Physicians and surgeons who treat bladder cancer must be able to judge the validity and relevance of clinical trials if they are to make appropriate recommendations to their patients. To do so, they must be familiar with the design of clinical trials and with the many problems that may confound their results and limit their validity. The present article has reviewed aspects of the design, analysis, and reporting of clinical trials with particular reference to bladder cancer. Nonrandomized studies are important in establishing the feasibility of new treatments and in providing preliminary information about effectiveness. However, selection of patients, together with many other types of bias, prevents a useful comparison of the results obtained in such trials with those obtained using alternative treatment strategies. Establishment of the value of a new type of treatment will usually require the conduct of a large randomized trial in which patients receive experimental or standard therapy by random assignment. While such trials may influence practice in an appropriate way, they must likewise be reviewed critically to ensure the validity of the results and generalizability to patients seen in a nontrial setting.