Correlations between patient-perceived outcome and objectively-measured biomechanical change following Total Knee Replacement.

Only 22 patients were analysed after surgery, so statistical power was limited and findings may be unstable. Patients had to be able to walk 10 m without a walking aid, so results may not apply to people who are more limited before surgery. Follow-up was at one time point, with timing varying from 9.3 to 22.8 months, so it does not show how function and gait change over time. The healthy group used to train the gait classifier was younger and had a lower body mass index than the osteoarthritis group, so recovery toward healthy gait may partly reflect age and weight differences. The pain measure was added late and was missing for some patients, leaving only 15 patients for pre-post pain comparison. The classifier involved data reduction and training, with a risk of overfitting despite splitting the training data and cross-validation. There is no established target for acceptable gait recovery after total knee replacement, so the size of objective gait change is hard to interpret clinically.

The authors declared no conflicts of interest.

Do not read the strong correlation between patient-reported function and gait pattern as proof that changing gait causes better perceived function. It was a small pre-post association in 22 patients, and pain change did not track gait change.