Sleep Disorders Are Associated with Mental Health, Quality of Life and Stigma in an Italian Cohort of People Living with HIV
Round 1
Reviewer 1 Report
Comments and Suggestions for AuthorsI thank the opportunity to review the entiteled manusctript Sleep disorders are associated with mental health, quality of life and stigma in an Italian cohort of people living with HIV, submitted to BrainSci.
This is a very interesting research work which present a concise method with clinical results for better understanding PLWH health conditions. In the following lines I may now deliver some considerations and comments for the authors.
The authors present the Ethics IRB approval at the end of the document, I also suggest including it in the methods section.
2.7 Statistical analysis
U Mann Whitney was used to compare means? Or rank means?
Line 185, is the p value <0.100 correct?
There is a blank page line 239
Lines 220-232. This paragraph presents results of univariate analysis, but it is not clear if the authors refer to Mann Whitney´s U? I would suggest considering to present the results from this paragraph in a table, presenting significant and non-significant results, but also presenting other statistical values that may better express the tests results.
Instruments section. Lines 129, 149-150, the authors make a very important precision regarding the interpretation of the instruments´results, did they make the necessary adjustments to make the interpretations? for example in PSQI higher scores represent poor sleep quality, and Quality of life higher scores represent higher QOL, meaning they have opposite score direction. Could the authors please further explain how this was considered for the statistical analysis? Table 5 needs some formatting, also please consider further explaining in the results text why some variables do not present any results, for example Sex (Female vs Male), Years from HIV diagnosis, Time from starting first ART regimen, Self-reported health status, among others. Discussion section, the variable social problems, previously described in lines 95 and 96, receive a different name, now is described as social stigma, and other lines presenting the variable as "internalizing stigma". I suggest considering unifying the variable´s name in the same way along the different sections to prevent a different interpretation of the concept. Limitations of the study The authors make a very good self analysis of limitations and future directions. Please state if family and social support was measured or not, in case it was not, I consider that should be a future variable to use for controlling sample's variances. I hope these comments may help to improve the research report presentation, and wish the researchers a successive result with it.
Author Response
This is a very interesting research work which present a concise method with clinical results for better understanding PLWH health conditions. In the following lines I may now deliver some considerations and comments for the authors.
The authors present the Ethics IRB approval at the end of the document, I also suggest including it in the methods section.
We appreciate your suggestion. We have now included the Ethics IRB approval statement in the Methods section to enhance clarity and transparency.
2.7 Statistical analysis
U Mann Whitney was used to compare means? Or rank means?
We acknowledge the reviewer’s observation. The Mann-Whitney U test was used to compare rank means rather than arithmetic means. We have now clarified this in the Methods section to improve precision.
Line 185, is the p value <0.100 correct?
We thank the reviewer for pointing this out. This was a typographical error, and we have now corrected the threshold to p < 0.005, in line with our statistical approach.
There is a blank page line 239
We appreciate the reviewer’s observation. The blank page at line 239 has been removed to ensure proper formatting and readability.
Lines 220-232. This paragraph presents results of univariate analysis, but it is not clear if the authors refer to Mann Whitney´s U? I would suggest considering to present the results from this paragraph in a table, presenting significant and non-significant results, but also presenting other statistical values that may better express the tests results.
We thank the reviewer for the suggestion. We have clarified that the univariate analysis was conducted using the Mann-Whitney U test. Additionally, we have reformatted the results into a table (Table 3), which now presents both significant and non-significant results, along with additional statistical values to improve clarity.
Instruments section. Lines 129, 149-150, the authors make a very important precision regarding the interpretation of the instruments´results, did they make the necessary adjustments to make the interpretations? for example in PSQI higher scores represent poor sleep quality, and Quality of life higher scores represent higher QOL, meaning they have opposite score direction. Could the authors please further explain how this was considered for the statistical analysis?
We thank the reviewer for this important observation. We confirm that the interpretation of instrument scores was correctly aligned in the statistical analysis. Given that PSQI and SF-12 have opposite score directions, we carefully considered this in reporting the effect estimates. We have now added a clarification in the Statistical Analysis section to explicitly state how this was managed.
Table 5 needs some formatting, also please consider further explaining in the results text why some variables do not present any results, for example Sex (Female vs Male), Years from HIV diagnosis, Time from starting first ART regimen, Self-reported health status, among others.
In the multivariate models, some variables such as Sex (Female vs Male), Years from HIV diagnosis, Time from starting first ART regimen, and Self-reported health status were not included due to their non-significant association in univariate analysis or high collinearity with other variables. This decision was made to ensure model stability and avoid redundancy in the predictors. Instead, we prioritized variables that showed stronger associations with sleep quality and were more relevant to our research aims. Additionally, we would like to inform the reviewer that Table 5 is now Table 6, as we have introduced an additional table earlier in the manuscript. This renumbering ensures a logical flow of data presentation in line with the revised structure of the manuscript.
Discussion section, the variable social problems, previously described in lines 95 and 96, receive a different name, now is described as social stigma, and other lines presenting the variable as "internalizing stigma". I suggest considering unifying the variable´s name in the same way along the different sections to prevent a different interpretation of the concept.
We appreciate the reviewer’s suggestion and have now standardized the terminology throughout the manuscript. Specifically, we have replaced variations of the term (social problems, social stigma, internalizing stigma) with a single, consistent term (internalized stigma) based on its conceptual relevance to our study. This adjustment ensures clarity and prevents potential misinterpretations.
Limitations of the study The authors make a very good self analysis of limitations and future directions. Please state if family and social support was measured or not, in case it was not, I consider that should be a future variable to use for controlling sample's variances.
We appreciate the reviewer’s valuable insight. Family and social support were not directly assessed in our study, although they represent important factors that could influence sleep quality and mental health in PLWH. We have now added this point in the Limitations section and suggested it as a future research direction.
Reviewer 2 Report
Comments and Suggestions for AuthorsThis paper focuses on the quality of sleep, mental health, and quality of life (QOL) among People Living with HIV (PLWH), as well as their perceptions of stigma. Although there is already a considerable body of research on sleep disturbances and QOL in PLWH, this study is notable for examining the complex relationship with internalized stigma. Through its findings, the paper underscores the importance of reinforcing interventions for stigma and psychological support—alongside existing HIV treatments, stress reduction measures, and sleep hygiene education. This work suggests expanded clinical approaches by highlighting connections between exercise habits, social support, and sleep disturbances. Such a multifaceted assessment could offer new perspectives for future research and care for individuals with HIV.
On the other hand, there are several shortcomings in the current version of the manuscript. In particular, the lack of critical information on data analysis needs to be addressed. Please consider revisions centered on the following points:
- Positioning of Univariate Analysis
If univariate analysis was conducted solely to select variables for multivariate analysis, please state that purpose clearly in the manuscript. Otherwise, specify the criteria for determining which variables would be included in the multivariate analysis. For instance, even if “Education” is not statistically significant in univariate analysis, it could be an important socioeconomic factor. Carefully consider whether it should be included as a variable.
- Notation for Univariate Analysis
While “mean difference” is commonly used for univariate analysis, this paper uses “mean change.” If the study did not conduct a regression analysis, the term “mean change” may not be appropriate. Because “mean change” is frequently used in reference to regression coefficients (β), it may be more accurate to employ “mean difference” or “median difference.”
- Reference Group for Mean Change
“Mean change” generally indicates a comparison against a reference, but no control group is identified here. It is necessary to clarify which group or baseline is serving as the reference. Clearly identify the reference group or time point in the manuscript.
- Model Fit in Multivariate Analysis
Regarding the results of the multivariate analysis, you have not provided model fit indices such as R² or AIC/BIC. Without these, the adequacy of the model cannot be evaluated. Please include these indicators to substantiate the validity of your analysis results.
- Potential for Multicollinearity and Interaction
Please present the Variance Inflation Factor (VIF) results in the multivariate analysis. For example, the combination of Depression and Anxiety in the DASS-21 could lead to interaction effects. Depending on the findings, integrating or adjusting certain variables may be appropriate. Please examine this point and incorporate it into your discussion.
Author Response
-
Positioning of Univariate Analysis
If univariate analysis was conducted solely to select variables for multivariate analysis, please state that purpose clearly in the manuscript. Otherwise, specify the criteria for determining which variables would be included in the multivariate analysis. For instance, even if “Education” is not statistically significant in univariate analysis, it could be an important socioeconomic factor. Carefully consider whether it should be included as a variable.
We appreciate the reviewer’s insightful comment. We have now clarified in the Statistical Analysis section that variable selection for the multivariate analysis was based on both statistical significance (p < 0.005 in univariate analysis) and theoretical/clinical relevance. This ensures that important predictors, such as Education as a socioeconomic factor, are appropriately considered in the model, even if not statistically significant in univariate tests. This clarification enhances the transparency and rationale of our analytical approach.
-
Notation for Univariate Analysis
While “mean difference” is commonly used for univariate analysis, this paper uses “mean change.” If the study did not conduct a regression analysis, the term “mean change” may not be appropriate. Because “mean change” is frequently used in reference to regression coefficients (β), it may be more accurate to employ “mean difference” or “median difference.”
We appreciate the reviewer’s insightful observation. We acknowledge that "mean change" is typically used in regression analyses, whereas "mean difference" and "median difference" are more appropriate for univariate tests. In response, we have revised the manuscript to ensure that "mean difference" or "median difference" is used in univariate analyses (t-test, Mann-Whitney U), while "mean change" is retained in multivariate regression models. This distinction has been explicitly stated in the Statistical Analysis section, and terminology has been corrected accordingly throughout the manuscript.
-
Reference Group for Mean Change
“Mean change” generally indicates a comparison against a reference, but no control group is identified here. It is necessary to clarify which group or baseline is serving as the reference. Clearly identify the reference group or time point in the manuscript.
We appreciate the reviewer’s comment and have now explicitly defined the reference groups in the Statistical Analysis section. Specifically, we clarify that categorical predictors were analyzed in relation to a reference category (e.g., male participants, low stigma individuals, non-pathological DASS-21 scores), while continuous predictors reflect changes per unit increase. Additionally, we have incorporated a clarification in the Results section to ensure transparency in the interpretation of "mean change".
-
Model Fit in Multivariate Analysis
Regarding the results of the multivariate analysis, you have not provided model fit indices such as R² or AIC/BIC. Without these, the adequacy of the model cannot be evaluated. Please include these indicators to substantiate the validity of your analysis results.
We appreciate the reviewer’s valuable suggestion regarding the evaluation of multivariate model adequacy. In response, we have implemented several methodological refinements and incorporated additional statistical details in the manuscript:
-
Methods (Statistical Analysis): We have now explicitly stated that we calculated model fit indices (R², AIC, BIC) to assess the adequacy of the multivariate regression. Additionally, we have detailed our approach to handling multicollinearity using the Variance Inflation Factor (VIF) and applied Principal Component Analysis (PCA) to reduce the redundancy among highly correlated variables (DASS-21 Depression, Anxiety, and Stress).
-
Results: We have reported the model fit indices (R² = 1.00, AIC = -259.54, BIC = -261.50) and discussed their implications, noting that the model demonstrated high explanatory power but may be affected by collinearity. Additionally, we present a comparison of VIF values before and after PCA, showing how dimensionality reduction improved the model’s stability.
-
Potential for Multicollinearity and Interaction
Please present the Variance Inflation Factor (VIF) results in the multivariate analysis. For example, the combination of Depression and Anxiety in the DASS-21 could lead to interaction effects. Depending on the findings, integrating or adjusting certain variables may be appropriate. Please examine this point and incorporate it into your discussion.
We appreciate the reviewer’s insightful suggestion. We have now incorporated the Variance Inflation Factor (VIF) analysis in the Methods, Results, and Discussion sections. Specifically, we applied Principal Component Analysis (PCA) to reduce multicollinearity among DASS-21 Depression, Anxiety, and Stress, creating a single Psychological Distress Factor. We have also reported VIF values before and after the adjustment and discussed residual collinearity among Stigma and Quality of Life measures. These updates enhance the robustness of our analysis and clarify our statistical approach.
Round 2
Reviewer 2 Report
Comments and Suggestions for AuthorsThis second version of the paper is a great improvement, the authors are to be commended.