Randomised Trials

#Why allocate by chance?

A randomised trial assigns participants, or sometimes whole groups, to different study conditions using a chance-based process. The aim is to make the groups comparable before the intervention begins. This reduces the influence of both measured and unmeasured factors that could otherwise distort the comparison.

Randomisation does not guarantee perfectly matched groups, especially in small studies. The allocation process also needs protection so that recruiters cannot predict or influence the next assignment. This protection, called allocation concealment, is different from keeping participants or researchers unaware of assignments after allocation.

#What is the comparison?

A comparison group may receive another active intervention, usual care, a placebo or no additional intervention, depending on the question and ethical requirements. The result describes a comparison between those particular options. Showing benefit against one comparator does not establish benefit against every available alternative.

Blinding means keeping some people unaware of which intervention was assigned. Where feasible, it can reduce differences in expectations, care or outcome assessment. Some interventions cannot be blinded, such as many surgical or behavioural approaches. Researchers can still use safeguards, including independent outcome assessment and consistent follow-up.

#What can the results establish?

An intention-to-treat analysis compares people according to their original assignment, even if they did not fully follow it. This helps preserve the advantages of randomisation and estimates the effect of assignment under trial conditions. Missing outcomes can still undermine that comparison and require careful handling.

Well-conducted trials can provide strong evidence about intervention effects, but their conclusions have boundaries. Duration, participant selection, adherence and outcome choice all matter. A trial may be too small to detect uncommon harms, and findings may not transfer directly to populations or settings that were not studied.

#Common misunderstandings

“Randomised” describes how participants are assigned to groups, not how carefully a study is planned. Allocation by chance helps reduce systematic differences between groups, but it does not guarantee that the groups will be identical, especially in small trials.

A comparison group does not always receive a placebo or no treatment. It may receive usual care or another active treatment. Randomisation is also different from blinding: participants and researchers may know which intervention was assigned, even when allocation was random.

A positive result does not mean everyone benefited, or that the benefit was large enough to matter in everyday life. Likewise, a result showing no clear difference does not necessarily prove that treatments are equivalent; the trial may have been too small to detect an important difference.

Randomised trials can provide strong evidence about cause and effect, but missing results, treatment switching and selective reporting can still weaken conclusions.

#Questions worth asking a clinician

• How were participants randomly assigned, and were there important differences between the groups at the start despite randomisation?
• How was allocation concealed before assignment, and who was blinded afterward to reduce bias?
• Why was this comparator chosen, and how does it compare with the care I would otherwise receive?
• Do the chosen outcomes measure benefits and harms that matter to patients, rather than only changes in tests?
• How much outcome data was missing, could this change the findings, and was follow-up long enough to detect lasting benefits and delayed harms?