Randomized controlled trial

A randomized controlled trial (RCT) tests an intervention by using a chance mechanism, such as a random number generator, to assign participants to receive the treatment or a comparison such as placebo or usual care. Because assignment is random, the groups tend to start out similar in both measured and unmeasured characteristics, which lets researchers attribute outcome differences to the intervention rather than to pre-existing group differences. When reading an RCT, check that randomization was genuine, that the baseline table shows balanced groups, and how many participants were lost to follow-up, since heavy dropout can undo randomization's benefits. Also note whether it was blinded and analyzed by intention-to-treat. Example: to learn whether a drug lowers heart attacks, an RCT randomizes thousands to drug or placebo and compares event rates, isolating the drug's effect. Ask whether the trial was large enough and long enough to detect the outcome that matters.