The clinical development of cancer chemoprevention agents is similar to that of cytotoxic agents. On the basis of promising preclinical studies, agents with potential efficacy are introduced to the clinic as a phase I trial to study the dose-toxicity relationship of the agent and its human pharmacology. If doses projected to have biological activity have acceptable side effects, the agent proceeds to a phase II trial, which enrolls a larger number of participants and administers the agent for a longer time period (up to 5 years). Phase IIb trials test the effect of these agents on potential biomarkers of carcinogenesis to get some indication if the agent has activity. The phase II trial also pilots study design, mechanisms of recruitment, and adherence for future phase III trials. The phase III trial of a chemopreventive agent determines its effect on cancer incidence. Close attention is also paid to long-term side effects. The unique features of cancer prevention agents and their evaluation must be considered in this stepwise clinical evaluation. These features include (i) populations recruited to prevention trials are healthy; (ii) in this population there is a low tolerance for side effects; and (iii) the clinical end point of the trial, cancer, is statistically an uncommon event. The evaluation of beta-carotene and retinol as studied in the Carotene and Retinol Efficacy Trial (CARET) is used to illustrate this stepwise clinical development.