For the purpose of prevention of hospital-acquired infection caused by antibiotic-resistant bacteria, we examined a method to establish an appropriate time period for the administration of antibiotics to compromised hosts. Using these antibiotics we monitored patients who received instruction about the drug regimen in the Blood and Respiratory Diseases Department ward. We monitored a) third-generation cephalospolins, b) Imipenem/Cilastatin, and c) antibiotics used against methicillin-resistant Staphylococcus aureus. When the antibiotics were administered over 14 days, pharmacists notified physicians of the current duration of administration using a confirmation form, and confirmed their future administration schedule. We examined the antibiotic usage regimen of all the patients in this ward before and after the confirmation form was adopted. Patients given the same antibiotics within 14 days significantly increased in percentage from 82% to 91% after the confirmation form was adopted (p < 0.05). The median duration of antibiotic administration decreased from 7 days to 5 days. The case with antibiotic administration for the longest duration was a patient with leukemia who received vancomycin for 116 days after adoption of the confirmation form. This patient died 4 days after his antibiotic was changed. Only 16% of the patients administered antibiotics in this ward were monitored for the duration of antibiotic administration after adoption of the confirmation form. When the pharmacists positively provided physicians with information on some patients concerning the prolongation of antibiotics administration, the number of patients administered antibiotics for less than 14 days significantly increased throughout this ward without interfering with the treatment of patients who required long-term administration of antibiotics.