Abstract: Each large observational database contains specific data elements. The number of data elements are chosen carefully to cover the essential needs of the database as well as to avoid excessive burden of collection. Frequently, an important study question cannot be answered because one database does not contain some essential data elements. This deficiency may be present because the proposed study is cross-disciplinary, because the study requires more granular information on a specific topic than is practical to collect in a broad-based registry, or because the relevant questions, and hence essential data elements, have changed over time. An obvious way to overcome some such challenges, when one database contains some of the information and another contains the further needed data, is to link different databases. While the prospect of linking databases is appealing, the practicalities of doing so often are daunting. Challenges may be practical (information-technology barriers to crosstalk between the registries), legal, and financial.

In the first section of this paper, Dr. Nathanael Heckmann discusses linking large orthopaedic databases, focusing on linking databases with detailed, short-term data to those with longer-term longitudinal data. In the second part of this paper, Nathan Glusenkamp discusses efforts to link the American Joint Replacement Registry (AJRR) to other data sources, an ambition not yet fully realized but one that will bear fruit in the near future.