It is unfortunate that Dr. Romano believes that third parties will use information in the Annals supplement to hurt patients and their physicians. Our intent in publishing this supplement was to begin to educate internists about the potential benefits and risks of using large latabases to evaluate health care. Dr. Romano fears that managed care organizations, the federal government, and administrators use such data to justify withholding funds for patient care, and they sometimes do. However, research done by using large databases often supports the directing of resources toward certain clinical activities. One such example is given in this issue [1]. By using large databases, McCarthy and colleagues show a clear association between regular use of mammography and diagnosis of early-stage breast cancer. The results of this large database study support spending money on strategies to increase mammography use among older black women rather than withholding resources for care.

Nobody wants their care to be "dictated by abstract averages," but it is safe to say that everyone wants care that is evidence-based. Sometimes the best evidence results from careful analyses of large databases. We hope that the supplement on large databases illustrates the potential strengths of these databases as well as alerting readers to ways in which they might be misused. If Dr. Romano and others truly believe that third parties are using large databases against them and their patients, a good understanding of the pitfalls of large-database research will enable a more successful attack on "the enemy."