We conducted a survey of the prescription of antibiotics among the outpatient clinics of the Instituto Nacional de la Nutrición "Salvador Zubirán", a third level hospital in México City. We made an auditory of the medical record and prescriptions given to every patient treated for an infectious episode, accounting for at least six questions to evaluate the quality of the prescription: 1) if the patient should had received antibiotics; 2) if the antibiotic prescribed was adequate; 3) if the dose was sufficient; 4) if the frequency of administration was correct; 5) if the route of administration was adequate; and 6) if the length of the treatment was sufficient. We validated the concordance among two evaluators and found that it was 89% for the whole questionnaire. In the evaluation we found that the patients should had received antibiotic 94% of the time: the antibiotic selected was a right choice 80% of the time; the dose was adequate 45% of the time; the frequency of administration was adequate 70% of the time; the route of administration was adequate 79% of the time; and the length of treatment was adequate 38% of the time. The worst findings were seen in two of the most important issues of antibiotic prescription which directly affect the appearance of antibiotic resistance.