PIP: Observation was used to measure the extent of errors in prescribing and administering drugs in a medical ward of the London Hospital. A working-party with medical, nursing, pharmaceutical, work study, and administrative representation was established to examine the problem. An error incidence of 15.3% was found. Variations on the established system were introduced to determine if this incident rate could be reduced. 4 experiments were conducted to test the more important of these improvements. In the 1st experiment a pharmacist went to the ward each morning and inspected each drug sheet. Where new prescriptions required dispensing, the drug label was written in the ward and taken to the pharmacy. The pharmacy service was withdrawn in the 2nd experiment, and a redesigned drug sheet was introduced. The 3rd experiment tested the effects of the items in the 2nd experiment with the addition of a locking trolley. A combined drug-prescription and recording sheet were introduced in the 4th experiment to reduce clerical work on the part of nurses by eliminating the duplication of data onto recording documents and to provide doctor, nurse, and pharmacist with a feedback of information at the bedside. The introduction of a redesigned drug sheet along with other modifications appeared to be associated with a reduction of errors in prescribing and administering drugs to 4.2% of the drugs given.