Acessibilidade / Reportar erro

Health in cities: is a systems approach needed?

Saúde nas cidades: precisamos de uma abordagem sistêmica?

Salud en las ciudades: ¿es necesario un enfoque sistémico?

Abstract

This paper reviews the potential utility of using the concepts and tools of systems to understand and act on health in cities. The basic elements of systems approaches and the links between cities as systems and population health as emerging from the functioning of a system are reviewed. The paper also discusses implications of systems thinking for urban health including the development of dynamic conceptual models, the use of new tools, the integration of data in new ways and the identification of data gaps, and the formulation of different kinds of questions and identification of new policies. The paper concludes with a review of caveats and challenges.

Health Inequalities; Social Determinants of Health; Cities; Urban Health

Resumo

O artigo faz uma revisão da utilidade potencial de conceitos e ferramentas sistêmicos para compreender e agir sobre a saúde urbana. A revisão inclui os elementos básicos das abordagens sistêmicas e os elos entre cidades enquanto sistemas e a saúde da população, a partir do funcionamento de um sistema. A autora também discute as implicações do pensamento sistêmico para a saúde urbana, incluindo o desenvolvimento de modelos conceituais dinâmicos, o uso de novas ferramentas, novas formas de integração de dados e a identificação de lacunas nos dados, e a formulação de tipos diferentes de questões e a identificação políticas novas. O artigo conclui com uma revisão das limitações e desafios.

Desigualdades em Saúde; Determinantes Sociais da Saúde; Cidades; Saúde Urbana

Resumen

Este trabajo analiza las posibilidades de uso de los conceptos e instrumentos de la teoría de sistemas para entender y actuar en el ámbito de la salud urbana. Se revisan los elementos básicos del enfoque sistémico y los vínculos entre las ciudades como sistemas y la salud poblacional como fenómeno emergente del funcionamiento del sistema. Se discuten también las implicaciones del pensamiento sistémico para la salud urbana, incluyendo la elaboración de modelos dinámicos conceptuales, el uso de nuevos instrumentos, la integración innovadora de los datos, la identificación de carencias de información, el planteamiento de problemáticas nuevas y la elaboración de nuevas políticas. El artículo concluye con una reflexión sobre las dificultades y problemas que todo ello plantea.

Desigualdades en la Salud; Determinantes Sociales de la Salud; Ciudades; Salud Urbana

Levels of urbanization have grown exponentially worldwide. Today over half of the world’s population lives in cities, and it is expected that 60% will live in cities by 203011. World Health Organization. Urban population growth. http://www.who.int/gho/urban_health/situation_trends/urban_population_growth_text/en/.
http://www.who.int/gho/urban_health/situ...
. Urbanization rates in some regions of the world are exceptionally high. For example Northern America is the most urbanized: 82% of the population lived in urban areas in 2010, and its population is projected to be almost 89% urban by 2050. About 79% of the population of Latin America lived in urban areas in 2010. By the middle of the 21st century Latin America’s population is projected to be 87% urban 22. Population Division, Department of Economic and Social Affairs, United Nations. World urbanization prospects, the 2011 revision. New York: United Nations; 2012.. As urban populations have grown, the health issues faced by city residents have expanded beyond the traditional urban health concerns linked to infectious diseases and toxic environmental exposures to also encompass chronic diseases linked to poor diets, sedentary life styles, and obesity, as well as physical and mental health issues linked to violence, poverty, and unemployment. In addition, because city residents are often very diverse in race/ethnicity and socioeconomic circumstances, cities typically have large inequalities in health across social groups that are often manifested spatially as pronounced differences in health across neighborhoods. These inequalities are in turn reinforced by important differences across neighborhoods in physical and social environments important to health.

Cities also present enormous opportunities for health improvement There are many examples of the ways in which various urban policies – ranging from the design of cities, to the location of food stores and recreational facilities, to improving access to early childhood education, to the more traditional public health policies of infection control and mitigation of environmental hazards – can have measurable beneficial impacts on health. Despite this potential, urban dwellers continue to experience many adverse health outcomes and large health inequalities. The recent emphasis on multilevel approaches to research and intervention in population health has highlighted the ways in which a range of policies – and more generally many of the ways in which we organize or fail to organize ourselves as a society – can have important health consequences. Although this is no different in urban than in rural areas, cities by force of their sheer density and diversity of population (with consequent environmental and social implications) have always presented unique challenges as well as opportunities to health.

Cities as systems

There is a long tradition of conceptualizing cities as systems. Early on systems approaches to cities involved objectives like modeling the development, evolution, and decay of cities over time. For example, over 40 years ago, Forrester, often described as one of the founders of system dynamics, used a growth model to generate the life cycle of an urban area from its founding through its decay 250 years later. He also used systems modeling to explore how various policies (related to employment, housing, and industry) might affect the evolution of a depressed city over the next 50 years 33. Forrester J. Urban dynamics. Boston: Pegasus Communications; 1969.. More recently, in his book The New Science of Cities , Batty 44. Batty M. The new science of cities. Boston: MIT Press; 2014. argues that to understand cities we must view them as systems of networks and flows, and uses the concepts and tools of complex systems to discuss the structure of cities and how they function.

Although multiple definitions of systems and complex systems exist, key elements of systems include (1) the presence of factors at multiple levels of organization (e.g. city level factors related to governance, neighborhood factors such as environmental or social characteristics, various institutions and their organizational characteristics, families, and individuals); (2) heterogeneous units, be they individuals, institutions, neighborhoods etc.; (3) dependencies between units, e.g. norms being transmitted from person to person or neighborhoods close together in space influencing each other; and (4) positive and negative feedbacks (environmental policies influencing residential segregation and segregation in turn reinforcing environmental policies) 55. Galea S, Hall C, Kaplan GA. Social epidemiology and complex system dynamic modelling as applied to health behaviour and drug use research. Int J Drug Policy 2009; 20:209-16.,66. Diez Roux AV. Complex systems thinking and current impasses in health disparities research. Am J Public Health 2011; 101:1627-34.,77. Sterman JD. Learning from evidence in a complex world. Am J Public Health 2006; 96:505-14.. Acting together, these factors and processes result in the emergence of patterns that may not be easily understandable or predictable if one is not cognizant of the functioning of the system as a whole. Arguably, identifying the most appropriate intervention or policy to move the systems in a particular direction or predicting the impact of an intervention or policy can be difficult (or perhaps even impossible) via reductionist approaches that attempt to isolate the effect of a given factor while keeping all others constant. This is because the effect of a given intervention on the system depends on the relationships and state of all other key system components.

Systems thinking in health

Separately from the long tradition of conceptualizing and modeling cities as systems, there has been increasing interest in applying the concepts and tools of systems thinking and systems analysis to population health problems 55. Galea S, Hall C, Kaplan GA. Social epidemiology and complex system dynamic modelling as applied to health behaviour and drug use research. Int J Drug Policy 2009; 20:209-16.,66. Diez Roux AV. Complex systems thinking and current impasses in health disparities research. Am J Public Health 2011; 101:1627-34.,88. Pearce N, Merletti F. Complexity, simplicity, and epidemiology. Int J Epidemiol 2006; 35:515-9.,99. Auchincloss AH, Diez Roux AV. A new tool for epidemiology: the usefulness of dynamic-agent models in understanding place effects on health. Amn J Epidemiol 2008; 168:1-8.. Population health in the context of cities is no exception. Systems approaches go beyond the recognition that health is affected by distal factors or factors defined at multiple levels of organization, as has been the tradition in public health for a long time. They explicitly allow for dynamic processes including feedbacks as well as interdependencies and interactions between individuals and between individuals and environments over time. These feedbacks and interdependencies can result in non-linear relations and unanticipated effects distant in space or time. In going beyond the traditional emphasis on isolating the “independent” effect of specific factors to an understanding of the functioning of the system as a whole, the use of systems approaches implies a paradigm shift in the way in which population health is conceptualized, studied, and intervened on.

The study of infectious diseases (because of the contagious nature of many infections) has long embraced and incorporated some features of systems in analytical approaches (such as the need to account for transmission from individual to individual and consequent dependencies), but only recently have systems approaches garnered attention in other areas of health 1010. Auchincloss AH, Riolo RL, Brown DG, Cook J, Diez Roux AV. An agent-based model of income inequalities in diet in the context of residential segregation. Am J Prev Med 2011; 40:303-11.,1111. Yang Y, Diez Roux AV, Auchincloss AH, Rodriguez DA, Brown DG. A spatial agent-based model for the simulation of adults’ daily walking within a city. Am J Prev Med 2011; 40:353-61.,1212. Cerda M, Tracy M, Ahern J, Galea S. Addressing population health and health inequalities: the role of fundamental causes. Am J Public Health 2014; 104 Suppl 4:S609-19.. Two important factors have contributed to this trend. One factor is the increasing frustration with traditional approaches used in population health research, essentially the randomized experiment and the use of observational data to approximate the randomized experiment to the extent possible. Although these strategies have yielded much useful information, there is a sense that they do not allow us to fully understand or identify the best way to act upon many big outstanding questions in health, including key drivers of population health trends and disparities in health across social groups 55. Galea S, Hall C, Kaplan GA. Social epidemiology and complex system dynamic modelling as applied to health behaviour and drug use research. Int J Drug Policy 2009; 20:209-16.,66. Diez Roux AV. Complex systems thinking and current impasses in health disparities research. Am J Public Health 2011; 101:1627-34.. It is believed that novel conceptualizations and tools are necessary to transcend impasses in our understanding. A second major factor has been increasing availability of the tools needed to build and simulate appropriate formal models of systems. This has led for example to an explosion of interest in agent-based models, as an accessible and often intuitive approach to systems modeling for the novice 99. Auchincloss AH, Diez Roux AV. A new tool for epidemiology: the usefulness of dynamic-agent models in understanding place effects on health. Amn J Epidemiol 2008; 168:1-8.,1313. Marshall BDL, Galea S. Formalizing the role of agent-based modeling in causal inference and epidemiology. Am J Epidemiol 2015; 181:92-9..

Implications for understanding health in cities

It is important to consider how the long tradition of studying urban systems and the more recent interest in applying systems approaches to the study of population health problems can be brought together to enhance our understanding of the drivers of health in cities and the plausible impact of various policies and interventions. In considering the utility of systems approaches for the study of urban health it is important to move beyond metaphorical discussions to concrete applications that illustrate the informativeness and utility of systems approaches to answer meaningful questions about the drivers of urban health or the plausible impact of various actions. When used with scientific rigor, systems approaches may yield new insights into old problems, identify new important questions, and point to new kinds of data that need to be collected or processed in order to advance the goal of improving health in cities. But what are some of the specific ways in which systems thinking can transform research and action in urban health?

One major implication of applying systems approaches to urban health problems (perhaps the most important implication) is the development of conceptual models of the processes leading to health in urban settings that explicitly incorporate dynamic relations, i.e. dependencies and feedbacks, as appropriate to the problem at hand. These models must move beyond generic depictions to very specific conceptualizations relevant to a given health problem or research question. The development of these models may involve input from stakeholders as well as scientists as appropriate to the research problem and context. A major challenge in developing these dynamic conceptual models is setting the bounds and including only the elements fundamental to understanding the process at hand. It is important to remember that a systems model (or a complex systems model) is not necessarily a very complicated model. Intelligent abstraction is key if these models are to be useful in advancing knowledge. It is also important to recognize that often many aspects of these models can be investigated using traditional approaches (i.e. formal simulation of a system is not always necessary). The dynamic conceptual model in itself serves to place findings in context regardless of whether formal simulation follows. Moreover, the process of developing these models often starkly illuminates where knowledge about the underlying processes is thin or absent, and points to new directions for inquiry. The knowledge gained about components of the dynamic model using traditional approaches can then be used to refine the conceptual model and inform formal simulation of the system in the future.

A second implication is the need to employ new tools, specifically the development of a formal simulation model that can then be used to explore the functioning of system and the plausible impact of various interventions on the system. Various types of formal simulation models exist including agent-based models and systems dynamic models. These formal models involve the creation of virtual worlds, hopefully informed by empirical data that can then be manipulated to better understand the health implications of the various relations encoded in the model. It is important to emphasize that the extent to which conclusions derived from the model are valid in the real world depends on the extent to which the model appropriately captures the essential processes. The model is not the real world, but merely our effort to mimic how it functions 1414. Diez Roux A. The virtual epidemiologist: promise and peril. Am J Epidemiol 2015; 181:100-2.. Therefore, simulation modeling will never replace rigorous observation, although it can help us understand the implications of several observed facts acting in concert.

Third, the use of systems approaches also has implications for the ways we collect and integrate data. By definition, the development of a formal systems simulation model requires integrating various kinds of data, including both qualitative and quantitative data. Meaningful systems models allow us to understand the implications and consequences of various pieces of quantitative and qualitative information in ways that we would not necessarily have predicted if we did not do the formal simulation. This is where the real value of systems modeling lies. In addition the process of conceptualizing the dynamic model and then formalizing the simulation model often leads to the identification of important gaps in data or suggests new questions that can be answered using traditional approaches. The collection of new data and the responses to new questions can also subsequently serve to improve our understanding of how the system works and improve the validity of the simulation model for answering other questions.

Lastly, the use of systems approaches also has implications for the types of questions we ask, and ultimately for the types of policies we can identify as important to improving health in cities. For example, a traditional urban health question might be “Are neighborhood characteristics independently associated with health after accounting for the individual-level socioeconomic position of residents?” whereas the related question that we would answer with a systems approach would be “To what extent (and under what conditions) could residential segregation generate, and reinforce, health disparities by race?” In another example, the traditional question would be “Is proximity to supermarkets (as a proxy for healthy food access) associated with better diet after adjustment for individual-level characteristics of residents?”, whereas a more systems oriented question would be “What is the plausible impact on dietary health inequalities of a strategy to subsidize the location of supermarkets in certain areas under various spatial patterning scenarios?” 1010. Auchincloss AH, Riolo RL, Brown DG, Cook J, Diez Roux AV. An agent-based model of income inequalities in diet in the context of residential segregation. Am J Prev Med 2011; 40:303-11.. These different kinds of questions allow for the possibility not only of directly assessing the plausible impact of various policies but also of identifying potentially useful policies that we might not have imagined (or identified as relevant) before performing the simulation modeling. Here the emphasis on “plausible” and “potential” is key, because as we have noted, the model only allows us to identify implications of an action under the set of conditions encoded in the model. Often the insights obtained from simulation modeling still need to be tested in the real world.

Caveats and conclusion

Current interest in systems thinking in population health presents important opportunities for creative thinking in the area of urban health. Aside from the crucial factor of encouraging the development of dynamic conceptual models of health in cities, systems approaches provide us with new analytical tools (specifically simulation modeling) that can be used to enhance insight and identify new areas for additional exploration using a range of analytical approaches. Systems thinking also provides formalized ways for urban health researchers to link to other disciplines interested in the functioning of cities and allows input from various stakeholders and communities in the formulation of conceptual models and in the interpretation of results of simulation modeling. However, in order to be truly transformative these new approaches need to be employed to answer very specific questions and yield new insights. Moving beyond metaphorical discussions of urban systems to concrete applications that yield new information will be essential. At their best, the use of systems approaches will allow us to see and understand patterns and trends that researchers, communities, and policymakers may not otherwise be aware of. They may also allow us to identify the best ways to modify these patterns. Training in these approaches and the development of exemplar applications will be important. Identification of specific types of situations in which systems approaches may be most useful will also be valuable. Systems approaches will never replace traditional quantitative and qualitative empirical approaches in urban health research, but if integrated with other strategies, they may help advance our understanding of the determinants of urban health as well as enhance our ability to intervene to improve health and reduce health inequalities in urban settings.

References

  • 1
    World Health Organization. Urban population growth. http://www.who.int/gho/urban_health/situation_trends/urban_population_growth_text/en/.
    » http://www.who.int/gho/urban_health/situation_trends/urban_population_growth_text/en/
  • 2
    Population Division, Department of Economic and Social Affairs, United Nations. World urbanization prospects, the 2011 revision. New York: United Nations; 2012.
  • 3
    Forrester J. Urban dynamics. Boston: Pegasus Communications; 1969.
  • 4
    Batty M. The new science of cities. Boston: MIT Press; 2014.
  • 5
    Galea S, Hall C, Kaplan GA. Social epidemiology and complex system dynamic modelling as applied to health behaviour and drug use research. Int J Drug Policy 2009; 20:209-16.
  • 6
    Diez Roux AV. Complex systems thinking and current impasses in health disparities research. Am J Public Health 2011; 101:1627-34.
  • 7
    Sterman JD. Learning from evidence in a complex world. Am J Public Health 2006; 96:505-14.
  • 8
    Pearce N, Merletti F. Complexity, simplicity, and epidemiology. Int J Epidemiol 2006; 35:515-9.
  • 9
    Auchincloss AH, Diez Roux AV. A new tool for epidemiology: the usefulness of dynamic-agent models in understanding place effects on health. Amn J Epidemiol 2008; 168:1-8.
  • 10
    Auchincloss AH, Riolo RL, Brown DG, Cook J, Diez Roux AV. An agent-based model of income inequalities in diet in the context of residential segregation. Am J Prev Med 2011; 40:303-11.
  • 11
    Yang Y, Diez Roux AV, Auchincloss AH, Rodriguez DA, Brown DG. A spatial agent-based model for the simulation of adults’ daily walking within a city. Am J Prev Med 2011; 40:353-61.
  • 12
    Cerda M, Tracy M, Ahern J, Galea S. Addressing population health and health inequalities: the role of fundamental causes. Am J Public Health 2014; 104 Suppl 4:S609-19.
  • 13
    Marshall BDL, Galea S. Formalizing the role of agent-based modeling in causal inference and epidemiology. Am J Epidemiol 2015; 181:92-9.
  • 14
    Diez Roux A. The virtual epidemiologist: promise and peril. Am J Epidemiol 2015; 181:100-2.

Publication Dates

  • Publication in this collection
    Nov 2015

History

  • Received
    07 Nov 2014
  • Accepted
    19 Nov 2014
Escola Nacional de Saúde Pública Sergio Arouca, Fundação Oswaldo Cruz Rua Leopoldo Bulhões, 1480 , 21041-210 Rio de Janeiro RJ Brazil, Tel.:+55 21 2598-2511, Fax: +55 21 2598-2737 / +55 21 2598-2514 - Rio de Janeiro - RJ - Brazil
E-mail: cadernos@ensp.fiocruz.br