The Use of Randomisation in Clinical Trials
Randomisation Controlled Trials (RCT) is the process of randomly assigning participants to groups that receive different treatments, which ensures that each individual has an equal chance of receiving any of the treatments used in the clinical trial. It allows the trial staff to make unbiased and informed decisions about new treatments, and is usually considered the best way to evaluate the efficacy of a new treatment and the effectiveness of interventions.
Methods of randomisation include using a table of random numbers or a computer program which generates random numbers. Other methods like assignment by date of birth or alternating assignment can be more prone to bias.
Benefits of Randomisation
- Eliminates selection bias (when the researcher decides who is going to receive which medication).
- Ensures that all participants are treated with the same respect and that there is no bias.
- Allows one treatment to be directly compared to another to discover its effectiveness.
- Considered the ‘gold standard’ of clinical trials (makes the results more publishable).
- Any difference in the outcome of the clinical trial can be attributed to the treatment itself, and no other factors.
Disadvantages of Randomisation
- Results may not always reflect a real life situation.
- Participants can’t be ethically randomised unless all treatments have the support of medical professionals.
- Trials which test for effectiveness are typically more expensive.
Different Types of Randomisation
Simple randomisation is based on a single sequence of random assignments. Participants are allocated to groups by flipping a coin, throwing dice or shuffling cards. Simple Randomisation is the most simple and easy to implement randomisation style. It minimises any bias by eliminating predictability and allows participants to maintain their independence during the clinical trial. It is especially useful to use in larger clinical trials as most often it results in an equal number of participants for each group.
Problems with Simple Randomisation
- It doesn’t work as well with smaller clinical trials (the group sizes won’t always be as balanced).
- If the recruitment of participants is ongoing, it can lead to inaccurate results between the treatment groups as different participants are being enrolled over time.
Designed to assign participants into groups that result in equal numbers. The block size is decided by the Investigator and is often determined by multiplying the number of groups (if there are two treatment groups, there should be 4, 6 or 8 blocks). Blocks are kept on the smaller side with predetermined group assignments which helps to keep the treatment groups a similar size throughout the clinical trial. After the block sizes have been decided, the equal group sizes from each block must be determined. Blocks are then randomly selected to sort the participants into groups.
Problems with Block Randomisation
- Selection bias is more common as the researcher can easily predict the treatment allocation of the group.
Separating participants into groups depending on different factors (gender). If sex is the chosen factor, then the number of strata is two and randomisation is applied to each stratum. This form of randomisation can reduce the imbalances in characteristics found in treatment groups and increase statistical power. This method makes it easier to generate block randomisation lists for different prognostic factors.
Problems with Stratified Randomisation
- While the aim is to remove selection bias, it also means that the groups won’t always have the same important characteristics.
Occurs when the allocation of participants is changed according to the progress of the clinical trial. It can be used to minimise the imbalance between treatment groups and can adjust the treatment groups depending on the effect the treatment has on participants. Prognostic factors come into play and are used to define the participants and influence treatment/the course of the clinical trial.
Minimisation is used as a tool to balance out the prognostic factors in a clinical trial. The first participant is allocated to a treatment group using simple randomisation and the rest of the participants are then assigned to balance out the prognostic factors, based on previous participants and their placement in the clinical trial. Designed to overcome the challenges of stratified randomisation.
Problems with Minimisation
- It doesn’t meet all the requirements of randomisation.
The allocation of participants into treatment groups is kept secret until the moment of assignment. This helps to reduce selection bias and prevents researchers from influencing the assignment of participants into certain groups.
Ethics of Randomisation
- There are ongoing questions about whether the treatment method for participants should be determined by probability instead of medical professionals.
- Possible risks of a clinical trial.
- Incomplete information provided (participants not shown the full picture).