The study was designed to compare antibiotic use, cost and consumption before and after an initiation of an antibiotic-restriction policy in our hospital. The policy was applied in 2003, and the prescription of two groups of antibiotics (intravenously used and expensive antibiotics) was restricted. A prescription for the restricted antibiotics could be obtained with approval by an infectious disease specialist (IDS). All the hospitalized patients who received antibiotics were evaluated by a cross-sectional study with standard criteria. The annual cost and consumption of antibiotics were evaluated. After restriction, the rate of antibiotic use decreased from 52.7 to 36.7% (P < 0.001), and the appropriate use increased from 55.5 to 66.4% (P < 0.05). Appropriate use was higher for restricted antibiotics (88.4%) than for unrestricted ones (58.2%) (P < 0.001), and higher in the presence of ID consultation (97.5%) than in the absence of consultation (55.7%) (P < 0.001). Culture-based treatment was increased, and appropriate use in such cases (93.0%) was higher than empirical treatment (33.3%) (P < 0.001). After the restriction policy, consumption of antibiotics belonging to the restricted groups was decreased by 44.8%. Total expenditure of all antibiotics was decreased by 18.5%, and the savings were US$332,000 per year. This restriction policy was effective in promoting rational antibiotic prescription and lowering antibiotic cost and consumption in our hospital.