Socioeconomic Differences between Sexes in Surgically Treated Carpal Tunnel Syndrome and Ulnar Nerve Entrapment
Round 1
Reviewer 1 Report
Dear authors,
The subject of the study is interesting. The number of patients who were included in the study is large so the statistical results are representative.
I recommend reading the instructions for the authors regarding the editing of the article.
I recommend reviewing the conclusions with reference to this study.
I consider it exists an excessive number of self-citations - 6
Author Response
Thank you! We have reviewed the conclusions and the citations, please see the reference list.
Author Response File: 
Author Response.docx
Reviewer 2 Report
In this paper, the authors describe the different socioeconomic factors between sexes for patients treated surgically for CTS and UNE, as well as the influence of these inequalities on surgical prognosis.
The data for this study are extracted from the HAKIR registry, a national hand surgery registry. It has the strength of being a national registry, with wide representation in the country. However, it is not set up for the purposes of the study, therefore the results have to be evaluated with caution.
In the review process, we have identified some aspects of the manuscript that could be improved.
1) Introduction
You should write correctly the objectives of the study, the objective number 1 of the study "socieconomic differences between..." does not indicate if it is to describe, evaluate or compare.
2) Material and Methods
The major methodological issue with this article is that there is no stated sample size calculation based on the objectives of the study. The authors should indicate, based on the population data, which sample of the register is necessary and sufficient to be able to evaluate the differences between the socio-economic factors between sexes and which is representative of the target population, in order for the study to have external validity.
Confounding factors
One of the major confounding factors for socio-economic levels, besides the level of education, is the occupation of patients. Information on jobs should be included, as well as differences in socio-economic status by occupation, in order to rule out that this is not a confounding factor.
Social assistance should be described more broadly and measured not only by the number of times it is obtained. We propose an objective parameter such as the total amount of social assistance. This variable, also should be shown if are or not influenced by different occupation.
Sick days are also influenced by the type of occupation of the patients. And the observed differences may not be due to inherent gender causes, but to occupation. This is why sick days are also shown for different occupations.
3) Results
For the treatment of UNE, three different surgical procedures are included. Information should be included on whether there are differences in the performance of the procedures between the sexes and whether this leads to a different final outcome as measured by the QuickDASH score.
Figure 1 is not referenced in the text; this should be indicated before point 3.1.
The preoperative and postoperative value of the QuickDASH score should be shown in a table for the different confounding factors between the sexes.
Tables 1 and 3 show 95% confidence intervals for the means, but not for the population values of the proportions. Confidence intervals for proportions should be provided for the population value.
In tables 1 and 3 the first comparison column is Women, while in tables 2 and 4 it is Men. Please unify for a better reading of the results.
With linear regression coefficients, it is difficult to interpret and assess the magnitude of the observed differences. Therefore, instead of performing a linear regression with the QuickDASH questionnaire score on a continuous basis, it would be more informative to perform a logistic regression. The suggested cut-off point is 20 points on the QuickDASH questionnaire. These 20 points are the minimal clinical important difference (MCID) stated on the questionnaire website https://www.dash.iwh.on.ca/about-quickdash.
4) Discussion
The authors should modify the results discussion, and rewrite it. There are entire paragraphs that are practically identical and go round in circles about the same ideas.
Are the observed differences relevant? Are the observed differences different from the target population? Are these differences the same in other surgeries? What information does the study provide that could be useful in public health? Are questions, do not response in deep in the discussion.
5) References
There is an excessive number of self-citations, almost 25% of the total number of citations in the article. Replace these citations with other articles relevant to the topic.
Author Response
1) Introduction
You should write correctly the objectives of the study, the objective number 1 of the study "socieconomic differences between..." does not indicate if it is to describe, evaluate or compare.
Reply: We have changed the phrasing of the objectives to clarify this.
2) Material and Methods
The major methodological issue with this article is that there is no stated sample size calculation based on the objectives of the study. The authors should indicate, based on the population data, which sample of the register is necessary and sufficient to be able to evaluate the differences between the socio-economic factors between sexes and which is representative of the target population, in order for the study to have external validity.
Reply: As this is a registry study, we wanted to include all available data. However, to detect a difference between sexes regarding mean income (power of 80%, p=0.05) an estimated sample size of approximately 100 patients would be needed for carpal tunnel syndrome and 150 patients for ulnar nerve compression.
Looking at outcome, to detect a difference in QuickDASH between sexes (power of 80%, p=0.05) an estimation of 1130 individuals with carpal tunnel syndrome and 250 individuals with ulnar nerve compression are needed.
Confounding factors
One of the major confounding factors for socio-economic levels, besides the level of education, is the occupation of patients. Information on jobs should be included, as well as differences in socio-economic status by occupation, in order to rule out that this is not a confounding factor.
Reply: We agree that occupation is important in this context and that it most likely is a confounder. Since the primary focus of this study was to evaluate several socioeconomic factors, we limited occupation to the classical stratification of manual and non-manual work. This might be a little blunt, but we think it is difficult and might be misleading to do a more extensive stratification.
Social assistance should be described more broadly and measured not only by the number of times it is obtained. We propose an objective parameter such as the total amount of social assistance. This variable, also should be shown if are or not influenced by different occupation.
Reply: We have changed the parameter as suggested, please see section 3.1, table 1, section 3.2, table 3 and the regression analysis. We adjusted the regression analysis for manual occupation.
Sick days are also influenced by the type of occupation of the patients. And the observed differences may not be due to inherent gender causes, but to occupation. This is why sick days are also shown for different occupations.
Reply: We agree with the reviewer and this is an interesting view. We have added results on sick leave in manual and non-manual occupations stratified by sex, please see section 3.1 and 3.2.
3) Results
For the treatment of UNE, three different surgical procedures are included. Information should be included on whether there are differences in the performance of the procedures between the sexes and whether this leads to a different final outcome as measured by the QuickDASH score.
Reply: There is to our knowledge no studies on performance on different surgical procedures for UNE between sexes. Women generally score their symptoms and disability higher in the QuickDASH than men, which is also true in UNE and also in this study population. Whether this reported difference in the QuickDASH represents a true clinical difference remains controversial. We have added a comment on this in the results section 3.2 and in the discussion. We think that further analysis of sex differences between different surgical methods would be interesting, but that it is outside the scope of this study.
Figure 1 is not referenced in the text; this should be indicated before point 3.1.
Reply: Figure 1 is referenced in point 3.1 after the corresponding results in the text.
The preoperative and postoperative value of the QuickDASH score should be shown in a table for the different confounding factors between the sexes.
Reply: We have added the QuickDASH scores as suggested, please see table 1 and table 3.
Tables 1 and 3 show 95% confidence intervals for the means, but not for the population values of the proportions. Confidence intervals for proportions should be provided for the population value.
Reply: We are not sure that we understand what the reviewer is asking for. Data in table 1 and table 3 are presented as mean (95% confidence interval) or number of patients and percent (%) or median [interquartile range]. It is not possible to provide a confidence interval for the number of patients, that is an exact number.
In tables 1 and 3 the first comparison column is Women, while in tables 2 and 4 it is Men. Please unify for a better reading of the results.
Reply: Thank you, we have changed this as suggested.
With linear regression coefficients, it is difficult to interpret and assess the magnitude of the observed differences. Therefore, instead of performing a linear regression with the QuickDASH questionnaire score on a continuous basis, it would be more informative to perform a logistic regression. The suggested cut-off point is 20 points on the QuickDASH questionnaire. These 20 points are the minimal clinical important difference (MCID) stated on the questionnaire website https://www.dash.iwh.on.ca/about-quickdash.
Reply: This is an interesting suggesting. The MCID for the QuickDASH has been debated, and for CTS there is one study that found 20 as the MCID (Clement, 2016). We have not found any studies evaluating MCID for QuickDASH in UNE, however, it is stated on the QuickDASH website that a 20 point difference could be considered MCID in all evaluated syndromes. Unfortunately, in our dataset, there is not enough power to perform a logistic regression as suggested using the MCID, at least not for UNE, since there simply are not enough patients who have responded both to the preoperative questionnaire and the postoperative at 12 months. We therefore humbly suggest to keep the current regression analyses.
4) Discussion
The authors should modify the results discussion, and rewrite it. There are entire paragraphs that are practically identical and go round in circles about the same ideas.
Are the observed differences relevant? Are the observed differences different from the target population? Are these differences the same in other surgeries? What information does the study provide that could be useful in public health? Are questions, do not response in deep in the discussion.
Reply: We have modified and rearranged the discussion as suggested.
5) References
There is an excessive number of self-citations, almost 25% of the total number of citations in the article. Replace these citations with other articles relevant to the topic.
Reply: We have reviewed the citations, please see the reference list.
Reviewer 3 Report
1. Are you referring to sick leave days immediately after the operation for CTS/UNE or before the operation but connected with the diagnoses mentioned above (G56.0, G56.2)?
2. It is interesting that no correlation was found between manual occupation and operative treatment results. How do you interpret that? Maybe fewer sick leave days would be connected with less physically demanding work (earlier return to white-collar work as opposed to manual work)?
3. There is no information as to which colours represent men and women in Figure 2.
Author Response
- Are you referring to sick leave days immediately after the operation for CTS/UNE or before the operation but connected with the diagnoses mentioned above (G56.0, G56.2)?
Reply: We studied sick leave over time in this population, i.e. all sick leave, not only sick leave related to the treatment of CTS/UNE. Data was available from 1994-2016 and mean sick leave was calculated as net days per employed year over 20 years of age. The reason for this was that we thought it would be interesting in the socioeconomical setting to evaluate whether long-term sick leave would affect outcome in CTS and UNE.
- It is interesting that no correlation was found between manual occupation and operative treatment results. How do you interpret that? Maybe fewer sick leave days would be connected with less physically demanding work (earlier return to white-collar work as opposed to manual work)?
Reply: We also think this is interesting. Normally, individuals with manual work tasks require longer sick leave than individuals without manual work tasks. We have added a comment on this in the discussion.
- There is no information as to which colours represent men and women in Figure 2.
Reply: Only the characteristics of women are shown, we have clarified this in the figure legend.
Round 2
Reviewer 2 Report
The authors have significantly improved the work after revision. There are only a couple of aspects that should be finally considered.
1) Include the calculation of the sample size stated in the response to the editor in the section on materials and methods.
2) The social assistance variable has been changed in Tables 1 and 3 but not in the tables corresponding to the multivariate models. Please change them and include the new coefficients obtained in the model.
Author Response
Dear Editor and Reviewer,
We have included the power calculation in the methods section, and recalculated the regression models with the new variable.
Best regards,
Malin Zimmerman
Corresponding author
