Chronic diseases represent the major illness burden of developed nations. A chronic disease databank system consists of parallel longitudinal data sets from diverse locations describing the courses of thousands of patients with chronic illness over many years. Illustrated by ARAMIS (The American Rheumatism Association Medical Information System), such data resources facilitate analysis of long term health outcomes and the factors associated with particular outcomes. A model for clinical investigation of contemporary disease is presented, based on the overwhelming prevalence of chronic illness, the variability, complexity, and uniqueness of the individual patient course, the difficulties of traditional univariate reductionist approaches, and the time span required for study. In this model, data are systematically accrued and continually analyzed, and the data collected are gradually modified based upon evolving anticipation of future needs. The strategies underlying the development of ARAMIS are described, investigational results summarized, and future directions outlined.