April 15, 2014
A colleague of mine described a map of Baltimore City as a butterfly, while another colleague saw angel’s wings. I saw a heart. Regardless of our interpretations, it’s a pretty pattern…until you realize that this “butterfly” really comes from dots of crime locations and corner stores in Baltimore City. Sadly, this butterfly/angel’s wings/heart pattern reemerges when looking at areas of poverty, and elevated rates of chronic diseases such as cardiovascular disease and stroke, and food insecurity in Baltimore City. Looking at these maps side by side, it seems obvious that there is some relationship between all of these factors, and a series of maps saliently conveys this pattern far better than words or numbers could. However, while maps are very powerful tool within a public health practitioner’s arsenal, it’s important to remember that maps are also only a tool – one that requires careful and thoughtful use – and in many cases are really only the first step in a much more complex inquiry.
A map was an instrumental to John Snow in his investigation of London’s cholera outbreak in 1854. By mapping the locations of cholera cases, he deduced that the Broad Street pump was the source of the outbreak. More than 150 years later, maps – and more broadly geographic information systems (GIS) – are still a very valuable tool to public health practitioners, from looking at the spread of infectious diseases and the distribution of chronic diseases, investigating environmental health risks, assessing contextual risks (such as crime or liquor store locations) in urban neighborhoods that may adversely affect health, and identifying resource poor areas that could benefit from additional health care infrastructural investment. One GIS application that particularly interests me is in the area of obesity prevention: how GIS can help us understand how neighborhood factors, such as the food retail environment, may contribute to high obesity rates. With tools like ArcGIS, we can create maps that visually display rates of obesity, and overlay these maps with the locations of fast food restaurants, grocery stores, parks as well as hospitals, schools, and churches; distribution of income; or car ownership levels…the list of potential things to investigate is nearly infinite! Maps are a convenient, compelling, and powerful way to visualize these relationships.
Mapping is also a very important tool for policy makers. In some cases, maps can tell a much more compelling store than odds ratios and percentages; and in other cases, they tell a story that standard regression alone could not tell. The 2008 Baltimore City Health Commissioner, Dr. Joshua Scharfstein, saw the first map of the availability of healthy and unhealthy foods in Baltimore City food retail stores and noted that the stratification of unhealthy food availability by race was “unacceptable.” This promoted him to form a task force to look into the issue (and that eventually led to the hiring of the city’s first Food Policy Director and the Baltimore Food Policy Initiative). In addition to conveying information, policy makers use maps as a tool to guide policy decisions. Maps are particularly useful in identifying areas of high need. In Baltimore City, the Baltimore Food Policy Initiative has employed GIS to identify underserved areas that could benefit from zoning codes that would increase access to healthy food. The Chicago Development Fund, a non-profit founded by the city of Chicago, provides financial incentives to projects and businesses that benefit low-income communities. Their website has a map identifying target census tracts in Chicago for potential projects.
However, with all this excitement over the potential of mapping and GIS capabilities, it’s important to remember that maps are a tool and like any tool, require thoughtful consideration about how to best use it. In particular, maps should be thought of as the start rather than the end product of a research question. Maps are incredibly useful for generating hypotheses about potential spatial relationships. Looking at a map, we may notice that certain outcomes cluster in a particular area, or that different outcomes tend to cluster in the same place. However, maps are subject to interpretation: just as I saw a heart and my coworker saw a butterfly, different people can interpret the same map differently. Maps, themselves, do not provide any statistical inferences about whether spatial correlations actually exist.
Additionally, researchers employing GIS have to consider whether this method is truly appropriate given the research question: what research questions can maps answer? Are the findings we see consistent with our theories? What are the underlying assumptions of the maps?
Maps may be able to hint at correlations, but we have to careful when we use information from maps to create interventions. For example, maps may suggest a strong correlation between crime and obesity rates in Baltimore city, but an intervention that includes more policing to reduce crime in “hot spot” neighborhoods may not have an effect on obesity. Perhaps it is a third, underlying factor, such as poverty or inequality, that causes both crime and obesity. Moreover, we can manipulate maps to tell different stories based upon our how we present our information. Should we present average rates for a health outcome of interest at the city, the county, or the state level? For example, if we were interested in the variation in average income in Maryland, very different stories would emerge if we looked at average income county, city, census tract, or even neighborhoods.
The Center for a Livable Future has undertaken a large mapping project to better understand its local food system. The Maryland Food System Map Project collects data on all aspects of the food system, in addition to health and demographic data, and maps them in an online geodatabase. Realizing that we are just beginning to understand geographic relationships in the food system and how they impact public health, the project is available for presentations as a way to spark conversations about understanding the root causes of these relationships, and inspiring further research into opportunities to strengthen the food system.
Whether it’s a heart, a butterfly, or angel’s wings, maps have enormous potential in the realm of public health. Their use in research and policy has become increasingly common, but it is also important to remember that they, like all other tools, also require careful consideration in how they are used. They are a start into a much more complex and fascinating inquiry.