Review Reports

Round 1

Reviewer 1 Report (Previous Reviewer 3)

Comments and Suggestions for Authors

This is a very interesting article.  The authors have satisfactorily addressed all of my comments.

Author Response

Dear reviewer, 

 

Thank you for your comments and suggestions that improved our manuscript. 

 

Sincerely, 

 

Eduardo Rocha

 

Reviewer 2 Report (Previous Reviewer 2)

Comments and Suggestions for Authors

Thank you for submitting the article with the title “Big Data on climatic and environmental parameters associated with acute ocular surface symptoms and therapeutic assessment: Eye drops sales, Google Trends and Environmental changes” to MDPI vision.

Please respond to the following comments and questions and adequately address the changes in the manuscript:

The analysis has different confounders which are problematic. Is there a way to resolve the issues? Should more fundamental conceptual changes be considered to improve validity?

Conditions that lead to dry eye disease will naturally increase use of treatments against the condition. Please provide justification and explain why the research is relevant. Could the hypothesis be modified or strengthened?

Can you please address potential gaps in the data and how these exactly influence the analysis.

L59: Is it correct to use the mean value from the most populated areas or could this introduce a bias?

Author Response

Comments 01: The analysis has different confounders which are problematic. Is there a way to resolve the issues? Should more fundamental conceptual changes be considered to improve validity?"

Response 01: The confounding factors in the analysis are inherent to Big Data and the observational nature of the study, but the methodology used effectively mitigates them, which reinforces the structure and methodology of the article. Principal Component Analysis (PCA) was applied to resolve the problem of multicollinearity among the correlated environmental variables, which increases the interpretability and reliability of the results, despite limitations like geographical incongruence and monthly frequency. Therefore, the methods employed in the article already address the problems of validity and confounding factors within the available data constraints, establishing a robust proof of concept.

"Comments 02: Conditions that lead to dry eye disease will naturally increase use of treatments against the condition. Please provide justification and explain why the research is relevant. Could the hypothesis be modified or strengthened?"

Response 02: The premise is correct: the worsening of dry eye symptoms leads to an increase in the use of symptomatic treatments, as demand for immediate relief drives sales and searches. The relevance lies in the indirect and population-level monitoring of ocular surface (OS) symptoms, establishing a low-cost, high-scale public health surveillance methodology that utilizes valid proxies for treatment demand and the prevalence of acute ocular surface symptoms. The moderate correlations found, especially between ozone (O3) and UV Radiation, with the increase in sales of decongestant eye drops, reinforce the association of these factors as environmental triggers for OS symptoms, with sales/searches being a reflection of the symptomatic demand, which reinforces the structure and methodology of the article. The hypothesis was detailed in a previous article published by our team, which discusses the strengths and weaknesses of the method (the paper is cited in the present manuscript and is attached here). 

"Comments 03: Can you please address potential gaps in the data and how these exactly influence the analysis?"

Response 03: The data in the article are robust and should be evaluated pragmatically. The geographical incongruence (data from the Metropolitan Region vs. the entire State) and the low temporal resolution (monthly) do not invalidate the findings, but merely weaken the correlation signal. The fact that the study found moderate and significant correlations, such as between O3 and UV Radiation with eye drop sales, despite these limitations, reinforces the association potency of these factors as population triggers.

"Comments 04: L59: Is it correct to use the mean value from the most populated areas, or could this introduce a bias?"

Response 04: The methodology already included the cross-correlation test to compensate for the monthly frequency and search for time lags. The use of the mean value from the most populated areas was a pragmatic and necessary measure to bypass data collection discontinuity, making it the best available proxy for the environmental condition affecting the majority of the population in the study region. Therefore, the correlations found establish a powerful proof of concept, despite the limitations of the Big Data.

Author Response File: Author Response.pdf

Round 2

Reviewer 2 Report (Previous Reviewer 2)

Comments and Suggestions for Authors

Thank you very much for the responses to the reviewer’s comments and corrections.

Please respond to the following comments and corrections and adequately address the changes in the manuscript:

Comment 01: The authors responded that all methodological shortcomings and confounder were removed with statistical measures. However, I disagree with the author’s view and would like to ask for further elaboration. Specifically, the issues highlighted in comment 2 of the last revision round (conceptual deficit) indirectly introduce confounding and reduce validity. I do not see how these conceptual problems can be removed with statistical measures. Please explain in detail.

Comment 02: Thank you for the explanation. However, the response to comment 02 does not make it very clear what the advantage of the indirect monitoring (“indirect and population-level monitoring of ocular surface (OS) symptoms”) is, if similar effects can be measured directly and reliably. Please explain.

Comment 03: What is meant with a “pragmatic evaluation”? What is meant with “weaken the correlation signal”? I apologise for not understanding these phrases, but the collocations are not generally used in this way in scientific and academic writing.

Comment 04: In comment 4, the authors admit that there was “data collection discontinuity”, which is contradictory to the response to comment 03 where the authors state that “the data was robust”. Please explain the discrepancy.

Author Response

Comment 01: The authors responded that all methodological shortcomings and confounders were removed with statistical measures. However, I disagree with the author’s view and would like to ask for further elaboration. Specifically, the issues highlighted in comment 2 of the last revision round (conceptual deficit) indirectly introduce confounding and reduce validity. I do not see how these conceptual problems can be removed with statistical measures. Please explain in detail. 

Response 1: The methodological shortcomings and confounders were mitigated, but statistical measures did not remove them. The purpose of statistical analysis was not to remove shortcomings or confounders; in general, and here, the goal is to organize and summarize the vast numerical data, test hypotheses of climatically associated dry eye symptoms based on correlations among eyedrop sales, pollutants, and climatic parameters in a specific region. The work identified trends and relationships, concluding that environmental changes are indirectly correlated with dry eye symptoms through eye drop sales in a larger population over four years. It is an indirect data bank analysis that concluded based on previous premises (dry eye symptoms change in association with climatic changes, and dry eye symptoms induce lubricant and other eye drop sales), but it has limitations. Principal Component Analysis (PCA), an additional statistical analysis included following the reviewers' suggestions, improved the analysis's perspective by addressing the power of the variables tested and assigning an influence of 87.3% to the variables accessed. We agree that the variables chosen do not explain all (100%) of the correlations. Still, they confirm that the method, although with the observed limitations, is relevant for monitoring the influence of weather on dry symptoms, based on its correlations with the group of eyedrop sales during the same period.  We also agree with the reviewer that our methodological proposal cannot eliminate some specific confounders, more closely related to the presentation of dry eye symptoms, such as age, indoor environment, short-term symptoms, and chronic systemic diseases. To address that, a discussion about the limitations of the method that we are defending here was included in the Discussion section (page 10, paragraph 01):

 

“It does not account for other measures people might take for eye discomfort or confirm that increased sales were due to discomfort, dry eye, or ocular surface disease. Since DED is often chronic, immediate responses, such as searching for information or purchasing eye drops, may be less likely than with acute illnesses. Because of that, we tested the time lag and found  some correlations. Additionally, using monthly intervals and only considering outdoor environmental factors prevents detecting short-term correlations or responses to indoor conditions, such as air-conditioned spaces. Also, the demographic, socio-economic, ethnic, psychological, genetic, and medical backgrounds of the users providing data by Google Trends remain largely unknown. Moreover, this analysis was not performed on patients formally diagnosed with DED but at the population level, which introduces potential confounders and limits the ability to establish direct causal associations.”  

 

Comment 02: Thank you for the explanation. However, the response to comment 02 does not make it very clear what the advantage of the indirect monitoring (“indirect and population-level monitoring of ocular surface (OS) symptoms”) is, if similar effects can be measured directly and reliably. Please explain.

Response 2: The advantage of indirect monitoring using big data, as observed here and in other studies, is that it provides a valuable tool for estimating the frequency of common symptoms in a large population,  indicating changes in the observed pattern, and helping to raise possible associations that could be better investigated. It is helpful because it enables healthcare policymakers to consider whether a change in population behavior patterns constitutes a public health problem, based on their rates and correlations. 

In the Discussion, we mention some previous studies that employed an indirect big data approach to monitor epidemiological events, including a study from our group that evaluated the seasonality of fungal keratitis and its possible climatic and labor-related risk factors, based on anti-fungal eyedrops in Brazil. The present work introduces the novelty of double-checking correlations (Google search and eyedrop sales) and studying a complex, multifactorial condition, such as dry eye, which will be helpful for future research on chronic diseases, not just dry eye.  We also discussed their limitations (page 09, paragraph 03): 

 

“Previous studies have used Google searches to track conjunctivitis outbreaks [34], predict seasonal influenza [35], and assess population mental health dynamics [36]. While these studies suggest that search trends can anticipate disease outbreaks, the selected terms or study population may not have strongly correlated with environmental changes or caused discomfort at a detectable frequency or intensity.”

 

34. Deiner, M.S. et al. Google Searches and Detection of Conjunctivitis Epidemics Worldwide. Ophthalmology, 2019. 126(9): p. 1219-1229. 
35. Zhang, Y. et al. Using Google Trends and ambient temperature to predict seasonal influenza outbreaks. Environ Int, 2018. 117: p. 284-291. 
36. Adam-Troian, J. et al. Using absolutist word frequency from online searches to measure population mental health dynamics. Scientific Reports, 2022. vol. 12 (1) 2619. 
37. Ibrahim, M.M. et al. A new method to predict the epidemiology of fungal keratitis by monitoring the sales distribution of antifungal eye drops in Brazil. PLoS One, 2012. 

 

Comment 03: What is meant with a “pragmatic evaluation”? What is meant with “weaken the correlation signal”? I apologise for not understanding these phrases, but the collocations are not generally used in this way in scientific and academic writing.

Response 3: We appreciate your request for clarification on our previous response. We also apologise for using colloquial words in the reply letter. In the last round, the reviewer requested that we clarify the potential gaps presented in the model. In our work and in previous studies aimed at understanding the correlation between dry eye and climatic variables, gaps were recognized. Those gaps are primarily associated with two itens: 1.  the geographical location of climatic monitors and the extension of the area they cover. It would be more accurate if more stations covered smaller areas with more details, including climatic and pollutant variations. 2. The time interval of one month to get the mean values of climatic and eyedrop sales. We evaluated the data retrospectively over a 4-year period; however, it is acceptable that short-term peaks in climatic changes and eye drop sales would strengthen the statistical analysis. However, what we mean by “pragmatic evaluation” in the previous reply was that in the “real World”, the big data available related to those climatic monitor stations is not broadly distributed, and the eyedrop sales are not available on a daily or weekly basis.  Ideally, those finer trackers would be desirable, as they would generate an exponentially higher volume of data. However, we still do not have them, and they would help confirm the concept that supports our hypothesis. We hope to revisit this hypothesis in the future with more sophisticated tools. 

Comment 04: In comment 4, the authors admit that there was “data collection discontinuity”, which is contradictory to the response to comment 03, where the authors state that “the data was robust”. Please explain the discrepancy. 

 

Response 4: We would also like to thank the reviewer for their comment and the careful attention dedicated to our manuscript.  We aim to ensure that the data is robust, based on reliable sources, the time length of the observations, and the amount of data, which reveals sinusoidal curves similar for every year studied, and significant correlations, as highlighted in the manuscript results. 

The “data collection discontinuity” mentioned in the Methods (page 2, line 59) is a methodological limitation associated with the interruption of climatic and pollution data collection for weeks or months in certain areas during the evaluation period. These discontinuities were observed in some climatic stations, caused by technical problems and power outages that shut down monitoring stations during the period, and therefore, the values were left empty for those periods. This limitation was relevant and was presented in the discussion (page 9, line 176) to inform future research in climatic sciences on those events. We believe the data is robust, despite those discontinuities, because the climatic data was presented as the mean of monthly based available data, and the standard deviation was not distinct from the other fulfilled months.

We hope to have, in the future, more stable climatic data sources to offer data that overcome the limitations indicated in Comments 3 and 4. 

Author Response File: Author Response.docx

This manuscript is a resubmission of an earlier submission. The following is a list of the peer review reports and author responses from that submission.

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

This study is feasible and somewhat interesting, but the composition needs a lot of improvements. The Introduction seems like piling up background information without leading readers to the heart of the study. The data collection is fine, but the statistical approaches are just simple correlation analyses, which cannot distinguish possibly true influential factors from confounders. Multivariate regression analysis is suggested to solve the problem. Besides, the correlation among air pollutants has to be shown to find out any possibility of collinearity while conducting multivariate regression analysis.

Author Response

Comment 1: This study is feasible and somewhat interesting, but the composition needs a lot of improvements.

Reply: Thank you for the encouraging comment. 

Comment 2: The Introduction seems like piling up background information without leading readers to the heart of the study.

Reply: Thank you for the constructive feedback. An updated version of the introduction has been prepared and updated in the revised Manuscript, as highlighted in yellow. 

Comment 3: The data collection is fine, but the statistical approaches are just simple correlation analyses, which cannot distinguish possibly true influential factors from confounders. Multivariate regression analysis is suggested to solve the problem. Besides, the correlation among air pollutants has to be shown to find out any possibility of collinearity while conducting multivariate regression analysis.

Reply: Thank you for your comment. Now, a Principal Component Analysis (PCA) was employed to reduce the dimensionality of the environmental dataset, which comprised several correlated variables, by transforming them into a smaller set of uncorrelated principal components. The principal components explaining the highest proportion of variance were retained for further analysis. Multiple regression models were then constructed using these principal components as predictors to assess their combined and individual effects on the response variables of interest. This approach not only addresses potential multicollinearity among the environmental variables but also enhances the interpretability and reliability of the regression analysis.

The novel multivariate analysis revealed the most relevant variables in terms of environmental influences (Table 1) and the most sensitive responses in population terms captured by IQVia and Google Trends data (Table 2). Those tables are cited in the Results session and highlighted in yellow in the revised version of the manuscript. 

Reviewer 2 Report

Comments and Suggestions for Authors

Thank you for submitting the article with the title “Big Data on climatic and environmental parameters associated with acute ocular surface symptoms and therapeutic assessment: Eye drops sales, Google Trends and Environmental changes” to MDPI vision.

Please respond to the following comments and questions and make appropriate changes in the manuscript:

Major points:
The analysis has different confounders which are problematic and inadequately addressed. Please identify them and provide a solution on how to deal with them. In any case these have to be discussed as limitations; but does the research have sufficient validity to be published?

The hypothesis and research question: It is rather obvious that condition which perpetuate a dry eye disease in a population will increase the consumption of remedies against the condition. What is the justification and relevance of the research? How does it increase the evidence base, provide a new perspective on the topic or cover a research gap?

Methods: Is a regression analysis necessary for the research?

Conflict of interest: “There are no more conflicts of interest to disclose”. Is there a conflict of interest? Any conflict of interest of any kind has to be declared. What is the relationship with IQVIA?

 

Minor points:

In the abstract and introduction, please define the terms “environment” and “environmental factors” with specific descriptors.

Similarly, in L37 “the rising prevalence of disease”; please make sure to use specific descriptors.

L54 cont.: “The inability to obtain and use climate and pollution data for the entire state of Sao Paulo was due to data collection failures at some monitoring stations in the state’s interior during the study period, as well as the high variability in data across regions during the same periods, given the state’s vast territorial extension and regional differences.” Please revise the sentence (grammar).

 

Author Response

Comment 1: Thank you for submitting the article with the title “Big Data on climatic and environmental parameters associated with acute ocular surface symptoms and therapeutic assessment: Eye drops sales, Google Trends and Environmental changes” to MDPI vision.

Response 1: Thank you for the attention and suggestions to improve our work. 

Comment 2: The analysis has different confounders which are problematic and inadequately addressed. Please identify them and provide a solution on how to deal with them. In any case these have to be discussed as limitations; but does the research have sufficient validity to be published?

Response 2: Thank you for your pertinent comment. Now, a Principal Component Analysis (PCA) was employed to reduce the dimensionality of the environmental dataset, which comprised several correlated variables, by transforming them into a smaller set of uncorrelated principal components. The principal components explaining the highest proportion of variance were retained for further analysis. Multiple regression models were then constructed using these principal components as predictors to assess their combined and individual effects on the response variables of interest. This approach not only addresses potential multicollinearity among the environmental variables but also enhances the interpretability and reliability of the regression analysis.

Comment 3: The hypothesis and research question: It is rather obvious that condition which perpetuate a dry eye disease in a population will increase the consumption of remedies against the condition. What is the justification and relevance of the research? How does it increase the evidence base, provide a new perspective on the topic or cover a research gap?

Response 3: The present work identifies and proves that eyedrop sales are an alternative method to detect the increased DED frequency in a population set, associated with climatic variations. The study is relevant because DED is a prevalent condition with many presentations and diagnostic tools limited to hospital facilities, and presenting low predictive value. Our work intends to contribute to epidemiological studies. 

Comment 4: Methods: Is a regression analysis necessary for the research?

Response 4: Yes, the suggestion is very interesting. It was incorporated in the analysis, and the text was included in the Methods, Results, and Discussion as highlighted in yellow. 

Comment 5: Conflict of interest: “There are no more conflicts of interest to disclose”. Is there a conflict of interest? Any conflict of interest of any kind has to be declared. What is the relationship with IQVIA?

Response 5: IQVIA and other companies were contacted to provide the data to be used in the project. IQVIA gently provided the information, free of cost and avoiding the mention of any pharmaceutical brand. Therefore, no commercial relationships were established. 

Comment 6: In the abstract and introduction, please define the terms “environment” and “environmental factors” with specific descriptors.

Response 6: The term environment was expanded to the external environment to clarify its association with ecological factors associated with climate and outdoor pollution. It was defined in the introduction, and the term environmental factors was also defined in the introduction to clarify the meaning of climate and pollution indices measured by outdoor stations.  Those descriptions are highlighted in the Introduction session. 

Comment 7: Similarly, in L37 “the rising prevalence of disease”; please make sure to use specific descriptors.

Response 7: The expression was replaced by "a rising incidence" to correct the misuse of the term prevalence. It was highlighted in yellow. 

Comment 8: L54 cont.: “The inability to obtain and use climate and pollution data for the entire state of Sao Paulo was due to data collection failures at some monitoring stations in the state’s interior during the study period, as well as the high variability in data across regions during the same periods, given the state’s vast territorial extension and regional differences.” Please revise the sentence (grammar).

Response 8: The sentence was revised and replaced by a clear edited one, and the text was highlighted in yellow as follows: 

We could not obtain climate and pollution data from the whole São Paulo state and the entire analysis period. The reasons were discontinuity of data collection in some monitoring stations and vast areas where the climate and environment data are not monitored. We used the mean values of reliable data from the most populated areas to get around this limitation. 

Reviewer 3 Report

Comments and Suggestions for Authors

This is an interesting and novel use of big data for an observational ecological study.

1. Please provide references for the second sentence of the introduction about observed prevalence/incidence of dry ranging from 3%-75%.
2. It would be good to add mention in the background or discussion about the side effects of long-term use of artificial tears.
3. It would be good to include mention of next steps, perhaps for further research and/or for education.

Author Response

Comment 1: This is an interesting and novel use of big data for an observational ecological study.

Response 1: Thank you for the positive comment and for your attention to our manuscript.

Comment 2: Please provide references for the second sentence of the introduction about observed prevalence/incidence of dry ranging from 3%-75%.

Response 2: The references were added and included in the References list. 

 Stapleton F, Argüeso P, Asbell P, Azar D, Bosworth C, Chen W, Ciolino J,
Craig JP, Gallar J, Galor A, Gomes JAP, Jalbert I, Jie Y, Jones L, Konomi K, Liu
Y, Merayo-Lloves J, Oliveira FR, Quinones VAP, Rocha EM, Sullivan BD, Sullivan
DA, Vehof J, Vitale S, Willcox M, Wolffsohn J, Dogru M. TFOS DEWS III Digest
Report. Am J Ophthalmol. 2025 Jun 3:S0002-9394(25)00276-4. doi:
10.1016/j.ajo.2025.05.040. Epub ahead of print. PMID: 40472874.

Pereira LA, Arantes LB, Persona ELS, Garcia DM, Persona IGS, Pontelli RCN,
Rocha EM. Prevalence of dry eye in Brazil: Home survey reveals differences in
urban and rural regions. Clinics (Sao Paulo). 2025 Jan 28;80:100578. doi:
10.1016/j.clinsp.2025.100578. PMID: 39879905; PMCID: PMC11814509.

 

Comment 3: would be good to add mention in the background or discussion about the side effects of long-term use of artificial tears. 

Response 3: The recommendation is very interesting and it was added in the Discussion session as follows: 

It is important to observe that excessive and continuous use of eye drops may cause ocular surface discomfort similar to DED, because of the toxicity of eyedrop conservants (ref). 

Comment 4: It would be good to include mention of next steps, perhaps for further research and/or for education.

Response 4: The suggestion was welcome and it was attended including the following phrase in the Discussion:

In the future, monitoring eyedrop sales and internet searches for ocular symptoms may help identify climate changes and DED incidence in epidemiological studies in big population cities and remote geographic regions.

Round 2

Reviewer 1 Report

Comments and Suggestions for Authors

The authors added more analyses, which made the results look more convincing; however, none of the new data were described in Abstract. Introduction still needs editing; too many paragraphs there are difficult for readers to follow.

The logic of the article does not seem good. The authors have to re-arrange the whole thing to make it sound better.

Author Response

Comment 1: The authors added more analyses, which made the results look more convincing; however, none of the new data were described in Abstract.

Response 1: Thank you for your encouraging comments. The novel analysis was included in the Abstract  (lines 9-10). 

Comment 2: Introduction still needs editing; too many paragraphs there are difficult for readers to follow.

Response 2: We rewrote the Introduction session, summarizing the background, hypothesis, and objective. The new version was limited to lines 16-45, and it is highlighted in yellow.

Comment 3: The logic of the article does not seem good. The authors have to re-arrange the whole thing to make it sound better.

Response 3: Thank you for the comment. The work is interdisciplinary, involving four major fields of daily life that are the clinical aspects of dry eye, the climatic parameters, the trend searches on the internet, and the market sales of eye drops.  The manuscript's purpose is to investigate a potential association of the parameters of those fields. We believe it will be useful to integrate those areas in eye disease, but also in other medical fields in the future. 

We hope this revision addresses our intention to make it clear for readers and future researchers in the area.  

Author Response File: Author Response.pdf

Reviewer 2 Report

Comments and Suggestions for Authors

Thank you very much for the responses to the reviewer’s questions.

Please respond to the following comments and questions and adequately address the changes in the manuscript:

Confounders: “Thank you for your pertinent comment. Now, a Principal Component Analysis (PCA) was employed to reduce the dimensionality of the environmental dataset, which comprised several correlated variables, by transforming them into a smaller set of uncorrelated principal components. The principal components explaining the highest proportion of variance were retained for further analysis. Multiple regression models were then constructed using these principal components as predictors to assess their combined and individual effects on the response variables of interest. This approach not only addresses potential multicollinearity among the environmental variables but also enhances the interpretability and reliability of the regression analysis.”

Does this change adequately address the issues regarding the confounders? How could the authors execute the more substantial changes in the research design that may be needed?

Research question: “ The present work identifies and proves that eyedrop sales are an alternative method to detect the increased DED frequency in a population set, associated with climatic variations. The study is relevant because DED is a prevalent condition with many presentations and diagnostic tools limited to hospital facilities, and presenting low predictive value. Our work intends to contribute to epidemiological studies.” Despite the comment, the hypothesis remains still weak because the same conclusions can be drawn with sensible reasoning. How do the authors justify the need for the study? How can the authors strengthen the hypothesis?

Conflict of interest: Having a conflict of interest, such as in the present study can be problematic. Can the authors discuss the pros and cons of such a conflict of interest? Could it be avoided? How can the authors mitigate against the risks of the conflict of interest?

“We used the mean values of reliable data from the most populated areas to get around this limitation.” Is this methodologically correct?

Author Response

Comment 1: Thank you very much for the responses to the reviewer’s questions.

Please respond to the following comments and questions and adequately address the changes in the manuscript:

Response 1: Thank you very much for your attention, time to read our work, and constructive suggestions. 

Comment 2: Confounders: “Thank you for your pertinent comment. Now, a Principal Component Analysis (PCA) was employed to reduce the dimensionality of the environmental dataset, which comprised several correlated variables, by transforming them into a smaller set of uncorrelated principal components. The principal components explaining the highest proportion of variance were retained for further analysis. Multiple regression models were then constructed using these principal components as predictors to assess their combined and individual effects on the response variables of interest. This approach not only addresses potential multicollinearity among the environmental variables but also enhances the interpretability and reliability of the regression analysis.”

Does this change adequately address the issues regarding the confounders? How could the authors execute the more substantial changes in the research design that may be needed?

Response 2:  Thank you for this observation. The present work identified the key variables associated with actions induced by ocular discomfort. Future studies can be designed to investigate the parameters in shorter periods of time (weekly or daily), at more climatic stations, and to compare climatic variables indoors and outdoors.  Those improvements in the study design will depend on the government's understanding of the relevance of public health and financial investment. The relevance of the present work lies in informing the scientific community about the importance of environmental variables in daily health and in stimulating future research in this field.   

Comment 3: Research question: “ The present work identifies and proves that eyedrop sales are an alternative method to detect the increased DED frequency in a population set, associated with climatic variations. The study is relevant because DED is a prevalent condition with many presentations and diagnostic tools limited to hospital facilities, and presenting low predictive value. Our work intends to contribute to epidemiological studies.” Despite the comment, the hypothesis remains weak because the same conclusions can be drawn with sensible reasoning. How do the authors justify the need for the study? How can the authors strengthen the hypothesis?

Response 3. Thank you for the comment. We revised the hypothesis statement in the Introduction session (lines 32-34), emphasizing that a third and fourth element can confirm and detect the relationship among ocular discomfort signs and climate. They are the internet searches and the eyedrop sales. Therefore, we propose extraclinical parameters and independent confirmatory associations. They are useful in this disease, but also could be applied to other diseases, as an epidemiological tool. 

Comment 4: Conflict of interest: Having a conflict of interest, such as in the present study can be problematic. Can the authors discuss the pros and cons of such a conflict of interest? Could it be avoided? How can the authors mitigate against the risks of the conflict of interest?

Response 4: Thank you for the comment. We rewrote the Conflict of Interest statement to make clear that we harvested information from private companies, but we did not establish any commercial or financial relationships with those companies. As medical equipment used in clinical research and supplies used in biomedical studies are mentioned in the projects and in the papers, here we mention the tools used to obtain data needed to prove our hypothesis. All the relationships are clearly stated in the methods and in the conflict of interest statement. 

Comment 5: “We used the mean values of reliable data from the most populated areas to get around this limitation.” Is this methodologically correct?

Response 5: The reviewer is correct in mentioning that limitation. The lack of more climatic stations and stations better distributed in the geographic area of our study made us select the most populated regions to combine data from internet searches, eyedrop sales, and climate variables. The limitation is addressed in the Discussion session, lines 189-199. We hope that in the future, the climate stations will be able to cover more details of this association. 

Author Response File: Author Response.pdf