When new treatments for patients with rheumatic diseases become available, how are physicians to determine whether these therapies are efficacious? In initial trials of a drug or therapy, the treatment may be given to patients to determine its tolerability and safety; these studies will likely be open trials and will not use comparison or control groups of untreated or placebo-treated subjects. Subsequently, the therapy will need to be compared with either placebo and/or the standard therapy to determine its effectiveness. This type of clinical trial will involve the identification of appropriate comparison or control groups and the selection of valid study design. In this article, we will review 1) the types of comparison groups that can be used to assess the effectiveness of a treatment; 2) elements in the design of prospective, controlled clinical trials, including concerns regarding comparability of the treatment groups and generalizability of the study results to different patient populations; 3) differences between parallel and crossover trials; 4) current concepts of outcome measures; and 5 ) issues regarding the statistical analyses used in these trials, including inferences based on significance testing. The reader is also referred to other reviews (1,2) that have addressed similar issues in relation to rheumatic disease.