Geography GIS and Health
Yujie Hu, Steven Reader
  • LAST MODIFIED: 26 November 2019
  • DOI: 10.1093/obo/9780199874002-0211


The capability of GIS to be able to store, retrieve, display, and analyze large quantities of spatially referenced data has facilitated the rapid growth of geographic-based research within various health fields, including epidemiology and health care provisioning. This is reflected in the numerous texts focused entirely on GIS and health, as well as texts focused on geography and health that include substantial material on GIS. There are also journals where GIS-based health research frequently appears. This bibliography lists a number of these texts and journals. GIS-based health research is wide ranging, but several major themes can be identified. In this bibliography the first major theme presented is GIS-based visualization of health information, a topic which involves geocoding, disease-mapping methodologies, and alternative cartographic schemes of representation. The second major theme is GIS-derived measures for health research, where the focus is on how GIS has transformed how accessibility to health care is measured and enabled complex forms of environmental exposures to be derived for use in epidemiologic studies. The final theme is GIS-enabled analysis for health research where a deliberate choice was made to focus on those analytical areas which are only feasible through the specific spatial analytical capabilities of GIS.

General Overviews

A large body of literature exists in the field of GIS and health. Overviews of how GIS can be beneficial to health studies can be found in several popular texts and review articles. For example, Gatrell and Loytonen 1998 is perhaps one of the earliest texts in discussing GIS and its applications in health research. Melnick 2002, Cromley and McLafferty 2011, and McLafferty 2003 provide comprehensive reviews of GIS applications in public health. Other texts focus on summarizing GIS applications in the more general social sciences, including health studies, such as Ballas, et al. 2017 and Wang 2015. Some texts in the field of medical geography, such as Brown, et al. 2009; Emch, et al. 2017; and Gatrell and Elliott 2014 also cover some health GIS topics.

  • Ballas, Dimitris, Graham Clarke, Rachel S. Franklin, and Andy Newing. GIS and the Social Sciences: Theory and Applications. New York: Routledge, 2017.

    DOI: 10.4324/9781315759326Save Citation »Export Citation »E-mail Citation »

    This is a recent GIS book, where Part 1 introduces fundamental concepts in GIS including data query, spatial analysis, visualization, and network analysis. Chapter 9 in Part 2 discusses GIS applications in health care planning.

    Find this resource:

    • Brown, Tim, Sara McLafferty, and Graham Moon. A Companion to Health and Medical Geography. Malden, MA: John Wiley & Sons, 2009.

      DOI: 10.1002/9781444314762Save Citation »Export Citation »E-mail Citation »

      Part 2 discusses the applications of GIS in mapping disease spatial patterns and modeling disease-spreading patterns. Part 4 covers the utilization of GIS and spatial models in understanding environmental health risk and neighborhood health.

      Find this resource:

      • Cromley, Ellen K., and Sara L. McLafferty. GIS and Public Health. London: Guilford Press, 2011.

        Save Citation »Export Citation »E-mail Citation »

        A well-known text in GIS and public health. It covers topics like GIS and spatial data (chapters 1, 2, 3); mapping health patterns (chapter 4); identifying spatial clustering patterns of health events (chapter 5); analyzing environmental hazards (chapter 6); infectious disease patterns (chapters 7, 8); and health accessibility (chapter 9).

        Find this resource:

        • Emch, Michael, Elisabeth D. Root, and Margaret Carrel. Health and Medical Geography. London: Guilford Press, 2017.

          Save Citation »Export Citation »E-mail Citation »

          The fourth edition of the popular textbook formerly titled Medical Geography. Chapter 5 covers GIS fundamentals (spatial data, geocoding, and scale effect) and the use of GIS in mapping and analyzing patterns (spatial clustering and spatial statistics) of health events. Chapter 13 discusses health care access, regionalization, and location-allocation.

          Find this resource:

          • Gatrell, Anthony C., and Susan J. Elliott. Geographies of Health: An Introduction. Malden, MA: John Wiley & Sons, 2014.

            Save Citation »Export Citation »E-mail Citation »

            Chapter 3 discusses GIS methods (visualization and exploratory spatial data analysis) and some health applications.

            Find this resource:

            • Gatrell, Anthony, and Markku Loytonen, eds. GIS and Health. GISDATA 6. London: CRC Press, 1998.

              Save Citation »Export Citation »E-mail Citation »

              One of the earliest books on health GIS research. It covers the various applications of GIS in health studies such as spatial analysis, spatial statistical models, time geography, and Bayesian models.

              Find this resource:

              • McLafferty, Sara L. “GIS and Health Care.” Annual Review of Public Health 24.1 (2003): 25–42.

                DOI: 10.1146/annurev.publhealth.24.012902.141012Save Citation »Export Citation »E-mail Citation »

                In this review article, McLafferty provides an in-depth review on the health GIS literature from the following aspects: (1) analyzing population demand, (2) measuring health access and identifying inequalities in access, (3) evaluating geographic variations in health resource utilization, and (4) locating health services.

                Find this resource:

                • Melnick, Alan L. Introduction to Geographic Information Systems in Public Health. New York: Aspen Publishers, 2002.

                  Save Citation »Export Citation »E-mail Citation »

                  This textbook covers GIS data (chapter 2), mapping techniques (chapter 3), analyzing disease risk (chapter 4), and disease transmission (chapter 5).

                  Find this resource:

                  • Wang, Fahui. Quantitative Methods and Socio-Economic Applications in GIS. London: CRC Press, 2015.

                    Save Citation »Export Citation »E-mail Citation »

                    Chapter 4 discusses the delineation of trade areas of public hospitals in Louisiana. Chapter 5 describes the popular Two-Step Floating Catchment Area (2SFCA) and its application in measuring health care accessibility, a big theme of health GIS studies. Chapter 9 focuses on the small population problem and the regionalization methods. Part of Chapter 11 covers health care resource optimization.

                    Find this resource:

                    Reference Resources

                    There are many journals that publish health GIS papers—some focus specifically on this particular subfield, while others focus on more general disciplines such as geography, urban planning, and health science. Here, only a few key journals directly related to this subfield are listed: Applied Geography, Environment and Planning B: Urban Analytics and City Science, Health and Place, International Journal of Health Geographics, Journal of Health Economics, Social Science & Medicine, and Spatial and Spatiotemporal Epidemiology.

                    Theme 1. GIS-Based Visualization of Health Information

                    Mapping health events is among the most common tasks in GIS and health. Health data often come in two general formats—point locations (e.g., residences of people with health problems) and area rates (e.g., the number or rate of a health event in an area). As reported in McLeod 2000, the earliest example of mapping point locations of health data, perhaps, can be traced back to Dr. John Snow’s dot map of cholera deaths, where each incident is mapped as a dot in the map. Health event locations recorded by a georeferenced coordinate pair, such as longitude and latitude, can be simply mapped in GIS. Locations represented by street addresses may need to be geocoded before mapping in GIS, and Zimmerman, et al. 2007 provides a comprehensive review on this task. Most of the times, health data are available only for areas such as zip codes and census tracts. Such area count or rate data can be mapped by dot density or choropleth maps in GIS. According to Winn 2005, a well-known example is the Long Island Breast Cancer Study Project, where GIS is used to make choropleth maps for understanding the high breast cancer rate in the region. More details about the mapmaking process and guidelines such as map elements, scale, projection, and symbolization are discussed in Lawson and Williams 2001 and Centers for Disease Control and Prevention (CDC). The commonly used choropleth maps have limitations, however. They can mislead the reader when there is a significant variation in the size of background areas. For a less confusing representation, one can use the so-called cartogram, in which the areas are distorted, and their sizes are proportional to their values being mapped. More detail can be found in Tobler 2004. Going a step further, some GIS methods—more precisely spatial smoothing techniques—can summarize disease point patterns and generate a smoothed surface that shows the disease density patterns. From this smooth density surface map, one can simply identify disease hotspots. More detail is discussed in Rushton 2003.

                    Theme 2. GIS-Derived Measures for Health Research

                    In addition to the prevalent role in visualizing disease spatial patterns, GIS also finds its applications in deriving measures for health research. This includes several topics and the focus here is on how GIS has facilitated deriving measures for (1) health care accessibility; (2) walkability, built, and food environment; and (3) air pollution and environmental health exposure.

                    Health Care Accessibility Measures

                    As noted in Guagliardo 2004, health care accessibility represents the relative ease by which residents in an area can obtain health care. This topic is often tied to policymaking because there are always some people who face significant barriers in accessing health care. As discussed in Ikram, et al. 2015, the barriers can be geographic (e.g., travel distance or cost) or socioeconomic (e.g., income or language); hence health care accessibility is generally related to two aspects: spatial accessibility (where you are) and aspatial accessibility (who you are). Due to its spatial focus, GIS has been primarily used to measure spatial accessibility. There are several ways to measure spatial accessibility in GIS. The simplest approach is to calculate a provider-to-population ratio in a geographic zone such as state, county, or census tract. For example, Schonfeld, et al. 1972 uses this measure to estimate the number of physicians that an accessible primary medical care should have. Guagliardo, et al. 2004 proposes a kernel-based provider-to-population ratio to account for the edge effect (cross-border care-seeking patterns) that the original method failed to consider. Travel costs to the nearest provider is another intuitive spatial accessibility measure. This method computes the travel costs, such as travel distance or time, from a patient’s residence or a residential zone’s centroid to the closest care provider, where the lower the value, the better the accessibility. Using this measure, Onega, et al. 2008 examines the spatial accessibility to cancer care in the United States. The third type of spatial accessibility model is the gravity-based method, which models the potential spatial interaction between any population and provider. The potential is assumed to be positively related to provider attractiveness such as number of physicians and hospital beds and negatively affected by the geographic distance between them. A great example of this line of work is Joseph and Bantock 1982. A more recent advance in spatial accessibility modeling is the two-step floating catchment area (2SFCA) method designed by Luo and Wang 2003. It is also a weighted provider-to-population ratio, yet it is simpler to interpret and easier to implement in a GIS environment. There are also some efforts to further improve this method, such as Luo and Qi 2009. Wang 2012 provides a methodological review on measuring spatial accessibility of health care. Most recently, Wang 2018 proposes an inverted 2SFCA to measure potential service crowdedness in facilities.

                    Walkability, Built, and Food Environment Measures

                    There is little doubt that the ongoing development of GIS over recent decades, including enhanced tools and programming interfaces, has enabled more sophisticated and nuanced measures of the many contextual and neighborhood factors that have increasingly become part of the analytical framework in public health studies and epidemiology as hierarchical or multilevel statistical modeling approaches have flourished. GIS has become, in large part, an integral technology that many public health researchers routinely use for data management and visualization, but in this section the focus is on studies where the inherent functional capabilities of GIS allow for the measurement of contextual environmental measures that would otherwise be very difficult to derive. There is a myriad of approaches to measuring the built environment through GIS and so review articles are particularly useful. In this regard, general reviews such as Brownson, et al. 2009; Frank, et al. 2012; and Jia, et al. 2017 provide excellent overviews of both the range and heterogeneity of such measures. Other review articles provide deeper insight with a narrower focus, such as Nordbø, et al. 2018 on environmental measures and exposures related to mental health and mental activity participation in children and adolescents; Sweeney, et al. 2016 on food environment measures; and Stewart, et al. 2016 on walkability measures. Applied studies where built environment measures are related to public health outcome measures through statistical analyses are numerous. Bödeker, et al. 2018 uses a multilevel Poisson regression model to relate active travel to a derived walkability index, while Todd, et al. 2016 uses a latent profile analysis of built environment measures to relate these profiles to physical activity through generalized linear mixed models. The capture of study participants’ exposure to measures of the built environment through use of GPS is another increasingly common research activity, and Burgoine, et al. 2015 demonstrates this for walk-to-school routes. Finally, Kwan 2018 in a highly prescient contribution draws attention to the many issues and limits surrounding the measurement and use of neighborhood effects in public health research.

                    • Bödeker, Malte, Emily Finne, Jacqueline Kerr, and Jens Bucksch. “Active Travel Despite Motorcar Access: A City-Wide, Gis-Based Multilevel Study on Neighborhood Walkability and Active Travel in Germany.” Journal of Transport & Health 9 (2018): 8–18.

                      DOI: 10.1016/j.jth.2018.03.009Save Citation »Export Citation »E-mail Citation »

                      This paper uses GIS to estimate a neighborhood walkability index based on connectivity, land use mix, retail floor area, and household density with minor adaptations to this US-based index for the European context. The index is used in a multilevel Poisson regression analysis focused on active travel. Available online by subscription or purchase.

                      Find this resource:

                      • Brownson, Ross C., Christine M. Hoehner, Kristen Day, Ann Forsyth, and James F. Sallis. “Measuring the Built Environment for Physical Activity: State of the Science.” American Journal of Preventive Medicine 36.4 (2009): S99–S123.

                        DOI: 10.1016/j.amepre.2009.01.005Save Citation »Export Citation »E-mail Citation »

                        GIS-derived measures are one of three categories of built environment data collection methods reviewed. A comprehensive review of the variability in definition and operationalization of such measures is discussed.

                        Find this resource:

                        • Burgoine, Thomas, Andy P. Jones, Rebecca J. Namenek Brouwer, and Sara E. Benjamin Neelon. “Associations between BMI and Home, School and Route Environmental Exposures Estimated Using GPS and GIS: Do We See Evidence of Selective Daily Mobility Bias in Children?” International Journal of Health Geographics 14.1 (2015): 8.

                          DOI: 10.1186/1476-072X-14-8Save Citation »Export Citation »E-mail Citation »

                          In this paper, GIS is used to map walk-to-school routes from GPS data, and then GIS buffers or inverse-distance weighting techniques are used to quantify food environment measures and physical activity/walkability measures.

                          Find this resource:

                          • Frank, Lawrence D., Brian E. Saelens, James Chapman, et al. “Objective Assessment of Obesogenic Environments in Youth: Geographic Information System Methods and Spatial Findings from the Neighborhood Impact on Kids Study.” American Journal of Preventive Medicine 42.5 (2012): e47–e55.

                            DOI: 10.1016/j.amepre.2012.02.006Save Citation »Export Citation »E-mail Citation »

                            The authors describe a set of GIS-based built environment indictors related to physical activity and nutrition, with a view to their use in obesity studies. The indicators include established neighborhood walkability indices; the presence, accessibility, and quality of parks; and the presence and density of fast-food and supermarket food outlets. Available online by subscription or purchase.

                            Find this resource:

                            • Jia, Peng, X. Cheng, H. Xue, and Y. Wang. “Applications of Geographic Information Systems (GIS) Data and Methods in Obesity-Related Research.” Obesity Reviews 18.4 (2017): 400–411.

                              DOI: 10.1111/obr.12495Save Citation »Export Citation »E-mail Citation »

                              This paper reviews studies at the interface between obesity research and GIS prior to 2016. The main areas of GIS application are found to be in data visualization and the construction of obesogenic environmental indicators. The paper reviews types of geographic data available, GIS methods, and the range and heterogeneity of indicators. Available online by subscription or purchase.

                              Find this resource:

                              • Kwan, Mei-Po. “The Limits of the Neighborhood Effect: Contextual Uncertainties in Geographic, Environmental Health, and Social Science Research.” Annals of the American Association of Geographers 108.6 (2018): 1482–1490.

                                DOI: 10.1080/24694452.2018.1453777Save Citation »Export Citation »E-mail Citation »

                                This important paper draws attention to the many issues that researchers need to be aware of when using neighborhood contextual effects in their analyses and makes some suggestions for mitigating some of these issues. Issues such as the multidimensionality of measures, the spatial-temporal frame of measurement used, and population mobility are all discussed. Available online by subscription or purchase.

                                Find this resource:

                                • Nordbø, Emma Charlott Andersson, Helena Nordh, Ruth Kjærsti Raanaas, and Geir Aamodt. “GIS-Derived Measures of the Built Environment Determinants of Mental Health and Activity Participation in Childhood and Adolescence: A Systematic Review.” Landscape and Urban Planning 177 (2018): 19–37.

                                  DOI: 10.1016/j.landurbplan.2018.04.009Save Citation »Export Citation »E-mail Citation »

                                  A comprehensive and highly detailed review of operational definitions of GIS-derived built environment measures and operational definitions of exposure focused on measures for mental health and activity participation among children and adolescents. Available online by subscription or purchase.

                                  Find this resource:

                                  • Stewart, Orion T., Heather A. Carlos, Chanam Lee, et al. “Secondary GIS Built Environment Data for Health Research: Guidance for Data Development.” Journal of Transport & Health 3.4 (2016): 529–539.

                                    DOI: 10.1016/j.jth.2015.12.003Save Citation »Export Citation »E-mail Citation »

                                    This paper provides a chronicle of methods used to obtain and attempt to standardize a range of built environment measures related to neighborhood walking using readily available local data sources for nine US towns. Nine different domains of these measures are discussed in detail, highlighting issues raised by the heterogeneity in data sources across the various towns. Available online by subscription or purchase.

                                    Find this resource:

                                    • Sweeney, Glennon, Michelle Hand, Michelle Kaiser, Jill K. Clark, Christy Rogers, and Colleen Spees. “The State of Food Mapping: Academic Literature Since 2008 and Review of Online GIS-based Food Mapping Resources.” Journal of Planning Literature 31.2 (2016): 123–219.

                                      DOI: 10.1177/0885412215599425Save Citation »Export Citation »E-mail Citation »

                                      This is a review paper of GIS-based methodologies for mapping the food environment from 2008 to 2015 and it provides exhaustive summaries of thirty-four academic papers in this domain, and seventy web-based food mapping projects. Available online by subscription or purchase.

                                      Find this resource:

                                      • Todd, Michael, Marc A. Adams, Jonathan Kurka, et al. “GIS-measured Walkability, Transit, and Recreation Environments in Relation to Older Adults’ Physical Activity: A Latent Profile Analysis.” Preventive Medicine 93 (2016): 57–63.

                                        DOI: 10.1016/j.ypmed.2016.09.019Save Citation »Export Citation »E-mail Citation »

                                        This paper uses GIS-derived built environment measures for walkability, public transit access, and recreation facility access. A latent profile analysis using these measures is used to assign participants to one of four profiles. These profiles are then related to measures of physical activity while controlling for a range of covariates using generalized linear mixed models.

                                        Find this resource:

                                        Air Pollution and Environmental Health Exposure Measures

                                        Measures for air pollution and other environmental health exposures present many challenges due to their often dynamic nature and their variable but often continuous spatial extent, either as areal data or focused along linear pathways. Given these complexities, GIS has increasingly become adopted to provide exposure metrics for these phenomena. Chakraborty, et al. 2011 provides a good general overview of the spatially based methods employed for environmental exposure measurement. GIS has also been used in innovative approaches to environmental exposure measurement such as the use by Habran, et al. 2019 of a web-based GIS to allow different assessments of the environmental burden of pollutants and noise from different aggregation and weighting schema and by Huang and London 2016 to map environmental hazards and vulnerability through a participatory GIS approach. A more general overview of the breadth of application of GIS in an environmental exposure context is provided in Akkus and Ozdenerol 2014 who outline five different categories of GIS use in a review article focused on childhood lead exposure studies. A focus of GIS use in air pollution studies has been on the performance of GIS-based air pollution models. Jerrett, et al. 2005 reviews six different classes of GIS-based models for intraurban air pollution exposure; Coudon, et al. 2019 compares a GIS-based model to a validated dispersion model; and Khan, et al. 2019 describes an update to the exposure modeling system AirGIS and assesses its performance. Pilla and Broderick 2015 demonstrates how GIS can be customized via Python to operationalize various air pollution models.

                                        Theme 3. GIS-Enabled Analysis for Health Research

                                        The final theme is focused on the analytical areas in health research that are only feasible through the specific spatial analytical capabilities of GIS. Three areas are chosen and discussed here: Service Area Analysis for delineating health service delivery regions, Location-Allocation Analysis for optimizing health care supply and demand, and Spatial Epidemiological Analysis for investigating health outcomes and health risks.

                                        Service Area Analysis

                                        Oftentimes, researchers use political boundaries or census units to study health patterns. Such units are not specifically designed for health studies and thus may not be appropriate. Based on the 1992–1993 Medicare inpatient discharge records, the Dartmouth Institute for Health Policy and Clinical Practice (Dartmouth Atlas of Health Care) designs two units—Hospital Service Area (HSA) and Hospital Referral Region (HRR)—to delineate the local health services delivery region and provides both units at Dartmouth Atlas of Health Care for open access. HSA is delineated by merging zip codes where most residents receive local hospital care, while HRR is defined by merging HSAs where most residents are referred for two specialized surgical services—cardiovascular and neurosurgical services. The two units have been widely used by researchers and policymakers to examine geographic variations in health care expenditures and service utilization. A seminal paper is Zhang, et al. 2012 that looks into the spatial variations of health care spending at both HSA and HRR levels. Using a similar delineation process, Klauss, et al. 2005 defines HSA and investigates the geographic variations of the health care system in Switzerland. However, the Dartmouth HSA and HRR are not free of concerns. Jia, et al. 2015 finds that they are outdated. Guagliardo, et al. 2004 argues that they are not representative of other patient groups other than Medicare patients. Kilaru, et al. 2015 reports that the units do not accurately represent the local health care markets across the nation. Hu, et al. 2018 further questions the delineation process of the Dartmouth method and proposes an automated optimization, data-driven approach to redefine both units that have the highest local hospitalization patterns in each unit.

                                        • Dartmouth Atlas of Health Care.

                                          Save Citation »Export Citation »E-mail Citation »

                                          The Dartmouth Institute for Health Policy and Clinical Practice defines the HSA and HRR and makes both units available online at the Dartmouth Atlas of Health Care. It provides the data in a GIS data format (shapefile) for open access, and details how the two units are defined. Several selected research projects that are based on the two units can also be found there.

                                          Find this resource:

                                          • Guagliardo, Mark F., Kathleen A. Jablonski, Jill G. Joseph, and David C. Goodman. “Do Pediatric Hospitalizations Have a Unique Geography?” BMC Health Services Research 4.1 (2004): 2.

                                            DOI: 10.1186/1472-6963-4-2Save Citation »Export Citation »E-mail Citation »

                                            Dartmouth HSAs are defined based on the discharge records of Medicare inpatients. This paper questions the representativeness of the widely used Dartmouth HSAs and finds that Dartmouth HSAs are not appropriate for all age groups and service types such as pediatric.

                                            Find this resource:

                                            • Hu, Yujie, Fahui Wang, and Imam M. Xierali. “Automated Delineation of Hospital Service Areas and Hospital Referral Regions by Modularity Optimization.” Health Services Research 53.1 (2018): 236–255.

                                              DOI: 10.1111/1475-6773.12616Save Citation »Export Citation »E-mail Citation »

                                              This paper identifies three major issues of the Dartmouth units—outdated, unrepresentative of the overall population, and involving arbitrary choices and manual adjustments. Accordingly, it proposes an automated optimization method that is often used in complex networks to refine both units, which is up-to-date, representative of the general population, and most importantly, has the optimal hospital localization patterns.

                                              Find this resource:

                                              • Jia, Peng, Imam M. Xierali, and Fahui Wang. “Evaluating and Re-demarcating the Hospital Service Areas in Florida.” Applied Geography 60 (2015): 248–253.

                                                DOI: 10.1016/j.apgeog.2014.10.008Save Citation »Export Citation »E-mail Citation »

                                                Based on the 2011 Medicare inpatient discharge records in Florida, this research reconstructs the Dartmouth HSAs and HRRs using the same method as Dartmouth’s and identifies the changes in terms of the unit configuration over the twenty years. It also proposes a Huff-model based approach to delineating HSAs. Available online by subscription or purchase.

                                                Find this resource:

                                                • Kilaru, Austin S., Douglas J. Wiebe, David N. Karp, Jennifer Love, Michael J. Kallan, and Brendan G. Carr. “Do Hospital Service Areas and Hospital Referral Regions Define Discrete Health Care Populations?” Medical Care 53.6 (2015): 510–516.

                                                  DOI: 10.1097/MLR.0000000000000356Save Citation »Export Citation »E-mail Citation »

                                                  This paper uses three indices—localization index, market share index, and net patient flow—to evaluate whether the Dartmouth HSAs and HRRs accurately represent the hospitalization patterns. It finds an inconsistent performance of Dartmouth HSAs and HRRs in representing patients’ hospitalization patterns in three selected states: Washington, Arizona, and Florida.

                                                  Find this resource:

                                                  • Klauss, Gunnar, Lukas Staub, Marcel Widmer, and André Busato. “Hospital Service Areas—A New Tool for Health Care Planning in Switzerland.” BMC Health Services Research 5.1 (2005): 33.

                                                    DOI: 10.1186/1472-6963-5-33Save Citation »Export Citation »E-mail Citation »

                                                    Similar to Dartmouth’s delineation process, this article constructs HSAs for Switzerland. The three hospital localization indices—localization index, market share index, and net patient flow—are designed in this study.

                                                    Find this resource:

                                                    • Zhang, Yuting, Seo Hyon Baik, A. Mark Fendrick, and Katherine Baicker. “Comparing Local and Regional Variation in Health Care Spending.” New England Journal of Medicine 367.18 (2012): 1724–1731.

                                                      DOI: 10.1056/NEJMsa1203980Save Citation »Export Citation »E-mail Citation »

                                                      This seminal paper finds significant geographic variations of healthcare spending and utilization in both HSAs and HRRs across the nation. Therefore, it questions the effectiveness of geographic-based healthcare policies (e.g., payment reform focusing on HRRs).

                                                      Find this resource:

                                                      Location-Allocation Analysis

                                                      Locating health facilities is a critical task in health research and decision making. Location-allocation models in GIS have been developed and applied to this task. Depending on the objectives of locating facilities, these models can be grouped into three basic categories—the p-median problem, location set covering problem (LSCP), and the maximal covering location problem (MCLP). Hakimi 1965 generalizes the p-median problem that aims to find from a set of candidate facility sites a given number of facilities (usually p) so that the total travel distance or time to serve the demands assigned to the facilities is minimized. There exist many applications of this particular optimization problem in health research. For example, Achabal, et al. 1978 applies the p-median problem to find the optimal location for a new hospital among five candidate sites. Quite differently, Toregas, et al. 1971 introduces the LSCP that attempts to find the minimum number and locations of facilities that can provide service to all demand areas within a defined maximal travel distance or time. Horner and Mascarenhas 2007 applies LSCP in a study of dental facilities in Ohio to identify the service coverage inequalities. A limitation of LSCP is that the cost to build the facilities required to cover all the demand areas may exceed the budget available, and thus an alternative solution is to maximize the demand coverage subject to limited budget. Church and ReVelle 1974 develops MCLP to locate a pre-specified number (p) of facilities such that the demand coverage within a maximal travel distance or time is maximized. Eaton, et al. 1985 employs MCLP to determine locations for siting emergency medical service in Austin, Texas. One may refer to such review papers as Brotcorne, et al. 2003; Daskin and Dean 2005; and Rahman and Smith 2000 on the methodological improvements and health-related applications in these location-allocation models. Besides these three traditional location problems, other novel optimization problems have emerged. An intriguing example relates to the issue of spatial accessibility and equity. Specifically, this line of work aims to achieve the optimal health system layout where different population groups have the most equal spatial accessibility to health services. Luo, et al. 2017 reports such as example in rural China.

                                                      • Achabal, Dale D., Harold Moellering, Jeffrey P. Osleeb, and Ralph W. Swain. “Designing and Evaluating a Health Care Delivery System through the Use of Interactive Computer Graphics.” Social Science & Medicine. Part D: Medical Geography 12.1 (1978): 1–6.

                                                        DOI: 10.1016/0160-8002(78)90002-3Save Citation »Export Citation »E-mail Citation »

                                                        Applies the p-median problem to locate the best site for a hospital among five candidate sites so that the total travel distance for all demand areas is minimized. Available online by subscription or purchase.

                                                        Find this resource:

                                                        • Brotcorne, Luce, Gilbert Laporte, and Frederic Semet. “Ambulance Location and Relocation Models.” European Journal of Operational Research 147.3 (2003): 451–463.

                                                          DOI: 10.1016/S0377-2217(02)00364-8Save Citation »Export Citation »E-mail Citation »

                                                          Reviews two location models—the location set covering problem and the maximal covering location problem—and their applications in ambulance-siting research. Available online by subscription or purchase.

                                                          Find this resource:

                                                          • Church, Richard, and Charles ReVelle. “The Maximal Covering Location Problem.” Papers of the Regional Science Association 32.1 (1974): 101–118.

                                                            Save Citation »Export Citation »E-mail Citation »

                                                            The first study that proposes MCLP and provides its formulation. Available online by subscription or purchase.

                                                            Find this resource:

                                                            • Daskin, Mark S., and Latoya K. Dean. “Location of Health Care Facilities.” In Operations Research and Health Care. Edited by Margaret L. Brandeau, William P. Pierskalla, François Sainfort, 43–76. Boston: Springer, 2005.

                                                              Save Citation »Export Citation »E-mail Citation »

                                                              Reviews the applications of p-median, LSCP, and MCLP models in health location planning.

                                                              Find this resource:

                                                              • Eaton, David J., Mark S. Daskin, Dennis Simmons, Bill Bulloch, and Glen Jansma. “Determining Emergency Medical Service Vehicle Deployment in Austin, Texas.” Interfaces 15.1 (1985): 96–108.

                                                                DOI: 10.1287/inte.15.1.96Save Citation »Export Citation »E-mail Citation »

                                                                This paper uses MCLP to select locations for emergency medical service in Austin, Texas. Available online by subscription or purchase.

                                                                Find this resource:

                                                                • Hakimi, S. Louis. “Optimum Distribution of Switching Centers in a Communication Network and Some Related Graph Theoretic Problems.” Operations Research 13.3 (1965): 462–475.

                                                                  DOI: 10.1287/opre.13.3.462Save Citation »Export Citation »E-mail Citation »

                                                                  This research generalizes the p-median problem that aims to locate p facilities on a graph such that the total distances between nodes of the graph and the facilities is minimized. Available online by subscription or purchase.

                                                                  Find this resource:

                                                                  • Horner, Mark W., and Ana Karina Mascarenhas. “Analyzing Location-based Accessibility to Dental Services: An Ohio Case Study.” Journal of Public Health Dentistry 67.2 (2007): 113–118.

                                                                    DOI: 10.1111/j.1752-7325.2007.00027.xSave Citation »Export Citation »E-mail Citation »

                                                                    It applies the location set covering problem to identify regional inequalities in a dental provider location in Ohio. Available online by subscription or purchase.

                                                                    Find this resource:

                                                                    • Luo, Jing, Lingling Tian, Lei Luo, Hong Yi, and Fahui Wang. “Two-step Optimization for Spatial Accessibility Improvement: A Case Study of Health Care Planning in Rural China.” BioMed Research International (2017): 1–12, 2094654.

                                                                      Save Citation »Export Citation »E-mail Citation »

                                                                      Reports a location-allocation problem that aims to achieve an equal health accessibility across population groups

                                                                      Find this resource:

                                                                      • Rahman, Shams-ur, and David K. Smith. “Use of Location-Allocation Models in Health Service Development Planning in Developing Nations.” European Journal of Operational Research 123.3 (2000): 437–452.

                                                                        DOI: 10.1016/S0377-2217(99)00289-1Save Citation »Export Citation »E-mail Citation »

                                                                        With a specific focus on developing countries, this paper reviews the applications of the three location-allocation models in locating health service. Available online by subscription or purchase.

                                                                        Find this resource:

                                                                        • Toregas, Constantine, Ralph Swain, Charles ReVelle, and Lawrence Bergman. “The Location of Emergency Service Facilities.” Operations Research 19.6 (1971): 1363–1373.

                                                                          DOI: 10.1287/opre.19.6.1363Save Citation »Export Citation »E-mail Citation »

                                                                          Introduces the location set covering problem and applies it to locate emergency service facilities. Available online by subscription or purchase.

                                                                          Find this resource:

                                                                          Spatial Epidemiological Analysis

                                                                          GIS is a common tool used by spatial epidemiologists undertaking ecological studies where access to population data in such spatial units as census divisions, health service areas, and postal codes is required. Many of these studies largely draw upon the GIS functionalities of data management and visualization rather than the spatial measurement and analytical capabilities of GIS. In this section the focus is on those types of study which rely on the latter to derive measurements, both geometrical and topologic, for use in spatially based statistical methods. Since many of these methods have often first appeared in noncommercial, nonmainstream software environments, a broad view of GIS is adopted with the knowledge that such developments are gradually infiltrating into more mainstream GIS software. One enduring area of emphasis within spatial epidemiological studies of the type considered here is that of point-based cluster detection methods and Mclafferty 2015 provides a review of recent developments in disease cluster detection methods. Cai, et al. 2012 presents and validates a number of enhancements to point-based cluster detection methods, while Jacquez, et al. 2006 outlines an approach to modeling both residential histories and covariates while undertaking point-based cluster detection. Spatial regression models and geographically weighted regression (GWR) are also approaches that rely on GIS-derived spatial inputs such as spatial weights. Chakraborty 2012 presents a comparison between traditional nonspatial regression and a spatial error model for the investigation of spatial and social inequities related to air pollution while Tabb, et al. 2018 uses spatial regression models and GWR to investigate spatial heterogeneity in US County Health Rankings data. In other GWR applied research, Cabrera-Barona, et al. 2015 investigates spatial heterogeneity in the relationships between deprivation indices and distance to health services and the percentage of mothers that have never had a live birth, and Goovaerts, et al. 2015 investigates prostate cancer incidence in Florida and identifies local areas with disparate rates. One area destined to impact spatial epidemiology in terms of big data, disaggregated data, and alternative approaches to analysis is CyberGIS. Shi and Wang 2015 discusses the emergence of CyberGIS and outlines four characteristics which distinguish it from traditional approaches, while Wang, et al. 2013 provides a wide-ranging discussion of the challenges and opportunities of CyberGIS and discusses a possible roadmap for the integration of its various components. Finally, spatial epidemiology has been impacted by the Bayesian paradigm and Lawson 2013 is a seminal work in the field.

                                                                          back to top