Self-controlled observational study designs, such as the case-crossover design and the self-controlled case series, are reviewed, and their respective rationale, strengths and limitations are compared. Although no single design is generally superior to the others, they share the trait of being robust towards confounders that are stable over time. The self-controlled designs can be particularly useful when using secondary healthcare data for pharmacoepidemiological research and might be useful in screening for adverse drug effects. The main limitations of self-controlled designs are that they are amenable only to transient effects; some may be inefficient with long-term exposure; and they may be sensitive towards trends in exposure.