Despite the large availability of diagnostic tests and procedures and their importance to medicine, evidence for their appropriate use is often limited. Unfortunately, the evaluation of tests is difficult and susceptible to several forms of bias. This paper describes the standard methods for characterizing and comparing the accuracy of diagnostic tests and addresses initiatives such as STARD and QUADAS to improve the quality of diagnostic studies. We inform the reader how to critically appraise study results and when to rely on results of diagnostic accuracy studies or randomized clinical trials.