It is established good practice that hip and knee replacements should have regular follow-up and for the past seven years at the North Hampshire Hospital a local joint register has been used for this purpose and we compare this with results of the Swedish and UK national and the Trent Regional registries.

Since March 1999, all primary and revision hip and knee arthroplasties performed at North Hampshire Hospital, Basingstoke have been prospectively recorded onto a database set up by one of the senior authors (JMB). Data from patients entered in the first five years of the register were analysed. All patients have at least one year clinical and radiological review then a minimum of yearly postal follow-up. 3266 operations (1524 hips and 1742 knees) were performed under the care of 13 consultants. Osteoarthritis was the most common primary diagnosis in over 75% of hips and knees. Our revision burden was 7.5% (10.2% hips and 3.5% knees). As of 31/12/2006 6.2% of patients had died and 5.5% were lost to follow-up.

Revision rates were 1.5% and 1.4% for primary total hip and knee replacements respectively.

Our data analysis of revisions and patello-femoral replacements has allowed us to change our practice following local audit which is ongoing. Oxford scores at 2 years had improved from a mean of 19 and 21 pre-operatively to 40 and 39 for primary hips and knees respectively. Our costs are estimated at approximately £35 per patient for their lifetime on the register.

Compared to other registries:

Our dataset is more complete and comprehensive

Magnification of anteroposterior radiographs of the pelvis is variable. To improve the accuracy of templating, reliable and radiographer-friendly methods of scaling are necessary. We assessed two methods of scaling digital radiographs of the pelvis: placing a coin of known diameter in the plane of interest between the patient’s thighs, and using a caliper to measure the bony width of the pelvis. A total of 39 patients who had recently undergone hemiarthroplasty of the hip or total hip replacement were enrolled in the study. The accuracy of the methods was assessed by comparing the actual diameter of the head of the prosthesis with the measured on-screen value. The coin method was within a mean of 1.12% (0% to 2.38%) of the actual measurement, the caliper group within 6.99% (0% to 16.67%). The coin method was significantly more accurate (p < 0.001). It was also reliable and radiographer friendly. We recommend it as the method of choice for scaling radiographs of the pelvis before hip surgery.