When monitoring spontaneous reports of adverse reactions to registered drugs, it is important to detect any change in the number of reported adverse reactions in the course of time. Sales adjusted adverse drug reaction rates are usually compared in order to be able to take drug exposure into account. Here we review the so-called arithmetic and some statistical procedures which could form the basis for an alert system. The advantages and disadvantages of each of these methods are discussed. The importance of data requirements and the problems which arise when using an alert system are pointed out and then clarified with the help of the example of diphtheria/tetanus vaccine.