A new study suggests scientists should change their methods and test mice in deliberately varying environmental conditions. Just as no two humans are the same, treating mice more as individuals in laboratory testing could cut down on erroneous results.
Researcher Joseph Garner said that human testing uses a broad range of subjects, giving scientists an idea of how a drug or treatment might affect different types of people. But scientists often use mice that are basically genetically identical and try to limit internal and external environmental factors such as stress, diet and age to eliminate variables affecting the outcome. When mouse testing creates a false positive, leading a researcher to believe a drug has worked, the drug could be sent to further animal testing and human clinical trials.
Garner said there is no practical way to ensure that all environmental conditions are the same with mice, however, because they respond to cues humans cannot detect. For example, a researcher's odor in one lab might cause more stress for a mouse than another researcher's odor in a second lab with different mice, giving different results. But scientists, unaware of the odor difference, may believe a treatment worked when the mice were actually responding to an environmental cue, giving a false positive.
The study used three different strains of mice from previously published data and compared their behavioural characteristics against each other. The observations were done in three different labs, two different types of cages and at three different times to make 18 different replicates of the same experiment. Traditional testing theories say the results should have been the same in all those experiments. Once the results were compared, however, the researchers found many false positives, or instances when one strain appeared to act differently from another when it actually should not.
The researchers suspected the problem was in the traditional lab experiment design. So they re-evaluated the data, picking a mouse of each strain from each environment - similar to matching pairs in human clinical trials - and found only the same number of false positives as would be expected by chance.
MEDICA.de; Source: Purdue University