Virtually all recognized adverse drug reactions result in discontinuation or dose reduction of the suspected drug. In a study of adverse drug reactions, a pharmacist‐based surveillance system was used to monitor changes in drug orders. Drug reactions prompted 7.3 per cent of drug discontinuations or dose reductions. Compared with a physician‐based system of daily monitoring, the pharmacist‐based system detected more drug reactions, provided a more uniform method of study, and had greater acceptance by house physicians. Among 939 patients studied on the medical service of a university hospital, 48 (5.1 per cent) had adverse drug reactions on admission, and 99 (10.5 per cent) developed adverse drug reactions in the hospital.