Registry review

Orthopaedic registries were introduced as a way of collecting information on joint replacements at a population (often national) level. They allow for a combination of monitoring both clinician and implant performance. Initially they only included failure rates, but more recently data collected on patient-reported outcome measures (PROMs) have provided a further insight into the success rates and outcomes following knee arthroplasty.

There are now 31 members of the International Society of Arthroplasty Registers (ISAR) that collect data on patients undergoing knee arthroplasty.¹

The majority of these are national registries (such as the Swedish Registry that first started collecting data in 1975), although in the United States there are a number of regional or privately funded registries also collecting data (ten in total). There is very little standardisation between registries; not all require compulsory entry, not all collect outcome data other than failure rates (such as PROMs), and not all produce public reports. Before any data from registries are analysed, it is important to understand the registry design and methodology behind the process of data collection for each one (for instance the compliance rate, data entry time and entry clerk varies significantly between individual registries). Many registries collect data at a population level and despite the inclusive nature and huge volume of data collected, they are still subject to possible bias as previously described in our first registry review.³Table I and table II describe what features make a good registry, and what their main purposes are.