Contextual Factors Matter: The Role of Social Support in Work-Related Difficulties and Employment Status in Persons with Multiple Sclerosis

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

 

General comment

Overall, the paper addresses an important research question and employs a comprehensive methodology but the presentation of the results is a bit chaotic and should be improved. Overall a more detailed discussion of limitations (direction of the potential biais) would strengthen the paper further.

Specific points

I am not 100% convinced that the SDMT is the only test sensible to detect cognitve impairment in MS patients, you should rephrase section 2.2.3.

Was the BDI-II validated by Brenlla and Rodriguez or only translated?

Table 1: p-value is NEVER 0 since it's a probability! Please modify table 1 accordingly. The table is quite difficult to read, present the results in one column per category mean (std), it will increase the readability

 

Did you perform sample size estimation? How was the sample size defined?

Statistics: a lot of testing have been done, did you ajdust the p-value for multiple testing? Correlogram would be better than Table 2.

What are the conclusion and the clinical implications?

The results of the mediation analysis shoud be further discussed

Minor

Lines 74-79: Please use i, ii and iii instead of number as it may be confusing with the references

Figure caption should be bellow the figure, not on top

3.4. Italic

 

Comments on the Quality of English Language

Ok

Author Response

We tried our best to incorporate all the reviewer suggestions. The words and phrases changed or added have been marked in red. Our responses to the editor are listed below.

 

Reviewer: 1

Overall, the paper addresses an important research question and employs a comprehensive methodology but the presentation of the results is a bit chaotic and should be improved. Overall a more detailed discussion of limitations (direction of the potential biais) would strengthen the paper further.

Response: We appreciate the comment of the reviewer and we tried to improve and have worked on making the manuscript easier to read.

 

Specific points

I am not 100% convinced that the SDMT is the only test sensible to detect cognitive impairment in MS patients, you should rephrase section 2.2.3.

Response: We find this comment very accurate. The sentence was modified.

 

Was the BDI-II validated by Brenlla and Rodriguez or only translated?

Response: In Argentina the BDI II was adapted and validated by Brenlla and Rodriguez. In section 2.2.4 the term validated was added.

 

Table 1: p-value is NEVER 0 since it's a probability! Please modify table 1 accordingly. The table is quite difficult to read, present the results in one column per category mean (std), it will increase the readability.

Response:  We consider this comment to be appropriate, for this reason, we have added the requested data to Table 1.

 

Did you perform sample size estimation? How was the sample size defined? Justificar con bibliografía. Correr simulación de modelos. 

Response:  The sample was circumstantially defined, given the local conditions (low prevalence of MS and limited resources). As recommended by Giner-Sorolla et al. (2024) for these cases, we performed an effect-size sensitivity analysis.  Using the simr R package we conducted a series of simulations, which suggested that for power over 80% and 0.05 alpha, the target effect size for the main effects of interest was around Z = 3. Although not ideal, this still provides evidence for strong effects, which are ultimately those that will have more impact and clinical relevance. This should be considered when drawing conclusions about the absence of effects as smaller effects could exist. This was considered in the study limitations.

Giner-Sorolla, R., Montoya, A. K., Reifman, A., Carpenter, T., Lewis, N. A., Jr, Aberson, C. L., Bostyn, D. H., Conrique, B. G., Ng, B. W., Schoemann, A. M., & Soderberg, C. (2024). Power to Detect What? Considerations for Planning and Evaluating Sample Size. Personality and social psychology review. Advance online publication. https://doi.org/10.1177/10888683241228328

 

Statistics: a lot of testing have been done, did you adjust the p-value for multiple testing? Correlogram would be better than Table 2.

Response: P-values were not adjusted for multiple testing. Given the number of tests, it is possible that there were false positives so we added this clarification so that conclusions can be drawn with caution.

What are the conclusion and the clinical implications?

Response: Because of the importance of prioritizing the relationship between research data and clinical practice, we highlight this comment. We have added the following sentences: “These findings have implications on clinical and rehabilitation practice. First, our finding that processing speed and depression directly influence ES suggests that cognitive training interventions [51] and effective treatments for depression [48] can be utilized to improve these MS symptoms and, in consequence, reduce their negative influence on job outcomes. Furthermore, social support may act as a facilitator of employment by the other two variables. These results underscores the importance of social support in the design of effective vocational interventions that involve direct interaction with the person’s social environment and can improve PwMS’ job outcomes. In vocational rehabilitation, social support can potentially be increased by training PwMS on making clear and active requests for help from the social environment and to better communicate social support needs, including providing suggestions on how help could be provided [14]. Also, both family members and friends can be included in educational programs that can help improve their management of issues and complications related to the disease [52]. The importance of health specialists considering social support in clinical settings is emphasized, as it is a significant resource available to PwMS to cope with the challenges of daily life, as well as the close relationship between social support and depression”.

 

The results of the mediation analysis should be further discussed -

Response: In terms of the mediation models the following hypotheses were added: “In conclusion these results support the existing evidence of the importance of social support and social connections for mood disturbance, cognition, and particularly for employment. Concerning ES, the analyses suggest that there is mediation between the variables, that is, that social support might affect ES indirectly, through processing speed and depression. It is possible that this environmental resource has a positive effect on depressive symptoms, which in turn have a positive effect on ES. It may also be likely that an enriched social environment promotes better cognition, which improves ES. These hypotheses should be tested in future studies. For the other occupational outcomes, social support seems to be more relevant. Based on these data, we highlight the importance of studying the intersection of the individuals and environmental-contextual factors to understand how these variables interact and generate consequences on an individual`s occupational performance [20,14].”

 

Minor

Lines 74-79: Please use i, ii and iii instead of number as it may be confusing with the references

Response: The changes were done.

 

Figure caption should be bellow the figure, not on top

Response: The requested change was made.

 

3.4. Italic

Response: The requested change was made.

Reviewer 2 Report

Comments and Suggestions for Authors

This is a very well written and interesting manuscript. I have some suggestions and comments which may improve the quality of this paper.

Introduction

The main critique is that the introduction would benefit from more clearly situating the study and stating the gap or problem this particular research addresses. The authors introduce social support as a less studied factor and state more information is needed, but the connection back to employment could be more clearly articulated. Why focus on social support and employment in MS specifically? What unique contribution can this study provide? Sharpening the purpose and motivation may strengthen the rationale leading into the specific aims.

Additionally, the aims themselves come across as slightly broad and unfocused. Aim 1 about studying the general relationship between social support and employment status is vague. More specificity about the hypothesized relationship or a more narrowly defined research question could help. Aim 3 also lacks detail on the hypothesized relationship between social support and work accommodations.

Methods

Overall, the methods are clearly described.

A few critiques and suggestions for improvement:

1. More justification for the sample size of 90 could be provided. Was a power analysis conducted to determine this was an adequate number to detect effects?
2. The data analysis section is on the brief side. More detail could be provided on the specific statistical tests used to analyze differences between groups, the regression/mediation analysis approaches, and how variables were entered into models.
3. Related to the above point, the authors could elaborate on their rationale for using generalized linear models and logit models to analyze the relationships between variables. Why were these approaches selected?
4. When describing the SDMT, the authors mention it is the "sole cognitive measure" used due its ability to predict employment status. However, the results/analysis does not seem to incorporate SDMT performance. Clarifying if/how this variable factored into the analysis would be helpful.
5. The inclusion/exclusion criteria around employment status is unclear - for example, were students excluded from all analysis or only certain parts? Defining the groups more clearly would help avoid confusion.

Results

Overall, the results are clearly presented with helpful summary tables and figures. The authors include relevant statistics and effect sizes to support their findings. Dividing the analysis by specific research aims facilitates comprehension.

A few suggestions to strengthen the results:

1. Including some effect size metrics (e.g. Cohen's d) for the between groups differences in Table 1 would further quantify meaningful differences beyond statistical significance.
2. When describing the mediation analysis, the authors could explain the choice to use a quasi-Bayesian approach and bootstrapping. How do these methods add rigor?
3. For the generalized linear models examining social support's relationship to absenteeism, harassment, negative work events, and accommodations, reporting some model fit statistics would bolster interpretation of the results.
4. The accommodations analysis is difficult to interpret without more context. The authors state no predictors had a significant effect initially, but then MOS family did in a univariate model. Clarifying the value of the univariate test and reconciling the discrepant findings would help the reader make sense of the results.
5. There are some loose ends from the methods that are not addressed in the results, such as how the SDMT data factored into the analysis and employment status groups. Tying up these loose threads could improve flow.

Discussion

Overall, the discussion sets up a nice conceptual synthesis of the study variables and articulates why examining social support matters for employment and work experiences in MS. Addressing the following minor points below would further enhance the quality of the discussion and maximize its value for the field.

1. Though speculation about causality is limited due to the cross-sectional design, the authors could still put forth some hypothesized models that might explain the direction of effects based on theory and prior research. This would strengthen the conceptual framing.
2. When describing the effects of social support, inclusion of effect sizes could help quantify the magnitude of differences and associations for the reader.
3. Additional limitations could be noted, such as the limited generalizability due to the specific sample and lack of comparison group of healthy controls or patients with other conditions.
4. The implications for clinical practice and future research could be expanded on more. In particular, explicit suggestions for translating findings through assessments, interventions, or vocational rehab programs would make an applied contribution.

 

 

 

 

Author Response

We tried our best to incorporate all the reviewer suggestions. The words and phrases changed or added have been marked in red. Our responses to the editor are listed below.

Reviewer: 2

Comments and Suggestions for Authors: This is a very well written and interesting manuscript. I have some suggestions and comments which may improve the quality of this paper.

Response: We appreciate the reviewer's comments and will do our best to make improvements on the basis of all feedback.

 

Introduction

The main critique is that the introduction would benefit from more clearly situating the study and stating the gap or problem this particular research addresses. The authors introduce social support as a less studied factor and state more information is needed, but the connection back to employment could be more clearly articulated. Why focus on social support and employment in MS specifically? What unique contribution can this study provide? Sharpening the purpose and motivation may strengthen the rationale leading into the specific aims.

Response: This review seems to us to be very accurate, so we have added the following sentences to improve understanding of the relationship between social support and employment: “Employment and social support may be related [12]. In women with MS, family support, including the opportunity to delegate household tasks, receive help with childcare and understanding from family members, has been linked to job retention and better work-life balance [20]. It has also been observed that employed PwMS report greater perceived social support [21, 22], although the influence of other variables that may be relevant to this relationship remains to be studied. In people with other chronic illnesses, receiving advice, encouragement and motivation from friends and family, as well as feeling understood, is associated with a higher likelihood of returning to work after sickness absence [23]. Therefore, being employed can provide meaningful social networks for the person with a chronic illness that provide contention at the workplace, while social support outside work can also influence the individual's behavior related to ability to maintain work role functioning, including the implementation of reasonable workplace adjustments”.

 

Additionally, the aims themselves come across as slightly broad and unfocused. Aim 1 about studying the general relationship between social support and employment status is vague. More specificity about the hypothesized relationship or a more narrowly defined research question could help. Aim 3 also lacks detail on the hypothesized relationship between social support and work accommodations.

Response: We have written a few sentences to clarify the objectives.

 

Methods

Overall, the methods are clearly described.

Response: We thank the reviewer for his comments.

 

A few critiques and suggestions for improvement:

1. More justification for the sample size of 90 could be provided. Was a power analysis conducted to determine this was an adequate number to detect effects?

Response: The sample was circumstantially defined, given the local conditions (low prevalence of MS and limited resources). As recommended by Giner-Sorolla et al. (2024) for these cases, we performed an effect-size sensitivity analysis.  Using the simr R package we conducted a series of simulations, which suggested that for power over 80% and 0.05 alpha, the target effect size for the main effects of interest was around Z = 3. Although not ideal, this still provides evidence for strong effects, which are ultimately those that will have more impact and clinical relevance. This should be considered when drawing conclusions about the absence of effects as smaller effects could exist. This was considered in the study limitations.

Giner-Sorolla, R., Montoya, A. K., Reifman, A., Carpenter, T., Lewis, N. A., Jr, Aberson, C. L., Bostyn, D. H., Conrique, B. G., Ng, B. W., Schoemann, A. M., & Soderberg, C. (2024). Power to Detect What? Considerations for Planning and Evaluating Sample Size. Personality and social psychology review. Advance online publication. https://doi.org/10.1177/10888683241228328

 

1. The data analysis section is on the brief side. More detail could be provided on the specific statistical tests used to analyze differences between groups, the regression/mediation analysis approaches, and how variables were entered into models.

:

1. Related to the above point, the authors could elaborate on their rationale for using generalized linear models and logit models to analyze the relationships between variables. Why were these approaches selected?

 

1. When describing the SDMT, the authors mention it is the "sole cognitive measure" used due its ability to predict employment status. However, the results/analysis does not seem to incorporate SDMT performance. Clarifying if/how this variable factored into the analysis would be helpful.

 

Response: Points 2., 3., and 4. were expanded in section 2.4 “Data analysis” section (p. 4), we hope it is now clearer.

 

1. The inclusion/exclusion criteria around employment status is unclear - for example, were students excluded from all analysis or only certain parts? Defining the groups more clearly would help avoid confusion.

 

Response: To clarify the confusion, we have added in section 2.4 “Data analysis”, the following sentences: “PwMS were classified into employed if they had a paid job, both full-time or part-time and as employee or self-employed; or into unemployed if they were unemployed because they cannot find work or were unable to work. Some of them received disability benefits and others did not. In all cases we excluded those who were retired due to age, homemakers, and students”.

 

Results

Overall, the results are clearly presented with helpful summary tables and figures. The authors include relevant statistics and effect sizes to support their findings. Dividing the analysis by specific research aims facilitates comprehension.

Response: We thank the reviewer for his comments.

 

A few suggestions to strengthen the results:

1. Including some effect size metrics (e.g. Cohen's d) for the between groups differences in Table 1 would further quantify meaningful differences beyond statistical significance.

 

Response: We consider this comment to be appropriate, for this reason, a column with Cohen's d results has been added in Table 1.

 

1. When describing the mediation analysis, the authors could explain the choice to use a quasi-Bayesian approach and bootstrapping. How do these methods add rigor?

Response: We added this information in 2.4 “Data analysis” section  (p. 4).

 

1. For the generalized linear models examining social support's relationship to absenteeism, harassment, negative work events, and accommodations, reporting some model fit statistics would bolster interpretation of the results.

Response: For each model we reported the corresponding deviance goodness-of-fit test. The test for the harassment model was not significant (.07). As described in the manuscript “While variance inflation factor (VIF) values were not high, which does not suggest multicollinearity, these measures are moderately correlated with each other, so we explored univariate models. In this case, MOS total had a significant effect (OR = 0.34, 95% CI [0.10, 0.93], p < .05) and deviance difference (χ2 (3) = 6.90, p = .07)”. We included this information because we thought it could be informative given the lack of research on this topic, clearly stating in the Discussion section that conclusions around this result should be approached with due caution and be subjected to replication by future studies which could consider it as a hypothesis.

1. The accommodation analysis is difficult to interpret without more context. The authors state no predictors had a significant effect initially, but then MOS family did in a univariate model. Clarifying the value of the univariate test and reconciling the discrepant findings would help the reader make sense of the results.

Response: We believe that in the case of accommodations, as with harassment, this result may be clinically relevant and can be tested in future research with a larger sample of PwMS and using causal studies.

1. There are some loose ends from the methods that are not addressed in the results, such as how the SDMT data factored into the analysis and employment status groups. Tying up these loose threads could improve flow.

Response: We hope the improvements made in section 2.4 “Data analysis”  will make this point clearer.

Discussion

Overall, the discussion sets up a nice conceptual synthesis of the study variables and .why examining social support matters for employment and work experiences in MS. Addressing the following minor points below would further enhance the quality of the discussion and maximize its value for the field.

Response: With the responses from the comments, we have tried to improve the conclusions.

 

1. Though speculation about causality is limited due to the cross-sectional design, the authors could still put forth some hypothesized models that might explain the direction of effects based on theory and prior research. This would strengthen the conceptual framing.

Response: It has been added in relation to the difference in social support between employed and unemployed PwMS: “A key finding was that MOS Total, emotional/informational support, and affective support significantly differed between employed and unemployed PwMS with an estimated effect size between medium to large. Employed PwMS showed greater social support, and their scores were similar to those reported by healthy control subjects [41]. Emotional/informational support may be particularly important for ES because it refers to advice and information received from the social environment that helps PwMS to solve and anticipate problems, which may be about work. This advice can also come from the work environment, both coworkers and supervisors. Affective support, which refers to the demonstrations of care and empathy, can be relevant to ES as it can function as a buffer against stressful events generated by work pressures.”

And in terms of the mediation models, the following hypotheses were added: “It is possible that this environmental resource has a positive effect on depressive symptoms, which in turn have a positive effect on ES. It may also be likely that an enriched social environment promotes better cognition, which improves ES. These hypotheses should be tested in future studies. For the other occupational outcomes, social support seems to be more relevant. Based on these data, we highlight the importance of studying the intersection of the individuals and environmental-contextual factors to understand how these variables interact and generate consequences on an individual`s occupational performance [20,14].”

 

1. When describing the effects of social support, inclusion of effect sizes could help quantify the magnitude of differences and associations for the reader.

Response: We hope that the responses to comment 11 will also clarify this comment 12.

 

1. Additional limitations could be noted, such as the limited generalizability due to the specific sample and lack of comparison group of healthy controls or patients with other conditions.

Response: We agree with this comment and have therefore added the following sentence: “Moreover, the low prevalence of MS in Argentina and the scarce socioeconomic resources limit the possibility of obtaining a larger sample size and this must be taken into account when making conclusions. Also the lack of comparison groups of healthy controls or patients with other conditions limit the generalizability of the results”.

 

1. The implications for clinical practice and future research could be expanded on more. In particular, explicit suggestions for translating findings through assessments, interventions, or vocational rehab programs would make an applied contribution.

 

Response: As our response to reviewer 1, because of the importance of prioritizing the relationship between research data and clinical practice, we highlight this comment. We have added the following sentences: “These findings have implications on clinical and rehabilitation practice. First, our finding that processing speed and depression directly influence ES suggests that cognitive training interventions [51] and effective treatments for depression [48] can be utilized to improve these MS symptoms and, in consequence, reduce their negative influence on job outcomes. Furthermore, social support may act as a facilitator of employment by the other two variables. These results underscores the importance of social support in the design of effective vocational interventions that involve direct interaction with the person’s social environment and can improve PwMS’ job outcomes. In vocational rehabilitation, social support can potentially be increased by training PwMS on making clear and active requests for help from the social environment and to better communicate social support needs, including providing suggestions on how help could be provided [14]. Also, both family members and friends can be included in educational programs that can help improve their management of issues and complications related to the disease [52]. The importance of health specialists considering social support in clinical settings is emphasized, as it is a significant resource available to PwMS to cope with the challenges of daily life, as well as the close relationship between social support and depression”.

Round 2

Reviewer 2 Report

Comments and Suggestions for Authors

The authors responded to my comments and suggestions very well. Thank you.