To develop the maps in the Atlas, we used three national datasets to obtain information on hospitals, nursing homes, and demographics in each of the 50 States and the District of Columbia. We underscore the fact that data representations are cross-sectional snapshots for a given period of time.
The locations of hospitals and data related to total beds, admissions, and staffing were obtained from the American Hospital Association (AHA) 2005 Annual Hospital Survey. The survey represents both AHA member and nonmember hospitals and gathers information on organizational structure, facilities and services, utilization, finances, and staffing. The average annual response rate is 83 percent; information for hospitals not responding to the survey is estimated using data from previous years.
Although the AHA Annual Hospital Survey serves as the most comprehensive inventory of hospitals, not all hospitals are included. An institution is listed by the AHA as a hospital if it is accredited as a hospital by the Joint Commission on Accreditation of Healthcare Organizations (JCAHO), is certified as a provider of acute services under Title 18 of the Social Security Act, or is licensed as a hospital by the appropriate State agency (AHA, 2005). Only acute care general hospitals are included in the Atlas. Long-term care, specialty, rehabilitation, and psychiatric hospitals were excluded. Counts of facilities, beds, admissions (total and emergency department), and full-time equivalent (FTE) registered nurses were created by summing the totals for all facilities located within a county.
The locations of nursing homes and data related to beds and staffing were obtained from the Centers for Medicare & Medicaid Services (CMS) December 2004 Provider of Services file. These data are collected by CMS through the Online Survey Certification and Reporting (OSCAR) data collection network, which contains the most comprehensive source of facility-level information on the operations, patient census, and regulatory compliance of nursing facilities.
A nursing home located within a hospital is included as a separate nursing home facility. At times, the reader may notice that hospitals and nursing home symbols are located directly on top of one another on a map. This indicates that these facilities are colocated. Other types of post-acute care facilities, such as long-term care hospitals, psychiatric facilities, and inpatient rehabilitation facilities, were not included in the Atlas. Counts of facilities, beds, and FTE registered nurses were created by summing the totals for all facilities located within a county.
Population counts classified by age, sex, and urban versus rural residence were obtained from the Census 2000 Summary File 3, available on the U.S. Census Bureau's Web site (http://www.census.gov). These data were used to calculate the percentage of the population aged 65 or older and the percentage of the population living in an urban area that were used to create the base maps in the Atlas. Population pyramids were also developed to show the age and sex distribution of the population by summing population counts into 5-year age groups by sex.
International, State, county, and census tract boundaries, as well as roads and major water bodies, were obtained from ESRI in Redlands, California. To support its software, ESRI has developed a wide range of datasets that define basic geopolitical boundaries (often referred to as base maps), transportation networks, and other geographic features. Much of ESRI's data are available online free of charge at http://www.geographynetwork.com.
The boundaries for the Health Resources and Services Administration (HRSA) Bioterrorism Hospital Regions, Metropolitan Medical Response System (MMRS), Emergency Medical Services (EMS) Regions, Emergency Management (EM) Regions, Hazardous Materials (HAZMAT) Regions, and Trauma Coordination Regions were developed using information collected through telephone conversations with State emergency planners and Internet searches.
Most of the maps in the Atlas are choroplethic maps, in which geographic units (e.g., counties) are shaded to represent different values of a variable. When the data are continuous in nature (e.g., percentage or ratio), data values are grouped into several ranges, or "classes." The legends on each map page indicate the range of data values that each color represents. These maps are useful when data have been scaled or normalized in some way. When data are nominal in nature, data values are grouped into discrete categories, and a different color is assigned to each category.
An important feature of maps is the ability to represent multiple datasets. Each map synthesizes information from different sources into a comparative visual display. For example, many of the maps in the Atlas integrate data from the AHA (red cross symbols for the location of hospitals) and the CMS (purple circle symbols for the location of nursing homes). These symbols are graduated (change size) proportional to the number of beds in a facility (their bed size) in a given area. The legend on each map page indicates the range of bed sizes that each size symbol represents.
Most map titles list the geographic reference (e.g., State name) and the theme of the map. Legend captions provide a more detailed explanation of the data displayed on the map. The reader should begin by reading the title to understand the general topic of the map, followed by the legend for explicit details about the data, and finally closely examine the map.
There are many different methods for deciding class breaks, or data ranges, for choroplethic maps. In the Atlas, we enable comparison between maps by using the same class breaks across maps. For example, the percentage of the population living in an urban area has the same class breaks across all of the 51 State maps as well as the case study maps that use this variable.
The class breaks used for the maps were developed using the "natural break" method. The natural breaks classing method identifies logical breaks, or groupings, that naturally exist in the data. These logical breaks are identified by making a bar graph showing the frequency distribution of the data (e.g., a histogram). For example, the frequency distribution of the data on the percentage of the population living in an urban area showed a gap in the data values between 8 and 10 percent. There were many areas with 8 or 10 percent of the population living in an urban area but fewer areas with 9 percent of the population living in an urban area. A natural break in the data occurs between 8 and 10 percent, so 9 percent becomes the end of one class break and the beginning of another.
The natural breaks for the six case study States were visually examined and combined to create a consistent class system for all of the maps in the Atlas. Both the demographic data and the facilities data were classed using this same method.
Color assists the reader in understanding the data and should reflect the subject matter of the map as well as the values of the data. The appropriate use of color is important because it reduces the chances of seeing patterns that do not exist or missing patterns that do.
The Atlas uses color schemes developed by Cynthia Brewer (1994). The color schemes were developed to be accessible to individuals who are color-blind and can be photocopied in black and white without loosing contrast. These color schemes can be accessed through a Web-based application called ColorBrewer at http://www.colorbrewer.org.
On many of the maps, colors are arranged from dark to light, with dark colors representing high values and light colors representing low values. For example, dark blue corresponds to a high percentage of an elderly population living in an area. On the regional maps (i.e., HAZMAT, Trauma, EMS, and EM), color schemes were used to delineate the different regions. These colors do not correspond to a value; rather, they are chosen to make identifying each region as easy as possible.
For example, in North Carolina, EMS Regions 1 and 2 are adjacent to each other, so contrasting colors of purple were chosen to represent each area. Finally, pre-press files were prepared for fourcolor separation according to the U.S. Government Printing Office (GPO) guidelines for best practices for submitting electronic design and pre-press files.
Occasionally, areas on a map are hatched, or symbolized by diagonal lines. The addition of hatched lines is used to convey an additional piece of information, such as the presence of an MMRS, without masking other data on the map, such as color.
The maps in the Atlas were created using ArcView 9 [for Windows®, 2004] geographic information and mapping software from ESRI. Data on the total beds and location of hospitals and nursing homes and demographic characteristics of the population were loaded into the ArcView software and used in conjunction with ESRI's geographic data (i.e., geopolitical boundaries and transportation networks) to develop the maps in the Atlas. Publication of the Atlas was prepared using the following software: Adobe® InDesign CS 2, Adobe® Illustrator CS 2, and Adobe® Photoshop CS 2.
Brewer CA. Color use guidelines for mapping and visualization. In: MacEachren AM, Taylor DRF, eds. Visualization in Modern Cartography. Oxford, England: Pergamon Press; 1994:123-47.
American Hospital Association. AHA Guide to the Health Care Field. Washington, DC: American Hospital Association; 2005.