Comparison of patient-reported and clinician-assessed outcomes following total knee arthroplasty.

BACKGROUND Although the necessity of long-term follow-up after total knee arthroplasty is unquestioned, this task may become burdensome as greater numbers of total knee arthroplasties are performed. We sought to use comparisons with clinician-assessed values to determine whether patients could reliably assess their own outcome with use of a combination of American Knee Society Score and Oxford Knee Score questionnaires and self-reported knee motion. We hypothesized that patients would self-report worse pain and function and a similar range of knee motion than clinicians would. METHODS One hundred and forty patients (181 knees) scheduled for routine follow-up at two centers after primary total knee arthroplasty were mailed American Knee Society Score and Oxford Knee Score questionnaires, a set of photographs illustrating knee motion in 5° increments for comparison with the patient's range of knee motion, and a goniometer with instructions. The patient's American Knee Society Score, Oxford Knee Score, and knee motion were then independently assessed within two weeks of the self-evaluation by one of three clinicians who had not been involved with the surgery. Patient-reported and clinician-assessed measures were compared with use of a paired-sample t test and the Spearman correlation coefficient. RESULTS The mean patient-reported American Knee Society pain subscore was 4 points worse than the clinician-assessed score, and the function subscore was 10 points worse (p < 0.001 for both). The mean Oxford Knee Score did not differ significantly between the patient self-assessment and the clinician assessment (p = 0.05). The mean maximum flexion reported by the patient with use of the photographs differed by <1° from the mean value reported by the patient with use of the goniometer or the mean value measured by the clinician; these differences were not clinically important. CONCLUSIONS Patients' self-reported American Knee Society pain and function subscores were worse than the corresponding clinician assessments, but the two Oxford Knee Scores were similar. Range of knee motion may reasonably be self-assessed by comparison with photographs. Long-term follow-up of patients after total knee arthroplasty may be possible with use of patient-reported measures, alleviating the burden of clinic visits yet maintaining contact, but further studies involving other validated instruments is warranted.