Most organizations generate or collect data that includes street addresses, ZIP Codes, or other georeferences. GIS is able to spatially enable this data — that is, add geographic coordinates or make data records joinable to boundary maps. Geocoding, also known as address matching, uses street addresses as input and assigns point coordinates to address records on or adjacent to street centerlines, such as in the TIGER/Line street maps. Geocoding uses a sophisticated program that has built-in intelligence — similar to that of a postal delivery person delivering your mail — that can interpret misspellings, variations in abbreviations, and rearrangement of certain address components.
Policy, planning, and research activities often require data aggregated over space and time, rather than individual points. For example, in a study of demand patterns for locating a satellite medical clinic, it may be desirable to aggregate patient residence data to counts per census tract or ZIP Code boundaries for a recent year. GIS has the unique capacity to determine the areas in which points lie, using a spatial join or overlay function that allows the analyst to count points or summarize their attributes by area, using sums or averages.
This leads to the last GIS capability to be described in this section, proximity analysis. As an example, we conducted a study to determine whether the decline in the use of senior centers was affected by the distance from the seniors’ residences (Johnson et al. 2005). We geocoded the senior centers that provided human services and where the senior citizens lived, placing them as points on a map, and included the total target population using census data by city block. Then we used ArcGIS to create buffers around the facilities using a radius of 0.5 mi, 1 mi, 1.5 mi, 2 mi, and so on. We used spatial joins to assign buffer identifiers to residence points and blocks, count clients by buffer area, and sum up the population by block group.
Next, through careful subtraction of counts and sums, we were able to get the total number of clients and the target population in each ring around facilities (such as 0.5 mi, 1 mi, 1.5 mi), calculate use rates for each ring by dividing the count of clients by the sum of the target population, and plot the relationship of use rate versus mean distance of each ring from facilities. We found that the use rate by clients, who were elderly users of the senior center facilities in our county, declined rapidly with distance, presumably because of the seniors’ mobility limitations. The policy implication is that senior centers need to be located in areas that have high densities of elderly populations so that more of the elderly will be able to use them. You will conduct similar buffer analyses in chapters 7, 9, and 10 in a variety of health contexts.
Summary of chapters
Following is a brief summary of chapters. Chapters 1 – 9 include tutorials for learning GIS and chapters 10 and 11 are optional case studies that reinforce the skills learned in the previous chapters.
Chapter 1: Introducing GIS and health applications
Health care is a large, growing, and complex sector of economies around the world. The United States, like other countries, faces many challenges in providing the best health care for its citizens and at the lowest possible cost. Currently, many health informatics systems are manual and/or nonintegrated (The Economist 2005) in that interacting organizations have in-house information systems that are not connected to each other in ways that would allow beneficial sharing of data. This has led to top-heavy administrations, high costs of transactions (Hagland 2004), medical errors, and duplication of efforts, such as unnecessary medical tests (Protti 2005). Additional health policy issues can arise from too much emphasis on the treatment of sickness and not enough on the prevention of illness (Kennedy 2004), enormous numbers of uninsured persons, the overuse of emergency rooms, nursing shortages, and the lack of preparedness for bioterrorism (Featherly 2004).
One clear trend in health policy, research, planning, and management is the increasingly important role of health informatics. More systems will become automated and integrated, at a large cost but with even larger benefits. What does this mean for GIS applications? One consequence is that there will be even more data available for possible input to GIS — additional data on patients, facilities, programs, and events that include disease incidence, medical diagnostics, and treatments. Much of the additional data will have street addresses, ZIP Codes, or other location elements that will make it applicable to GIS processing.
In addition to ArcGIS for Desktop, there are online and mobile applications that allow you to view, download, and share basemaps. ArcGIS.com map viewer is a free web application that allows sharing and searching of geographic information, as well as content published by Esri, ArcGIS users, and other authoritative data providers. ArcGIS.com map viewer allows users to create and join groups and to control access to items shared publicly or within groups. For those wanting to explore using GIS on mobile devices, ArcGIS for iOS is a free application that runs on an iOS device such as iPad or iPhone. You can download this free application to your iOS device to view web maps created in ArcGIS Online and perform simple GIS functions such as measuring distances and areas.
In this chapter, you begin to explore the online sites by creating layer packages in ArcMap, one of the primary components of ArcGIS for Desktop, to upload to ArcGIS.com as well as explore health content already available on this website.
The book overall contains a sampling of health GIS applications that cut a wide path across the landscape we have just described.
Chapter 2: Visualizing health data
Chapter 2 uses a simple public health application for visualizing breast cancer mortality rates at the state and county levels across the United States. In the tutorials, you work with breast cancer data. While valuable for providing a snapshot of cancer mortality, the primary purpose of the application is to get you comfortable with map navigation in ArcGIS. ArcGIS provides many ways to change scales and views of a map in search of information, as well as ways to work with the data records behind the map features. You will get experience with some of them in this chapter.
Chapter 3: Designing maps for a health study
In chapter 3, you learn about uninsured populations in a state and their health-care financial needs. In the tutorials, you analyze where to locate programs that provide health-related financial support for uninsured populations in Texas counties. You use an advanced map with a bivariate analysis to contrast the magnitude of uninsured populations with measures of poverty by county.
GIS can provide many kinds of outputs. In this chapter, you build stand-alone map layouts that contain components such as multiple map frames, legends, and scale bars for use in presentations and reports.
Chapter 4: Projecting, downloading, and using spatial data
What are the most appropriate ways to map disease incidence (numbers of cases) versus prevalence (numbers of cases per population of 10,000)? In the first health application in chapter 4, you visualize a communicable disease — HIV infection and AIDS. You build two kinds of map layers to encode data: one for disease incidence based on polygon centroids (points) using size-graduated point markers and the other for disease prevalence represented by the fill color for polygons.
A second application teaches you how to download international HIV/AIDS data from a website for a selected country.
A third application uses local data to pursue the problem of child