Data Envelopment Analysis (DEA) has been widely studied in the literature since its inception with Charnes, Cooper and Rhodes work in 1978. The methodology behind the classical DEA method is to determine how much improvements in the outputs (inputs) dimensions is necessary in order to render them efficient. One of the underlying assumptions of this methodology is that the units consume and produce real valued data. This paper deals with the extension of this methodology for the case of integer-valued data. Based on an additive DEA model, a mixed integer linear programming model is proposed for setting integer-valued targets. An empirical example illustrates the approach.