Oxford Knee Score Calculator

Twelve questions cover the two things that drive a patient to consider knee replacement: pain and the loss of specific movements. The items ask about usual knee pain, washing and drying the whole body, getting in and out of a car or onto public transport, walking distance before pain becomes severe, standing up from a chair after a meal, limping, kneeling and rising again, night pain in bed, interference with work including housework, the knee giving way, shopping unaccompanied, and walking down stairs. Two of these are distinctly knee problems rather than general leg problems. Kneeling loads the patellofemoral joint through deep flexion and remains the single activity most often still impossible after a technically successful replacement, and a sense of the knee giving way reflects ligamentous or quadriceps insufficiency that has no equivalent at the hip. Every item is scored 0 for the worst option to 4 for the best, and the twelve are summed to 0 through 48. Totals are usually read as 0 to 19 severe, 20 to 29 moderate to severe, 30 to 39 mild to moderate and 40 to 48 satisfactory. The minimally important change is commonly quoted at around five to seven points, and average improvement after total knee replacement runs roughly sixteen to eighteen points, consistently smaller than the gain seen after hip replacement, which mirrors the well documented finding that a lower proportion of knee replacement patients report themselves satisfied. Limitations to keep in mind are the four-week recall period, the ceiling effect in patients who do well, and the fact that the score was validated for arthroplasty populations rather than for young patients with ligament or meniscal injury.

The Oxford Knee Score was published in 1998 by Jill Dawson, Ray Fitzpatrick, David Murray and Andrew Carr in the Journal of Bone and Joint Surgery, British volume, two years after the same Oxford group produced the corresponding hip questionnaire. It was built the same way, by interviewing patients waiting for knee replacement, letting the item pool come from what those patients described as limiting them, and then reducing that pool statistically and testing it in a surgical cohort. The intention was to break with the clinician-scored knee ratings then in use, which mixed patient symptoms with measured flexion arc, alignment and stability and so could improve on paper while the patient noticed nothing. As originally published each item was scored 1 to 5 for a total of 12 to 60 with low numbers indicating a better knee. The 2007 revision by Murray and colleagues restated the scoring as 0 to 4 per item and 0 to 48 overall with high numbers better, aligning it with the hip version and ending a long period in which published means could not be compared without checking which convention an author had used. From April 2009 the English National Health Service required the score to be collected before and after every publicly funded knee replacement under its Patient Reported Outcome Measures programme, which has since generated hundreds of thousands of paired records. It has been translated into many languages and is licensed through Oxford University Innovation.