ABSTRACT. Successful pharmacotherapy is dependent on a long chain of considerations, decisions and actions. Some routines in the prescription, recording and administration procedures for drugs were found to be inadequate in our hospital. New procedures were developed, based on the following principles: 1) Drug prescriptions, written and signed by the doctors/nurses, included the indication(s) for starting or stopping the treatment. 2) No transcriptions of drug prescriptions were allowed. 3) The medication times were adapted to pharmacokinetic principles. 4) All doses taken by the patients were recorded and countersigned by a nurse. The system was tested during a three‐month period in two medical wards. The 274 patients obtained 1653 prescriptions, 83% by physicians. Of the prescriptions, 7% were never administered. One half of the patients did not take all of the prescribed drug therapy. This accounted for an average of one‐tenth of the prescribed doses. An average number of 5.1 “drug exposures” were prescribed and 4.7 given. These figures are lower than those reported from other countries. Corresponding prescribing errors of the physicians were 4 and 42% with the test and the old system, respectively. We conclude that the principles of the test system allow increased safety and accuracy in hospital drug handling. The recording of drug intake makes the evaluation of pharmacotherapy more reliable. A thorough system of drug handling can be used in the education of patients and students.