The COVID-19 pandemic highlighted the complex interplay between built, biotic, social, and health processes. Existing research acknowledges the connections between these factors, but this paper aims to develop a conceptual framework that adequately represents the built environment's causal relevance, which has been limited in current scientific models and data practices. The authors draw on architecture and design, philosophy of science, and systems biology to explore how the built environment regulates these multi-process pathways. The paper's analysis is both diagnostic and prescriptive. The diagnostic aspect focuses on identifying failures in representing these complex processes in COVID-19 data and modeling, highlighting problems with missing data, idealized measurements, and the use of shallow surrogate measures. The prescriptive aspect focuses on developing new conceptual tools—like the ideas of 'context', 'nudge', 'affordance', and 'interface'—to more accurately represent the built environment's multifaceted role and suggest practical applications for ongoing public health practices.

The paper reviews existing literature on the junctures between built environments, ecological factors, social structures, and health. It notes the thorough research already conducted but emphasizes the inadequacy of current representational frameworks. The authors cite works highlighting the connections between built environments and health outcomes, including studies on the impact of urban sprawl, pollution exposure, and access to green spaces. However, the review points to the need for more nuanced representations that account for the complex interactions between these factors, particularly the moderating role of the built environment. The limitations of existing integrative public health concepts, such as 'One Health' and 'EcoHealth', are also discussed, noting their anthropocentric biases and limited focus on socio-ecological interconnections.

The paper's methodology is primarily conceptual and analytical. It uses a relational approach to explore the intricate connections between the built environment and other factors influencing health. The authors present three intertwined problems with existing scientific representations of COVID-19: missing data (especially on race, ethnicity, and social determinants of health), idealized measurement outcomes that fail to capture the complexity of the interactions, and the use of shallow surrogate measures that oversimplify causal relationships. They analyze counterintuitive causal relationships, such as the unexpected link between high-density built environments and low transmission rates, highlighting the need to consider contextual factors. The paper proposes using concepts from various disciplines—'nudge', 'affordance', and 'interface'—to organize and analyze the built environment's multiple regulatory functions. It suggests visualizing these interactions using three-dimensional representations inspired by Waddington's landscape metaphor, allowing for the depiction of complex, multivariable interactions and emergent properties. The paper further emphasizes the importance of community participation in data gathering and model building, advocating for a more equitable and inclusive approach to scientific representation. Finally, the authors present visual models (3D representations and landscape diagrams) as tools to improve the ontological, epistemological, and methodological frameworks for studying the built environment.

The paper's key findings center around the limitations of current scientific practices in representing the complex interplay between built, biotic, social, and health processes during the COVID-19 pandemic. The authors highlight three major issues: 1) **Missing data:** Current data collection methods frequently lack information on race, ethnicity, and social determinants of health, leading to overlooked disparities in COVID-19 morbidity and mortality. 2) **Idealized measurement outcomes:** Existing models often simplify complex relationships, overlooking how built environments and social factors moderate health outcomes. For instance, research showing links between air pollution and COVID-19 mortality often lacks details on race, ethnicity, or accumulated lifetime exposure. 3) **Shallow surrogate measures:** The reliance on surrogate measures in predictive models can obscure the underlying causal complexity. This is particularly problematic when dealing with social and built environment factors. The paper reveals counterintuitive relationships between factors, challenging simplistic assumptions. For example, high-density environments don't always correlate with high transmission rates; instead, social beliefs and public health behaviors play crucial roles. The paper proposes a relation-based framework where the built environment is viewed as an interface that dynamically reorganizes multiple causal landscapes, simultaneously deregulating some factors while leaving others unaffected. This framework necessitates incorporating contextual factors (e.g., social norms, cultural contexts, environmental racism) to understand the varying impact of the built environment on health. Visual representations, such as 3D landscapes and multivariable diagrams, are suggested as tools for representing these complex interactions, emphasizing the need for interdisciplinary collaboration and the incorporation of community perspectives. The authors suggest that the inadequacy of 'nudging' as a sole framework for influencing public health behaviors is due to an oversimplification of the causal connections involved, ignoring the contextual nuances that shape the success of such interventions.

The paper's findings directly address the limitations of current scientific practices in representing the complex interactions between built, biotic, social, and health processes. The proposed relation-based framework offers a more nuanced and accurate understanding of how the built environment shapes health outcomes. The emphasis on contextual factors and multidimensional visualization provides a more holistic perspective, moving beyond simple variable correlations. The inclusion of community participation in data gathering and model development is crucial for achieving equitable and socially just outcomes. The integration of concepts like ‘affordances’ and the representation of the built environment as an interface allows researchers to move beyond the limitations of existing frameworks and capture the complexity of these interactions. The paper's recommendations for improved data collection and modeling techniques have significant implications for public health interventions, particularly in addressing health disparities and improving community resilience. The insights gained can inform the development of more effective and equitable public health policies.

The paper concludes that adequately representing the built environment's role in public health requires a shift away from simplistic models towards a relation-based framework that considers multiple interacting factors and their contextual nuances. The proposed three-dimensional visualizations and multivariable landscape diagrams offer valuable tools for representing these complex interactions. Future research should focus on developing robust methodologies for incorporating community perspectives into data collection and model building. This inclusive approach is crucial for achieving equitable outcomes and addressing health disparities. Further investigation into the interactions between different factors (e.g., pollution, social norms, access to resources) within the built environment will yield a deeper understanding of the complex causal pathways that influence public health. The development of tools and methods for visualizing and analyzing these interactions will be essential for creating more effective and equitable public health interventions.

While the paper presents a comprehensive conceptual framework, it acknowledges some limitations. The proposed 3D visualization models are still conceptual, and further work is needed to translate them into practical tools for data analysis and modeling. The paper largely focuses on the conceptual framework and visualization, with limited discussion of the specific statistical or quantitative methods required for implementing these ideas. The call for greater community participation in research requires careful consideration of practical and ethical implications.