OBJECTIVES: To estimate the long-term health and economic impact of treating patients with Alzheimer disease with galantamine (Reminyl) in different countries. METHODS: A pharmacoeconomic model, The Assessment of Health Economics of Alzheimer's Disease (AHEAD), was used to predict the time until Alzheimer's disease patients require full-time care and the associated costs. Full-time care was the consistent requirement for a significant amount of care giving and supervision each day. Efficacy data were obtained from three clinical trials comparing galantamine with placebo. For each country, local data were obtained on service use, balance of care between community and institutions, and relevant unit costs. Analyses were completed for The Netherlands, Sweden, Finland, Germany, UK, Canada and New Zealand. Forecasts were made for up to ten years. Costs are reported in 2001 currencies and determined from a perspective somewhat broader than that of a comprehensive payer, including the cost to a national health service as well as other relevant stakeholders such as providers of social care services. Both health benefits and costs were discounted at 3%. Sensitivity analyses were carried out on key input parameters and combinations of these parameters. RESULTS: In each country, full-time care was estimated to account for at least two thirds of the cost of caring for patients over ten years, and more than 60% of this cost was from providing institutional care. Galantamine is predicted to reduce the duration of full-time care by almost 12%. Approximately 5 patients need to be placed on treatment to avoid one year of full-time care, resulting in incremental savings in the majority of countries. CONCLUSIONS: Delaying the time to full-time care is expected to produce savings in the majority of health care systems. Galantamine was considered to be an economically dominant strategy compared to no treatment in most health care systems.