Abstract
Implant registries are set up to register implants. They therefore collect information about both primary and revision joint replacements. If a revision is linked to a primary it is then possible to determine the revision rate of the primary. This information is, however, of limited value as detailed information that affects the revision rate such as indications for the primary and the revision, and surgical technique used are not recorded. As a result comparisons of different implant designs and implant types are not reliable. For example implants that are commonly used in young or active patients are likely to have higher revision rates than those used in elderly sedate patients even though they may be better. Similarly, implants that are easy to revise will have higher revision rates than those more difficult to revise even if they provide better functional results. Finally, implants that are commonly used by more experienced surgeons will tend to have lower revision rates than those used by less experienced surgeons. Data from registries are therefore useful for identifying hypotheses that can formally be tested in other ways.
