Determinants of patient satisfaction one year after total knee arthroplasty: an item-level analysis of the Oxford Knee Score.

Abstract

To identify which individual Oxford Knee Score (OKS) questions were most strongly associated with patient satisfaction one year after total knee arthroplasty (TKA) through an item-level analysis. This retrospective single-centre cohort study included patients undergoing primary TKA with preoperative and one-year postoperative OKS (with at least 10 of 12 responses available) and level of satisfaction. OKS items were selected for regression modelling based on the strength of association with satisfaction. There were 2688 (1618 [60.2%] female; mean age 70.0 years) TKA included. Satisfaction showed a weak correlation with preoperative OKS (r = 0.133, p < 0.001), but correlated strongly with postoperative OKS (r = 0.601, p < 0.001) and moderately with the change in score (r = 0.466, p < 0.001). At the item level, preoperative OKS questions demonstrated uniformly weak correlations with satisfaction (r < 0.150). Postoperative items were more strongly associated, particularly postoperative pain and ability to work (r = 0.543, p < 0.001). All changes (baseline to one-year) in the individual OKS items correlated significantly (p < 0.001) with satisfaction, with standing (r = 0.511) showing the strongest association and washing (r = 0.167) the weakest. In multivariable regression (R2 = 0.346), postoperative limping (β = 5.332, p < 0.001), working (β = 4.529, p < 0.001), pain (β = 2.351, p < 0.05), and standing (β = 1.901, p = 0.014) and change in standing (β = 1.709, p < 0.001) were independently associated with satisfaction. Exploratory models confirmed that postoperative pain and core functional activities (standing, limping, working) were the strongest independent determinants of satisfaction. Postoperative pain, limping, working ability, and standing were the most important independent determinants of satisfaction following TKA. Focusing on these in patient counselling, expectation setting, and questionnaire development may enhance patient experience and outcome measurement efficiency. Retrospective study, Level III.