Improving Health Outcomes in Women Who Use Traditional Open Fire Cookstoves by Addressing Cooking Behaviors: A Longitudinal Cohort Study

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

This appears to be an important study on this topic. Longitudinal studies like this are few and far between. Thank you!  These findings will make an important contribution to the literature.  The study overall seems great, but the presentation in the manuscript was difficult in places.  Please accept the following comments as constructive feedback for what appears to be a very important paper. 

Highlights:

CHAMP-BC is defined in the third bullet point, and again in the fourth bullet point.  Only needs to be defined the first time it’s used.

The highlights seem unusually long.  My recommendation is to condense them down to a single, brief, well-written paragraph.

Abstract:

No recommendations? 

Introduction:

CHAMP-BC is defined as “Community Health Action Model….” at least four times in the first two pages of the manuscript.  

• Background – The two paragraphs in this section are hard to follow. For example, it is not clear where the “foundational studies” were conducted without looking at the reference list.  It would help the reader if you could provide more context. Were these foundational studies conducted in communities similar to your study population?

The sentence starting with “At the time of this study in Nicaragua…” is hard to follow. There are several ideas presented in this sentence.  It would help the reader if this was broken up into two or three smaller sentences.

The flow of ideas is bumpy in the two paragraphs in section 1.1. There does not appear to be a clear focus for either paragraph. It would help significantly if each paragraph started with a clear focus sentence (your claim), followed by several sentences providing evidence/interpretation to support your claim.  For example, the first paragraph in 1.1 makes the claim that foundational studies show a link between household air pollution and health effects, but this claim is only followed by one very long sentence with one reference to back up the claim. Paragraph 2 in 1.1 seems to restate the same claim as paragraph 1.

Overall, the background section lacks a clear thesis, and the ideas do not flow logically from one idea to another.

• Target Population – This section suffers from some of the same problems identified in 1.1. It seems customary for paragraphs to have closer to 4-6 sentences, starting with a topic sentence, followed by several sentences providing supportive evidence or interpretation. There are several 2-sentence paragraphs in these sections that do not have a clear thesis or logical flow of ideas. It makes it hard to follow. The first three paragraphs in 1.1 and 1.2 only have two sentences each. The supporting information seems to jump around between these paragraphs, and does not flow easily from one idea to the next. 

As an example, the first paragraph in 1.2 gives a list of reasons why health inequities exist in Nicaragua, namely accessibility (I’m assuming you mean to healthcare), acceptability (not sure what this means), affordability (do you mean to clean fuels?), and access to health promoting materials (not sure what this means).  Based on this claim, I would expect to see several sentences backing up these claims.  However, there is only one sentence of supporting information that focuses on women being vulnerable due to roles/responsibilities as primary cook for the family. So, the first sentence makes a bunch of claims that are not supported in the paragraph. 

The second paragraph in 1.2 has similar problems.  

• Behavioral and Environmental Risk Factors – This paragraph mentions factors that seem to be specific to your population. It’s not clear where this information came from.  It would help the reader if this paragraph explained how this information was collected, and from whom, and when. It almost seems like it belongs in the study results.

 

• Literature Review – In this paragraph, can you add a sentence providing some context for the reader. It would be helpful to know how long some of these prior studies were conducted for.  Was it months?  Years?  It’s not clear what you mean by “long-term” if we don’t have some numbers to compare. 

 

• Indoor Air Pollution – This is another two-sentence paragraph that is hard to follow. After reading the first sentence, I would expect to see several sentences to back up the claim that improved cookstoves improve indoor air quality (and there are many studies that support this). However, there is only one sentence providing supporting information.  I like that the sentence focuses on Guatemala.  The closer the supporting information is to your study population, the better.  The ideas in this paragraph should be more fully developed, with at least 2 – 4 more sentences providing additional supporting information and interpretation.

 

It would recommend using “Bruce et al. (year)” rather than “Bruce and colleagues…” with no year.

Also, the way the results from Bruce are presented is distracting. There is a direct quote from Bruce et al., rather than a succinct summary of what they found. Did they really measure PM3.5, or is that a typo?  PM3.5 is not very common.  Please provide your own summary of Bruce et al.’s findings, rather than a long, wordy quote. Then, please add several other studies to back up your claim. There is probably a good meta-analysis or two out there that could be referenced here.

This section just seems thin for such a cool longitudinal study that you all conducted. 

• Improved Health Outcomes – Please change author citations to “Diaz et al. (year)…”; “Smith-Sivertsen et al. (year)”; etc….

There are a lot of direct quotes in this manuscript. The second sentence starting with “At 18 months…”, tor example, and again later in the paragraph starting with “For all respiratory…”.  Please summarize findings in your own words, rather than copying from the articles.  Direct quotes should be used sparingly.  Much of the information presented in the direct quotes from Bruce et al. and Diaz et al. can be summarized much more succinctly anyway, making it flow better for the reader.

• Acceptability of Cookstoves – Please change author citation to “Mukhodaphyay et al. (year)…”

Ok, this would be a great place to orient the reader to what you mean by “long-term”.  You make a compelling case for cook stove acceptability based on the Mukhodaphyay study (wish I knew when this was published), but there is no mention of how long the participants used the cookstoves before post-measurements were made. That would be really good to know for this paper, since it is the longest study to date on this topic.  Honestly, this theme should run throughout the intro, and hopefully the discussion section of the paper. It would really help the reader if the issue of time was clear throughout.  As I written, I don’t know what I’m comparting your study to.  Also, another use of a direct quote here.

Also, the first paragraph in this section only uses one study as evidence. This seems like a really big issue that is very relevant to your paper, so I’m surprised to see only one paper to back up your claim.  Cookstove acceptability is a huge issue – is there not a meta-analysis on this, or a good review paper that can be cited here, or at least some other studies with participants similar to yours? 

1.7         Community Health Action Model for Participatory Behavior change is defined again as “CHAMP-BC”.  I’m not sure it needs to be defined so many times in the paper.

Materials and Methods.

CHAMP-BC is defined again here. 

It’s hard to see in the methods where your survey instruments are described. For example, in the implementation action phase you state that baseline data was collected.  Can you add the survey to the supplement, or to the methods section?  How many and what types of health questions did you ask?  This seems like an important part of the study, but the measures are not well described.  What was the basis of your health questionnaire?  Did this come from a pre-existing study, and did you modify it for your population, or was it developed from scratch?  What about the other measures you collected – where did the surveys come from? 

2.1 Partnership-Preparation Phase – who were the community partners, specifically?  If there a many, can you include a table in the supplement?  It would be helpful to know who you collaborated with.

Results

CHAMP-BC is defined again in the first paragraph of both the results and discussion sections.

It seems like there is a lot of data from this study that is not presented in the results. It’s referred to, but it seems vague.  Here are some examples: 

"the key informants voiced that it was difficult to prioritize their primary health concerns; thus all health concerns were recorded (n=39).”  It’s not clear if this means 39 key informants, or 39 health concerns.  How was this data coded, and can it be included in the supplement?  It’s just not clear what happened during this part of the research, but it seems important. Part of the problem is that the first and second paragraphs go together – not sure why they are separate paragraphs.

“Similar to the key informants, the community members had difficulty prioritizing their primary health concerns; thus, to adequately…”  Where is the data for this?  Can it be included in a supplement?  How was it coded? 

How many people do you think attended the community forums?  It’s hard to picture what this was like, but it seems important to the research.  Who attended, and how many? 

You state that “the community forums raised the level of consciousness of individuals…”, but there is not data to back this up.  How did you measure this?  How do you know this “further ignited their interest in changing their cooking behavior”?  Do you have data to back this up, or is this assumed to have happened? 

There are several more instances of this throughout the results, where statements are made, but there is not data to back it up.  This should be fixed throughout.

“During the community readiness survey, community members linked their traditional cooking method, over open fire, to their respiratory health. By self-evaluating their knowledge and beliefs about cooking over open fire and its impact on their health, com-munity members were able to progress toward preparing for action”.  How do you know this?  Is there data to back it up? 

Discussion

CHAMP-BC is defined again in the first paragraph of both the results and discussion sections.

RESPIRE is again defined in the discussion section (previously defined).

The ideas presented in the Discussion seem somewhat superficial, considering the importance of the study. For example, you state earlier in the paper that this is the longest longitudinal study on this topic; however, the discussion does not do a great job of tying your results back to previous work. How long were the other studies conducted for compared to yours?  What factors in your study design were different from what has been done before, and how do these factors help explain the durability of your findings over 7 years?  Just looking at the approach you used, the CBPR, is there any literature showing how CBPR leads to better long-term health or exposure outcomes, not just for air pollution, but for other environmental or public health-related problems.  You conducted a really cool study, but the discussion does not provide an adequate interpretation of your results in the context of previous work on this topic. 

The following sentences are found in the Limitations paragraph.  These do not seem like limitations, but rather strengths of your study:  “To date, cookstove science has largely been atheoretical, which has likely contributed to poor sustainability of improved cooking behaviors overtime. Replication of the CHAMP-BC framework has the potential to move cookstove science forward, particularly in areas similar to rural Nicaragua where previous cookstove research has not effected significant change in the prevalence of traditional open-fire cooking behaviors. By focusing on cooking behaviors through the use of the CHAMP-BC framework, longitudinal improvements in air pollution related symptoms were actualized in the target population for up to seven years.”

Conclusions

CHAMP-BC is defined again here…

Ok, I can see where you have one sentence in the Conclusions about the benefits of CBPR.  This should be discussed more fully higher up in the discussion section, and then maybe reiterated in one or two sentences in your conclusions.  It seems strange to have this as a conclusion from your study when it wasn’t really discussed. 

Author Response

Dear Reviewer,

Thank you for the thoughtful and constructive feedback on our manuscript. We appreciate the time and care taken in reviewing our work. We have revised the manuscript substantially in response to these comments. Below, we provide a point-by-point response and indicate where changes were made in the manuscript and supplementary materials.

Reviewer 1

Highlights

Comment 1

Reviewer comment 1: CHAMP-BC is defined in the third bullet point, and again in the fourth bullet point.  Only needs to be defined the first time it’s used.

Response 1: Thank you for pointing this out. We agree with this comment and have revised the manuscript to reflect the definition of CHAMP-BC in its first occurrence in the highlights while using the acronym thereafter. The revised text now appears in the Highlights section at the beginning of the manuscript.

Comment 2

Reviewer comment 2: The highlights seem unusually long. My recommendation is to condense them down to a single, brief, well-written paragraph.

Response 2: Thank you for this helpful recommendation. While we appreciate the recommendation for a single, brief, well-written paragraph, we were uncertain how to do this while maintaining each of the following requisite parts: relevance, significance, and implications. As such, we elected to condense the Highlights section into shorter bullet points to improve readability and better align with journal expectations. The revised text now appears in the Highlights section at the beginning of the manuscript.

Introduction

Comment 3

Reviewer comment 3: CHAMP-BC is defined as “Community Health Action Model….” at least four times in the first two pages of the manuscript.

Response 3: Thank you for pointing this out. We agree with this comment and have revised the manuscript to reflect the definition of CHAMP-BC in its first occurrence in both the abstract and the body of the paper while using the acronym thereafter.

In response to this and related comments, we reviewed the manuscript and substantially reduced the use of the full definition of CHAMP-BC and replaced this with use of only the acronym. With one exception, we retained the full definition in the title of section 1.7 where CHAMP-BC is described in detail then continued to use the acronym in the body of this section and, again, moving forward. The revisions now appear in the Abstract and throughout the Body of the Paper (except the title of Section 1.7)

Comment 4

Reviewer comment 4: Background – The two paragraphs in this section are hard to follow. For example, it is not clear where the “foundational studies” were conducted without looking at the reference list.  It would help the reader if you could provide more context. Were these foundational studies conducted in communities similar to your study population?

The sentence starting with “At the time of this study in Nicaragua…” is hard to follow. There are several ideas presented in this sentence.  It would help the reader if this was broken up into two or three smaller sentences. The flow of ideas is bumpy in the two paragraphs in section 1.1. There does not appear to be a clear focus for either paragraph. It would help significantly if each paragraph started with a clear focus sentence (your claim), followed by several sentences providing evidence/interpretation to support your claim.  For example, the first paragraph in 1.1 makes the claim that foundational studies show a link between household air pollution and health effects, but this claim is only followed by one very long sentence with one reference to back up the claim. Paragraph 2 in 1.1 seems to restate the same claim as paragraph 1.

Overall, the background section lacks a clear thesis, and the ideas do not flow logically from one idea to another.

Response 4: Thank you for this helpful comment. We substantially revised Section 1.1. Background to improve clarity, coherence, and flow. Specifically, we now identify that the foundational cookstove studies were conducted in rural biomass-fuel-dependent communities, particularly in Guatemala, and explain their relevance to our rural Nicaraguan study population. We also divided the previously long sentence describing the Nicaraguan disease burden into shorter sentences and reorganized the section so that each paragraph begins with a clear focus statement followed by supporting evidence. These revisions can be found in Section 1.1. Background.

Comment 5

Reviewer comment 5: Target Population – this section suffers from some of the same problems identified in 1.1. It seems customary for paragraphs to have closer to 4-6 sentences, starting with a topic sentence, followed by several sentences providing supportive evidence or interpretation. There are several 2-sentence paragraphs in these sections that do not have a clear thesis or logical flow of ideas. It makes it hard to follow. The first three paragraphs in 1.1 and 1.2 only have two sentences each. The supporting information seems to jump around between these paragraphs and does not flow easily from one idea to the next.

As an example, the first paragraph in 1.2 gives a list of reasons why health inequities exist in Nicaragua, namely accessibility (I’m assuming you mean to healthcare), acceptability (not sure what this means), affordability (do you mean to clean fuels?), and access to health promoting materials (not sure what this means).  Based on this claim, I would expect to see several sentences backing up these claims.  However, there is only one sentence of supporting information that focuses on women being vulnerable due to roles/responsibilities as primary cook for the family. So, the first sentence makes a bunch of claims that are not supported in the paragraph. 

The second paragraph in 1.2 has similar problems.

Response 5: Thank you for this thoughtful comment. We revised Section 1.2. Target Population to improve paragraph structure, logical flow, and clarity. Specifically, we reorganized the section so that each paragraph begins with a clear topic sentence followed by supporting evidence and interpretation. We also clarified the structural factors contributing to inequities among women in rural Nicaragua and more explicitly linked those factors to the rationale for selecting the study population. These revisions can be found in Section 1.2. Target Population.

Comment 6

Reviewer comment 6: Behavioral and Environmental Risk Factors – This paragraph mentions factors that seem to be specific to your population. It’s not clear where this information came from.  It would help the reader if this paragraph explained how this information was collected, and from whom, and when. It almost seems like it belongs in the study results.

Response6: Thank you for this helpful observation. We revised Section 1.3. Behavioral and Environmental Risk Factors to clarify that the factors described were identified during the formative community assessment phase of the study. The revised text now specifies that these data were collected from key informants and community members through interviews, community forums, and readiness surveys conducted from 2007 to 2010, and that these findings informed the intervention design and conceptual model. We also added language directing readers to the Methods and Results sections for additional detail. These revisions can be found in Section 1.3. Behavioral and Environmental Risk Factors.

Comment 7

Reviewer comment 7: Literature Review – In this paragraph, can you add a sentence providing some context for the reader. It would be helpful to know how long some of these prior studies were conducted for.  Was it months?  Years?  It’s not clear what you mean by “long-term” if we don’t have some numbers to compare.

Response 7: Thank you for this helpful recommendation. We revised the Literature Review section to provide specific temporal context for prior cookstove studies. The revised text now clarifies that many of the foundational studies cited in this section evaluated outcomes over relatively short periods, generally several months to approximately 18 months, and notes that many intervention studies include two years or less of follow-up. This revision helps clarify what we mean by “long-term” in relation to the present seven-year study. These changes can be found in Section 1.4. Literature Review.

Comment 8

Reviewer comment 8: Indoor Air Pollution – This is another two-sentence paragraph that is hard to follow. After reading the first sentence, I would expect to see several sentences to back up the claim that improved cookstoves improve indoor air quality (and there are many studies that support this). However, there is only one sentence providing supporting information.  I like that the sentence focuses on Guatemala.  The closer the supporting information is to your study population, the better.  The ideas in this paragraph should be more fully developed, with at least 2 – 4 more sentences providing additional supporting information and interpretation.

It would recommend using “Bruce et al. (year)” rather than “Bruce and colleagues…” with no year.

Also, the way the results from Bruce are presented is distracting. There is a direct quote from Bruce et al., rather than a succinct summary of what they found. Did they really measure PM3.5, or is that a typo?  PM3.5 is not very common.  Please provide your own summary of Bruce et al.’s findings, rather than a long, wordy quote. Then, please add several other studies to back up your claim. There is probably a good meta-analysis or two out there that could be referenced here.

This section just seems thin for such a cool longitudinal study that you all conducted. 

Response 8: Thank you for this valuable recommendation. We substantially revised Section 1.4.1. Indoor Air Pollution to improve clarity and depth. Specifically, we replaced the direct quotation from Bruce et al. with a concise summary of the study findings, identified the Guatemalan setting explicitly, and expanded the paragraph to include additional interpretation and supporting evidence from review-level household air pollution intervention literature. The Bruce et al. study used PM 3.5, so we included “fine particulate matter” in our revisions to add clarity while broadly representing smaller particles. We also revised the narrative citation style to use standard author-year phrasing. These revisions can be found in Section 1.4.1. Indoor Air Pollution.

Comment 9

Reviewer comment 9: Improved Health Outcomes – Please change author citations to “Diaz et al. (year)…”; “Smith-Sivertsen et al. (year)”; etc…

There are a lot of direct quotes in this manuscript. The second sentence starting with “At 18 months…”, for example, and again later in the paragraph starting with “For all respiratory…”.  Please summarize findings in your own words, rather than copying from the articles.  Direct quotes should be used sparingly.  Much of the information presented in the direct quotes from Bruce et al. and Diaz et al. can be summarized much more succinctly anyway, making it flow better for the reader.

 

Response 9: Thank you for this helpful recommendation. We revised Section 1.4.2. Improved Health Outcomes to use standard narrative author citations (for example, Diaz et al. (2008) and Smith-Sivertsen et al. (2009)) and removed the direct quotations. We replaced the quoted material with concise summaries of the study findings in our own words to improve readability and flow. These revisions can be found in Section 1.4.2. Improved Health Outcomes.

Comment 10

Reviewer comment 10: Acceptability of Cookstoves – Please change author citation to “Mukhodaphyay et al. (year)…

Ok, this would be a great place to orient the reader to what you mean by “long-term”.  You make a compelling case for cook stove acceptability based on the Mukhodaphyay study (wish I knew when this was published), but there is no mention of how long the participants used the cookstoves before post-measurements were made. That would be really good to know for this paper, since it is the longest study to date on this topic.  Honestly, this theme should run throughout the intro, and hopefully the discussion section of the paper. It would really help the reader if the issue of time was clear throughout.  As I written, I don’t know what I’m comparting your study to.  Also, another use of a direct quote here.

Also, the first paragraph in this section only uses one study as evidence. This seems like a really big issue that is very relevant to your paper, so I’m surprised to see only one paper to back up your claim.  Cookstove acceptability is a huge issue – is there not a meta-analysis on this, or a good review paper that can be cited here, or at least some other studies with participants similar to yours?

Response 10: Thank you for this insightful recommendation. We substantially revised Section 1.4.3. Acceptability of Cookstoves to improve context and strengthen the literature synthesis. Specifically, we changed the narrative citation style, removed the direct quotation, and added Central American evidence from studies conducted in Guatemala and Honduras to better align the literature review with the geographic and cultural context of our study. We also clarified that many prior acceptability studies evaluated stove use over relatively short periods or cross-sectional assessments, which helps define what we mean by “long-term” in relation to our seven-year follow-up. These revisions can be found in Section 1.4.3. Acceptability of Cookstoves.

Materials and Methods

Comment 11

Reviewer comment 11: It’s hard to see in the methods where your survey instruments are described. For example, in the implementation action phase you state that baseline data was collected.  Can you add the survey to the supplement, or to the methods section?  How many and what types of health questions did you ask?  This seems like an important part of the study, but the measures are not well described.  What was the basis of your health questionnaire?  Did this come from a pre-existing study, and did you modify it for your population, or was it developed from scratch?  What about the other measures you collected – where did the surveys come from?

Response 11: Thank you for this important comment. We agree that the original manuscript did not describe the study instruments in sufficient detail. In response, we revised the Methods section to more explicitly describe the interviewer-administered baseline and follow-up household surveys, including the health symptom measures used in the study. We now clarify that the household surveys were adapted for the study context and included items on household characteristics, fuel use and acquisition, kitchen and ventilation characteristics, smoking exposure, stove use, and women’s and children’s health and well-being. We also specify that the health section included open-ended questions regarding perceived effects of smoke exposure and prompted symptom items related to eyes, cough, chest illness, shortness of breath, and headache. To improve transparency while maintaining a concise supplement, we added Supplementary Table S1, which summarizes all study instruments, their domains, sample items, response formats, and the study phase in which each was administered. These revisions can be found in Sections 2.1, 2.2, 2.4, 2.5, and 2.6, and in Supplementary Table S1.

Comment 12

Reviewer comment 12: 2.1 Partnership-Preparation Phase – who were the community partners, specifically? If there are many, can you include a table in the supplement? It would be helpful to know who you collaborated with.

Response 12: Thank you for this helpful suggestion. In response, we added Supplementary Table S2, which summarizes the characteristics of the key informants involved in the partnership-preparation phase, including age range, sex, educational background, and occupational representation. We also added text in Section 2.1 directing readers to this table and clarifying the range of community partners involved in the early partnership-preparation work. These revisions can be found in Section 2.1. and 3.1 Partnership-Preparation Phase and Supplementary Table S2.

Results

Comment 13

Reviewer comment 13: It seems like there is a lot of data from this study that is not presented in the results. It’s referred to, but it seems vague… Here are some examples:

Response 13: Thank you for this important comment. We agree that the original Results section did not always clearly distinguish between the number of participants interviewed and the number of responses recorded, and that several passages were written more strongly than the formative data could support. In response, we revised Section 3.1 to clarify the formative assessment findings, distinguish participant counts from response counts, and improve the flow of the opening paragraphs. We also revised language throughout Section 3 to better distinguish measured findings from theory-informed interpretations within the CHAMP-BC framework. To improve transparency, we added supplementary tables summarizing both participant characteristics and response distributions from the formative phases. Specifically, Supplementary Table S2 summarizes the characteristics of the key informants, Supplementary Table S3 presents the primary health concerns identified by key informants, Supplementary Table S4 summarizes the characteristics of the community health needs assessment participants, and Supplementary Table S5 presents the primary health concerns identified by community members as counts and percentages of total responses. The manuscript now reports the original summary findings while using the supplement to provide clearer descriptive context. These revisions can be found in Sections 3.1 Partnership-Preparation Phase, 2.6 Data Collection and Analysis Plan, and in Supplementary Tables S2–S5.

Comment 14, Example 1

Reviewer comment 14.1: “the key informants voiced that it was difficult to prioritize their primary health concerns; thus all health concerns were recorded (n=39).” It’s not clear if this means 39 key informants, or 39 health concerns. How was this data coded, and can it be included in the supplement? It’s just not clear what happened during this part of the research, but it seems important. Part of the problem is that the first and second paragraphs go together – not sure why they are separate paragraphs.

Response 14.1: Thank you for this helpful comment. We agree that the original text did not clearly distinguish between the number of key informants interviewed and the number of health-concern responses recorded. In response, we revised Section 3.1 to clarify that 13 key informants were interviewed and that 39 total health-concern responses were recorded across those interviews. We also combined and revised the opening paragraphs of Section 3.1 to improve flow and clarify how the key informant findings informed the subsequent community health needs assessment. To improve transparency, we added Supplementary Table S2, which summarizes the characteristics of the key informants, and Supplementary Table S3, which presents the primary health concerns identified by key informants as counts and percentages of total responses. The manuscript now reports the original summary findings and provides structured supplementary tables to clarify participant characteristics and response distributions. These revisions can be found in Section 3.1. Partnership-Preparation Phase, Section 2.6. Data Collection and Analysis Plan, and Supplementary Tables S2 and S3.

Comment 14, Example 2

Reviewer comment 14.2: “Similar to the key informants, the community members had difficulty prioritizing their primary health concerns thus to adequately…” Where is the data for this? Can it be included in a supplement? How was it coded?

Response 14.2: Thank you for this helpful comment. We agree that the original text did not clearly explain what these data represented. In response, we revised Section 3.1 to clarify that the community health needs assessment included 77 individuals from 47 households and that 186 total health-concern responses were recorded across those interviews. We now specify that respondents often identified more than one concern, so all concerns mentioned were recorded as responses rather than limiting participants to a single priority concern. To improve transparency, we added Supplementary Table S4, which summarizes the characteristics of the community health needs assessment participants, and Supplementary Table S5, which presents the primary health concerns identified by community members as counts and percentages of total responses. The manuscript now reports the original summary findings from this formative phase and provides structured supplementary tables to clarify the participant characteristics and response distributions. These revisions can be found in Section 3.1. Partnership-Preparation Phase, Section 2.6. Data Collection and Analysis Plan, and Supplementary Tables S4 and S5.

Comment 14, Example 3

Reviewer comment 14.3: How many people do you think attended the community forums? It’s hard to picture what this was like, but it seems important to the research. Who attended, and how many? You state that “the community forums raised the level of consciousness of individuals…”, but there is not data to back this up… How did you measure this?  How do you know this “further ignited their interest in changing their cooking behavior”?  Do you have data to back this up, or is this assumed to have happened.

Response 14.3: Thank you for this thoughtful comment. We agree that the original manuscript described the community forum too briefly and used language that could be interpreted as implying direct measurement of forum-related behavioral change processes. In response, we revised Section 3.1 to describe the forum more concretely and cautiously. We now specify that 32 community members attended the March 2010 community forum and clarify that the forum served as a setting in which community members shared and discussed perceived links between cooking over open fire and health symptoms, as well as possible local resources to address smoke exposure. We also revised the text to avoid implying that constructs such as increased consciousness or increased interest in behavior change were directly measured outcomes of the forum itself. Instead, these discussions are now described more conservatively and interpreted within the CHAMP-BC framework rather than presented as empirically measured forum outcomes. These revisions can be found in Section 3.1. Partnership-Preparation Phase and Section 2.6. Data Collection and Analysis Plan.

Comment 14, Example 4

Reviewer comment 14.4: There are several more instances of this throughout the results, where statements are made, but there is not data to back it up. This should be fixed throughout.

Response 14.4: Thank you for this important observation. We agree that several statements in the original Results section were written too strongly and could be interpreted as implying direct empirical support where the underlying data were descriptive, theory-informed, or not retained in a form that supported additional reanalysis. In response, we conducted a manuscript-wide review of the Results section and revised wording throughout to better distinguish measured findings from interpretive statements. Specifically, we removed or softened unsupported causal language, clarified when statements were being interpreted within the CHAMP-BC framework, and revised text so that the Results section more closely reflects the data available for analysis. These revisions were made throughout Section 3 and, where appropriate, were complemented by clarifying language in Section 2.6. Data Collection and Analysis Plan.

Comment 14, Example 5

Reviewer comment 14.5: “During the community readiness survey, community members linked their traditional cooking method, over open fire, to their respiratory health. By self-evaluating their knowledge and beliefs about cooking over open fire and its impact on their health, community members were able to progress toward preparing for action.” How do you know this? Is there data to back it up?

Response 14.5: Thank you for this important comment. We agree that the original wording in Section 3.2 was too strong and could be interpreted as implying direct measurement of progression in stage of change. In response, we revised this passage to distinguish more clearly between what was directly supported by the community readiness survey and what was interpreted within the CHAMP-BC framework. The revised text now states that community members discussed the relationship between open-fire cooking and health during the readiness assessment and that these responses were interpreted, within the theoretical framework, as being consistent with readiness for action rather than as directly measured evidence of progression to a new stage of change. We also clarified the description of the community readiness survey in the Methods section and summarized the instrument in Supplementary Table S1. Because detailed item-level records from the formative assessments were not retained in a form that supported additional reanalysis, the manuscript reports the original summary findings. These revisions can be found in Sections 2.2, 3.2, and Supplementary Table S1.

Discussion

Comment 15

Reviewer comment 15: The ideas presented in the Discussion seem somewhat superficial, considering the importance of the study. For example, you state earlier in the paper that this is the longest longitudinal study on this topic; however, the discussion does not do a great job of tying your results back to previous work. How long were the other studies conducted for compared to yours?  What factors in your study design were different from what has been done before, and how do these factors help explain the durability of your findings over 7 years?  Just looking at the approach you used, the CBPR, is there any literature showing how CBPR leads to better long-term health or exposure outcomes, not just for air pollution, but for other environmental or public health-related problems.  You conducted a really cool study, but the discussion does not provide an adequate interpretation of your results in the context of previous work on this topic?

Response 15: Thank you for this thoughtful and constructive comment. We agree that the original Discussion did not sufficiently situate our findings within the broader cookstove and participatory public health literature. In response, we substantially strengthened the Discussion in three ways. First, we now compare the duration of prior women’s symptom outcome studies more precisely with the present study. Specifically, we clarify that among the peer-reviewed studies we identified that followed women’s symptoms or symptom-related health outcomes after improved cookstove intervention, follow-up generally extended from approximately 1 year to 18 months, whereas the present study followed women for up to seven years. Second, we revised the manuscript to avoid an overly broad “longest study in the world” claim and instead use a more precise formulation. The revised text now states that to our knowledge, this study is among the longest longitudinal improved cookstove studies to report sustained symptom outcomes among women, with follow-up extending up to seven years. Third, we expanded the Discussion to better interpret what may explain the durability of our findings and added discussion of broader CBPR literature to support the interpretation that participatory approaches may strengthen intervention relevance, trust, uptake, and long-term sustainability in environmental and public health interventions. These revisions can be found in Section 4. Discussion, with related wording updates in the Introduction and Conclusion to ensure consistency.

Comment 16

Reviewer comment 16: The following sentences are found in the Limitations paragraph.  These do not seem like limitations, but rather strengths of your study:  “To date, cookstove science has largely been atheoretical, which has likely contributed to poor sustainability of improved cooking behaviors overtime. Replication of the CHAMP-BC framework has the potential to move cookstove science forward, particularly in areas similar to rural Nicaragua where previous cookstove research has not effected significant change in the prevalence of traditional open-fire cooking behaviors. By focusing on cooking behaviors through the use of the CHAMP-BC framework, longitudinal improvements in air pollution related symptoms were actualized in the target population for up to seven years.”

Response 16: Thank you for this helpful observation. We agree that these statements were misplaced in the Limitations section. In response, we removed this language from Section 4.1. Limitations and relocated its core ideas to the Discussion and Conclusion, where they are more appropriately framed as interpretation of the study’s contribution rather than as study limitations. We also revised the wording to ensure that claims about the CHAMP-BC framework are presented cautiously and in alignment with the data. These revisions can be found in Section 4.1. Limitations and the revised Discussion and Conclusion.

Conclusions

Comment 17

Reviewer comment 17: Ok, I can see where you have one sentence in the Conclusions about the benefits of CBPR. This should be discussed more fully higher up in the discussion section, and then maybe reiterated in one or two sentences in your conclusions. It seems strange to have this as a conclusion from your study when it wasn’t really discussed.

Response 17: Thank you for this helpful comment. We agree that the original manuscript mentioned the potential benefits of CBPR too briefly in the Conclusions without fully developing this interpretation in the Discussion. In response, we expanded the Discussion to more fully address the potential contribution of CBPR to the durability of the intervention, drawing on both the design features of the present study and the broader participatory public health literature. We now discuss how formative community engagement, shared decision-making, local stove selection, use of local masons, household material contributions, and sustained community involvement may have supported acceptability, ownership, and longer-term use of the improved cookstoves. We also revised the Conclusions so that the statements about CBPR are now brief and consistent with the fuller discussion presented earlier in the manuscript. These revisions can be found in Section 4. Discussion and Section 5. Conclusions.

Supplementary Materials Added

• Supplementary Table S1. Summary of survey instruments, domains, sample items, response format, and study phase in which each measure was administered.
• Supplementary Table S2. Characteristics of key informants involved in the partnership-preparation phase.
• Supplementary Table S3. Key informant interview findings: primary health concerns identified during the partnership-preparation phase.
• Supplementary Table S4. Characteristics of community health needs assessment participants.
• Supplementary Table S5. Community health needs assessment findings: primary health concerns identified during the partnership-preparation phase.

We thank you, again, for the thoughtful feedback, which strengthened the manuscript substantially.

Reviewer 2 Report

Comments and Suggestions for Authors

The authors could significantly improve the introduction and literature review sections and better position their work within the pre-existent body of work. Several statements the authors make could benefit from a more careful analysis. For instance:

1. While it is generally true that “household air pollution from solid fuel and kerosene use is responsible for…”, solid fuels are significantly more dangerous to human health than kerosene use. The statement as written is misleading. For instance, Kim et al. (2023) show clear proof that the use of kerosene stoves significantly reduce stunting in Indonesian children, relative to the use of solid fuels.
2. There is a much larger literature out there on the effects of air pollutants on human health. Air pollution is responsible for way more than acute respiratory infections and COPD. Also there are also studies that show way more than simple associations. I recommend the authors to perform a thorough literature review and specifically focus on studies that use causal inference methods and that estimate true causal effects of pollution on human health. Papers such as Jayachandran (2009), Kim et al. (2017, 2020), etc. should be mentioned in the literature review section as documentations of true causal impacts on several dimensions of health.
3. The authors mention the combined effects of poverty and household air pollution without acknowledging the likely linkage between these two. It is likely that poverty itself is a leading cause for using solid fuels in the first place.
4. The claim that “women are the most vulnerable…” needs to be explained more. The mere fact that women are usually the ones to cook is not enough proof. There are studies who show that children are actually the most vulnerable.
5. The “literature review” section (1.4) is extremely inadequate. The reader cannot properly assess where this study fits in the literature. The authors need to significantly expand and improve this section and be clear on the position of their study within the larger literature. Especially section 1.4.2 that refers to improved health outcomes is severely lacking with many important studies missing from the discussion. It seems the authors only looked at a handful of studies from neighboring countries, but the discussion should be larger and particularly focused on studies that document objective measures of health in larger samples, not just self-rated health.
6. As it relates to the “acceptability of cookstoves” the authors need to explain what are the barriers that prevent the wide acceptability of these stoves over the long-term. Without a proper discussion, it is hard to understand why certain interventions such as those used by the authors are appropriate. It seems the authors focus primarily on psychological channels, but arguably the socio-economic barriers are way more important here.

 

The Methods section is not sufficiently detailed:

1. The authors mention Nicaragua in the introduction and “community partners” in the methods section, but it is not clear how many communities, what kind of communities, how were these communities chosen, etc. Is this a nationally representative sample of communities? Or maybe just a rural community? What is the scope of the study? It is not clear at all.
2. It is similarly not clear who answered some of the surveys the authors refer to. For instance the community readiness survey in 2010. Was that a household level survey? A community survey?
3. The “raffle” mechanism needs to be explained in detail. Since this is the basis of the randomization process, the authors need to explain precisely how the raffle was organized.
4. What was the basis on which the “convenience sample” was chosen? The idea of a “sample” (which should be random) is not consistent with the idea of “self-selection”. If this sample was indeed random, what do the authors mean by self selection. If the sample was chosen based on people volunteering into the program, the entire study is null.

 

The Results section presents several important issues:

1. I have significant reservations regarding the sample size which is too small to allow for robust scientific conclusions.
2. Although it is not entirely clear, it seems to me that the authors simply compare outcomes before and after the stove adoption and the entire sample of households had access to a stove. This empirical setup does not allow one to draw conclusions regarding the effects of the community health action model, which the authors claim as their main contribution. The effects captured here could simply be the effects of having access to a better stove, or they could simply be due to improvements in socio-economic factors over time that lie completely outside of the stove adoption decision. Why don’t the authors compare the “treated group” (those who were randomly given a stove) against the “control group” (those who were not given one)?
3. Why aren’t there any controls included in the analysis? It seems there was enough data collected in the baseline surveys to control for different demographic and socio-economic confounders. Why do the authors not use this data and simply report symptoms over time?
4. Overall, the claimed results cannot be justified without a more careful data analysis.

Author Response

Dear Reviewer,

Thank you for the thoughtful and constructive feedback on our manuscript. We appreciate the time and care taken in reviewing our work. We have revised the manuscript substantially in response to these comments. Below, we provide a point-by-point response and indicate where changes were made in the manuscript and supplementary materials.

Reviewer 2

Introduction

Comment 1

Reviewer comment 1: While it is generally true that “household air pollution from solid fuel and kerosene use is responsible for…”, solid fuels are significantly more dangerous to human health than kerosene use. The statement as written is misleading. For instance, Kim et al. (2023) show clear proof that the use of kerosene stoves significantly reduce stunting in Indonesian children, relative to the use of solid fuels.

Response 1: Thank you for this comment. We agree that the sentence could be phrased more precisely. Our study focuses on women using traditional biomass cookstoves in rural Nicaragua, where wood is the dominant household cooking fuel. To improve accuracy and avoid implying identical risk across fuel types, we revised the Introduction to refer more broadly to polluting household fuels while retaining the focus of the manuscript on biomass-based household air pollution. This revision appears in Section 1. Introduction.

Comment 2

Reviewer comment 2: There is a much larger literature out there on the effects of air pollutants on human health. Air pollution is responsible for way more than acute respiratory infections and COPD. Also there are also studies that show way more than simple associations. I recommend the authors to perform a thorough literature review and specifically focus on studies that use causal inference methods and that estimate true causal effects of pollution on human health. Papers such as Jayachandran (2009), Kim et al. (2017, 2020), etc. should be mentioned in the literature review section as documentations of true causal impacts on several dimensions of health.

Response 2: Thank you for this comment. We agree that the health effects of household air pollution extend beyond acute respiratory infections and chronic obstructive pulmonary disease. Current framing in the Introduction reports household air pollution is the leading environmental contributor to noncommunicable diseases and contributes to millions of deaths annually. However, the aim of the present manuscript is not to provide a comprehensive review of all air pollution health effects or of the broader causal-inference literature on pollution and health. Rather, the manuscript is focused on a specific household air pollution intervention in rural Nicaragua and on symptom outcomes relevant to women using traditional open fire cookstoves. In light of the scope of the study, we retained a focused literature review centered on household air pollution, improved cookstoves, and women’s symptom outcomes.

Comment 3

Reviewer comment 3: The authors mention the combined effects of poverty and household air pollution without acknowledging the likely linkage between these two. It is likely that poverty itself is a leading cause for using solid fuels in the first place.

Response 3: Thank you for this important comment. We agree that the relationship between poverty and household air pollution should be stated more explicitly. In response, we revised Section 1.2. Target Population to clarify that poverty is not only a contextual inequity associated with women’s exposure, but also a likely driver of continued reliance on wood and other polluting household fuels because cleaner fuels and improved cooking technologies may be financially inaccessible or inconsistently available in rural settings. This revision strengthens the conceptual linkage between poverty and household air pollution in the Introduction and is in Section 1.2 Target Population.

Comment 4

Reviewer comment 4: The claim that “women are the most vulnerable…” needs to be explained more. The mere fact that women are usually the ones to cook is not enough proof. There are studies who show that children are actually the most vulnerable.

Response 4: Thank you for this comment. In the revised manuscript we no longer state that women are “the most vulnerable.” Instead, Section 1.2 now states that women in rural Nicaragua are disproportionately exposed to household air pollution because of gendered cooking roles and that, in many rural households, women are the primary cooks and therefore bear a significant burden of daily smoke exposure. This wording was intentionally revised to avoid an overly absolute claim while maintaining the manuscript’s focus on the study population. Because the present study specifically examined adult women as the target population for the intervention, we retained this exposure-focused framing in the revised manuscript in the Introduction and is in Section 1.2 Target Population.

Comment 5

Reviewer comment 5: The “literature review” section (1.4) is extremely inadequate. The reader cannot properly assess where this study fits in the literature. The authors need to significantly expand and improve this section and be clear on the position of their study within the larger literature. Especially section 1.4.2 that refers to improved health outcomes is severely lacking with many important studies missing from the discussion. It seems the authors only looked at a handful of studies from neighboring countries, but the discussion should be larger and particularly focused on studies that document objective measures of health in larger samples, not just self-rated health.

Response 5: Thank you for this comment. In the revised manuscript, we substantially strengthened this section to better position the study within the relevant cookstove and household air pollution literature. Specifically, we added clearer temporal context regarding prior follow-up periods, expanded the discussion of indoor air pollution, improved health outcomes, and cookstove acceptability, and incorporated additional evidence from studies conducted in settings that are geographically, culturally, and exposure-relevant to our study population, including Central America. The revised text now also distinguishes our study by emphasizing its focus on women’s symptom outcomes over up to seven years, which, to our knowledge, is among the longest longitudinal improved cookstove studies to report sustained symptom outcomes among women. We also note that the purpose of our literature review is not to provide an exhaustive review of all air pollution and health literature, nor all possible cookstove health endpoints, but rather to situate the present manuscript within the literature most relevant to its research question: a longitudinal, community-based participatory improved cookstove intervention in women using traditional open-fire cookstoves in rural Nicaragua. Accordingly, we prioritized literature on household air pollution, improved cookstove interventions, women’s health and symptom outcomes, sustained stove use, and participatory implementation. With regard to Section 1.4.2, we acknowledge that objective health outcomes are important in the broader literature. However, the present study specifically evaluates self-reported symptom outcomes in the target population, and the revised literature review reflects that focus while still situating the work in relation to prior intervention studies on improved health outcomes. We therefore believe the revised Section 1.4 Literature Review is appropriately scoped for the aims of the manuscript and adequately positions the study within the most relevant literature to the current study.

Comment 6

Reviewer comment 6: As it relates to the “acceptability of cookstoves” the authors need to explain what are the barriers that prevent the wide acceptability of these stoves over the long-term. Without a proper discussion, it is hard to understand why certain interventions such as those used by the authors are appropriate. It seems the authors focus primarily on psychological channels, but arguably the socio-economic barriers are way more important here.

Response 6: Thank you for this comment. In the revised manuscript, Section 1.4.3. Acceptability of Cookstoves was substantially expanded to address the broader literature on long-term adoption and sustained use. This section now explicitly discusses barriers beyond user perceptions, including household, social, and economic factors, and cites review-level evidence indicating that improved cookstove acceptance is influenced not only by stove performance, but also by cost, compatibility with household needs, fuel practices, and broader socio-economic context. In addition, the manuscript now links these broader barriers directly to the rationale for the present intervention. Section 1.2. Target Population explains that women’s exposure is shaped by poverty, geographic isolation, and limited access to cleaner household technologies, and notes that women may have few realistic options for reducing exposure without community-based, affordable, and locally acceptable interventions. The revised Discussion also explains that the durability of our findings may relate not only to the stove itself, but also to the intervention design, including formative assessment with community members, local selection of stove design, use of local masons, and shared household material contributions, all of which speak directly to socio-economic feasibility and local acceptability. For these reasons, we believe the revised manuscript addresses the reviewer’s concern and appropriately explains why the intervention used in this study was suitable for this setting in Section 1.4.3. Acceptability of Cookstoves, Section 1.2. Target Population, and the Discussion.

Materials and Methods

Comment 7

Reviewer comment 7: The authors mention Nicaragua in the introduction and “community partners” in the methods section, but it is not clear how many communities, what kind of communities, how were these communities chosen, etc. Is this a nationally representative sample of communities? Or maybe just a rural community? What is the scope of the study? It is not clear at all.

Response 7: Thank you for this comment. We agree that study scope and setting should be described clearly. We note that this issue has been addressed in the revised manuscript. Specifically, Section 1.2. Target Population now states that the intervention was conducted among women aged 18 years and older living in a small rural village in the Rivas Political Department of Nicaragua, and Section 3 clarifies that the Results describe cooking behavior change in one small community in rural Nicaragua. We also clarify in Section 2.1. Partnership-Preparation Phase that community partners included local leaders, health personnel, educators, community members, and local technical resources involved in cookstove planning and implementation, with additional detail provided in Supplementary Table S2. Accordingly, the revised manuscript makes clear that this was a community-based study in a single rural community, not a nationally representative sample of communities. The revisions can be found in Section 1.2. Target Population, Section 2.1 Partnership-Preparation Phase, Section 3 Results, and Supplementary Table S2.

Comment 8

Reviewer comment 8: It is similarly not clear who answered some of the surveys the authors refer to. For instance the community readiness survey in 2010. Was that a household level survey? A community survey?

Response 8: Thank you for this comment. We agree that it is important to clarify who completed the formative surveys and what level of assessment they represented. We note that this issue has been addressed in the revised manuscript. Specifically, Section 2.2. Assessment-Contemplation Phase now states that the community readiness survey was completed via face-to-face interviews in March 2010 and was designed to assess community perspectives related to indoor air pollution across six domains: community efforts, community knowledge of efforts, leadership, community climate, knowledge about the issue, and resources for prevention efforts. This wording was intended to clarify that the community readiness survey was a community-level formative assessment, not a household baseline or follow-up outcome survey. In addition, the instrument is summarized in Supplementary Table S1, and its content further reflects a community-level orientation. Revisions can be found in Section 2.2. Assessment-Contemplation Phase and Supplementary Table S1.

Comment 9

Reviewer comment 9: The “raffle” mechanism needs to be explained in detail. Since this is the basis of the randomization process, the authors need to explain precisely how the raffle was organized.

Response 9: Thank you for this comment. We have clarified in the revision that the raffle was not the basis of allocation of stoves in the sense of an RCT. Rather, the raffle was used to randomize the order in which community members received a stove.  As described in the revised manuscript, community members reached consensus that if there was insufficient funding for all households to receive an improved cookstove, a community raffle would be an equitable method of allocating available cookstoves among interested households. Section 2.3. Planning-Preparation Phase explains that the raffle was held after community consensus and was used to distribute improved cookstoves across implementation cohorts, over time. We believe this clarifies the confusion about the intent and purpose of the raffle. We also note that the revised manuscript clarifies this point in the Methods and Results sections. Specifically, Section 2.3. Planning-Preparation Phase explains that the raffle was held after community consensus and was used to distribute improved cookstoves across implementation cohorts, while Section 2.6. Data Collection and Analysis Plan states that participants self-selected into the improved cookstove research program and were then selected by population distribution into groups during the community raffles. Because the study did not use randomization as part of an experimental trial design, we did not add further detail framing the raffle as a randomization procedure. These revisions are located in the revised manuscript as follows: Methods, Section 2.3. Planning-Preparation Phase, Section 2.6. Data Collection and Analysis Plan, and Results.

Comment 10

Reviewer comment 10: What was the basis on which the “convenience sample” was chosen? The idea of a “sample” (which should be random) is not consistent with the idea of “self-selection”. If this sample was indeed random, what do the authors mean by self-selection. If the sample was chosen based on people volunteering into the program, the entire study is null.

Response 10: Thank you for this comment. We respectfully disagree with the assertion that self-selection renders the study “null.” The present study was explicitly designed and reported as a longitudinal cohort study, not as a randomized controlled trial or a probability-based population survey. We were able to allocate improved cookstoves to 83% of the households; which would not have had the power to test an RCT in the remaining 17% of the households in the community. To maintain trust within the community, it would not have been ethical for us to randomize fewer than our max capacity when implementing an evidence-based intervention. As such, the term convenience sample refers to the way participants were initially recruited during home visits in the study community, whereas self-selection refers to the subsequent decision of eligible community members to participate in the improved cookstove program. These are not contradictory concepts in an observational, community-based intervention study. Rather, they describe two different aspects of recruitment and participation: initial identification/recruitment of potential participants and voluntary enrollment into the program. The revised manuscript clarifies this design. Specifically, Section 2.6. Data Collection and Analysis Plan states that “A convenience sample was recruited during home visits from October 2007 to March 2010 and subsequently self-selected into the improved cookstove research program,” and Section 4.1. Limitations explicitly acknowledge the resulting risk of selection bias due to our cohort design and the limits to generalizability beyond the target population. In other words, the manuscript does not present the study as randomized, nationally representative, or free from self-selection effects. Instead, it transparently identifies these features as limitations of the observational cohort design. We therefore do not agree that voluntary participation invalidates the study. Self-selection is common in community-based participatory and implementation research, particularly in real-world interventions where community engagement, acceptability, and feasibility are central to the research question. While self-selection may introduce selection bias and limits causal inference, it does not negate the value of longitudinal within-cohort findings on symptom trends, program feasibility, and sustained use over time. For this reason, we retained the study’s characterization as a longitudinal cohort study with a convenience sample and voluntary program participation, while also clearly acknowledging the associated limitations.

Results

Comment 11

Reviewer comment 11: I have significant reservations regarding the sample size which is too small to allow for robust scientific conclusions.

Response 11: Thank you for this comment. We respectfully disagree that the sample size is too small to support meaningful scientific conclusions in the context of the present study design. This study was conducted as a longitudinal cohort study in a single rural community, not as a large population-based survey or randomized controlled trial. The analytic sample included 167 participants at baseline, with follow-up data available for 147 participants at one year, 80 at three years, 38 at five years, and 19 at seven years, allowing evaluation of symptom trends over multiple post-intervention time points. The manuscript does not claim broad national representativeness or definitive causal inference; rather, it reports longitudinal findings from a community-based intervention and interprets them within the scope of that design. Importantly, despite the modest cohort size, the study identified statistically significant improvements over time in cough, chest illness, shortness of breath, headache, and eye irritation using the Cochran-Armitage Trend Test, which is a conservative, non-parametric test appropriate for small sample sizes. Moreover, from a statistical point of view, low power is a consequence of type 2 error, which is not a concern in the presence of statistical significance. In the presence of statistical significance, we are concerned with type 1 error, which we controlled at an alpha level of 0.05. Thus, our findings suggest that the sample was sufficient to detect meaningful longitudinal symptom patterns within the study population. At the same time, we explicitly acknowledge in Section 4.1. Limitations that the use of a convenience sample limits generalizability beyond the target population and that the cohort design does not permit the same level of inference as a randomized study. We therefore do not agree that the sample size renders the study scientifically invalid. Rather, the study should be interpreted appropriately as a community-based longitudinal cohort study that contributes evidence on feasibility, sustained use, and symptom outcomes over time in a real-world rural setting. The manuscript frames the findings within those boundaries and acknowledges the associated limitations.

Comment 12

Reviewer comment 12: It is not entirely clear, it seems to me that the authors simply compare outcomes before and after the stove adoption and the entire sample of households had access to a stove. This empirical setup does not allow one to draw conclusions regarding the effects of the community health action model, which the authors claim as their main contribution. The effects captured here could simply be the effects of having access to a better stove, or they could simply be due to improvements in socio-economic factors over time that lie completely outside of the stove adoption decision. Why don’t the authors compare the “treated group” (those who were randomly given a stove) against the “control group” (those who were not given one)?

Response 12: Thank you for this comment. We agree that the present empirical design does not permit definitive causal attribution of the observed symptom improvements specifically to the CHAMP-BC framework, independent of the improved cookstove itself or of other changes over time. We believe the revised manuscript acknowledges this limitation clearly. In the revised Discussion and Limitations sections, we explicitly state that although the participatory and theory-informed design of the intervention may help explain the durability of the findings, the study was not designed to isolate the independent contribution of CHAMP-BC. We state that, because all participants were drawn from the same geographical area and there was no unexposed comparison group with equivalent demographic and geographic characteristics, the study could not rule out secular changes over time. We also respectfully note that the reviewer’s characterization of a “treated group” and “control group” is not consistent with the design of this study. This was a longitudinal cohort study, not a randomized controlled trial. Participants self-selected into the improved cookstove program, and the raffle was a community-based allocation process among interested households rather than a trial randomization procedure designed to create experimental treatment and control groups. The revised Methods and Results sections clarify that women self-selected into the raffle, that households were then selected during community raffles across implementation cohorts, and that baseline measurements for cookstove recipients served as the within-subject comparison over time. Accordingly, the manuscript does not claim that the study demonstrates the isolated causal effect of CHAMP-BC. Rather, the contribution of the study is to report longitudinal symptom trends, sustained use patterns, and program implementation findings from a community-based, theory-informed cookstove intervention conducted in a real-world rural setting. In that context, we interpret CHAMP-BC as the guiding implementation framework for the intervention, not as an independently tested causal exposure. We therefore believe the revised manuscript addresses this concern appropriately by clearly distinguishing between the observed longitudinal findings and the limits of causal inference. For these reasons, we do not believe that framing the study as a treated-versus-control comparison would accurately reflect the actual design, and we have therefore retained the current observational cohort framing while explicitly acknowledging the corresponding inferential limitations. Please see the indicated revisions in the Methods, Results, Discussion and Limitations.

Comment 13

Reviewer comment 13: Why aren’t there any controls included in the analysis? It seems there was enough data collected in the baseline surveys to control for different demographic and socio-economic confounders. Why do the authors not use this data and simply report symptoms over time?

Response 13: Thank you for this important comment. We agree that the baseline survey collected demographic and household information that is important for describing the study population. However, the purpose of the present analysis was to evaluate longitudinal symptom patterns within the cohort over time, rather than to estimate an adjusted causal treatment effect in a comparative multivariable model. Consistent with that aim, the manuscript reports symptom trends across repeated post-implementation time points (T0–T4) using the Cochran-Armitage Trend Test, which was selected to assess linear trends in binomial symptom proportions over time. The analysis therefore focused on whether symptoms changed across follow-up periods within the observed cohort, rather than on modeling between-group differences or adjusted treatment effects. We also respectfully note that the study design limits the role that baseline covariate adjustment could play in resolving the reviewer’s broader concern. This was a longitudinal cohort study in which participants self-selected into the improved cookstove program, and there was no unexposed comparison group with equivalent demographic and geographic characteristics. As acknowledged in the revised Limitations section, the cohort design does not allow the study to rule out secular changes over time, and baseline values for cookstove recipients functioned as the within-subject comparison rather than as a randomized or matched control group. Under these circumstances, including demographic or socio-economic covariates in an adjusted model would not overcome the fundamental design limitation of the absence of a true comparison group. Accordingly, we chose to present the analysis transparently as a descriptive longitudinal cohort analysis of symptom trends over time, and we framed the findings within those inferential limits. The revised manuscript acknowledges that the observational design limits causal inference and generalizability beyond the target population. We therefore believe the current analytic approach is appropriate for the study design and research question. For these reasons, we did not add adjusted covariate models, because such models would not address the primary inferential limitation of the study design and could imply a level of causal control that the data do not support.

Comment 14

Reviewer comment 14: Overall, the claimed results cannot be justified without a more careful data analysis.

Response 14: Thank you for this important comment. We respectfully disagree that the reported results “cannot be justified” without a different or more complex data analysis. The current manuscript presents the study transparently as a longitudinal cohort study in a single rural community, not as a randomized controlled trial or a causal effect evaluation. The analytic approach was selected to match that design and the specific research question. As described in Section 2.6. Data Collection and Analysis Plan, symptom outcomes were analyzed across repeated follow-up time points (T0–T4) using the Cochran-Armitage Trend Test, an appropriate nonparametric method for assessing linear trends in binomial proportions over ordered time points. This method directly addresses the question posed in the manuscript: whether symptom patterns changed over time within the observed cohort after implementation of the improved cookstove program. We also note that the revised manuscript does not claim that the analysis establishes definitive causal effects of the cookstove intervention or of the CHAMP-BC framework independent of other factors. The revised Discussion and Limitations sections explicitly acknowledge that the observational cohort design, self-selection into the program, and absence of an unexposed comparison group limit causal inference and do not allow the study to rule out secular changes over time. The manuscript therefore frames the findings appropriately as longitudinal symptom trends, sustained use patterns, and implementation findings observed within a real-world community-based intervention. Accordingly, we believe the current analysis is appropriate for the scope and design of the study, and that the revised manuscript presents the findings with the necessary caution regarding interpretation. We therefore did not make additional analytic revisions in response to this comment but did revise the manuscript to add clarity in the Discussion and Limitations sections. 

Supplementary Materials Added

• Supplementary Table S1. Summary of survey instruments, domains, sample items, response format, and study phase in which each measure was administered.
• Supplementary Table S2. Characteristics of key informants involved in the partnership-preparation phase.
• Supplementary Table S3. Key informant interview findings: primary health concerns identified during the partnership-preparation phase.
• Supplementary Table S4. Characteristics of community health needs assessment participants.
• Supplementary Table S5. Community health needs assessment findings: primary health concerns identified during the partnership-preparation phase.

We believe these revisions address the reviewers’ recommendations without altering the results while importantly remaining consistent to the available data and the stated aims.

As Corresponding Author, I confirm that the revisions to the manuscript have been reviewed and approved for resubmission to IJERPH by all the listed authors.

We thank you, again, for the thoughtful feedback, which strengthened the manuscript substantially.

Reviewer 3 Report

Comments and Suggestions for Authors

The article is novel and addresses a topic of interest. Although the English is understandable, it needs to be revised. At times, the authors create overly long sentences that sound unnatural. In some paragraphs, certain words are repeated too frequently.

While the review included in the introduction is thorough, there are too many subsections, and it should be restructured. Furthermore, the “Literature Review” section is not usually included in the introduction of an original research article.

On another note, the “Target Population” section should be placed within the methodology, under a more general subsection that could be titled “Study Area.”

Within the methodology, it would be useful to include a diagram showing the phases of the project, the population involved in each phase, and the timeline of their development.

The discussion section is not adequate. The authors should discuss each block of results, contrast them with existing literature, propose hypotheses regarding their findings, highlight the novel contributions, and explain the relevance and usefulness of their results.

Author Response

Dear Reviewer,

Thank you for the thoughtful and constructive feedback on our manuscript. We appreciate the time and care taken in reviewing our work. We have revised the manuscript substantially in response to these comments. Below, we provide a point-by-point response and indicate where changes were made in the manuscript and supplementary materials.

Reviewer 3

Comment 1

Reviewer comment 1: The article is novel and addresses a topic of interest. Although the English is understandable, it needs to be revised. At times, the authors create overly long sentences that sound unnatural. In some paragraphs, certain words are repeated too frequently.

Response 1: Thank you for this helpful comment. We agree that the manuscript benefits from additional language polishing. In response to comments from multiple reviewers, we substantially revised the manuscript to improve clarity, reduce overly long sentences, remove repetitive wording, and improve overall flow. These revisions were made throughout the manuscript, particularly in the Introduction, Methods, Results, and Discussion. We have also carefully reviewed the revised manuscript again for style and readability and made additional edits to improve natural English phrasing.

Comment 2

Reviewer comment 2: While the review included in the introduction is thorough, there are too many subsections, and it should be restructured. Furthermore, the “Literature Review” section is not usually included in the introduction of an original research article.

Response 2: Thank you for this important comment. We revised aspects of the introduction for the purposes of this manuscript. In the revised version, Section 1.4 was substantially expanded and reorganized to better position the study within the most relevant cookstove and household air pollution literature. Specifically, the section now distinguishes prior evidence related to indoor air pollution, improved health outcomes, and cookstove acceptability, and it provides clearer temporal context regarding the shorter follow-up periods in prior studies compared with the present seven-year study. We believe this structure helps readers understand the specific gap addressed by the manuscript and how the present study contributes to the existing literature. We also note that the manuscript is not using Section 1.4 as a stand-alone, exhaustive “literature review” in the sense of a review article. Rather, this section functions as a focused review of prior cookstove intervention evidence within the Introduction, which is directly relevant to the study’s aims and hypotheses. The revised subsections improve clarity and readability after substantial restructuring prompted by earlier reviewer feedback. Accordingly, we did not remove this section from the Introduction or collapse all subsections into a single undifferentiated block of text.

Comment 3

Reviewer comment 3: On another note, the “Target Population” section should be placed within the methodology, under a more general subsection that could be titled “Study Area.”

Response 3: Thank you for this suggestion. In the manuscript, Section 1.2. Target Population serves an introductory purpose by explaining why adult women in a small rural village in the Rivas Political Department of Nicaragua were selected as the focus of the study, and by linking that choice to the broader public health problem of household air pollution, gendered cooking roles, poverty, and limited access to cleaner household technologies. We believe that this framing is appropriately placed in the Introduction, where it helps justify the study population and situate the research question before the methods are presented. We also note that the Methods section provides setting-specific information relevant to study implementation, including the partnership-preparation process in the community, the community-based nature of the intervention, and the recruitment and data-collection procedures used across study phases. In that sense, the current structure separates the rationale for the target population in the Introduction from the operational details of implementation in the Methods, which we believe improves readability and aligns well with the aims of the manuscript. Accordingly, we retained the Target Population section in the Introduction and did not move it into the Methods. While we appreciate the reviewer’s stylistic preference, we believe the current organization most clearly communicates both the public health relevance of the study population and the scope of the community-based intervention.

Comment 4

Reviewer comment 4: Within the methodology, it would be useful to include a diagram showing the phases of the project, the population involved in each phase, and the timeline of their development.

Response 4: Thank you for this helpful suggestion. We agree that a visual summary of the project phases could be useful for readers. However, we note that the revised manuscript presents this information in the text. Specifically, the opening of Section 2. Materials and Methods identifies the five cyclic phases of the CHAMP-BC framework and explains that the project progressed through Partnership-Preparation, Assessment-Contemplation, Planning-Preparation, Implementation-Action, and Evaluation-Maintenance. The populations involved in the formative phases are described in Section 3.1, including the 13 key informants, 77 individuals from 47 households in the community health needs assessment, and the 29 community members who attended the March 2010 community forum. The timeline is also described across Sections 2 and 3, including the key dates for formative assessment, implementation cohorts, and follow-up through 2018, with cookstove follow-up timing summarized in Table 5. In addition, Supplementary Table S1 summarizes the study instruments and study phases. For these reasons, we believe the revised manuscript contains the requested information in textual and tabular form, and we did not add an additional diagram.

Comment 5

Reviewer comment 5: The discussion section is not adequate. The authors should discuss each block of results, contrast them with existing literature, propose hypotheses regarding their findings, highlight the novel contributions, and explain the relevance and usefulness of their results.

Response 5: Thank you for this comment. We agree that the Discussion is a critical section for interpreting the results and situating the study within the broader literature. This concern has been substantially addressed in the revised manuscript. In the current version, Section 4. Discussion was expanded to more fully interpret the major findings in relation to prior cookstove intervention studies, particularly those reporting women’s symptom outcomes in Guatemala and Bolivia. The revised Discussion now contrasts the follow-up duration of prior studies with the present up to seven-year study, highlights the manuscript’s novel contribution as one of the longest longitudinal improved cookstove studies to report sustained symptom outcomes among women, and proposes several hypotheses for the durability of the findings, including the participatory design of the intervention, local stove selection, use of local masons, household contributions, and repeated follow-up over time. The revised Discussion also explains the broader relevance of the findings and addresses the potential contribution of CBPR to long-term intervention sustainability. Accordingly, we believe this concern has been addressed in the revisions.

Supplementary Materials Added

• Supplementary Table S1. Summary of survey instruments, domains, sample items, response format, and study phase in which each measure was administered.
• Supplementary Table S2. Characteristics of key informants involved in the partnership-preparation phase.
• Supplementary Table S3. Key informant interview findings: primary health concerns identified during the partnership-preparation phase.
• Supplementary Table S4. Characteristics of community health needs assessment participants.
• Supplementary Table S5. Community health needs assessment findings: primary health concerns identified during the partnership-preparation phase.

We believe these revisions address the reviewers’ recommendations without altering the results while importantly remaining consistent to the available data and the stated aims.

As Corresponding Author, I confirm that the revisions to the manuscript have been reviewed and approved for resubmission to IJERPH by all the listed authors.

We thank you, again, for the thoughtful feedback, which strengthened the manuscript substantially.

Round 2

Reviewer 1 Report

Comments and Suggestions for Authors

Thank you, and good luck in your future research on this important topic. 

Author Response

Dear Reviewer,

Thank you for the thoughtful comments on our manuscript and well wishes for our future work. We appreciate the time and care taken in reviewing our work and especially appreciate your recognition of the importance of our work to the body of evidence. 

As Corresponding Author, I confirm that the revisions to the manuscript have been reviewed and approved for resubmission to IJERPH by all the listed authors.

We thank you, again, for your thoughtful review.

Best,

Amy Buckenmeyer 

Reviewer 2 Report

Comments and Suggestions for Authors

The authors' revisions address some of the comments, but I remain concerned about the main empirical setting. Selection bias is a major problem here, even if the study is presented as a purely correlational study with no causality claims. We simply don't know what is the responsible channel of the observed changes over time. There is practically zero external validity with a non-random sample based on a single community in Nicaragua and without any controls being used. The only possible use for such a study is either to serve as a low-cost pilot or purely internal applications to that specific community in Nicaragua.

Author Response

Dear Reviewer,

Thank you for the feedback on our manuscript. We appreciate the time taken in reviewing our work. 

Reviewer 2, Round 2 Comment 1: "The authors' revisions address some of the comments, but I remain concerned about the main empirical setting. Selection bias is a major problem here, even if the study is presented as a purely correlational study with no causality claims. We simply don't know what is the responsible channel of the observed changes over time. There is practically zero external validity with a non-random sample based on a single community in Nicaragua and without any controls being used. The only possible use for such a study is either to serve as a low-cost pilot or purely internal applications to that specific community in Nicaragua."

Response 1: Thank you for these comments. We appreciate the reviewer’s continued concern regarding selection bias, external validity, and the absence of a comparison group. However, we respectfully disagree with the characterization that the study has little to no value. This research serves as an early feasibility study and practice-based case example. Novel intervention models, particularly those grounded in CBPR and implemented in under-resourced rural communities, often must begin with carefully bounded, real-world studies before larger comparative designs become possible. We therefore maintain that the manuscript offers a valuable contribution to the literature precisely because it provides new longitudinal evidence from a participatory intervention context that has been rarely documented.

We would also note that, at this stage, the reviewer’s remaining concerns appear to rest on requests for analyses that the present dataset and study design cannot support. We have revised the manuscript extensively to clarify the observational, community-based, and non-randomized nature of the study; to acknowledge selection bias, limited generalizability, and the absence of a control group; and to avoid unsupported causal claims. We would be fully willing to make additional revisions if more specific and feasible recommendations were provided. However, the current comment does not identify a concrete revision that can be implemented with the available data beyond the limitations already acknowledged in the manuscript.

For these reasons, we believe the revised manuscript already addresses the reviewer’s concerns appropriately by presenting the study design and limitations transparently, avoiding unsupported causal claims, and interpreting the findings within the proper scope of an observational, community-based, longitudinal cohort study. Accordingly, we agree with Reviewer 1 (Round 1 and 2), Reviewer 3 (Round 1), and the Editor ("Accept in current form”) and have not made additional revisions in response to these comments. 

As Corresponding Author, I confirm that the revisions to the manuscript have been reviewed and approved for resubmission to IJERPH by all the listed authors.

We thank you, again, for your review.

Best,

Amy Buckenmeyer