Controlled clinical trials in cancer research.

Knowledge of important aspects of the design and analysis of clinical trials is essential to clinical researchers and readers of medical literature. A brief description of proper trial design, including the contents of a trial protocol, as well as different strategies to avoid bias, is given. The concept of p-values is explained, and some commonly used statistical analysis methods are mentioned. Statistical power is defined, and two useful formulas and examples of estimating sample size are presented. The correct interpretation of trial results is emphasized, and misinterpretations and errors that frequently occur are dealt with. Various issues regarding multiple significance testing, such as interim analyses, multiple endpoints, and subgroup analyses, are addressed.