Identification of both the initial and the sustained clinical benefits achieved with symptomatic treatments for Alzheimer disease (AD) is a challenge. This commentary addresses a report by Wattmo et al. on 3-year follow-up of a cohort of patients with AD who were treated with donepezil. The investigators developed predictive regression models that can accurately calculate group mean Alzheimer's Disease Assessment Scale-cognitive subscale (ADAS-cog) and Mini-Mental State Examination (MMSE) scores. They determined that patients with mild to moderate AD have a mean 5-7-month cognitive improvement with donepezil treatment, with greater benefit in more-advanced disease. While these results are encouraging, this study has important limitations. Although the predictive models work well for determining group means, the authors note that they do not predict individual patient responses, which vary greatly. Additionally, the study had a drop-out rate of 62%, which might elicit survivorship bias and overestimation of treatment benefit. We remind clinicians that small improvements in cognition matter most when a concurrent measurable benefit is seen in daily function.