OBJECTIVES:  There are several disease-modifying treatments (DMTs) in development to slow progression of AD. Understanding the economic value of DMTs requires understanding their effects on long-term patient outcomes and how those effects differ from existing treatments that improve symptoms without altering disease progression. We evaluated long-term patient outcomes for simulated disease-modifying and symptomatic treatments that appear identical over 2 years. METHODS:  We simulated patients over 10 years using the AD Archimedes Condition-Event simulator (AD ACE) in a cohort of patients from the Alzheimer’s Disease Neuroimaging Initiative (ADNI) with mild AD. The AD ACE is constructed using predictive equations for rate of change in disease biomarkers (cerebrospinal fluid [CSF] amyloid-beta, CSF t-tau, FDG-PET, and hippocampal volume) and clinical scales (MMSE, CDR-SB, ADAS-Cog, and NPI) derived from statistical analyses of ADNI data. A hypothetical DMT was constructed which slowed the rate of biomarkers progression. Upon discontinuation, the DMT’s slowing effect was lost, with patients progressing as untreated from their current state. A hypothetical symptomatic was also constructed which directly improved cognition, in which discontinuation resulted in a return to clinical values as if the patient had never been treated. RESULTS:  Both the DMT and symptomatic were calibrated to yield a 3.3 point improvement in ADAS-Cog over a 2 year period. When treatments maintained efficacy for 10 years, patients gained 0.09 and 0.03 QALYs versus no treatment, with the DMT and the symptomatic respectively. As the treatment duration was shortened, the benefits of the symptomatic treatment decreased to 0.01 QALYs for a treatment duration of 2 years. In contrast, the benefits of the DMT were largely maintained, with a QALY gain of 0.08 for a DMT with a treatment duration of 2 years. CONCLUSIONS:  DMTs may have benefits after treatment discontinuation and comparisons with symptomatic treatments must be made in this context.