Some main problems eroding the credibility and relevance of randomized trials.

Randomized trials are an excellent research design with major advantages. However, randomized trials are not immune to biases, and inferences from them may be sometimes flawed or irrelevant. The present review addresses, in brief, some of the major threats to the credibility and relevance of the results of clinical trials: power problems, biases affecting internal validity (poor design, conduct, and analysis), biases affecting the total randomized evidence on a specific topic (publication bias and selective outcome and analysis reporting bias), lack of relevance, poor generalizability, and biases in the interpretation of the results.