OBJECTIVES: To assess the relationship between donepezil treatment and time to nursing home placement (NHP) for patients with Alzheimer's disease (AD). DESIGN: Observational follow-up of patient NHP and vital status. SETTING: Community. PARTICIPANTS: Patients previously enrolled in one of three randomized, double-blind, placebo-controlled clinical trials of donepezil and two subsequent open-label studies (total N = 1,115); 671 patients provided complete data for analysis. MEASUREMENTS: Data were obtained through follow-up interviews with caregivers and chart reviews of patients with AD. Comparison groups were defined by whether patients received an effective dose of donepezil (≥5 mg/d; ≥80% compliance) for specific numbers of weeks during the double-blind or open-label trial phase, in both phases, or in neither. Cox proportional hazards models were used to estimate risk ratios for NHP and survival curves from which median times to NHP were estimated for first dementia-related placement of longer than 2 weeks and permanent placement. The models were adjusted for age, sex, baseline Mini-Mental State Examination score, whether the caregiver was a spouse, caregiver continuity, and use of other cholinesterase inhibitors after the clinical trials. RESULTS: Use of donepezil of 5 mg/d or more was associated with significant delays in NHP. A cumulative dose-response relationship was observed between longer-term sustained donepezil use and delay of NHP. When donepezil was taken at an effective dose for at least 9 to 12 months, conservative estimates of the time gained before NHP were 21.4 months for first dementia-related NHP and 17.5 months for permanent NHP. CONCLUSION: Use of donepezil by AD patients resulted in significant delays in NHP. Long-term use of donepezil may help AD patients live longer in community settings, with consequent personal, social, and economic benefits.