Patient registries are essential tools for public health surveillance and research inquiry, and are a particularly important resource for understanding rare diseases. Registries provide consistent data for defined populations and can support the study of the distribution and determinants of various diseases. One advantage of registries is the ability to observe caseload and population characteristics over time, which might facilitate the evaluation of disease incidence, disease etiology, planning, operation and evaluation of services, evaluation of treatment patterns, and diagnostic classification. Any registry program must collect high quality data to be useful for its stated purpose. Registries can be developed for many different needs, and caution should be taken in interpreting registry data, which has inherent biases. We describe the methodological issues, limitations, and ideal features of registries to support various rare disease purposes. The future impact of registries on our understanding and interventions for rare diseases will depend upon technological and political solutions for global cooperation to achieve consistent data (via standards) and regulations for various registry applications.