Why Do Some Cities in the United States Integrate Health into Their Climate Plans More than Others?—Hypotheses and Tests

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

Peer Review Report

Manuscript Title: Why do some cities in the United States perform better at integrating health into their climate plans than others? hypotheses  and tests

Date: 05/09/2025

 

1. General Assessment
2. Structure and Clarity

• The paper is well organized. The objectives of the study are clearly defined. The language is clear

1. Abstract

• Abstract is well prepared and presents an interesting research topic. The research is well justified

 

• Introduction

• The introduction, in general, is well presented; even some changes can be made to improve the quality of the manuscript.
• Page 2, lines 46-49, add references to heatwaves and extreme weather in cities like:
• Hu, J., Zhou, Y., Yang, Y., Chen, G., Chen, W., & Hejazi, M. (2023). Multi-city assessments of human exposure to extreme heat during heat waves in the United States. Remote Sensing of Environment, 295, 113700.
• Page 2, lines 70-71, authors should mention in the manuscript the Global Framework of Climate Services (GFCS), stating the main priority areas, including the health sector. Also, note that climate services can be designed for other areas that are not prioritized by the GFCS. This framework could help to give more value to the present study, as it investigates the application of health adaptation and mitigation measures in cities. Some suggestions of reference:
• Lowe, R., Stewart-Ibarra, A. M., Petrova, D., García-Díez, M., Borbor-Cordova, M. J., Mejía, R., ... & Rodó, X. (2017). Climate services for health: predicting the evolution of the 2016 dengue season in Machala, Ecuador. The lancet Planetary health, 1(4), e142-e151.
• Page 2, lines 82- 91, this looks like more for the abstract and not for the introduction section. Consider removing it to other parts of the manuscript.

 

1. Literature review

The literature review is well presented. Even so, since there is no reference to studies on climate services and health in cities, the authors should explore this topic and incorporate it into the manuscript.

 

1. Hypotheses

Hypotheses in general are well presented; however, some of them are not really justified by the literature, or no explanation exists:

• H2a There is no justification for this hypothesis
• H4a There is no justification for this hypothesis
• H4b Is there any scientific reference for this hypothesis

1. Methods

The methods section, in general, is well presented. Even so, some improvements could be done:

• Figure 1 is well done
• You should mention the origin of the data that you use for the model. For example, the average monthly temperature (where the data comes from)
• The authors say that they studied 50 cities, and consider adding a map including the analysed cities.
• Results are well presented and provide interesting insights into the research topic

1. Discussion

Discussion is interesting, but the authors should also mention the inclusion of climate services for health in cities.

1. Conclusions

- Conclusions are well presented; the authors should state future research lines. Like study models that include whether there exist climate services for health in the studied city.

- Question for the authors: In steps 1, 2, and 3, you identified health linkages to sectoral actions. Is anyone linked with tourism and health? This should also be explored in climate action plans.

Author Response

Peer Review Report 1

Manuscript Title: Why do some cities in the United States perform better at integrating health into their climate plans than others? hypotheses  and tests

Date: 05/09/2025

General Assessment

Structure and Clarity

 

The paper is well organized. The objectives of the study are clearly defined. The language is clear

 

Thank you. 

 

Abstract

 

Abstract is well prepared and presents an interesting research topic. The research is well justified

Thank you. 

Introduction

 

The introduction, in general, is well presented; even some changes can be made to improve the quality of the manuscript.

Thank you. 

Page 2, lines 46-49, add references to heatwaves and extreme weather in cities like:

Hu, J., Zhou, Y., Yang, Y., Chen, G., Chen, W., & Hejazi, M. (2023). Multi-city assessments of human exposure to extreme heat during heat waves in the United States. Remote Sensing of Environment, 295, 113700.

 

Thank you.  This reference is now included.

 

Page 2, lines 70-71, authors should mention in the manuscript the Global Framework of Climate Services (GFCS), stating the main priority areas, including the health sector. Also, note that climate services can be designed for other areas that are not prioritized by the GFCS. This framework could help to give more value to the present study, as it investigates the application of health adaptation and mitigation measures in cities. Some suggestions of reference:

 

Lowe, R., Stewart-Ibarra, A. M., Petrova, D., García-Díez, M., Borbor-Cordova, M. J., Mejía, R., ... & Rodó, X. (2017). Climate services for health: predicting the evolution of the 2016 dengue season in Machala, Ecuador. The lancet Planetary health, 1(4), e142-e151.

 

 Thank you.  This reference is now included.

 

Page 2, lines 82- 91, this looks like more for the abstract and not for the introduction section. Consider removing it to other parts of the manuscript.

 

Thank you.  The text that feels like an abstract is removed.

 

Literature review

The literature review is well presented. Even so, since there is no reference to studies on climate services and health in cities, the authors should explore this topic and incorporate it into the manuscript.

 Thank you.  References are now made to climate services with appropriate citations.

 

Hypotheses

Hypotheses in general are well presented; however, some of them are not really justified by the literature, or no explanation exists:

• H2a There is no justification for this hypothesis
• H4a There is no justification for this hypothesis
• H4b Is there any scientific reference for this hypothesis

 

Thank you. The hypotheses have been modified and condensed in some places to strengthen relevant justifications. Relevant references have also been added to improve justifications.

 

Methods

The methods section, in general, is well presented. Even so, some improvements could be done:

• Figure 1 is well done
• You should mention the origin of the data that you use for the model. For example, the average monthly temperature (where the data comes from)

 

Thank you.  References are now included for the data.

 

• The authors say that they studied 50 cities, and consider adding a map including the analysed cities.

 

Thank you.  A map is now added.

 

• Results are well presented and provide interesting insights into the research topic

 

Discussion

Discussion is interesting, but the authors should also mention the inclusion of climate services for health in cities.

Thank you.  Climate services are now mentioned in the introduction and the literature review.

 

Conclusions

- Conclusions are well presented; the authors should state future research lines. Like study models that include whether there exist climate services for health in the studied city.

- Question for the authors: In steps 1, 2, and 3, you identified health linkages to sectoral actions. Is anyone linked with tourism and health? This should also be explored in climate action plans.

Thank you.  Tourism is now mentioned in the conclusion.

Reviewer 2 Report

Comments and Suggestions for Authors

*The introduction sets up the background well, but the motivation could be sharpened. Currently, it mixes justification (cities are important, health is important) with literature review. Restructure to present a clear gap: why previous studies did not test hypotheses systematically.
*  The main research question (why some cities connect health and climate better than others) is relevant, but the originality compared to [45] (noted as the lone exception) should be stressed.
*The review is broad but reads as a narrative summary. It needs more critical evaluation of existing works. What exactly is missing in prior studies?
* Some citations are dated; ensure inclusion of 2023–2025 studies on climate-health planning for stronger grounding.
* Hypotheses are clear but too many (10). This risks diluting the analysis. Consider prioritizing the most policy-relevant ones (plan age, plan type, political orientation, demographics). Some hypotheses (e.g., H3a GDP per capita) lack strong theoretical justification in the text. Strengthen or drop them.
* The operationalization of the dependent variable (health integration) through coding is innovative but raises concerns of subjectivity. More detail is needed on intercoder reliability, coding rules, and robustness checks.
*Reliance on keyword counts (“health” mentions) is weak as a second measure. Consider normalizing by plan length (pages/words), and present these results more centrally rather than in appendix.
* OLS regression may not be the most suitable given potential endogeneity (e.g., political orientation and plan adoption). Consider robustness with alternative models (Poisson/negative binomial for counts, or fixed-effects if data allows).
*  The sample of 50 largest cities is biased toward big metropolitan areas; results may not generalize to smaller cities. This limitation should be clearly acknowledged earlier.
* Tables are not fully interpretable; standard errors should be reported consistently. Significance codes (“\*\*\*, \*\*, \*”) should be explained in each table caption.
*  R² values are low (0.10–0.36), indicating weak explanatory power. This should be openly discussed, not hidden.
*  Many variables flip signs across models; the paper does not sufficiently discuss why. Provide interpretation or caution.
*  Figures (regression outputs) should be visualized with coefficient plots instead of dense tables for clarity.
* The discussion overstates findings. For instance, claiming “plan attributes are the most significant predictors” ignores that most other variables are insignificant and inconsistent. Tone down the claims.
* **Comment:** The role of federal policy changes (e.g., Inflation Reduction Act) is interesting but speculative. Provide evidence (citations, city plan references) before attributing causal influence.

Author Response

Peer Review Report 2

Review Report 2

*The introduction sets up the background well, but the motivation could be sharpened. Currently, it mixes justification (cities are important, health is important) with literature review. Restructure to present a clear gap: why previous studies did not test hypotheses systematically.

Thank you. The introduction has now been revised to underline the motivation and a gap: that is a systematic assessment of why cities vary in how much they integrate health co-benefits into planning.

*  The main research question (why some cities connect health and climate better than others) is relevant, but the originality compared to [45] (noted as the lone exception) should be stressed.

Thank you. We now highlight that we are the only study that is focusing on explain variation in the integration of health co-benefits (as opposed to more general development orientation).  We also highlight that the approach to assessing the dependant variable with two measures of integration distinguishes our work from previous studies.

*The review is broad but reads as a narrative summary. It needs more critical evaluation of existing works. What exactly is missing in prior studies?

Thank you for this.  We have now revised the review so that it highlights that studies have increasingly aimed to make there work more relevant to policymakers. That has gradually led to an interest in understanding what conditions will facilitate the integration of that work on health into urban climate plans. We further note that our work is unique in that we focus specifically on health as opposed to a broader development framing and that we have different measures of shallower and deeper health integration. Similar points are now highlighted in the abstract, intro and discussion section.

* Some citations are dated; ensure inclusion of 2023–2025 studies on climate-health planning for stronger grounding.

Thank you with have added in several new citations, focusing on work on climate services as well as new studies on differences between cities.

* Hypotheses are clear but too many (10). This risks diluting the analysis. Consider prioritizing the most policy-relevant ones (plan age, plan type, political orientation, demographics). Some hypotheses (e.g., H3a GDP per capita) lack strong theoretical justification in the text. Strengthen or drop them.

Thank you. We have reduced the number of hypotheses down to six across three categories following this suggestion. This entailed removing hypotheses related to climate conditions. These are now included as control variables in the modelling.

* The operationalization of the dependent variable (health integration) through coding is innovative but raises concerns of subjectivity. More detail is needed on intercoder reliability, coding rules, and robustness checks.

Thank you.  We have now elaborated on the methods by including an appendix on the coding rules; we have also added text to explain why we went through three rounds of coding and ultimately had a single coder review and recode.  That is, we found what we deemed inflated estimates of links based on a review of the number of links and the number of references to health.  We also note that this third phase helped to improve the inter-coding reliability; a test of the intraclass correlation (ICC) between the second phase (extracting the above coding that seemed inflated) and revised third phase coding suggests that there is statistically significant agreement among the two sets of codes (F(38,20.3) = 3.8 , p = 0.00109).

*Reliance on keyword counts (“health” mentions) is weak as a second measure. Consider normalizing by plan length (pages/words), and present these results more centrally rather than in appendix.

Thank you. We have now normalized this measure by using a word count offset for both the health reference as well as the health linkages variables.

* OLS regression may not be the most suitable given potential endogeneity (e.g., political orientation and plan adoption). Consider robustness with alternative models (Poisson/negative binomial for counts, or fixed-effects if data allows).

Thank you. We have modified the model and are no using negative binomial which is more appropriate for counts data with overdispersion.

*  The sample of 50 largest cities is biased toward big metropolitan areas; results may not generalize to smaller cities. This limitation should be clearly acknowledged earlier.

Thank you. This limitation is now highlighted in the methods section—and discussed more in the conclusion.

* Tables are not fully interpretable; standard errors should be reported consistently. Significance codes (“\*\*\*, \*\*, \*”) should be explained in each table caption.

Thank you: we have now added in the codes and standard errors into table in the appendix.  Following your other suggestion, we have used coefficient plots to improve the presentation of results.

*  R² values are low (0.10–0.36), indicating weak explanatory power. This should be openly discussed, not hidden.

We have moved to a negative binomial model and have now used discussions of the AIC for goodness of fit issues.

*  Many variables flip signs across models; the paper does not sufficiently discuss why. Provide interpretation or caution.

Thank you. The new models using the negative binomial specification generates more stable estimates.

*  Figures (regression outputs) should be visualized with coefficient plots instead of dense tables for clarity.

Thank you.  We have followed this suggestion and placed the tables in an appendix.

* The discussion overstates findings. For instance, claiming “plan attributes are the most significant predictors” ignores that most other variables are insignificant and inconsistent. Tone down the claims.

Thank you we have toned down and aimed to make the language in the discussion more precise. The updated modelling (using the negative binomial modelling) has nonetheless generated some results that has more significant predictors, however; we have tried to reflect that as well in the discussion.

* **Comment:** The role of federal policy changes (e.g., Inflation Reduction Act) is interesting but speculative. Provide evidence (citations, city plan references) before attributing causal influence.

Thank you. We have elaborated on this point by referring to specific city plans to make this claim. The influence of the Inflation Reduction Act is still not well understood in the published literature, but we do refer to the case of Milwaukee and Fresno to support these claims.

Reviewer 3 Report

Comments and Suggestions for Authors

This study investigates the factors influencing the integration of health co-benefits into climate action plans (CAPs) across 50 of the most populous U.S. cities. The authors test four sets of hypotheses—related to plan attributes, politics, demographics, and climate conditions—using ordinary least squares (OLS) regression models with two dependent variables: a coded measure of health linkages to sectoral actions and a simple count of the word "health." The main contribution lies in its systematic coding framework for assessing climate-health integration and its empirical testing of potential correlates. A key strength is the dual approach to measuring health integration, allowing for nuanced interpretation. However, there is a need for further improvements.

Abstract is well structured; however, it lacks clarity about data and methodology. Please clearly state the data and methods.

The research question (“why some cities integrate health better”) implies causal inference, but the methodology is correlational. The introduction should clarify that the study tests associations, not causality, especially given cross-sectional data.

H3a (GDP per capita) assumes wealthier cities invest more in health integration, yet GDP was log-transformed and used as a continuous variable without theoretical justification for nonlinearity.

H4c (climate zone vulnerability) uses dummy variables based on temperature zones, which may poorly capture actual climate risk exposure (e.g., flood-prone coastal vs inland areas). Vulnerability indices like those from [51] could have been incorporated directly.

The use of county-level election data (Per_Point Diff) to proxy city political leanings may misrepresent urban political dynamics, especially in counties dominated by suburban or rural votes (e.g., Dallas County vs. City of Dallas).

No justification is provided for excluding cities without climate plans beyond mentioning Fort Worth and Colorado Springs. Were there efforts to assess whether excluded cities differ systematically?

The coding process involved six researchers followed by re-coding by one coordinator. However, inter-coder reliability metrics (e.g., Cohen’s kappa) are missing, undermining confidence in the robustness of the health linkage variable.

While the F-test indicates improved model fit with additional variables, the lack of individual significance raises concerns about overfitting or multicollinearity. Variance inflation factor (VIF) analysis should have been included.

Overemphasis on the Inflation Reduction Act (IRA) as a driver of newer plans’ quality is speculative. While cited by some cities, no direct evidence links IRA funding to higher health integration scores.

The claim that “cities have agency to craft…plans” risks downplaying structural constraints (e.g., fiscal capacity, federal policy shifts) that limit local autonomy.

Spatial disconnect between climate risk and planning effort is noted, but no attempt is made to reconcile this with H4 hypotheses beyond acknowledging inconsistency.

Line 30–32: “The results further show that an unrestricted model with a complete set of regressors improves the fit of the data.” → This statement needs qualification. As shown in Tables 1 and 2, R² increases from ~0.1 to ~0.35, but none of the added variables (politics, demographics, climate) are individually significant. The improvement may reflect noise rather than explanatory power. Recommend revising to: “an unrestricted model shows improved overall fit, though most additional variables are not individually significant.”

Line 449–451: Claim that Fresno has “one of the lowest GDP per capita in our city sample” → This assertion lacks supporting data. Please provide comparative GDP statistics in appendix or revise.

Author Response

Peer Review Report 3

This study investigates the factors influencing the integration of health co-benefits into climate action plans (CAPs) across 50 of the most populous U.S. cities. The authors test four sets of hypotheses—related to plan attributes, politics, demographics, and climate conditions—using ordinary least squares (OLS) regression models with two dependent variables: a coded measure of health linkages to sectoral actions and a simple count of the word "health." The main contribution lies in its systematic coding framework for assessing climate-health integration and its empirical testing of potential correlates. A key strength is the dual approach to measuring health integration, allowing for nuanced interpretation. However, there is a need for further improvements.

Abstract is well structured; however, it lacks clarity about data and methodology. Please clearly state the data and methods.

Thank you. The abstract clearly states that the research uses a negative binomial regression on novel data set that includes count data.

The research question (“why some cities integrate health better”) implies causal inference, but the methodology is correlational. The introduction should clarify that the study tests associations, not causality, especially given cross-sectional data.

Thank you. The introduction now states that the study covers correlations.

H3a (GDP per capita) assumes wealthier cities invest more in health integration, yet GDP was log-transformed and used as a continuous variable without theoretical justification for nonlinearity.

Thank you. We now explain why this variable is log transformed to pull in possible outlying variables.

H4c (climate zone vulnerability) uses dummy variables based on temperature zones, which may poorly capture actual climate risk exposure (e.g., flood-prone coastal vs inland areas). Vulnerability indices like those from [51] could have been incorporated directly.

Thank you for this. We have now removed this hypothesis based on comments from reviewer 2; we include the same dummy variables as controls but mention this limitation in the conclusion.

The use of county-level election data (Per_Point Diff) to proxy city political leanings may misrepresent urban political dynamics, especially in counties dominated by suburban or rural votes (e.g., Dallas County vs. City of Dallas).

Thank you for this. We now mention this limitation in the presentation of the results as well as the conclusion.

No justification is provided for excluding cities without climate plans beyond mentioning Fort Worth and Colorado Springs. Were there efforts to assess whether excluded cities differ systematically?

Thank you for this.  While we debated whether to include these cities without plans we a score of 0 for health references, we ultimately decided that the study was focusing on city with plans that could be assessed. We now mention this limitation in the conclusion.

The coding process involved six researchers followed by re-coding by one coordinator. However, inter-coder reliability metrics (e.g., Cohen’s kappa) are missing, undermining confidence in the robustness of the health linkage variable.

Thank you.  We have now elaborated on the methods by including an appendix on the coding rules; we have also added text to explain why we went through three rounds of coding and ultimately had a single coder review and recode.  That is, we found what we deemed inflated estimates of links based on a review of the number of links and the number of references to health.  We also note that this third phase helped to improve the inter-coding reliability; a test of the intraclass correlation (ICC) between the second phase (extracting the above coding that seemed inflated) and revised third phase coding suggests that there is statistically significant agreement among the two sets of codes (F(38,20.3) = 3.8 , p = 0.00109).

While the F-test indicates improved model fit with additional variables, the lack of individual significance raises concerns about overfitting or multicollinearity. Variance inflation factor (VIF) analysis should have been included.

Thank you.  We have now included a discussion of multicollinearity and VIFs.  We have also toned down the importance of a fuller model.

Reviewer 4 Report

Comments and Suggestions for Authors

This is a well-designed, rigorously executed study with significant practical implications. The article systematically explores the key drivers influencing the integration of health factors into climate action plans in major US cities, filling an empirical gap in the existing literature regarding "why differences exist." The research methodology (hypothesis-driven, multivariate regression, dual dependent variable measurement) is appropriately applied, and the discussion section is insightful and thought-provoking. The overall quality of the paper is high and merits publication value. However, there is still room for improvement in writing clarity, depth of data analysis, and some details.

1. The recommendation of "regularly updating climate plans" in the abstract is strongly supported by the results section (Plan Age), but the recommendation for "linking mitigation-adaptation actions" primarily comes from the Health Count model, while the results of the Health Linkages model provide weaker support for this. The abstract could be slightly modified to more accurately reflect how different findings correspond to different recommendations, for example: "...while integrating mitigation with adaptation is associated with a greater quantity of health references."

2. Consider briefly mentioning the counterintuitive yet important finding that "political, demographic, and climate factors were not significant in the model" in the abstract.

3. The introduction's final paragraph, which outlines the paper's structure ("divided into six sections"), is somewhat redundant. It is common practice to omit this to allow for smoother writing. Readers can naturally grasp the structure from the headings.

4. The introduction could further emphasize this study's methodological contribution to "exploring causal associations" rather than merely "describing differences."

5. When mentioning literature that compares different cities, it could be more directly pointed out that these studies primarily describe "what is," whereas this paper aims to answer "why" through hypothesis testing, thereby more prominently highlighting the novelty of this research.

6. The formulations of hypotheses H1a and H1b both refer to "more overall health linkages." However, in the subsequent operationalization, "health linkages (Health Linkages)" is a specific term (referring to depth-coded connections), while "health word count (Health Count)" is another metric. The term "overall health linkages" here can easily be confused with the specific variable later. It is suggested to use more neutral terminology when proposing the hypotheses, such as "greater integration of health co-benefits."

7. When interpreting the R² values (e.g., 0.3567 for Model 4), a brief comment could be added, indicating that the model explains approximately 36% of the variance, which is a quite good level for cross-sectional studies at the city level, while also acknowledging that other important factors not captured by the model remain.

8. Regarding the result that "other variables were not significant," beyond stating the fact, a preliminary, conservative interpretation could be started to pave the way for the discussion section. For example, "It is surprising that factors such as mayoral party affiliation and GDP per capita did not show a significant influence, which might suggest...", but this could also be reserved for deeper exploration in the discussion section.

9. In the discussion, the results of "other variables not being significant" could be discussed more focusedly and in-depth. This is a very important finding, suggesting that cities' decisions regarding health integration may depend more on internal institutional capacity, the professionalism of the planning process, and the availability of technical tools (like CJEST), rather than external contextual factors. This should be clearly proposed as a key direction for future research.

10. When discussing the Fresno case study, its conclusion is very powerful. This could be more explicitly used as an argument to explain why traditional hypotheses like politics and economics were disproven in this study, while plan attributes and new policy tools (federal funding, screening tools) are more critical explanatory variables.

11. The section numbering appears incorrect; please check carefully. For example, shouldn't "3. result" actually be Chapter 5?

Author Response

Overemphasis on the Inflation Reduction Act (IRA) as a driver of newer plans’ quality is speculative. While cited by some cities, no direct evidence links IRA funding to higher health integration scores.

Thank you for this.  We have now toned down this inference and tried to be more specific about plans that refer directly to the IRA.

The claim that “cities have agency to craft…plans” risks downplaying structural constraints (e.g., fiscal capacity, federal policy shifts) that limit local autonomy.

Thank you for this.  We have now mentioned these structural constraints in the discussion of agency and toned the claims about agency with language suggesting that cities may be able to do more.

Spatial disconnect between climate risk and planning effort is noted, but no attempt is made to reconcile this with H4 hypotheses beyond acknowledging inconsistency.

Thank you.  We have removed the hypothesis on regional controls and remark on the limitations of this variable.

Line 30–32: “The results further show that an unrestricted model with a complete set of regressors improves the fit of the data.” → This statement needs qualification. As shown in Tables 1 and 2, R² increases from ~0.1 to ~0.35, but none of the added variables (politics, demographics, climate) are individually significant. The improvement may reflect noise rather than explanatory power. Recommend revising to: “an unrestricted model shows improved overall fit, though most additional variables are not individually significant.”

Thank you.  We have now moved to a negative binomial model and we discuss goodness of fit using the AIC. We no longer believe that this discussion is central to our story and have deemphasized it as such.  We also include the possibility that the improved fit is attributable to noise.

Line 449–451: Claim that Fresno has “one of the lowest GDP per capita in our city sample” → This assertion lacks supporting data. Please provide comparative GDP statistics in appendix or revise.

Thank you.  WE now include a listing of income and population in the appendix.

The recommendation of "regularly updating climate plans" in the abstract is strongly supported by the results section (Plan Age), but the recommendation for "linking mitigation-adaptation actions" primarily comes from the Health Count model, while the results of the Health Linkages model provide weaker support for this. The abstract could be slightly modified to more accurately reflect how different findings correspond to different recommendations, for example: "...while integrating mitigation with adaptation is associated with a greater quantity of health references."

Thank you for this.  We have rerun the models based on comments from reviewer 2 and have more support for the integrating mitigation and adaptation. We have reflected that modified inference in the abstract.

Consider briefly mentioning the counterintuitive yet important finding that "political, demographic, and climate factors were not significant in the model" in the abstract.

Thank you for this.  We have rerun the models based on comments from reviewer and have slightly more support for the politics variables. We nonetheless highlight the lack of relationships with other regressors and the randomness that this implies in the discussion.

The introduction's final paragraph, which outlines the paper's structure ("divided into six sections"), is somewhat redundant. It is common practice to omit this to allow for smoother writing. Readers can naturally grasp the structure from the headings.

Thank you for this.  We have now removed this subsection of the paper.

The introduction could further emphasize this study's methodological contribution to "exploring causal associations" rather than merely "describing differences."

Thank you. This point is now stressed in the abstract and the introduction.

When mentioning literature that compares different cities, it could be more directly pointed out that these studies primarily describe "what is," whereas this paper aims to answer "why" through hypothesis testing, thereby more prominently highlighting the novelty of this research.

Thank you. This kind of language is now inserted into the introduction and the literature review.

The formulations of hypotheses H1a and H1b both refer to "more overall health linkages." However, in the subsequent operationalization, "health linkages (Health Linkages)" is a specific term (referring to depth-coded connections), while "health word count (Health Count)" is another metric. The term "overall health linkages" here can easily be confused with the specific variable later. It is suggested to use more neutral terminology when proposing the hypotheses, such as "greater integration of health co-benefits."

Thank you. This more neutral terminology is now included in the hypotheses.

When interpreting the R² values (e.g., 0.3567 for Model 4), a brief comment could be added, indicating that the model explains approximately 36% of the variance, which is a quite good level for cross-sectional studies at the city level, while also acknowledging that other important factors not captured by the model remain.

Thank you.  We have now moved to a negative binomial model and we discuss goodness of fit using the AIC. We no longer believe that this discussion is central to our story and have deemphasized it as such.

Regarding the result that "other variables were not significant," beyond stating the fact, a preliminary, conservative interpretation could be started to pave the way for the discussion section. For example, "It is surprising that factors such as mayoral party affiliation and GDP per capita did not show a significant influence, which might suggest...", but this could also be reserved for deeper exploration in the discussion section.

Thank you. We do offer a little more discussion of this point, noting the lack of a systematic approach to assessing these benefits.

In the discussion, the results of "other variables not being significant" could be discussed more focusedly and in-depth. This is a very important finding, suggesting that cities' decisions regarding health integration may depend more on internal institutional capacity, the professionalism of the planning process, and the availability of technical tools (like CJEST), rather than external contextual factors. This should be clearly proposed as a key direction for future research.

Thank you. We followed this suggestion and highlighted the importance of making the planning process more frequent and systematic.

When discussing the Fresno case study, its conclusion is very powerful. This could be more explicitly used as an argument to explain why traditional hypotheses like politics and economics were disproven in this study, while plan attributes and new policy tools (federal funding, screening tools) are more critical explanatory variables.

Thank you. We do offer a little more discussion of this point but it is not as supported by the evidence after we reran the modelling. Thus, it is not as emphasized as much as the reviewer suggests.

The section numbering appears incorrect; please check carefully. For example, shouldn't "3. result" actually be Chapter 5?

Thank you. The numbering has been corrected.

 

 

 

Round 2

Reviewer 2 Report

Comments and Suggestions for Authors

The authors have carefully addressed all previous comments and substantially improved the clarity, structure, and scientific rigor of the manuscript. I am satisfied with the current version and recommend the manuscript for publication in its present form. 

Reviewer 3 Report

Comments and Suggestions for Authors

The authors have made significant changes.

Reviewer 4 Report

Comments and Suggestions for Authors

OK