A study of drug usage was conducted in five hospitals: two hospitals for patients with acute diseases, two for patients requiring long-term care, and one pediatric hospital. Partially due to the increased length of hospitalization, more drugs were used in long-term hospitals. The number of different drugs used to treat the same problem varied greatly. For some therapeutic indications, one particular drug was widely accepted by most hospitals; in others, there was a wide variance. More than one drug of the same pharmacological action was often prescribed for the same patient. Older patients and women received more drugs. Antibiotics were used in a high percentage of all hospital patients. There were many incomplete prescriptions and discrepancies between drug orders written by doctors and records of medications administered by nurses. These results indicate the need for an epidemiologic approach to therapeutics.