Chronic diseases are in many ways more complex than acute diseases. In chronic diseases, response times to environmental effects are long, and confounding variables are numerous and may fluctuate with time. Treatment schedules are complicated and may be unique to the individual patient. Controlled trials deal with rigid hypotheses and permit control of error rates; prospective registries permit access to homogeneous subgroups. There is now a need for a new methodology for the study of the treatment of chronic diseases, which combines the merits of both approaches.