Articles Currently Under Peer Review by the URISA Journal

PUBLIC HEALTH AND HUMAN SERVICES
(Version 1-24-00)

Gerard Rushton, Gregory Elmes, and Robert McMaster

Objective: The Challenge of Improving Health and Human Services Using GIS

Many activities to promote better health and to reduce disease are directed at changing the social, economic, political and physical environments in which people live. Using GIS, independently made observations of any of the above can be referenced to a common geo-spatial data framework. This permits different organizations to share spatial data about these phenomena. GIS promises to bring rich information data bases, linked to methods of spatial analysis, to determine relationships between geographical patterns of disease distributions and social and physical environmental conditions. As the core of a decision-support system GIS also has the potential to change the way the geographical allocations of resources are made to facilitate preventive health services and to control the burden of disease

Traditionally, administrative areas or other spatial units such as census-defined areas were the geographic units where health status and health outcomes were measured and where health resources were allocated. However, if the areas were small, for example, counties, the movement of people from one county to another made this spatial accounting scheme inaccurate for measuring the relationships between health status, health resources and health outcomes, and inappropriate as decision-making units. For decades, for example, the Federal Government has struggled to devise appropriate spatial accounting methods by which local communities deemed to be disadvantaged with respect to health resources could be assisted without spending scarce resources on other communities not so disadvantaged. (See Bureau of Primary Health Care, 1998; GAO, 1997; Lee, 1991) If the areas used to report disease data and resource allocations were larger than counties, however, differences within the areas were often large and they were inappropriate for both analysis and decision-making. In different regions of the U.S., of course, counties differ greatly in size and population. Other countries have administrative areas of different sizes, too.

It is clear that different health resources affect levels of health and disease over different local areas of different size. Health systems, that is, operate at a multitude of spatial scales that are constantly changing with the reorganization of health resources and the behavior of health-seeking populations. As many public health as well as private health organizations have discovered, to understand and to make decisions about this complex system of inputs and outcomes requires operational access to information at very local geographic scales. In this respect, health systems are no different than many other service systems for which people receive services from dispersed facilities. As with such systems the old model of a spatial accounting framework of counties or fixed service areas is now being replaced by a GIS-based model in which the geographic scale at which information is analyzed changes according to the kind of question being addressed. Using GIS, data from small local areas can now be flexibly aggregated to larger areas that are meaningful for the questions asked and the decisions to be made.

BACKGROUND

Federal, State and local institutions have recently developed a strong interest in GIS and health. The Centers for Disease Control (CDC) has supported the development of software for mapping diseases, (Dean, 1999). The National Cancer Institute (NCI) has supported the development of software for disease cluster identification, (Kulldorff, et al., 1998). CDC and ATSDR—a branch of EPA--have cooperated in sponsoring and organizing four annual meetings on GIS and Public Health. The most recent in San Diego in August, 1998. CDC also collaborates with the National Center for Health Statistics and the US Department of Health and Human Services on an annual conference on health statistics that increasingly has geographic information and internet technology components. In 1998, the National Cancer Institute (NCI) in collaboration with the National institute of Environmental Health Sciences (NIEHS), under Public Law 103-43 requested proposals to develop a health-related geographic information system (GIS) for Long Island. "The prototype health-related GIS will provide researchers a new tool to investigate relationships between breast cancer and the environment on Long Island, and to estimate exposures to environmental contamination." (NCI, 1999a, p.3).

Despite all this activity, there exists considerable skepticism in many quarters about the role of mapping and spatial analysis in the analysis of disease patterns and resource allocation. For example, Dennis Whalen, Executive Deputy Commissioner, New York State Department of Health, in testimony before a Committee of the NYS Assembly on March 8, 1999 noted in a discussion of "Cancer Mapping Challenges" that:

"Some experts say mapping itself, is ineffective—that maps will provide little additional information about cancer patterns, so resources should be directed to more promising research. They argue that maps presuppose a geographic link to cancer cases that cannot be proven, and in fact, may be completely irrelevant. For example, one would have to question the validity of overlaying a map of current environmental exposure data on a 1991 through 1996 cancer incidence map knowing that a particular type of cancer may have a latency period of 10 to 20 years, and that many of those diagnosed with a common cancer may not have lived in the area long enough for their cancers to have a common cause. Yet many mapping supporters expect that maps will definitively identify "hot spots". They expect that maps will demonstrate a cause and effect relationship between cancer cases and a particular risk factor or factors."

Mr. Whalen went on to argue that mapping might be a useful tool to focus further research and for generating hypotheses. Furthermore, maps can help target efforts such as increased physician education on available treatments. He concluded "We believe that cancer mapping is the next logical step to address the call by New Yorkers for more information about cancer cases in their communities." The New York State Health Department has formed an advisory committee on Cancer Surveillance and much of its research plan relates to GIS use.

There have been many reviews of the use of GIS in public health or in the provision or planning of health services—Briggs and Elliott, 1995; Clarke et al., 1996; Croner et al., 1996; Richards et al., 1999; Rickets et al., 1994; Rushton et al., 1997; Rushton,1999; Vine et al., 1997; Waller, 1996; Yasnoff and Sondik, 1999. These reviews, however, have focussed on the potential use of GIS as currently conceived. This review focusses on the educational and research needs to fulfill the potential of improving health with GIS. It is our contention that, although a great deal can be accomplished to improve health with GIS, significant challenges exist that only further research in GIS and health can solve.

THE UCGIS APPROACH—WHAT ARE THE KEY ISSUES?

Educational

Most public health workers hold the degree of MPH (Masters of Public Health). Teachers, researchers, and leaders of public health organizations hold more advanced degrees. Typical curricula for these degrees do not normally include GIS related subjects. As biostatisticians begin to use GIS, there are some common misunderstandings about what is needed to work effectively with GIS. Increasingly, they are investigating relationships between disease and environmental factors. In the case of major recognized polluted areas (Superfund sites), they often inherit spatially referenced data systems from legacy information systems. In such situations they is a tendency to confuse geographic information systems with geographic information science. Wanting to develop support systems for researchers working in this area, they often hire staff for their research institutes without stipulating in their personnel search that the person to be hired should know the concepts and terminology of GIS. The knowledge they think they need is experience and knowledge of geographic information systems—meaning particular software with which they intend to work. Only later do they discover that their new personnel do not know the basic data models and the conversion methods between data models. These may be raster or vector, TIN or network models. Their staff should know how to use recognized georeferencing and coordinate systems, including relative georeferencing and map projections. They should know the language and concepts of geometric and attribute accuracy. They should know about buffer zones for points, lines and areas as well as relational and hierarchical database systems and object oriented systems. They should know about positional data accuracy, hash functions, quad-trees and spatial logic operations. They should know the principles of aerial photo interpretation as well as supervised and unsupervised classifications. They should know some principles of surveying and remote sensing, street centerfile systems and address-matching; digitizing and scanning of spatial data. These matters are not covered within the typical curricula of departments of statistics, biostatistics, public health, or computer science. From where will practicing public health workers or academic scientists who study the relationships between environments, disease and health find people with this knowledge, (Bernhardsen, 1999).

There are serious educational needs for both researchers and practitioners in the health fields who are using GIS in their work and who are struggling to find the educational resources to meet their needs. CDC and ATSDR are currently developing distance learning modules on GIS and Public Health and they expect to broadcast these soon using satellite-based, video broadcasting systems. Rushton and colleagues, with a grant from the Department of Education, organized five, three day workshops for health professionals between 1993 and 1997. They also developed a web presence and CDROM on the subject of GIS and Public Health, (Rushton et al., 1997). A widely-held view is that more needs to be done to educate health professionals on the use of GIS in public health activities.

In the health sciences one common approach to educating advanced professionals in areas outside their area of traditional education is through focussed, post-doctoral training programs. NIH frequently supports such programs through its ongoing support of focussed research institutes. The Basic Science Research Program for Super Fund Sites of the National Institutes for Environmental health, for example, supports research and education units that include GIS among the core support areas of several of the Research Institutes they support. Such a program in GIS might contribute to the twin goal of preparing new teachers and researchers in GIS and health and in advancing critical research areas.

Research

As an application area that only recently recognized the potential contributions of GIS-based research, the health disciplines have not yet formulated a plan for research on GIS use in public health. The National Cancer Institute prepared, for its Long Island Breast Cancer Project, a glossary of GIS-based methods that have been used in investigations of the spatial distributions of disease and possible relationships with environmental factors. It is reproduced as Appendix A. For each application area, at least one citation to published research was provided. The list demonstrates the variety of ways in which GIS has been used in research in public health in the disease analysis area.

Many current applications of GIS in health are extremely wasteful of resources in that their ad hoc nature requires that costly GIS resources be developed to support single project plans. The recently developed, NCI supported, Long Island Breast Cancer Project attempts to address this problem by supporting, under contract, the development of a GIS utility for this region. The system, currently being developed by Aver Star Inc., will develop selected spatial coverages and will implement selected spatial analysis methods prioritized by NCI from the taxonomy of methods in Appendix A.

There are other research areas—such as the location of health facilities, for which no focussed reviews yet exist of work completed or problems not yet addressed, (Cromley and Shannon, 1986; Hirschfield et al., 1993; 1995; Mohan, 1983). Recent developments in the organization of health care through the development of managed care systems have strong geographical information and analysis components. Little research on this subject exists (Perkins, 1999).

In the case of diseases such as most cancers, exposures to agents that might increase the risk of disease often predate by 10 to 20 years the diagnosis of the disease. In such circumstances, location of diagnosis and location of probable exposure are unlikely to be the same. With a population that moves its residence so frequently, the challenge of estimating the places of likely exposure of people whose location at time of first diagnosis is known is formidable. Mark and Egenhofer (1998) and Mark et al., 1999 have recently begun geographic demographic research in the United States on methods to estimate the likelihood that a person whose current residence at time of first diagnosis is at x might have lived in exposure area y, t years ago. See the section "Temporal Aspects of GIS and Health" below. Research on possible prior exposure to risks is proceeding in Sweden for environmentally-linked leukemias and child-onset diabetes (Kohli et al 1997). In the United States calls for research on integrating lifeline analysis into health GIS have emerged (Platt 1995).

Several authors have argued the merits of exploratory spatial data analysis for health applications. Haining et al. (1998) illustrate a system of analysis, SAGE, that can undertake exploratory spatial analysis (ESDA) held in the ARC/INFO geographical information system. They illustrate the system with analyses of standardized incidence rates for cancer in Sheffield. Their system permits "brushing" of the region (identifying regions) and displaying relationships between variables for the data of the region brushed. Tools for regionalization are also developed in the SAGE system (Wise et al., 1997) as are computations of local statistics such as the widely used Getis-Ord (G_i*-) statistic (Getis and Ord, 1992; Ord and Getis, 1995). Anselin and Bao (1997) have also developed an interactive computational system that links many methods of spatial analysis to ArcView GIS.

Small area demographic data is crucial for many research applications of GIS in health, particularly for estimating the values of denominators in computing small area disease rates (Elliott et al., 1992; Martin, 1996). Quality demographic data for small geographic areas, especially publicly available data, frequently is not available, especially during inter-census periods.

PRIORITY AREAS FOR RESEARCH/RECOMMENDATIONS FOR RESEARCH

Improving Disease Surveillance Data Systems

There is general agreement that location variables have not been collected well in most current disease surveillance. Until recently, for example, the New York State Department of Health Cancer Registry registered the current address of people with cancer. When their residence changed, the new location replaced the old. In the Iowa Cancer Registry, for example, the locations of specific treatments are not coded even though they are available in the written record. There is a need for disease surveillance systems to adopt uniform methods for locational coding and to introduce quality assurance and quality testing standards for locations comparable to the standards they use for other data items they code. MacDorman and Gay (1999) recently reviewed state initiatives in geocoding vital statistics data. NIH and CDC are not unaware of this problem. In a recent report of the Surveillance Implementation Group (SIG) of the National Cancer Institute (NCI, 1999b) one of 11 research opportunities identified is:

"Research Opportunity 4
Explore the feasibility and utility of employing geographic information systems for geocoding surveillance data and reporting geographic relationships among screening measures, risk factors (including environmental exposures), and improved cancer outcomes. Methods need to be developed for assuring data confidentiality. (The cost of this effort is expected to be moderate; work should be initiated within the next 1-2 years.)
Research is needed on the utility of geographic information systems (GIS) as an innovative addition to the cancer surveillance infrastructure."

There is a need to develop methods of spatial analysis that can be routinely used for exploratory analysis of surveillance data. See Rushton, 1998; Rosenberg et al., 1999.

There is also a need for a national dialog on the improvement and standardization of the quality and quantity of spatial information associated with health statistics. This should include examination of existing national record and database systems e.g. HCFA Medicare / Medicaid Parts 1 and 2, Death Certificates etc.

In concert with the need to improve health surveillance systems calls have risen for better assessment of rural health and the health of minorities (Ricketts 1994; Bureau of Primary Health Care. 1998). The Office of Social Environment and Health Research (OSEARH) at West Virginia University in cooperation with CDC has published atlases of social environment affecting heart disease in Appalachia, and for women at the national and state scale (Barnett et al. 1998; Casper et. al. 1999). Gender and minority issues have not only been relatively neglected in the epidemiological literature but also raise elevated concern about confidentially.

Risk factors as contributors to disease and ill-health

Behavioral risk factors are often discovered through national health surveys. There is a need to link the findings from such national surveys to local socio-demographics to estimate local risk factors based on expected local behavior patterns, see Braden and Beauregard, 1994; Brown et al., 1991. Sometimes, attributes that can be observed in local administrative databases can be used as surrogates to estimate disease incidence rates. For example, the density of retail alcohol sites has been linked to local rates of alcohol abuse (Mackinnon et al., 1995). Haining et al., (1994) have investigated the relationship between material deprivation and rates of colorectal cancer.

It does seem clear that, with a few exceptions, theories of spatial diffusion and related spatial models, are rarely given serious consideration in CDC and NIH research activities. Gould (1993) described the lack of interest he encountered in NIH groups that discussed the spread of AIDS in the U.S. in the 1980s. Some examples of spatial diffusion and core disease areas as explanations for current patterns of disease can be found in the work of Becker et al., 1998 and Cook et al., 1999.

Ecological Studies of the Relationship Between Environmental Factors and Disease Transmission

As a World Health Organization Report recently noted (WHO, 1996), the spread of many infectious diseases is related to the climate, vegetation and socio-economic conditions in local areas.

Figure 1. Relationship between socio-economic conditions and physical environmental conditions and the spread of infectious diseases. (This model was developed by Jamil Kazmi and E. Lynn Usery in April, 1999, at the Department of Geography, University of Georgia, Athens, Georgia, USA.)

Available in Hard Copy only!

Kasmi and Usery note that although their model was developed for malaria, it is applicable to other vector-borne diseases as well. They write:

"The basic idea is that malaria is a three-factor disease which develops with the interaction of the vector (mosquito), parasite (plasmodium) and host (man). Absence of any of these three basic factors means the absence of malaria from various parts of the world. Each factor at the individual level has many contributing elements, for example, the vector has many physical and socio-economic elements which may contribute to the transmission and control of malaria. Therefore, the role of remote sensing and geographic information systems (GIS) as modern tools to study vector-borne diseases is to identify and interpret these contributing elements."

Recent research of Spear et al. (1998) illustrate the contribution of geographic information science to this area. An illustration of an ongoing project in South Africa which uses GIS in Malaria control activities can be seen at http://www.malaria.org.za/homested.htm. This project illustrates the sensitivity of malaria control activities to the geographic scale of surveillance activities.

With the possibility of significant climate change in many areas of the world, research is needed to project the likely human health effects of such changes. The frequency and magnitude of extreme events increases health risks (Smoyer, 1998; WHO, 1996).

Temporal Aspects of GIS and Health

People's movements through geographic space are a critical factor in exposures to environmental health hazards. Computational models that can account for the fact that people's locations in geographic space are dynamic rather than static will greatly enhance the power and potential of data analysis and reasoning methods for examining environmental exposures or discovering past clusters of currently-ill patients. Individuals navigate through space, they stay at locations where they meet other individuals and they perform regularly reoccurring tasks that involve variable or fixed locations in geographic space. These movements often expose people to environmental factors that can cause health problems at latency periods ranging for seconds to decades. For example, establishing whether a particular U.S. soldier was exposed to hazardous chemicals during Operation Desert Storm requires not only the space-time distribution of environmental risk, but also a record of the space-time behavior of the soldier. If the former is not known, space-time places of high risk might be inferred by comparing the space-time behaviors of soldiers showing symptoms of ill health with the behaviors of a control group of soldiers not showing symptoms.

For many health conditions, application of GIS has been hampered by the poor ability of commercial GISs to handle multitemporal geographic information or movement (Langran, 1992; Peuquet, 1994). This shortcoming severely impedes the utility for GIS to assist in understanding health problems with long latency periods, such as many forms of cancer, since with mobile populations, the location of the patient at the time of diagnosis or mortality may have little relation to the location of exposure to toxic substances or other environmental risks.

Recently, the National Institute of Environmental Health Sciences has supported a research project focused on the extraction of health-related information from geospatial lifelines, which capture individuals' locations in geographic space at regular or irregular temporal intervals (Mark and Egenhofer, 1998; Mark et al., 1999). The objectives of this project is to develop and test the theory of geospatial lifelines in the environmental health sciences by:

• developing methods to trace locations of individual people (patients, cases, or controls) back through time, to discover spatial clusters in the past or to determine past environmental exposures,
• designing, prototyping, and assess computational models that can deal with large sets of geospatial lifelines and environmental information, and
• examining the ethical and legal implications of recording individuals' geospatial lifelines in databases, and establishing procedures for appropriate restrictions on data analysis and dissemination.

Geospatial lifelines (Mark and Egenhofer, 1998) consist of series of discrete space-time samples over the domain of continuous movements, describing an individual's location in geographic space at regular or irregular temporal intervals. Methods for the analysis of and reasoning about movement in geographic space are based in theory outlined some three decades ago by Torsten Hägerstrand (1970). Hägerstrand's work has influenced conceptualizations of spatiotemporal constraints on human activities, but have rarely been implemented computationally (but see Miller, 1991). Geospatial lifeline data may be recorded at different resolutions, but in environmental health applications, researchers are mainly concerned with data over days to entire lifetimes, with a resolution of hours to years. The research will develop new methods for the analysis of geographically and temporally referenced medical information, and new methods for reasoning about environmental exposures and their consequences over space and through time (Mark et al., 1999). The methods also will be applicable to hazardous exposures of shorter time periods with more immediate impacts. Recent interest in moving points within the spatiotemporal database community further suggests that methods based on moving points (Erwig et al., 1998), and particularly on Hägerstrand's Time Geography model (Dumas et al., 1999; Fauvet et al., 1998, 1999), will become realistic tools in environmental health in the near future. Recently Forer (1998) has articulated time-space primitives, defining timelines, and activity volumes that have direct implications for locational histories in health research. Further articulation of time geography concepts in a GIS environment and their evaluation with empirical data is essential.

This research project also will examine statistical approaches for identifying clusters of hot spots of ill health in space-time. Often it is important to determine whether observations of some phenomenon are clustered in space or time or both. When trying to determine the causes of some outbreak or chronic pattern of ill health, analysts frequently plot the distributions of cases on maps. This method has been used at least since Dr. John Snow's now famous map of cholera deaths in London, England, which helped identify a particular public water pump as the source of the epidemic (Snow 1936). For infectious diseases with short incubation periods, analysis of the spatial distribution alone may be sufficient; however, there are problems with such methods in the study of environmentally-induced diseases with long latency periods, such as many forms of cancer, since the people could have moved several times since their exposure to environmental hazards, thus breaking up clusters and obscuring patterns.

Methods of analysis based on reasoning about geospatial lifelines of specific cases may reduce or eliminate this problem of cluster dispersion. If researchers have the data and information manipulation tools, this will allow them to roll cases back to places of residence or travel in the past when they might have been more markedly clustered. Clusters also could be identified directly in three dimensional space-time. A discussion of how Finnish census data is uniquely organized to permit the tracing of residences to places with known radon level measurements was described by Loytonen (1998).

Integrate the literature of spatial choice in geography and econometrics with the literature of preventive care choices.

Many critical choices are made by people that affect their health where the controlling factors are in a spatial context. The decision on when and where to seek health care is known to be affected by the geographical distribution of relevant resources. Hence the importance of geographic accessibility in seeking timely medical care. It is possible, for example, that the stage of a tumor’s development at the time of first diagnosis might relate to the choice of place and type of treatment. See Fortney et al., 1995; 1998.

Propose more systematic studies of access, health treatment choice, and health outcomes.

The traditional spirit of public health has always been a focus on the health of the public. Consequently, it is concerned whenever particular population groups experience a greater burden of disease (Townsend et al. 1988). See Cohen and Lee, 1985; Gober, 1997; McLafferty, 1988; Piette and Moos, 1996; Siegel et al., 1997.

Develop methods for targeting health resources.

There is common agreement that one important use of GIS is to target health resources to places where they are most in need (Bureau of Primary Health Care, 1998; Geronimus et al., 1996; Kerner et al., 1988; Larimore and Davis, 1995). The health science community is generally unaware of the extent of the development of general methods in geography and regional science for this purpose (Ayeni et al., 1987; McLafferty and Broe, 1990; Malczewski and Ogryczak, 1988; Walsh et al., 1997). The Federal Department of Health and Human Services, for example, assigns resources for reducing rates of infant mortality in areas of U. S. cities where infant mortality rates are double the U.S. rate. The program is called "Healthy Start" and now operates in selected areas within at least fifty U.S. cities. Methodologies for selecting these areas involve ad hoc methods of regionalization. Evaluating the geographical effects of policies designed to improve access to health services is another area of application of GIS in health. The problem is illustrated well in the General Accounting Offices’ review of the implementation of the Rural Health Care Centers Program of the Department of Health and Human Services, (GAO, 1997).

Improved methods of communicating with the public the results of research on health using GIS.

Maps of the locations of disease in local areas are difficult to interpret. Indeed they are open to misinterpretation by a public who do not know that disease rates based on information for small areas naturally exhibit marked geographic variation even when the true disease rates do not vary. They are also open to interpretation by groups whose purpose may include the deliberate manipulation of public opinion toward some end. The public often is suspicious that information is being withheld from their view. The challenge for the scientist is to assist in the interpretation of results from GIS-based analyses. Pickle and others (1995, 1997) have reported on the extensive perception research undertaken in conjunction with the National Mortality Atlas to reduce unintended interpretations. The public health community has had a focus on small area analysis for some time but the issue of misinterpretation is not resolved by efficient algorithms, for example those of Carvalho et al. (1996) or Elliot et al (1996). Headlines in the popular press can swiftly present conclusions that are not supported by scientific analysis. When the subject is health risks and the environment, it is important that information be presented to help the public sort through often conflicting material. Monmonier (1997) has written about risk communication in the health area. There have been a number of recent proposals to NIH for projects to support the development of web-based mapping that the public can access.

Maintaining the confidentiality of health records.

Personal health records are among the most sensitive and confidential pieces of information on individuals and many laws exist to ensure the privacy of individuals and the protection of information from others who have no right to see it. Releasing health information data for small areas may often fail to protect the privacy of individuals. Often, the desire to see health data in its geographic context is in conflict with protecting the confidentiality of individuals. Methods need to be developed for ensuring confidentiality while preserving the capability of geographical analysis. A report of the National Cancer Institute recognized this problem in the context of GIS, (NCI, 1999, p. 30):

"However, it is critical that mechanisms for protecting confidentiality be developed to maximize the utility of this technology. Spatial aggregation, which has been the standard method for preserving confidentiality of geographic data, will not suffice for health-related GIS activities. The SIG (Special Interest Group) recommends that research be conducted to develop alternative methods to guard the privacy of health records incorporated in GIS-based geographic analysis."

Only a small literature has addressed this problem. (Armstrong et al. 1999).

POLICY IMPLICATIONS

Reducing the burden of premature mortality and morbidity and providing health care for the elderly and uninsured are essential elements of a national health policy directed at reducing health care expenditures.

Peer review organizations, such as American Quality of Health Association (AQHA) – will increasingly use GIS for evaluation and assessment of the geographical equity of the use of procedures and interventions, especially in the assessment of the effectiveness of Medicare/Medicaid funds. (see Durch et al. 1997, for example)

Of more recent concern is the assessment of medical readiness in the face of mass casualty events, such as from weapons of mass destruction or natural disasters. Massive casualties anywhere in the United States would overwhelm the current medical care establishment and few places have been able to adopt lessons learned from the Oklahoma bombing, for example. A collaboratory was initiated in 1999 at West Virginia University with private partners such as Oracle, EDS, and others, to develop a national response capability through geographic information and telemedicine.

IMPORTANCE TO NATIONAL RESEARCH NEEDS

Health care represents approximately 13 percent of the GDP. Any contribution through the effective application of geographic information to reduce expenditures on health through improved surveillance, health care delivery, access to care or evaluation of outcomes of intervention projects will be of national significance.

ACKNOWLEDGEMENTS

We wish to thank David Mark and Max Egenhofer for contributing the section "Temporal Aspects of GIS and Health" and Jamil Kazmi and Lynn Usery for the section "Ecological Studies of the Relationship between Environmental Factors and Disease Transmission."

REFERENCES

Anselin, L. 1995. Local indicators of spatial association—LISA. Geographical Analysis 27:93-115.

Anselin, L. and Bao, S. 1997. Exploratory spatial data analysis linking SpaceState and Arc View. In Recent Developments in Spatial Analysis: Spatial Statistics., Behavioural Modelling and Neuro-computing (eds M. Fischer and A. Getis) Berlin: Springer, pp. 35-59.

Armstrong, M.P., G. Rushton and D.L. Zimmerman. 1999 Geographically Masking Health Data to Preserve Confidentiality. Statistics in Medicine 18:497-525.

Ayeni, B., G. Rushton, and M. McNulty. 1987. Improving the geographical accessibility of health care in rural areas: a Nigerian case study. Social Science and Medicine 25:1083-1094.

Becker, K.M., G.E. Glass, W. Brathwaite, and J.M. Zenilman. 1998. Geographic epidemiology of gonorrhea in Baltimore, Maryland, using a geographic information system. American Journal of Epidemiology. 147:709-716.

Bernhardsen, T. 1999. Geographic Information Systems: An Introduction. New York, John Wiley & Sons, Inc.

Braden, J. and K. Beauregard. 1994. Health status and access to care of rural and urban populations, (AHCPR Pub. No. 94-0031). National Medical Expenditure Survey Research Findings 18, Agency for health Policy and Research, Rockville, MD: Public Health Service.

Briggs, D.J. and P. Elliott. 1995. The use of geographical information systems in studies on environment and health. World Health Statistics Quarterly 48:85-94.

Brown, P.J. B., A. Hirschfield and P.W.J. Batey. 1991. Applications of geodemographic methods in the analysis of health condition incidence data. Papers in Regional Science: Journal of the Regional Science Association International 70:329-344.

Bureau of Primary Health Care. 1998. Procedures for Medically Underserved Areas and Medically Underserved Populations Designation. http://www.vdh.state.va.us/primcare/0143.htm

Carvalho M. S. O.G. Cruz, and F.F. Nobre. 1996. Spatial Partitioning Using Multivariate Cluster Analysis and a Contiguity Algorithm Statistics in Medicine 15:1885 - 1894.

Clarke, KJ. C., S.L. McLafferty and B.J. Templaski. 1996. On epidemiology and geographic information systems: a review and discussion of future directions. Emerging Infectious Diseases 2, no. 2:85-92.

Cohen, M., and H. Lee. 1985. The determinants of spatial distribution of hospital utilization in a region. Medical Care 23:27-38.

Cook, R.L., R.A. Royce, J.C. Thomas, and B.H. Hanusa. 1999. What’s driving an epidemic? The spread of syphilis along an interstate highway in rural North Carolina. American Journal of Public Health 89:369-373.

Cromley, E.K. and G.W. Shannon. 1986. Locating ambulatory medical care facilities for the elderly. Health Services Research 21:499-513.

Croner, C.M., J. Sperling, and F. Broome. 1996. "Geographic Information Systems (GIS): New Perspectives in Understanding Human Health and Environmental Relationships." Statistics in Medicine 15: 1961-1977.

Dean, A. G.1999. Epi Info and Epi Map: current status and plans for Epi Info 2000. Journal of Public Health Management and Practice 5 (4):54-57.

Dumas, M., M.-C. Fauvet, and P.-C. Scholl. 1999. TEMPOS: A Temporal Database Model Seamlessly Extending ODMG, Research-Report 1013-I-LSR-7, LSR-IMAG Laboratory, Grenoble, March.

Durch J. S., L. A. Bailey and M. A. Stoto (eds). 1997. Improving Health in the Community: A Role for Performance Monitoring, Institute of Medicine (US), Committee on Using Health Performance Monitoring to Improve Health, National Academy of Sciences, Washington D.C.

Elliott, P., J. Cuzick, D. English, and R. Stern, eds. 1992. Geographical and Environmental Epidemiology: Methods for Small Area Studies. Oxford: Oxford University Press.

Entwisle, B., R.R. Rindfuss, et al. 1997. Geographic information systems, spatial network analysis, and contraceptive choice. Demography 34:171-87.

Erwig, M., R. H. Güting, M. Schneider, and M. Vazirgiannis. 1998. Spatio-Temporal Data Types: An Approach to Modeling and Querying Moving Objects in Databases. FernUniversität Hagen, Informatik-Report 224, December 1997. Abstract, Paper.ps.gz (compressed postscript, 52 KB). Revised Version, August.

Fauvet, M.C., S. Chardonnel, M. Dumas, P.C. Scholl, and P. Dumolard. 1998. Analyse de données géographiques: application des bases de données temporelles, Révue Internationale de Géomatique 8 (No. 1-2): 149-65.

Fauvet, M.C., S. Chardonnel, M. Dumas, P.C. Scholl, and P. Dumolard. 1999. Applying temporal databases to geographical data analysis, In Proceedings of the DEXA Workshop on Spatio-Temporal Data Models and Languages, Florence (Italy), September, to appear.

Forer, P. 1998. Geometric Approaches to the Nexus of Time, Space, and Microprocess: Implementing a Practical Model for Mundane Socio-spatial Systems. In M. Egenhofer and R. Golledge, editiors, Spatial and Temporal Reasoning in Geographic Information Systems. New York: Oxford University Press, pp. 171-190.

Fortney, J., B. Booth, F. Blow, J. Bunn and C. Cook. 1995. The effects of travel barriers and age on the utilization of alcoholism treatment aftercare. American Journal of Drug and Alcohol Abuse 21:391-406.

Fortney, J., K. Rost, M. Zhang and G.R. Smith. 1998. A joint choice model of the decision to seek depression treatment and choice of provider sector. Medical Care, in press.

Gatrell A. C., and T. C. Bailey. 1966 Interactive Spatial Data Analysis in Medical Geography Social Science and Medicine 42:843-55.

Gatrell, C.C. and M. Loytonen. 1998. GIS and Health. London: Taylor & Francis Ltd.

General Accounting Office. 1997. Rural Health Clinics: Rising Program Expenditures not Focused on Improving Care in Isolated Areas. Washington, D.C. GAO/HeHS-97-24.

Geronimus, A., J. Bounnd, and L. Neidert. 1996. On the validity of using Census geocode characteristics to proxy individual socioeconomic characteristics. Journal of the American Statistical Association 91:529-537.

Glass, G.E., B.S. Schwartz et al. 1995. Environmental risk factors for Lyme disease identified with geographic information systems. American Journal of Public Health 85:944-8.

Gober, P. 1997. The role of access in explaining state abortion rates. Social Science and Medicine, 44:1003-10016.

Gould, P. 1993. The Slow Plague: A Geography of the AIDS Pandemic. Oxford: Blackwell.

Hägerstrand T. 1967. Innovation Diffusion as a Spatial Process. Chicago University of Chicago Press.

Hägerstrand, T. 1970. What About People in Regional Science? Papers, Regional Science Association 24: 1-21.

Haining, R. 1990. Spatial Data Analysis in the Social and Environmental Sciences, Cambridge University Press, New York.

Haining, R., Wise, S.M. and Blake, M. 1994. Constructing regions for small area analysis: material deprivation and colorectal cancer. Journal of Public Health Medicine 16: 429-438.

Haining, R., Wise, S. and Ma, J. 1998. Exploratory spatial data analysis in a geographic information system environment. The Statistician 47, Part 3: 457-469.

Haslett, J., G. Wills, and A.R. Unwin 1990. SPIDER—an interactive statistical tool for the analysis of spatially distributed data. International Journal Geographical Information Systems 4:285-296.

Hirschfield, A., P. Brown, and P. Bundred. 1993. Doctors, patients and GIS. Mapping Awareness 7(9):

Hirschfield, A., P. Brown, and P. Bundred. 1995. The spatial analysis of community health services on Wirral using geographic information systems. Journal of the Operational Research Society 46:147-159.

Jacquez, G.M. 1995. The map comparison problem: tests for the overlap of geographic boundaries. Statistics in Medicine 14:2343-61.

Kerner, J.F. H. Andrews, A. Zauber and E. Struening. 1988. Geographically based cancer controls: methods for targeting and evaluating the impact of screening interventions on defined populations. J. Clin. Epidemiology 41:543-553.

Kohli S., K Sahlén, O Löfman, Å Sivertun, M. Foldevi, E. Trell and O Wigertz. 1997. Individuals Living in Areas with High Background Radon: A GIS Method to Identify Populations at Risk. Computer Methods and Programs in Biomedicine 53:105-112.

Kulldorff, M. et al., 1998. SaTScan V. 2.1, Software for the spatial and space-time scan statistics. Bethesda, MD: National Cancer Institute.

Kulldorff, M. 1999. Geographic information systems (GIS) and community health: some statistical issues. Journal of Public Health Management and Practice 5 (2):100-106.

Langran, G. 1992. Time in Geographic Information Systems. London: Taylor and Francis.

Larimore, W., and A. Davis. 1995. Relation of infant mortality to the availability of maternity care in rural Florida. Journal American Board of Family Practice 8:392-399.

Lee, R. 1991. Current approaches to shortage area designation. The Journal of Rural Health 7:437-450.

Love, D. and P. Lindquist. 1995. The geographical accessibility of hospitals to the aged: a geographic information systems analysis within Illinois. Health Services Research 29:629-51.

Loytonen, M. 1998. GIS, Time Geography and Health. In A. C. Gatrell and M. Loytonen, GIS and Health. London: Taylor & Francis Ltd., pp. 97-110.

MacDorman, M. F.. and G. A. Gay. 1999. State initiatives in geocoding vital statistics data. Journal of Public Health Management and Practice 5 (2): 91-93.

Mackinnon, D.P., R. Scribner and K. A. Taft. 1995. Alcohol availability and alcohol problems database. Statistics in Medicine 14:591-603.

Mark, D. M. and Egenhofer, M. J. 1998. Geospatial Lifelines, In O. Guenther, T. Sellis, and B. Theodoulidis, editors, Integrating Spatial and Temporal Databases. Dagstuhl Seminar Report No. 228.

Mark, D. M., Egenhofer, M., Bian, L., Rogerson, P. A., and Vena, J., 1999. Spatio-Temporal GIS Analysis for Environmental Health. National Institute of Environmental Health Sciences, National Institutes of Health, Grant number 1 R 01 ES09816-01.

Martin, D. 1996. Geographic Information Systems: Socioeconomic Applications. London, Routledge.

McLafferty, S. 1988. Predicting the effect of hospital closures on hospital utilization patterns. Social Science and Medicine 27:255-262.

McLafferty, S., and D. Broe. 1990. Patient outcomes and regional planning of coronary care services: a location-allocation approach. Social Science and Medicine 30:297-304.

Malczewski, J., and W. Ogryczak. 1988. A multi-objective approach to the reorganization of health services areas: a case study. Environment and Planning, A 20:1461-1470.

Miller, H. J., 1991, Modeling Accessibility Using Space-Time Prism Concepts within a GIS, International Journal of Geographic Information Systems 5(3):287-301.

Mohan, J. 1983. Location-allocation models, social science and health service planning: an example from North East England. Social Science and Medicine 17:493-499.

Monmonier, M. 1997. Cartographies of Danger: Mapping Hazards in America. Chicago, The University of Chicago Press, see Chapter 13, John Snow’s Legacy.

National Cancer Institute. 1999a. The Long Island Breast Cancer Study Project. http://www-dccps.ims.nci.nig.gov/LIBCSP/index.html

National Cancer Institute. 1999b. Cancer Surveillance Research Implementation Plan. Surveillance Implementation Group, NCI, March, p. 30.

Ord, J.K. and A. Getis. 1995. Local spatial autocorrelation statistics: distributional issues and an application. Geographical Analysis 27:286-306.

Perkins, B.B. 1999. Re-forming medical delivery systems: economic organization and dynamics of regional planning and managed competition. Social Science and Medicine 48:241-251.

Peuquet, D.J., 1994, It's about time: a conceptual framework for the representation of temporal dynamics in geographic information systems. Annals of the Association of American Geographers 84 (3): 441-461.

Pickle L. W., M. Mungiole, G.J. Jones, and A. White, 1997 Atlas of United States Mortality, National Center for Health Statistics, Centers for Disease Control and Prevention,

Pickle L. W., and A. White. 1995. The choice of age-adjustment methods on maps of death rates, Statistics in Medicine 14:615 -627.

Piette, J.D. and R.H. Moos. 1996. The influence of distance on ambulatory care use, death, and readmission following a myocardial infarction. Health Services Research 31:573-591.

Ricketts T.C., Savitz, L. A., Gesler, W.M., Osbourne D. N. 1994, Geographic Methods for Health Services Research : A Focus on the Rural-Urban Continuum, University Press of America, Lanham MD.

Richards, T.B., C.M. Croner, G. Rushton, C.K.Brown and L. Fowler. 1999. Geographic information systems and public health: mapping the future. Public Health Reports 114:359-373.

Rosenberg, M.S., R.R. Sokal, N.L. Oden and D. DiGiovanni. 1999. Spatial autocorrelation of cancer in Western Europe. European Journal of Epidemiology 15:15-22.

Ross, N., M. Rosenber, and D. Pross. 1994. Siting a women’s health facility: a location-allocation study of breast cancer screening services in Eastern Ontario. The Canadian Geographer 38:150-161.

Rushton, G. (1998). "Improving the Geographic Basis of Health Surveillance Using GIS." in M. Loytonen and A. Gatrell, eds., GIS and Health, Taylor & Francis, London, pp. 63-79.

Rushton, G. 1999 "Methods to evaluate geographic access to health services." Journal of Public Health Management and Practice 5 (2): 93-100.

Rushton, G., M.P. Armstrong, C. Lynch and J. Rohrer. 1997. Improving public health through geographical information systems: an instructional guide to major concepts and their implementation [CDROM] Iowa City: University of Iowa Department of Geography. Available at : URL: http://www.uiowa.edu/~geog/health/

Siegel, C., A. Davidson, K. Kafadar, J.M. Norris, J.Todd and J. Steiner. 1997. Geographic analysis of pertussis infection in an urban area: a tool for health services planning. American Journal of Public Health 87:2022-2026.

Spear, R., P. Gong, E. Seto, Y. Zhou, B. Xu, D. Maszle, S. Liang, G. Dagis, and X. Gu. 1998. Remote sensing for Schistosomiasis control in mountainous areas in Sichuan, China. Geographic Information Sciences 4:14-22.

Smoyer, K.E. 1998. Putting risk in its place: methodological considerations for investigating extreme event health risk. Social Science and Medicine 47:1809-1824.

Snow, J., 1936. Snow on Cholera. Oxford University Press, London.

Vine, M.F., D. Degnan, C. Hanchette. 1997. Geographic information systems: their use in environmental epidemiologic research. Environmental Health Perspectives 105, no. 6:598-605.

Waller, L.A. 1996. Geographic information systems and environmental health. Health and Environment Digest 9:85-88.

Walsh, S., P. Page, and W. Gesler. 1997. Normative models and healthcare planning: network-based simulations within a geographic information system environment. Health Services Research 32:243-260.

Wise, S. M., R.P. Haining, and J. Ma. 1997. Regionalization tools for the exploratory spatial analysis of health data. In Recent Developments in Spatial Analysis: Spatial Statistics, Behavioural Modelling and Neuro-computing (eds M. Fischer and A. Getis), Berlin: Springer, pp. 83-100.

World Health Organization. 1996. Climate Change and Human Health. Geneva.

Yasnoff, W.A. and E. J. Sondik. 1999. Geographic information systems (GIS) in Public health practice in the new millennium. Journal of Public Health Management and Practice 5 (4): ix-xii.

APPENDIX A

"POSITION: Statistician GS-1530 -12/13 LOCATION: National Center for Chronic Disease Prevention and Health Promotion, Division of Adult and Community Health, Cardiovascular Health Branch, Atlanta, Georgia

Incumbent will design and/or apply standard and customized Geographic Information System (GIS) techniques to cardiovascular disease surveillance and analysis. Directs and designs the infrastructure of a GIS database to link multiple large and complex geocoded national datasets. Applies spatial statistical techniques to the analysis of the geographic variation of cardiovascular disease, Develops and maintains clear and comprehensive documentation of the GIS database. Produces high-resolution small area maps of the burden of cardiovascular disease, and prepares data and maps for publications and subsequent uploading to a website. Provides statistical consultation and technical advice to members of the Branch and others on the use of GIS for cardiovascular disease surveillance and research."

APENDIX B

GIS Functions Identified by The National Cancer Institute as either Necessary or Desirable Functions for its GIS for the Long Island Breast Cancer Project

REFERENCES

Copyright © 2000, Urban and Regional Information Systems Association
Comments/Questions:  info@urisa.org