Design, operation, and interpretation of clinical trials.

Randomized controlled clinical trials offer the best evidence for changing clinical practice and informing public health policy. Using examples from the literature, this paper reviews clinical trials for those who may be unfamiliar with their design, operation, and interpretation. In the design of a clinical trial, the question to be answered and a clinically meaningful outcome must be clearly defined. Ethics must be considered, sample size carefully estimated, and use of biomarkers and surrogate outcomes understood. Prominent issues in trial implementation include developing a manual of operations, trial registration, subject recruitment and retention, use of a data coordinating center, and data and safety monitoring. Interpretation of clinical trials requires understanding differences between efficacy and effectiveness; superiority, equivalence, and non-inferiority; intent-to-treat; primary and secondary analyses; and limitations of unregistered small clinical trials compared with large multi-center Phase III trials that are more likely to be representative of a population and change clinical practice or public health policy.