Dear Editor,
First of all, we thank the authors for their interest in our paper.
In our study, we used the procedures by Beaton et. al. and Guillemin et al. The recommendation mentioned in both studies already stated the necessity of making a pretest. It contains enough information to demonstrate that we are doing a pilot study. As a matter of fact, we also mentioned our pilot study in the discussion section. Beaton et al. recommended at least 30–40 patients. Guillemin did not show a minimum requirement for the number of cases taken for the pretest and developed a score to indicate that this was related to the level of evidence of the study. It would be true if the author’s critique was that we only considered the advice of Beaton et al.
We stated in our paper that the sample size was based on the general recommendations of Altman3; at least 50 subjects in a methods comparison study. There are numerous qualified reliability and validity studies conducted in accordance with this recommendation. According to another recommendation study, validation studies should be conducted with at least 5–10 times the number of survey items. However, considering the long structure of the New Knee Society Scoring System, it is not easy to reach these numbers in limited time periods of cross-sectional studies conducted in one center. It is seen that these numbers couldn’t be reached in German and French version studies too. It will be more accurate to state that the lower number of cases of these studies is a minor limitation of the study.
The user manual of the questionnaire emphasizes the necessity of using a different questionnaire for each knee in bilateral cases. Therefore, in a study involving individuals assessed for both knees, giving statistical results depending on the number of individuals will not be mathematically reasonable in some situations. We explained the reason and examples of the criticism you made about the number of cases just above.
Only the expectations section was analyzed separately because of differences in the preoperative and postoperative questionnaires. There were no difference in the other sub-sections and therefore this issue did not cause problem for statistical analysis.
The version studies of the New Knee Society Scoring System have been carried out with different methodologies and it is not possible to claim that they are wrong. In the other version studies, there is no consensus in terms of methodology and the comparison of the data (especially due to the use of different surveys for validity) was not compatible.
We agree, physical function domain of SF-36 should have ended in a positive correlation rather than a negative one when compared to the corresponding symptom domain of the new KSS. However, the results obtained from our study were in this direction. So, there was a negative correlation. This may suggest that the questionnaire differs from SF-36 because of its specificity to pathology. In the User’s guide of the questionnaire the expression “Part of the validation process involved confirmation that it was generally consistent with other” knee-specific scores. “ support our results. Furthermore, the fact that the questionnaire has a broad spectrum of functional activity sub-scales with four different levels indicates that it may be natural to be incompatible with KOOS and WOMAC subscales. The comparison of a questionnaire developed for the evaluation of preoperative and postoperative periods of total knee arthroplasty with a questionnaire developed for Knee Osteoarthritis should not lead to questioning the validity and reliability of the study.
In discussion section of our study we clearly stated some relationships between other version studies. Addressed paper (Letter to the editor) and reply to this comment showed that subjective interpretations are a product of different perspectives. Also, our expression “in conformity” with the literature does not have to show the conformity of all parameters.
Although the idea that multicenter study may be more effective in terms of increasing the level of evidence, it will be more accurate to consider with the advantages and disadvantages of its applicability. Advantages of multicenter study is already expressed in the literature. But, cost and time issues limits lots of studies to do so.
We have claimed that the questionnaire can be useful for Turkish-speaking individuals living abroad. Investigating the external validity in different countries could be better. Considering that the culture of each country is different, we think that it will be very hard to overcome this manner. It is just a concern that only one high quality version will be produced in one country for one given language will be more suitable, in our opinion.
Responsiveness analysis is a methodological choice in these studies. There is no rule that all studies are required to perform all sort of validation procedures. For a 34-item questionnaire specifically developed for Total Knee Arthroplasty, reliability analysis with Cronbach’s alpha value is much more important to identify the internal consistency. It would be appropriate to observe validity by comparing it with more than one questionnaire that is neglected in other studies. In conclusion, we reject neither revision nor any corrections on account of the replies above, but some issues like not conducting responsiveness analyses could be added as a suggestion for further studies.