Based on the instructions of the National Organization of Pharmaceutical Agents (Greece) from July 1, 2003, quinolones, 3( rd )and 4(th )generation cephalosporins, carbapenems, monobactams, glycopeptides, oxazolidinones, and streptogramins were considered as "restricted" antibiotics that could be used only with the approval of an Infectious Disease specialist. We analyzed the effect of the policy on the consumption and cost of antibiotics as a group and of specific classes, adjusted for the patient load, as well as on the antimicrobial resistance of isolated bacteria. We analyzed 5 trimesters (2 prior and 3 after the implementation of the new policy). A 20% and 16% reduction in adjusted consumption [in daily defined doses (DDDs)] and cost, respectively, of the restricted antibiotics was accomplished during the first trimester after implementation of the new policy. However, this was accompanied by a 36% and 56% increase in adjusted consumption and cost, respectively, of unrestricted antibiotics. A logistic regression model that we performed showed that the new policy had an independent positive effect on the in vitro antimicrobial susceptibility of Pseudomonas aeruginosa (p=0.051) but not of Acinetobacter baumannii and Escherichia coli isolates. Our data suggest that there are considerable limitations to the programs aiming to reduce the consumption of restricted antibiotics through the approval of their use by specialists, at least in some settings.