A study population with either a low or high exposure prevalence usually results in a less efficient study, compared with a study on a population with an intermediate exposure prevalence. We discuss a strategy for improving efficiency when exposure is rare: one can include in the study base, for each individual, only follow-up corresponding to a potential exposure-associated increase in risk. This strategy is achieved by choosing so-called pre inclusion and preexclusion periods, which, in turn, follow from relevant induction time assumptions.