About randomised clinical trials

Running reliable clinical trials

A 'clinical trial' is any type of medical research involving people. The most important result we want from clinical trials is that they give us answers we can trust. This is because we will use the results to treat future patients. The way in which a clinical trial is set up can have a great deal of impact on how reliable the results of that trial will be, so what are the features of reliable clinical trials? The following example will help to explain the problem, and the solution.

Imagine a room full of 1000 people with an illness for which there was no presently available treatment but you were going to do a research study on a promising new treatment. For this example we’ll say the treatment comes as a tablet. The obvious way to test the new treatment would be to split the 1000 people in to two groups of 500 people each. One half would take the treatment for a while and the other would not, and at the end of the study you would find out how the treatment worked.

So that people are not influenced by knowledge of which group they are in researchers commonly make up dummy pills (called placebos) that look exactly the same as the ones containing the medicine. They also do other things such as not revealing which patient is in which group until the whole study is finished. This avoids the doctors themselves introducing bias as they may subconsciously assess a person’s response differently depending on whether they know the patient is taking the real treatment or the placebo. These are all techniques that have been developed over the years that make the results of the trial more reliable. But they don’t deal with the first and biggest problem of all – how do you put people into the two groups in the first place in a completely even-handed way?

Bias – the problem

The problem is that in a group of 1000 people some will be older and some younger. There will be males and females, people with mild disease and others with severe disease, some with just this one illness but others with several illnesses, and so on. In short, human beings are all different, and these differences could influence the results of the study if they are allowed to distort the way the study is conducted.

Might you allocate the potentially new treatment more readily to some groups than others?

You may think that you could be completely unbiased in your allocation but studies have repeatedly shown that investigators almost always show some sort of bias when allocating people to treatment groups. The investigators themselves may not be aware of these biases but when the treatment groups are compared it is quite clearly present.

This is a very fundamental problem that can lead to trials having completely unreliable results. In the worst case these studies can show a treatment is beneficial when it is actually harmful!

This problem with clinical trials was known about for a long time before a solution was sorted out. The solution is to allocate the treatments randomly! If this is done properly then the group who get the potential new therapy will be very similar to those who get the dummy or placebo treatment. The 'play of chance' evens out the differences between the two groups. This means that at the end of the study the only difference between the two groups is the treatment that they got and any differences in how the groups fared reflect on how successful or otherwise the new treatment was against no treatment (dummy or placebo treatment).

Randomisation – the solution

Randomising people to the different parts of a clinical trial is therefore one of the key ways in which to achieve reliable results from the research. Nowadays this is all done by computer, which ensures the process is completely random. This is where the term Randomised Clinical Trial (RCT) comes from.

Randomised trials are the "gold standard" design and have paved the way for the introduction of many of the successful treatments that are available in modern medicine.

Without randomised clinical trials we could not reliably decide which treatments work and are safe. We need to do more of them.

Read more on taking part in a clinical trial.