One objective of the Alzheimer's Disease Cooperative Study (ADCS) is to develop new or improved instruments and assessment methods for evaluating treatment efficacy in Alzheimer disease (AD) clinical trials. The ADCS Instrument Committee has previously helped to define the state of the art in assessment for AD and Mild Cognitive Impairment clinical trials. We are now entering an exciting era of primary prevention trials to evaluate promising treatments that may delay disease onset and there is a need to develop appropriate instruments for these trials. The ADCS instrument committee has undertaken a project to develop instruments for prevention studies that assess domains known to be important in AD. Prevention trials are long and require large numbers of subjects, making them costly and requiring a high burden of participation for subjects. The current study focused on developing instruments that can be completed at home and in the clinic. The instruments are being evaluated in a cohort of nondemented elderly participating in a 4-year longitudinal study that simulates the design of a primary prevention trial. This report describes the design, baseline characteristics, and some longitudinal outcomes of the study cohort through the completion of the first 2 years of follow-up. We also describe the assessment domains to be measured with our new experimental instruments. This study recruited 644 subjects, 75 years of age and older. Participation in a "book club" that provided free books of interest to elders was offered as a recruitment incentive. Approximately 23% had some mild cognitive symptoms consistent with a Clinical Dementia Rating of 0.5. All subjects received a standardized in-clinic evaluation at baseline, which is repeated annually for 4 years to identify cases suspected of developing dementia and to measure longitudinal change on established clinical assessments. Subjects completed a set of self-administered experimental instruments at home or in the clinic designed to assess cognitive function and behavior, global change, activities of daily living, quality of life, and resource use. An additional "mail-in cognitive function questionnaire" was obtained separately by mail, 1 month before the other assessments. To evaluate the feasibility, efficiency, and validity of the home-based instruments in comparison with acquiring the same information during a clinic visit, subjects were randomized to 1 of 2 conditions in which the baseline and annual follow-up assessments are completed either at home ("home group") or at the study site during their clinic visits ("clinic group"). This initial report describes the ongoing 4-year longitudinal study and provides baseline results, which confirm the feasibility of obtaining home-based clinical information via mail or telephone. Initial results for the experimental instruments and for the book club are reported in separate accompanying articles.