The premise underlying the implementation of any intervention aimed at changing behavioral or other symptoms of Alzheimer's disease (AD) is that the change will be beneficial in some manner for the patient and the benefits will outweigh any negative consequences that may simultaneously result from the intervention. Consequently, the ultimate task of investigators studying the impact of proposed interventions is to identify and measure the benefit that stems from the intervention, as well as any negative effects. When interventions are intended to create organic changes, as is traditionally true with pharmacologic interventions, the identification of relevant outcomes is made relatively simple through indicators of organic functioning. However, when intended changes are behavioral, the identification of appropriate outcomes is far more complex. An individual's behavior ultimately has an impact on all the actors within his or her social setting, directly or indirectly. Furthermore, changes in an individual's behavior may have an impact on other persons in such a way that these secondary changes feed back to further influence the initial target individual. Hence, it is likely that the most appropriate study designs for assessing the impact of interventions aimed at behavioral problems will include multiple outcome measures for multiple actors.