The Ceiling Effects of Patient Reported Outcome Measures for Total Knee Arthroplasty

Abstract Background: Patient reported outcome measurements (PROMs) that exhibit a substantial ceiling effect show clustering of participant’s scores towards the upper limit of a scale and consequently have low discriminatory power among high end scores. This study aimed to compare ceiling effects at 1 and 2 years postoperatively across commonly used patient reported outcome measurements (PROMs) for TKA. Hypothesis: We hypothesized, that the analyzed PROMs differ substantially in regards to their ceiling effect. Patients and Methods: Patients that underwent a primary unilateral TKA and completed pre-operative and post-operative questionnaires were included in the analysis. Participants completed the KOOS, KOOS-12, KOOS-JR, KOOS-PS, WOMAC and OKS preoperatively, and completed the KOOS, KOOS-12, KOOS-JR, KOOS-PS, WOMAC,OKS and FJS postoperatively at 1 and 2 years. Results: 1-year and 2-year follow-up data was available for 380 and 193 patients, respectively. The preoperative mean age was 68.0 (8.5) and mean BMI was 31.4 kg/m2 (6.6), with a male to female ratio of 49.6% to 50.4%. At 1 year postoperatively, a ceiling effect was seen for the Pain and ADL subscales of the KOOS and the KOOS JR. The KOOS Pain, Symptoms, ADL and QoL subscales, the WOMAC Total and KOOS JR exhibited a ceiling effect at 2 years postoperatively. We found 9.0% and 14.8% of patients achieving a maximum score in the FJS at 1 and 2 years, respectively, indicating the absence of a substantial ceiling effect. Conclusion: The PROMs studied differ substantially with regards to their ceiling effect and consequently their ability to detect differences between well performing groups. The KOOS Pain, Symptoms, ADL and QoL subscales, the WOMAC Total and KOOS JR exhibited a substantial ceiling effect at 2 years postoperatively. We recommend using PROMs like the FJS and KOOS-12 with a more evenly distribution of scores across the scale when studying well performing cohorts. Level of Evidence: III