Effective management of drug therapy in the elderly is a challenge for primary-care physicians. There are 20,400 drugs approved for marketing in Canada. Most elderly patients will fill 33 prescriptions per year and take 5 different medications. To be a safe prescriber in the 1990s, physicians need to be aware that 33,000 drug interactions, 6,500 drug-disease contraindications, and 3,500 drug-allergy contraindications have been documented. Inappropriate prescribing is a problem in the elderly. At least one inappropriate prescription is given to 12% to 46% of seniors, and 25% of drug-related hospital admissions are due to prescribing errors. Half of all physicians will write at least one inappropriate prescription for an elderly patient each year, and one quarter of inappropriate prescriptions will be created by the presence of multiple prescribing physicians. Academic detailing is the most effective approach to improve physician prescribing. However, it is an expensive intervention that must be limited to a small number of drugs and conditions, and it must be continued to retain its effectiveness. Furthermore, it fails to address the problems created by multiple prescribers. In this project, we developed a prototype of the future office practice. Physicians are equipped with personal computers and expert prescribing-system software. This electronic academic detailer reviews all current medications for a patient, identifies therapeutic duplications, generates alerts for 50 prescribing problems that have been identified as clinically relevant by a Canadian expert panel, suggests suitable alternatives, and reviews all new prescriptions for potential problems. Information on all prescriptions received by the physicians' elderly patients is downloaded weekly from the provincial prescription claims database, so that the primary physician is able to coordinate and manage all drugs prescribed to their patients by all physicians. The effectiveness of this intervention is being evaluated in a randomized controlled trial of 110 physicians and approximately 16,000 elderly patients in Montreal. We will test whether the intervention reduces the rate of inappropriate prescribing, as well as the rate of drug-related injuries and hospitalizations among patients treated by physicians in the experimental group.