Prospective cohort studies are costly and time consuming yet appear to be the best means for understanding how genes interact with environmental risk factors to cause disease. This information is a necessary prerequisite for evidenced-based disease prevention, yet not all researchers agree about the importance of studying the interplay between genes and environments. They argue that we already know enough about which environmental 'exposures' can prevent most common diseases, for example, wholesome diet, adequate housing/income and access to good healthcare. Implicit is the notion that current disease categories (i.e. phenotypes) are 'real' and represent homogenous entities, and that identifying environmentally mediated risk is relatively straightforward. Other concerns relate to scientific basis, utility and ethics. These arguments are critically examined for a range of disorders, from diabetes, cancer and inflammatory bowel disease to depression. We refute the contention that incorporating the measurement of genotype into longitudinal-epidemiological studies is wasteful or unlikely to yield significant benefits.