The Relationship Between Urban Characteristics and Non-Communicable Diseases—Conceptual Framework of the HORUS Project

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

After reviewing the manuscript ''The relationship between urban characteristics and non-communicable diseases - conceptual framework of the HORUS project'', I provide my comments and suggestions for improvement below.

The review paper presents a conceptual framework, the HORUS project, aimed at understanding and ultimately preventing NCDs through the integration of advanced technologies in urban health research. The topis is very relevant, given the growing burden of NCDs in increasingly urbanized societies and the multitude of environmental, behavioral, and socioeconomic factors that shape population health and well-being.

The section on Urban area mapping is well-structured and the data sources cited are reliable and current. However, the authors do not address the dynamic nature of urban environments, as urban characteristics change over time, and the manuscript would benfit from a brief discussion of how temporal changes might affect mapping outcomes and the validaty of the framework over longer timeframes.

The section 2.2.: 3D modelling is referenced but not adequately explained. The authors should justify whether 3D modelling adds meaningful explanatory value over convential 2D mapping, and explicitly address how data gaps would be managed, as the complete 3D spatial coverage is rarely available across cities. Regarding virtual rality, the authors should discuss the feasibility of VR tools across diverse urban contexts, as the technologies are resource-intensive and their applicability may be limited in settings with fewer resources. More broadly, the paper would benefits from a critical discussion of whether these technologies together serve the public health framework or whether their combined complexity risks limiting real-world applicability.

The paper does not adequately address the generalizability of AI models trained on data from specific urban contexts to other cities or populations. As well, no clear criteria are presented for validating AI outputs against real health outcomes.

Throughout the paper, there is a notable absence of the limitations of each technology presented. That is essential to contextualize the framework's applicability and guide future research.

The paper focuses on four main NCDs without providing a clear rationale for this selection. The authors should justify why these diseases were prioritized and whether the exclusion of others wa intentional or limitation of the available evidence. Additionally, the role of socioeconomic context, more specifically differences across high, middle and low income urban settings, is inconsistently addressed in the manuscript. Especially given that income level may substantially influence both urban characteristics and NCD burden.

While the technologies underpinning the HORUS project are described in earlier sections, the implementation section would benefit from an integrative synthesis explaining how these tools function together as a coherent system.

The authors should expand the conclussion section, maybe include potential directions for future work, their perspectives on the long-term contribution of the HORUS framework…

Author Response

Comment 1: The section on Urban area mapping is well-structured and the data sources cited are reliable and current. However, the authors do not address the dynamic nature of urban environments, as urban characteristics change over time, and the manuscript would benfit from a brief discussion of how temporal changes might affect mapping outcomes and the validaty of the framework over longer timeframes.

Response 1: We completely agree with this observation. To address the dynamic nature of urban environments, we added a discussion in Section 2 noting that "urban spatial structures are not static, they undergo continuous physical, political, and demographic temporal shifts". We further clarified how the HORUS framework models this dynamism by "configuring rolling multi-temporal geospatial data tracks, actively calculating temporal data decay rates to prevent legacy artifacts from confounding longitudinal epidemiological validity over decade-long observation thresholds".

Comment 2: The section 2.2.: 3D modelling is referenced but not adequately explained. The authors should justify whether 3D modelling adds meaningful explanatory value over convential 2D mapping, and explicitly address how data gaps would be managed, as the complete 3D spatial coverage is rarely available across cities.

Response 2: Thank you for highlighting this omission. We have expanded Section 2.2 to explicitly justify the value of 3D modeling, explaining that it captures "environmental exposure verticality - such as localized atmospheric layer pooling, floor-level sound reverberation vectors, and localized shade indices within dense urban corridors". Regarding the management of spatial data gaps, we added that when "municipal 3D geometric coverage lacks complete city data files, the framework initiates iterative spatial imputation pipelines, blending satellite radar footprints with synthetic street profiles to patch coverage voids".

Comment 3: Regarding virtual rality, the authors should discuss the feasibility of VR tools across diverse urban contexts, as the technologies are resource-intensive and their applicability may be limited in settings with fewer resources. More broadly, the paper would benefits from a critical discussion of whether these technologies together serve the public health framework or whether their combined complexity risks limiting real-world applicability.

Response 3: Thank you for this information. We addressed the real-world applicability and resource considerations in Section 2.2 by shifting the described focus of VR from strictly clinical psychological diagnosis to democratizing urban architectural planning. We explained that combining immersive VR with real-time biometric eye-tracking allows researchers to quantify how biophilic modifications reduce ambient anxiety "across varying socioeconomic demographics before breaking ground," ensuring interventions are tested before expensive physical resources are committed.

Comment 4: The paper does not adequately address the generalizability of AI models trained on data from specific urban contexts to other cities or populations. As well, no clear criteria are presented for validating AI outputs against real health outcomes.

Response 4: Thank you for this point. We added a sentence in Section 2.3 clarifying that to "guarantee external validation generalizability across heterogeneous metropolitan environments, models incorporate standardized cross-geography domain transfer algorithms". Furthermore, to establish clear criteria for real health outcomes, we explicitly stated in the Conclusions (Section 5) that the outputs undergo "clinical validation against primary care digital health records".

Comment 5: Throughout the paper, there is a notable absence of the limitations of each technology presented. That is essential to contextualize the framework's applicability and guide future research.

Response 5: We agree that contextualizing the limitations is essential. Rather than creating a separate limitations section, we have woven these constraints directly into the respective technology subsections to provide immediate context. For example, we explicitly acknowledged the limitation of incomplete municipal data files in 3D mapping and introduced spatial imputation as the remedy. We also addressed the limitation of AI model generalizability by introducing cross-geography domain transfer algorithms to mitigate this risk.

Comment 6: The paper focuses on four main NCDs without providing a clear rationale for this selection. The authors should justify why these diseases were prioritized and whether the exclusion of others wa intentional or limitation of the available evidence. Additionally, the role of socioeconomic context, more specifically differences across high, middle and low income urban settings, is inconsistently addressed in the manuscript.

Response 6: Thank you for this valuable feedback. We have significantly expanded the Implementation section (Section 4) to justify targeting these four NCDs, explaining that the decision is "strategically driven by their shared reliance on modifiable built environment risk behaviors, offering the highest systemic public health ROI for structural environmental redesigns". We also thoroughly addressed the socioeconomic differences, explicitly noting that in low- and middle-income neighborhoods, the NCD burden is driven by an "absolute lack of infrastructure access and high exposure to uncontrolled industrial externalities," whereas in high-income settings, chronic diseases track more closely with "automated sedentary lifestyles and hyper-processed dietary access points".

Comment 7: While the technologies underpinning the HORUS project are described in earlier sections, the implementation section would benefit from an integrative synthesis explaining how these tools function together as a coherent system.

Response 7: We entirely agree. We have expanded the Conclusions section (Section 5) to include a clear, integrative synthesis of the closed-loop lifecycle. We explain how the system synchronizes four distinct operational stages: "GIS driven baseline data ingestion via macro digital twins, automated predictive vulnerability mapping using multimodal GeoAI foundation models, community-led neighborhood co-design via immersive VR, and clinical validation against primary care digital health records".

Comment 8: The authors should expand the conclussion section, maybe include potential directions for future work, their perspectives on the long-term contribution of the HORUS framework…

Response 8: Thank you. The newly expanded Conclusions section (Section 5) now better highlights the long-term contribution of the framework. We explicitly state that the results of the project will "advocate for a shift in urban policy that prioritizes environments that promote wellbeing, sustainability and inclusion," providing a clear direction for how this research bridges the gap between public health research and actionable municipal policy.

Reviewer 2 Report

Comments and Suggestions for Authors

1. Introduction

Line 77: change “health issues” to “health promotion”

Line 107: delete “a” to read “may get better insight and understanding…”

Line 109: change “simplify” to “simplifies”

2. Urban Area Mapping

Line 138: I am not sure whether it is mental health but mental wellbeing. The latter is more specific and would fit as a concept better for this paper. Mental health would include some severe mental health problems, and likely urban planning would not change those.

Line 140: Please specify “infective” disease outbreaks and tracking of “infective” disease spread.

2.1 GIS in Public Health

Line 152: NCDs are not spread per se, please use the term “emergence” – “affect the emergence of NCDs…”

2.3 Artificial Intelligence

Line 220: You use inequities and inequalities. These two terms are not interchangeable, and it might be clearer if you are addressing one of these concepts in the manuscript.

4. Implementation of the HORUS project

Line 272: Please change “citizens that” to “citizens who”

This is a great paper as written and I look forward to seeing the HORUS project produce urban planning results to improve health

Author Response

Comment 1: Line 77: change “health issues” to “health promotion”

Response 1: Thank you. We have revised the text to read "integrating health promotion into urban design".

Comment 2: Line 107: delete “a” to read “may get better insight and understanding…”

Response 2: We have removed the word "a" to ensure the sentence reads "may get better insight and understanding".

Comment 3: Line 109: change “simplify” to “simplifies”

Response 3: Thank you for catching this grammatical error. We have corrected "simplify" to "simplifies".

Comment 4: Line 138: I am not sure whether it is mental health but mental wellbeing. The latter is more specific and would fit as a concept better for this paper.

Response 4: Thank you for this important conceptual distinction. We agree and have replaced the term "mental health" with "mental wellbeing".

Comment 5: Line 140: Please specify “infective” disease outbreaks and tracking of “infective” disease spread.

Response 5: We have updated the text to specify "infectious disease outbreaks" and the tracking of "transmission pathways" to provide the precise epidemiological terminology required.

Comment 6: Line 152: NCDs are not spread per se, please use the term “emergence” – “affect the emergence of NCDs…”

Response 6: We completely agree. We have replaced the phrase "spread of NCDs" with "emergence of NCDs".

Comment 7: Line 220: You use inequities and inequalities. These two terms are not interchangeable, and it might be clearer if you are addressing one of these concepts in the manuscript.

Response 7: Thank you for pointing this out. We clarified this vital conceptual distinction in Section 2 by stating that effective mapping requires "strict conceptual clarity between mere spatial inequalities in resource distribution and actual systemic spatial inequities to ensure that mapping translates into actionable urban planning policies".

Comment 8: Line 272: Please change “citizens that” to “citizens who”

Response 8: To improve the grammatical flow and address the intent of your comment, we have completely restructured and rewritten this sentence. It now reads: "support urban residents, and above all those who are most vulnerable".

Comment 9: This is a great paper as written and I look forward to seeing the HORUS project produce urban planning results to improve health

Response 9: We sincerely appreciate your encouraging remarks and your attention to the manuscript's terminology, which has greatly improved the final text. Have a great day!