Not all questions are created equal: the weight of the Oxford Knee Scores questions in a multicentric validation study.

Abstract

Background: The Oxford Knee Score (OKS) has been designed for patients with knee osteoarthritis and has a widespread use. It has 12 questions, with each question having the same weight for the overall score. Some authors have observed a significant ceiling effect, especially when distinguishing slight postoperative differences. We hypothesized that each questions' weight will depend significantly on the patient's sociodemographic data and lifestyle.

Methods: In this international multicentric prospective study, we included patients attending a specialist outpatient knee clinic. Each patient filled out 3 questionnaires: (a) demographic data and data pertaining to the OKS, (b) the standard OKS, and (c) the patient gave a mark on the weight of the importance of each question, using a 5-point Likert scale (G OKS). Linear regression models were used for the analysis.

Results: In total 203 patients (106 female and 97 male) with a mean age of 64.5 (±12.7) years and a mean body mass index (BMI) of 29.34 (±5.45) kg/m² were included. The most important questions for the patients were the questions for pain, washing, night pain, stability, and walking stairs with a median of 5. In the regression models, age, gender, and driving ability were the most important factors for the weight of each of the question.

Conclusion: The questions in the OKS differ significantly in weight for each patient, based on sociodemographic data, such as age, self-use of a car, and employment. With these differences, the Oxford Knee Score might be limited as an outcome measure. Adjustment of the OKS that incorporates the demographic differences into the final score might be useful if the ceiling effect is to be mitigated.

Level of evidence: Level II prospective prognostic study.