1: Including a ‘control’ group.
At the heart of all interventional clinical trials is the idea of making a comparison. In most cases, the researchers aim to compare what happens to a group of people who receive a potential new medication (or device or procedure) with a similar group of people who do not.
- The oldest evidence of this sort of comparison was outlined in the Old Testament around 500 BC. Nebuchadnezzar, the King of Babylon, believed that a diet of meat and wine was beneficial to health but, according to the Book of Daniel, he tested this out by giving Daniel and his brothers a diet of vegetables and water for ten days. When he found they were noticeably healthier than his own men, he changed his views about nutrition.
- A Scottish physician called James Lind is credited with performing the first properly recorded clinical trial in 1747. While working as a surgeon on a ship he noticed the high rate of a fatal condition called scurvy among the crew. In those days no-one knew that scurvy is caused by lack of vitamin C. Dr Lind divided the sailors suffering from scurvy into small groups. All were given the same basic diet but each group was given one extra thing to eat or drink. It soon became clear that the men who were given oranges and lemons made a quick recovery, while the others did not.
In modern clinical research, it is essential to be sure that, when participants are given a potential new treatment, any effects on their health are really caused by that treatment. Sometimes people begin to feel better (or worse) due to the natural course of their condition over time. This is why many clinical trials need to include a ‘control group’.
Participants in the control group are not given the investigational medicine but, apart from that, they receive the same care as the other participants. Importantly, no trial participant will ever be at a known disadvantage because, at the start of a trial, researchers don’t yet know if the new treatment will be helpful or not.
2: ‘Blinding’ and ‘placebo’
However, it is often also important that none of the trial participants know which group they are in, as this knowledge can influence how well they do. Keeping this information secret is called ‘blinding’ and it avoids the beliefs and expectations of the participants affecting the results. Because the mind and body are closely connected, some people feel better (and even have better test results) if they think or believe a treatment they are taking will make them better. Even the colour or size of a pill can influence people and affect their progress.
So, to ensure that the trial participants can’t work out which group they are in, those in the control group must receive something that looks like the investigational medicine. This can either be an already approved medication or a ‘placebo’. A placebo is not a real medicine and has no active ingredients but is designed to look exactly the same as the treatment being tested.
- The idea of placebo was first introduced in 1863 when US physician Austin Flint planned a clinical trial that compared a “dummy” treatment to an active one. His trial, which focused on patients with rheumatism, gave some participants a herbal extract instead of the established remedy for rheumatism.
Often the doctors running the trial are also included in the blinding to avoid their own expectations influencing the outcome. In such cases, the trial is described as ‘double blind’. While blinding is an important design concept, not all trials can or should be blinded.
- The first double-blind trial was carried out in Britain in 1943 in an attempt to test a possible treatment for the common cold. This nationwide trial enrolled over a thousand British office and factory workers suffering from colds – quite a challenging mission in wartime.
3: ‘Randomisation’
Another key principle of clinical trial design is about how participants are assigned to their group in the first place. If researchers are free to pick and choose which participants get the potential new treatment, they have a natural tendency to assign people they think will benefit from it most. Allowing personal judgment to dictate this decision would mean that the group receiving the treatment would probably be at an advantage from the start and the results of the trial won’t be trustworthy.
The best way to achieve a fair comparison is to allocate participants to their group at random (like flipping a coin). Nowadays, this process of ‘randomisation’ is done by a computer. People interested in joining the trial are told that there is a chance (often 50%) that they will be put into the control group. It’s important to understand that being in the control group might not turn out to be a disadvantage and many trials allow control group participants to try the real treatment after the main trial is over.
- Randomisation was first introduced in 1948 by Sir Austin Bradford Hill, an English researcher and statistician. In a trial involving patients with the lung disease tuberculosis, Bradford Hill decided whether a patient should be treated with the antibiotic streptomycin plus bed rest, or bed rest alone, by using a table of random numbers. The investigators didn’t know which patient got each treatment; the details were in sealed envelopes.
4: The randomised, controlled, double-blind clinical trial – and beyond
The three key principles of clinical trial design (randomisation, including a control group and blinding of both participants and doctors) have given rise to the concept of the ‘randomised, controlled, double-blind trial’, which is universally seen as the best way to generate high quality, reliable proof that a potential new medicine is effective and safe for use in a particular medical condition.
However, many other research methods are increasingly being used to generate further evidence about the usefulness of new and existing treatments. The current digital age has opened up new ways to gain ‘real world’ evidence, which includes insights on particular healthcare approaches from patients in routine medical practice.
Advances in technology have also introduced the concept of ‘big data’ where information is gathered about very large numbers of people and analysed by artificial intelligence (AI). This is helping to build a more complete picture of the overall value of emerging and current medical treatments.