The KOOS‐12 shortform shows no ceiling effect, good responsiveness and construct validity compared to standard outcome measures after total knee arthroplasty
Lukas Eckhard, Selin Munir, David Wood, Simon Talbot, Roger Brighton, Bill Walter, Jonathan Baré
Abstract
PURPOSE: To investigate the validity, responsiveness and ceiling effect of the recently introduced KOOS-12 and compare its performance to the KOOS, OKS, WOMAC and UCLA activity scales. METHODS: Patients from an independent multicentre study examining a medially stabilized knee system prospectively completed the KOOS, OKS, WOMAC and UCLA preoperatively and at 1 year postoperatively. KOOS-12 scores were calculated from the full length KOOS data. Construct validity was assessed using Spearman's correlation analysis. The ceiling effect was evaluated by calculating the percentage of patients with a maximum score. If the percentage exceeded 15%, a ceiling effect was considered to be present. Responsiveness was evaluated by performing paired t tests on the changes in measures and calculation of Cohen's d. RESULTS: A ceiling effect was present for the KOOS Pain, ADL and QoL subscales and the KOOS-JR at 1 year postoperatively. No ceiling effect was observed for the KOOS-12. Correlation of the KOOS-12 was low (0.3 < r < 0.5) with the UCLA, moderate (0.5 < r < 0.7) with the KOOS symptoms, sports and WOMAC stiffness subscales and high (r > 0.7) with all other scores and subscales. Effect size of the UCLA activity scale was moderate (Cohen's d 0.2-0.8) whereas effect sizes of all other outcome measures were large (d > 0.8). CONCLUSION: The KOOS-12 does not exhibit a ceiling effect, has good convergent construct validity and is responsive to changes in pain, function, QoL and knee impact between preoperatively and 1 year postoperatively. LEVEL OF EVIDENCE: Diagnostic level III.