This article reviews longitudinal data collected from patients with Alzheimer's disease (AD) that are relevant to the design and interpretation of clinical treatment trials. Longitudinal data from patients tested with the Alzheimer's Disease Assessment Scale demonstrate that cognitive symptoms, including memory loss, dysphasia, and dyspraxia, worsen relentlessly over time with the rate of change depending upon baseline dementia severity. Noncognitive symptoms, such as agitation, depressed mood, and psychosis, are episodic, do not necessarily worsen over time, and tend not to be highly correlated with one another. The reliability of cognitive change measures increases with follow-up duration so that the likelihood of detecting drug effects on the rate of cognitive deterioration is greater with longer treatment trials. Functional measures of activities of daily living are difficult to standardize for AD patients but are important for determining the overall clinical and economic impact of AD treatments.