In Reply We appreciate the comments of Jiang and colleagues on our study. We are happy to reply to their concerns and address them in the order they were raised.

In this study, patient quality of life (QOL) was assessed for up to 3 years after surgery. Although the collection of as much data as possible was attempted by issuing queries for missing data, there were considerable missing data. As noted in the discussion, linear mixed models assess repeated measures and maximize the analysis of all data under the assumption of correlations within observations. In many previous studies, linear mixed models were used for the analysis of longitudinal QOL data by adding the baseline score without missing data as a covariate. We referred to the simulation results showing that accuracy does not decrease in mixed models without multiple imputations.¹ Also, the results showed that multiple imputation of missing end point data prior to linear mixed-effects models does not improve the accuracy of the estimated rate of change in the end point.² However, a recent study³ noted that sensitivity analyses for QOL can be performed using multiple imputation, which could enhance the robustness of the results for QOL. Therefore, we considered multiple imputation methods for missing values in further studies with long-term follow-up observations.