Littenberg and MacLean make very good arguments for the value of disease registries. We agree that registries are highly useful to monitor trends, look for patterns, and provide data for research. We also agree that chronic diseases, like diabetes, exact a tremendous toll on individuals, communities, and society. Thus, it is not surprising that there already exist many diabetes registries that are used for the goals outlined, and other registries could be developed that would be benign. All of the stated goals can be achieved using deidentified data. The registry that is our focus is one that uses the information to contact patients and their physicians directly about their health status. The authors never forthrightly address the significant risks we believe are associated with this opt-out registry, including risks to privacy, to the physician-patient relationship, to potential discrimination, and to patient autonomy.