Randomised Trials

#Why allocate by chance?

A randomised trial assigns participants to study conditions through a process that includes chance. Successful randomisation prevents known or unknown prognostic factors from determining the assignment. On average, this gives groups the same prognosis at the start, but it does not promise identical groups in a particular trial.

The allocation sequence also needs protection before and until assignment. If a recruiter knows the next assignment, selective enrolment can distort the comparison. Cochrane calls this allocation sequence concealment.

Evidence: Cochrane Handbook, Chapter 8: risk of bias in a randomized trial

#Concealment is different from blinding

Allocation concealment protects the assignment process. Blinding aims to prevent bias after assignment by keeping relevant people unaware of the assigned intervention. A study can be randomised without being blinded. Some comparisons, such as surgical versus non-surgical care, cannot conceal the intervention from everyone.

Ask which people were blinded and how outcomes were assessed, rather than relying on a broad label such as "double blind". Awareness can affect care, behavior or assessment, but its importance depends on the outcome and circumstances.

Evidence: Cochrane Handbook, Chapter 8: risk of bias in a randomized trial

#Read the analysis and missing outcomes

An intention-to-treat analysis aims to estimate the effect of assignment, including participants in their originally assigned groups regardless of whether they received treatment as intended. Cochrane warns that the term is used inconsistently, so read who was actually included.

Excluding people with missing outcomes can still introduce bias. A label saying "intention to treat" does not solve this problem by itself. The report should explain missing data and how they were handled, as CONSORT 2025 recommends.

Evidence: Cochrane Handbook, Chapter 8: risk of bias in a randomized trial / CONSORT 2025 explanation and elaboration

#Read a trial as a bounded comparison

A trial compares specified interventions under specified conditions. Read the protocol, comparator, outcomes, harms and follow-up before applying its result elsewhere. Complete reporting helps readers judge reliability; a reporting checklist is not a universal quality score.

Useful questions are: what comparison was tested, which outcomes were planned, who contributed to the analysis, and which limitations could change the interpretation? These questions support discussion with a clinician, not a treatment decision from this page.

Evidence: CONSORT 2025 explanation and elaboration / Cochrane Handbook, Chapter 8: risk of bias in a randomized trial

Source note

The sections above were checked against the linked sources. No clinical review has been performed. This is general research education, not a clinical guideline.