Randomized Controlled Trial

A randomized controlled trial is an experimental study design in which participants are allocated by chance to receive an intervention or a comparator, and outcomes are then compared between the groups. Random allocation is what distinguishes the trial from observational designs: by distributing both known and unknown characteristics evenly across groups on average, it allows differences in outcome to be attributed to the intervention, making the trial the reference design for evaluating intervention effects.

Definition

A randomized controlled trial is an experiment in which eligible participants are assigned by a chance mechanism to two or more groups receiving different interventions (or an intervention versus control), and outcomes are compared between groups to estimate the effect of the intervention.

Scope

The entry covers random allocation, allocation concealment, blinding, control groups, and the comparison of outcomes between arms, together with the design's main limitations — feasibility, ethics, cost, and generalizability. It treats the randomized controlled trial as a methodological topic within epidemiologic study designs, not as clinical guidance about any specific intervention.

Key concepts

• Random allocation
• Allocation concealment
• Blinding (masking)
• Control or comparison group
• Intention-to-treat analysis
• Confounding control through randomization
• Internal versus external validity

Mechanisms

Eligible participants are assigned to study arms by a chance mechanism, so that, on average and especially in large trials, the groups are balanced on both measured and unmeasured characteristics. This balance — exchangeability — is what lets a difference in outcomes be attributed to the intervention rather than to confounding, the feature that observational designs cannot guarantee. Two procedural safeguards protect this advantage: allocation concealment, which prevents foreknowledge of the next assignment and so prevents selection at entry, and blinding, which keeps participants, caregivers, or assessors unaware of group assignment to limit performance and detection bias. Analysis by intention to treat — keeping participants in their assigned group regardless of what they actually received — preserves the balance that randomization created. The trial's strength is high internal validity; its frequent trade-off is reduced external validity when eligibility criteria and settings are narrow.

Clinical relevance

Randomized trials and their syntheses provide the strongest evidence on whether interventions work, and appraising them is central to evidence-based practice and guideline development. This entry is a reference description of how trials generate and grade that evidence; it does not recommend or evaluate any particular treatment for an individual and is not a basis for clinical decisions.

Epidemiology

Randomized trials are the design of choice for questions about the effect of an intervention when randomization is ethical and feasible, including therapeutic, preventive, and public-health interventions. They are less applicable to harms, rare outcomes, or exposures that cannot be assigned, where observational designs or natural experiments are used instead, and their findings must be interpreted in light of how representative the enrolled population is.

Evidence & guidelines

Reporting of randomized trials is standardized by the CONSORT statement, which specifies the items needed to judge a trial's validity. In evidence hierarchies, well-conducted randomized trials and systematic reviews of them are generally placed at the top for questions of intervention effect, because randomization addresses confounding more directly than any observational design.

History

Controlled experimentation in medicine has older roots, but the modern randomized controlled trial is usually dated to the Medical Research Council's streptomycin trial for pulmonary tuberculosis, designed with Austin Bradford Hill and published in 1948, which Doll later called a watershed. Random allocation spread through clinical research in the following decades, and reporting standards such as CONSORT later codified how trials should be conducted and described.

Debates

How well do trial results generalize beyond the studied population?

Strict eligibility criteria and controlled settings raise internal validity but can limit external validity, so how far a trial's effect estimate applies to everyday populations and practice is a persistent question.

Why do allocation concealment and blinding matter so much?

Inadequate concealment or blinding can reintroduce the selection and detection biases that randomization is meant to prevent, and empirical work links these methodological weaknesses to exaggerated effect estimates.

Key figures

• Austin Bradford Hill
• Richard Doll
• Kenneth Schulz
• David Grimes
• Douglas Altman
• David Moher

Seminal works

• doll-1998-watershed
• schulz-2010-consort
• schulz-grimes-2002-allocation

Frequently asked questions

Why does randomization make a trial stronger than an observational study?

Allocating participants by chance balances both known and unknown characteristics across groups on average, so differences in outcome can be attributed to the intervention rather than to confounding. Observational designs cannot guarantee this balance because exposure is not assigned by the investigator.

What is the difference between allocation concealment and blinding?

Allocation concealment hides the upcoming assignment from those enrolling participants, preventing selection at entry; blinding hides the assigned group from participants, caregivers, or outcome assessors after enrollment, preventing performance and detection bias. They protect different stages of a trial.