The verbal Oxford Knee Score is not clinically different from the written score when assessed before or after total knee arthroplasty

The aim of this study was to assess the reliability of the Oxford Knee Score (OKS) collected verbally compared with the validated written score, using a population of patients who underwent total knee arthroplasty (TKR).

Ninety patients (mean age 70.6; (43–92), 56.7% female) undergoing TKR were prospectively assessed. One group (n = 45) completed written (standard) and verbal (over the telephone) OKS preoperatively, half (n = 23) performed the written questionnaire first followed by the verbal questionnaire, and the other half (n = 22) performed this in reverse. A separate group (n = 45) completed the same regime one year postoperatively.

A mean difference of 0.63 (95% CI − 0.985–2.23) points between verbal and written OKS was observed preoperatively, and of 1.36 (95% CI − 0.942–3.65) points was observed at one year postoperatively. Excellent reliability was observed using ‘average measures’ intra-class coefficient for the OKS preoperatively (r = 0.848) and at one year postoperatively (r = 0.970) in both groups who had written scores performed first, and those who had verbal scores performed first (preoperative r = 0.780, one year r = 892).

Bland and Altman plots demonstrated consistent correlation between patients reporting their preoperative score and one-year postoperative score verbally and written. There was no significant variation between groups who had written scores performed prior to verbal, compared with those who reported verbal scores prior to written.

Prospective written collection of OKS remains the benchmark. However, verbal recording of OKS is not clinically different to written score, and may be a useful alternative to OKS in patients who are unable to attend or complete written questionnaires.