Why is randomisation needed?
Comparisons of treatments are unfair if the patients in different treatment groups are not similar. The treatment groups must be composed of people who are similar to each other across all the groups – it’s no good if one treatment group is mostly full of, say, older people and another is mostly full of younger people, because then any comparison of the different treatments will not be fair. If doctors and patients choose treatments themselves, they can introduce biases – for example, older, sicker patients might choose one treatment and younger, healthier patients might choose another, leading to unfair comparisons. This can lead to dangerous and misleading results. For more information on randomisation and why it is needed, see “Testing Treatments”.
