Gastroenterology lags behind other specialties such as cardiology in the quality of its evidence base for clinical practice. One area where this is particularly evident is in cancer prevention, despite developments in chemoprevention strategies for high-risk patients. For chemoprevention strategies to be successful, we need appropriate clinical networks and translational science infrastructures, model chemoprevention agents and multiple, large, flexible and randomized clinical trials. Translational science must also be embedded into large-scale, long-term, randomized clinical trials that have hard endpoints, so that irrefutable evidence of the longevity of treatment efficacy can be gathered. We also need to be able to identify an individual's cancer risk using valid global patient populations, so that medical benefits can be applied to all, regardless of ethnicity, sex, economic status, age and comorbidities. The future success of gastrointestinal chemoprevention relies on fostering a closer link between basic pharmaceutical research and clinical applications, in a 'bench to bedside and back' manner. In this review we systematically assess the evidence for various cancer prevention strategies, especially chemoprevention, and highlight the obstacles to further exploitation of this knowledge base.