Abstract
To ascertain whether there is any relationship between the Oxford 12 scores gained six months post surgery and early revision for primary hip and knee arthroplasty.
The six month post surgery Oxford 12 scores were retrieved from the National Joint Registry (NJR) for the seven year period ending 31st December 2006. These were analysed in relation to revision of primary hip and knee procedures using three methods of statistical analysis; logistic regression, receiver operating characteristic (ROC) curve and direct plotting of groups of scores against the proportion of hips revised for that same group.
Logistic regression: For every one unit increase in the Oxford score there was an 11% increased risk of revision (hips) and 12% (knees) within the first two years of surgery, 5 and 6% respectively between two and four years and 3 and 4% risk respectively between four and six years (p> 0.001).
The ROC Curve Analysis: demonstrated that a patient with a score greater than 20 (hips) or 28.5 (knees) or 24 (uni knees) has eight times the risk of needing a revision within two years compared to a person with a score equal or less than the above numbers. Alternatively the ROC analysis predicted 73% of the revisions within three years for all three arthroplasty groups.
Plotting Scores Against Revisions: Plotting scores in groups of five demonstrated an incremental increase in the risk during the first two years. A person with a score greater than 40 has; for hips 24 times, knees 27 times and uni knees 69 times the risk of a revision within two years compared to a person with a score between 16 and 20.
Monitoring of the six month post surgical Oxford 12 score is another tool in the surgeon’s armamentarium for deciding which patients need closer monitoring following arthroplasty surgery.
Correspondence should be addressed to Associate Professor N. Susan Stott, Orthopaedic Department, Starship Children’s Hospital, Private Bag 92024, Auckland, New Zealand.
