We recently announced that we started to scale the System Graph beyond health and biomedical research into climate change and the environment. Today, we’re sharing that we have started to expand into sociology, economics, political science, and business, adding support for new statistical measures such as elasticity and Gini coefficient. This is an exciting milestone on our journey to relate everything to help the world see and solve anything as a system.

Here are a few examples of what’s now possible with System:

Poverty

As the World Bank notes, “Eradicating poverty requires tackling its many dimensions. Countries cannot adequately address poverty without also improving people’s well-being in a comprehensive way, including through more equitable access to health, education, and basic infrastructure and services, including digital.”

We can now use System to explore the social, behavioral, and environmental factors related to an increased risk of poverty. In the example above, we zoom in on five: psychological distress [1], out-of-pocket healthcare costs [2], lack of family planning use [3], environmental noise pollution [4], and food insecurity [5]. (Note that bidirectional connections with food security and psychological distress indicate that we’ve extracted relationships from studies that set those variables as the dependent variable in regression calculations, as opposed to poverty. They don’t indicate causal loops). Each of these factors was established in different study contexts and populations: Australian adults, households in rural Tanzania, inner city Berlin adults, and older adults living in social housing in Ontario.

Psychological distress is a broad concept, and intuitively can be caused by many factors. Study population is once again critical when analyzing this system; when examining the upstream factor family relations, [6] showed that British single mothers experience psychological stress due to financial hardship,  [7] found higher stress levels in individuals diagnosed with epilepsy that were also experiencing family problems, and [8] found that transgender female youth in the SF Bay Area had lower risk of stress if they were close to their parents.

Poverty has many negative downstream effects for both those experiencing it and those living in high poverty areas. [9] and [10] showed that county and neighborhood poverty levels were significant predictors of firearm violence and alcohol consumption, respectively, while crime was shown to be associated with both [11]. Low health literacy rates were associated with poverty (amongst hispanics living in California) [12], which, in turn, are associated with higher rates of smoking [13] and vaccination hesitancy (amongst people living in homeless shelters in France) [14].

These are just some of the many social and behavioral factors and paths found in the complex system of poverty. We hope that by illuminating the system and keeping it updated with new evidence, we can help policymakers and funders identify new strategies, policies, and research directions for their specific context. 

Development

The Human Development Index (HDI) is an important indicator used by the UN to measure the well-being of a population. It specifically takes into account three broad dimensions: life expectancy, education, and income. System can now be used to explore over 3,000 diverse factors that can directly and indirectly contribute to a country’s HDI.

Deforestation

In a previous post, we explored the system of deforestation, linking it to over 200 health outcomes via associations with sulfur dioxide and nitrogen dioxide. Using newly extracted data, we find these health outcomes to be linked to hundreds of downstream behavioral and social outcomes. For illustration, we chose three health (asthma, cardiovascular disease, and mental depression) and four behavioral/social (poverty, employment status, school attendance, food security) outcomes to analyze here:

• (red) Asthma has been shown to increase the risk of falling into multidimensional poverty [1], unemployment [2][3], and missing school [4]
• (yellow) Cardiovascular disease has been shown to increase the risk of unemployment [5] and food insecurity [6]
• (blue) Mental depression has been shown to (and its subclass of depressive disorder) increase the risk of income poverty [7], unemployment [8][9], and food insecurity [10]
• (green) Depression can also indirectly affect employment status and food security via the negative effects of social media addiction [11], specifically alcohol consumption [12][13][14] and smoking [15][16].

Not only does this map show how deforestation may affect important social and economic metrics at the individual level, but also how just a handful of topics can have many interdependencies. You can now use System to do a deep dive into any of these topics and see, for example, how mental depression can impact over 200 behavior and social outcomes. 

Equipped with these new insights, we can step back and examine an enriched system of deforestation, now linked to HDI: 

The central four social and behavioral outcomes are all linked to the human development index, each of which is influenced by upstream biomedical factors, and eventually deforestation, as previously described. What’s becoming clear with the new data in the System Graph is how deforestation can have potentially harmful health and economic outcomes at the micro and macro levels. 

It’s important to note that this map is far from complete; in the interest of illustrating one line of analysis, we’ve left out interactions among hundreds of factors. The contexts (study population, location, time period) of the studies that constitute the nodes and edges vary across the map, and we’ve yet to perform a quantitative analysis of effect sizes over our paths, which include both causal and non-causal edges. All of these need to be considered when making decisions, and we’re working to add all these functionalities to System. 

We hope this analysis underscores the importance of systems thinking. Whereas previous discussions around deforestation, for example, may have involved the question of whether deforestation could even affect food security or poverty, we believe the question should shift to: “Now that we see these things connected, can we continue to ignore it when making decisions?”

Release Risk: You’ll note that non-biomedical concepts currently carry a disclaimer, similar to what you might find on Wikipedia, regarding the completeness of this information. As we expand the source material that we extract data from, further fine-tune our AI, and validate the completeness of these systems against canonical benchmarks (as we have done in health), we will remove this message.