you will have to remove some animals that have a very high or low weight.