The First International Pharmacoeconomic Conference on Alzheimer's Disease (AD) was held in Amsterdam in July 1998. The meeting was held under the auspices of the International Working Group for Harmonization of Dementia Drug Guidelines (http://dementia.ion.ucl.ac.uk/harmon), bringing together academics, clinicians, purchasers, and representatives from industry. Presentations were given on the methodology of pharmacoeconomic studies in AD, particularly focusing on caregiver burden, quality of life (QOL), and resource utilization. Three economic models of AD were presented based on data from the United States, Canada, and the United Kingdom. In two studies, these data were then used to model the cost-effectiveness and effect on cost of treatment with donepezil. Both studies suggested a possible cost advantage for the use of donepezil, when compared with no placebo or treatment, particularly when donepezil is used appropriately in mild-to-moderate AD. These data need to be interpreted with care, as none of the cost or utility information were collected during the clinical trials. Additional data from a 2-year clinical trial of selegiline and vitamin E suggest that cognitive measures may be poor predictors of economic outcome, which is better measured directly. Both economic models of donepezil rely on short-term cognitive data to predict long-term outcome, a methodf that may not be useful in predicting economic savings. The issues facing pharmacoeconomists, researchers, clinicians, and families in the future were addressed in a series of workshops using a method of strategic futuring. The workshops attempted to see 7 years into the future for a range of areas, including consumer and caregiver use of pharmacoeconomic data; early detection and prevention; Japanese perspectives; activities of daily life and what will be daily life activities; caregiver burden; QOL at the end of life; new uses for new information and communication technology in clinical research; and physicians' use of pharmacoeconomic data. A range of exciting futures were predicted, although common themes that arose when considering barriers to achieving these futures included cost, education, political will, confidentiality, privacy, and ethics. The first conference was deemed to have been a success, having attracted more than 160 delegates and many distinguished speaker. A second conference is planned for the year 2000. Over the next 2 years, research needs to be broadened particularly in the methodological areas of resource utilization, QOL, and caregiver burden. Data from clinical trials with relevant economic and QOL outcomes will be needed by purchasers if drug treatments for dementia are to gain widespread use. It is also hoped that the models described at the meeting may become more freely available to politicians, purchasers, clinicians, and caregivers to help them make better decisions about treatment.