Respiratory Infections in Adults and Inequality: An Analysis of Deaths and Their Socioeconomic Determinants in Brazil

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

Dear Authors,

 

This manuscript lacks conceptual clarity and analytical rigor. Among them, main concerns are pointed out below.

 

First, what is the unit of analysis? Data used in this study were obtained from multiple sources, but some of them are individual-level data and others not clearly described in this manuscript. More concerningly, some individual-level data appear to be aggregated to a unknown unit and them as group-level data. Therefore, it is very difficult to understand the meaning of tables presented in this manuscript.

 

Second, multinomial logistic regressions are used in this study, but the coefficients in a multinomial logistic regression indicate how changes in the predictor variables are associated with changes in the probability of each outcome category relative to a reference category. In short, multinomial refers to the type of outcome variable (multiple categories), not a type of effect within the model itself. Therefore, the outcome of interest needs to be categorized into two (e.g., no/yes or 0/1) and binomial logistic regressions need to be performed on your data. In doing so, you will be able to utilize all mortality data and understand the probability of a cause-specific-death by estimating the relationship between that event and other causes of death. Otherwise, the results of this study are not informative and very misleading.

 

Third, the Gini index has been used to measure various types of inequality (e.g., income, wealth, consumption, and access to resources) so it is unclear what aspect of inequality the Gini index used in this study is referring to. In additional, both the Gini index and the human development index are a summary measure of the entire country. Since Brazil encompasses a total land area of 8,509,379.576 km² (line 79 on page 2), significant local variations within Brazil make it very difficult to understand the meaning of Table 4.

 

Fourth, all observational studies cannot establish cause-and-effect relationships (causality) due to the potential influence of confounding factors. In other words, the results of all observational studies are only useful for understanding the associations between explanatory variables and an outcome of interest (whether linear or binomial). Therefore, terms such as “influence” and “cause” cannot be used in your manuscript and are very misleading.

 

For the reasons pointed out above, the research question and the study design need to be reconsidered for conducting appropriate statistical analysis. As a minor comment, some tables are missing a footnote on abbreviations (e.g., what does “RF” in Table 2?) and need to be separated into two or three (e.g., Table 4 needs to be separated into two tables).

Author Response

Please see the attachment.

Author Response File: Author Response.docx

Reviewer 2 Report

Comments and Suggestions for Authors

This work analyzes the epidemiological profile related to the demographic and socioeconomic factors that contribute to death from respiratory infections in Brazil. These diseases can be caused by bacteria, viruses, fungi, or parasites. In the case of tuberculosis, the prevention and reduction of morbidity and mortality are addressed with antibiotic therapy, which incurs a significant cost reported in R$ (I suggest reporting it in dollars since most of us do not know what R$ means). Pneumonia is another of the very common infections that lead to complications, including COVID-19. Brazil has different socioeconomic factors that can cause inequality in health services. This work analyzes the impacts that these respiratory infections had between 2014 and 2016. Data on territorial extension, population, and political-administrative organization are provided. The health system of Brazil is based on the principles of universality, equality, and comprehensiveness in care for the population. A table presents the description of the classifications and deaths due to respiratory infections. Subsequently, the procedure and statistical analysis of the obtained data is rigorously described. A table presents the description of the classifications and deaths due to respiratory infections. Subsequently, the procedure and statistical analysis of the obtained data is rigorously described.  In summary, the difficulties in etiological differential diagnosis are described, as well as the factors influencing susceptibility to respiratory infections and the impact of socioeconomic factors. The references are adequate and current to support the observations and conclusions obtained.  It is a work worthy of publication and will serve as a reference to help address these health issues in Brazil.

Author Response

Please see the attachment.

Author Response File: Author Response.docx

Reviewer 3 Report

Comments and Suggestions for Authors

Introduction

- Please clarify what proportion deaths from respiratory infections have among all deaths, to justify this focus on respiratory infections

- Also specify how deaths from respiratory infections are particularly likely to be influenced by social status in Brazil (while there is a system of care management for the most disadvantaged) compared to other diseases, there too to explain the interest of a focus on respiratory infections

 

Methods

-Please explain your choice to limit the study to adults?

-It seems to me that the inclusion of the Covid pandemic period is penalizing for the study insofar as very significant resources have been mobilized in the country for the management of the Covid infection, thus decreasing those available for the management of other infections, particularly among disadvantaged patients, but, at the same time, anti-Covid prevention measures also decreased contamination for other infectious agents: this adds several confounding factors making the results of the survey more difficult to interpret

-The first 3 paragraphs of Method would be more in their place in the Introduction

-The 4th paragraph ('In this study.... Ethic Committee') should be placed after the description of the data extracted from the national databases rather than before.

- I understand that the association of individual social status indicator (educational level) and death fromrespiratory infection was assessed by multinomial logistic regression and the association between region social status indicators (Gini and HDI) was separately assessed by a correlation study (Table 4). I am not an expert in biostatistics but would it not have been more appropriate to include HDI and Gini in a multilevel multivariate analysis. This should be clarify by the authors

- In the paragraph « Participants », please explain the difference between « death by respiratory infection » and « death with respiratory infection »

- I don’t understand the sentence « However, deaths recorded with DAI that combined abscesses and granulomas in the same category were not taken into account » (or rather I don’t understand the reason for exclusion)

- Why include non-pulmonary tuberculosis deaths in a study about death from respiratory infections?

- Specify the variables present on the death certificate that are used in the analysis, otherwise we only learn in the Analysis section that the educational level is registered on death certificates?

- Please clarify the sentence “Variables with a low number of deaths and variables with no record of death as “not informed” were categorized as “ignored””

- In the results, please also provide the evolution of the share represented by these deaths by respiratory infection among all the deaths

- I don’t see in the Results section the deaths 'with' respiratory infection?

- I don’t understand the sentence « In 2023, there was an increase in most variables related to infectious diseases »

- In the sentence “Only increases in the incidence of the Northeast and North regions were detected in the pandemic”, do the author mean “Increases in the incidence of death by respiratory infections during the pandemic were detected only in Northeast nd North regions”,`

-In the sentence “In addition to these regions, the incidence of the Midwest region also increased in 2023.” Do the author mean “In 2023, the incidence of death by respiratory infections increased only in the Midwest region”?

- It seems to me that Table 3 provides too many descriptive dat: may be these data should be provided as supplementary material, and only the results from multivariate analysis would be provided in the manuscript

-DC is not defined when first used

 

Discussion

- I suggest an extensive rewriting of the Discussion section, which I fin too long and difficult to read.

- I think the authors should begin the Discussion section with a sentence summarizing their main results, in link with the objective of the study

- The discussion is quite long and begins with paragraphs (“Difficulties in etiological differential diagnosis”, “Factors influencing susceptibility to respiratory infections “,…) that are not provided as a comment and an interpretation of the main results but come as independent comments

- The discussion about the main result and objective of the study (“Impact of socio-economic factors”) comes after four and half pages of discussion about less important results: I think it should come first.

- The main and more surprising result is stated only at line 520: “In this study, deaths from influenza were not correlated with socioeconomic indicators.” I propose the discussion starts from this point and states first, why this result is reliable, and second, how it can be explained in comparison with literature data regarding health social inequities

The Conclusion section should also be shorten, avoiding repetitions of findings

 

 

 

Comments on the Quality of English Language

Some sentences are ambiguous (see comments above)

Author Response

Please see the attachment.

Author Response File: Author Response.docx

Round 2

Reviewer 1 Report

Comments and Suggestions for Authors

Table 4 has not been changed and is very difficult to read. Please improve the readability of Table 4.