HipLink: Registry-Retrieval Linkage

What problem did we seek to address?

There is now a wealth of information available to patients on joint replacements. One of the biggest sources is the National Joint Registry (NJR) for England, Wales and Northern Ireland, who have collected data on more than 1.8 million operations (1). They now publish on their website how different implants, surgeons and hospitals are performing. The purpose of our study was to answer the simple question: ‘How accurate are NJR reports?’

How did we tackle the problem?

First, we made sure there were no mistakes in our data. We took implants at the LIRC out of storage and re-analysed them to make sure that their details were correct. We then joined the two big databases at the LIRC and the NJR and used a process called ‘validation’ to check for errors.

What were the main findings?

We found that, where a procedure was reported to the NJR, they did an excellent job of recording it. Data completion was over 99.9% for all fields and the error rate was typically less than 5%. We also showed that NJR data was getting better, year-on-year. However, we found that a sizable number of procedures were missing from the NJR database and we could only identify 60.9% of the implants we held at the LIRC.

What does this mean for patients?

This study has shown that the NJR database is largely well completed and accurate. It is probably the best source of data we currently have to analyse how joint replacements perform.

Across the Registry as a whole, missing data probably ‘averages itself out’ to give a good idea of how an implant is performing. However, for individual surgeons (or even hospitals), where the number of cases is much smaller, this missing data could effect how we interpret their performance.