Randomisation

What it is

Randomisation is assigning trial participants to study groups by chance to reduce bias; in a 1:1 trial, each person has an equal allocation ratio. In clinical research, this means a participant is placed into one of the study arms, such as standard treatment versus a new treatment, using a preplanned random process rather than clinician or patient choice. The main purpose is to make the groups comparable at the start of the study so that differences seen later are more likely to be due to the intervention and not to pre-existing differences.

Randomisation is a core feature of randomised controlled trials, including many cancer drug, surgery, radiation, and supportive-care studies. It helps balance both known and unknown prognostic factors across groups, especially in larger trials. This does not guarantee that every characteristic will be perfectly equal, but it reduces systematic selection bias. A related but separate concept is allocation concealment, which means the person enrolling participants cannot predict or manipulate the next assignment.

Common approaches include:

Method	What it does	Typical use
Simple randomisation	Assigns each participant purely by chance	Larger trials
Block randomisation	Keeps group sizes similar at intervals	Small to medium trials
Stratified randomisation	Balances key factors such as stage or centre	Multicentre cancer trials
Minimisation or dynamic allocation	Balances several factors adaptively	Trials needing close balance

How it works

A proper randomisation process has two parts: generating an unpredictable sequence and implementing that sequence with concealment. The sequence may be created by a computer programme, random-number table, or central web or telephone system. Methods such as alternating patients, using date of birth, hospital number, or day of clinic visit are not true randomisation because they can be predicted.

In a cancer trial, investigators may stratify by factors that strongly affect outcomes, such as disease stage, treatment centre, or biomarker status. For example, a study may randomise patients 1:1 within strata so that each arm has a similar mix of early-stage and advanced-stage disease. Block randomisation may be used so that after every block, the numbers in each arm stay reasonably close.

Randomisation works best when combined with other trial safeguards:

1. Allocation concealment prevents foreknowledge of the next assignment.
2. Blinding reduces bias after assignment in outcome assessment or care.
3. Intention-to-treat analysis keeps participants in their assigned groups for analysis, preserving the benefit of randomisation.

These steps matter because randomisation alone cannot fix poor trial conduct, missing data, or biased outcome measurement.

Evidence and uses

Randomisation is used when researchers want the most reliable comparison between interventions. In oncology, it is common in phase 2 and phase 3 trials that compare a new drug, combination, schedule, or supportive-care strategy against standard care or placebo where appropriate. Regulators and guideline groups generally place randomised trials above observational studies when judging treatment benefit and harm.

The main value of randomisation is that it reduces confounding. In non-randomised studies, patients who receive one treatment may differ from those who receive another in ways that affect survival, toxicity, or quality of life. Random allocation lowers this problem by distributing prognostic factors by chance. This is why randomised controlled trials are often considered the reference standard for testing efficacy.

Still, randomisation has limits. Small trials can remain imbalanced by chance alone. Some interventions cannot be blinded easily, such as surgery or radiation techniques. Crossover, treatment non-adherence, and loss to follow-up can weaken the practical benefit of random assignment. For rare cancers or highly personalised treatments, randomised trials may also be difficult to run.

In India, the same principles apply in trials conducted in academic centres, government hospitals, and private cancer centres. Ethics committees, trial protocols, and registry reporting should describe how randomisation was generated and concealed. When reading a trial report, patients and families should look for details such as the allocation ratio, method used, and whether the study says "randomised" without explaining the process.

When to see a clinician

If you are being invited to join a clinical trial, ask the research team to explain whether the study is randomised and what that means for your care. It is reasonable to ask:

• What are the study groups?
• What is the allocation ratio, such as 1:1 or 2:1?
• Will I definitely receive active treatment, or could I receive standard care or placebo?
• Is the trial blinded?
• Can I leave the trial later if I change my mind?

For cancer patients, randomisation does not mean care is arbitrary or lower quality. Trials are reviewed by ethics committees, and each study arm should be considered acceptable based on current evidence. If you are unsure whether a randomised trial is right for you, discuss the expected benefits, risks, alternatives, costs, travel burden, and how the trial fits with your treatment goals.

Limitations and open questions

Randomisation reduces bias, but it does not make a study automatically trustworthy. A trial can still be misleading if allocation was not concealed, if many participants dropped out, if outcomes were changed after the study started, or if the sample size was too small. Published reports sometimes use the word "randomised" without enough detail to judge quality.

Another limitation is generalisability. Even a well-randomised trial may include selected patients who are younger, fitter, or treated at specialist centres, so results may not fully reflect routine practice. This is relevant in cancer care, where comorbidities, affordability, and access to treatment can differ across settings, including across India.

Newer trial designs also raise practical questions. Adaptive randomisation and platform trials can improve efficiency, but they are more complex to design, analyse, and explain to patients. Evidence in methods research supports these approaches in some settings, but they require careful planning to avoid operational bias and misinterpretation.

For readers assessing a study, the key question is not only whether it was randomised, but whether the randomisation was truly unpredictable, concealed, and preserved through analysis. Those details often determine how much confidence clinicians can place in the results.