Home Articles GIS in health

GIS in health

L. N. Balaji
Chief, Strategic Planning, Monitoring and Evaluation,
UNICEF, New Delhi
[email protected]

The ability of GIS to “overlay” different entities based on their common geographic occurrence that makes it a very valuable tool in epidemiological research.

The idea that place and location can influence health is a very old and familiar concept in Western Medicine. As far back as the time of Hippocrates, (460-370 BC), the father of Modern Medicine, physicians have observed that certain diseases seem to occur in some places and not in others. Hippocrates was constantly seeking the causes of disease. He studied such things as climate, water, clothing, diet, habits of eating and drinking and the effect they had in producing disease. The Hippocratic concept of health and disease stressed the relation between man and his environment. People have also been aware of the process of disease diffusion across geographic regions for centuries, even during times when the cause of the disease – aetiology – was a mystery. Frequently, attempts to understand why certain diseases seem to only occur in certain places and not others has led to new insights into the nature of the disease itself.

Fig. 1:Map of Cholera deaths in London, 1840s as done by John Snow
A historic milestone in the evolution of medicine is the “great sanitary awakening” which took place in England in the mid-nineteenth century and gradually spread to other countries. In a classic study undertaken in 1854, John Snow (see Fig.1) demonstrated the utility of mapping disease outbreaks to gain insights as to their cause – or to coin a phrase, to “get a handle on a disease and off the water pump”. Snow, an anaesthesiologist, mapped the distribution of cholera cases in Soho, London during an epidemic. He discovered that the highest density of cases occurred in households, which used the public pump on Broad Street as their water source.

Epidemiology and Medical Geography
As the study of epidemiology (epi=among; demos=people; logos=study) matured, simultaneously the discipline of Medical Geography (defined as the branch of Human Geography concerned with the geographic aspects of health (status) and health care (systems). Epidemiology, on the other hand, has been defined variously through times: as the branch of medical science which treats epidemics (Parkin, 1873); the study of disease, as a mass phenomenon (Greenwood, 1934); the study of the distribution and determinants of disease frequency in man (MacMohan, 1960) and now as the study of the distribution and determinants of health-related states and events in populations, and the application of this study to control health problems. Thus it can clearly be seen that both these disciplines have depended heavily on some form of mapping and understanding of the health, diseases and systems in different populations.

Inherent in the definition of epidemiology is measurement of ‘frequency’, ‘distribution’, ‘determinants’ of disease. All of these require GIS application to make the information easy to understand, interpret and take action. Until relatively recently, doctors and public health professionals measured health strictly in terms of indicators of ill-health such as morbidity and mortality. There is however, a long tradition in the allied fields of Medicine viz. Epidemiology, Public Health and Medical Geography to examine the distribution of disease and death at various geographic scales, in an attempt to determine if the presence or absence of particular illness is associated with some factor(s) in the social or physical environment. In the case of infectious diseases, there is the added dimension of examining the diffusion of disease through space over a given period of time. Although mapping of disease can be relatively straightforward, interpreting spatially referenced disease data can sometimes be challenging. Regardless of the difficulties in data acquisition, map representation, scale, statistical analysis, and the interpretation and utility of results, the study of disease distribution may well be the most challenging and fascinating research area.

   

Generally, the objectives of a GIS are the management (acquisition, storage and maintenance), analysis (statistical and spatial modelling), and display (graphics and mapping) of geographic data. GIS is a valuable tool to assist in health research, in health education, and in the planning, monitoring and evaluation of health programmes and health systems.

Buffer Zones Creation
Another process that is very useful in health research and planning is the benefit to the investigation of illness at or near pollution and other hazardous sites, is the ability to create buffer zones around the lines or points which represent those locations. The user can specify the size of the buffer and then intersect or merge this information with disease incidence data to determine how many counts of the illness fall within the buffer. This technique has been used extensively for many conditions – e.g. to understand the association between prevalence of childhood leukemia in northern England and the proximity to nuclear power stations. This technique also provides answers to questions such as “How many persons live within a 10 km radius from a particular PHC or within how many kilometres from a community is a First Referral Centre for attending to complex emergencies such as Caesarian section during childbirth.

The generation of distance/proximity surface, taking into account the distance, and the resistance to get there in terms of costs which may be measured in rupees or time or even transportation means that may not be available on the one hand, and the allocation modelling i.e. assignment of every point of an area to the nearest set of designated features, for example, the First Referral Centres are other geographic analysis tools that can be used in health research and planning.

Spatial Analytic Techniques
As discussed earlier, spatial variation in health related data is well known, and its study is a fundamental aspect of epidemiology. There are many types of spatial analysis : Point patterns – As the name implies, these are dot maps, which display the distribution of health events as data locations. A classic example is the identification of the source of cholera spread shown in Figure 1. An alternative is the use of dynamic graphics such as associating the dot map with a histogram of case occurrence. Selecting the upper tail of the histogram automatically highlights the corresponding cases on the map thus allowing characterisation of the regions by high incidence of the disease. Line patterns Vectors or lines that aid in the analysis of disease diffusion and patient-to-health care facilities flow. The lines can indicate the presence of flow between two sub-regions or if the arrows have widths proportional to the volume of flow, they can represent quantification. Francis and Schneider in 1984 designed an interactive graphic computer programme called FLOWMAP, to produce a variety of maps of origin-destination data. Area patterns Chloropleth maps are used to embed histograms into the polygons of maps. Another alternative is to use stem-and-leaf plots to classify data before area pattern analysis. Surface and contour patterns To overcome the limitation of administrative regions for mapping, surface, and contour pattern maps allow the variable under study as continuous process throughout the region. This analysis assumes that the event is a continuous process and thus we get the maps, which are called isoline or isopleths.

Temporal Analytic Techniques
Most of the techniques described earlier can be used not only in health but in other areas as well. However, surveillance of diseases requires continuous systematic collection and analysis of a series of quantitative measurements. The detection and interpretation of changes in the pattern of the constructed time series is very important and therefore this presents a major challenge to the public health systems, as late detection of the ‘disease’ may result in missed opportunities for intervention. Quality Control Charts are useful methods for mapping surveillance data. The methods commonly used for such mapping include – the Shewhart test, the simple cumulative sum test and the V-mask. These methods are based on a comparison of incoming values from the time series with constant values, usually defined empirically from historical data. The advantages of such methods are that they can provide graphic information, and as such can be incorporated into an information system, helping the public health systems in the process. These are techniques poorly developed in our country but are found widely used in the US, France and some of the European countries. Statistical Monitoring Epidemiologists like to use the concept of ratio of case numbers at a particular time to past case occurrence using the mean or median. Based on this concept, the Centers for Disease Control and Prevention of the US Government has developed a technique in which the expected values are calculated based on a five year average. Public health professionals directly report the incidence of diseases to a central registry through a dedicated extranet environment and the analysed maps are available readily, sometimes even in the public domain.

GIS for Advocacy and Communication in Health
Since mapping is an excellent means of communicating a message clearly even to those who are not necessarily familiar with the methodology, GIS can be used effectively with leadership at various levels – panchayats, nagarpalika, districts, states and national administration to convey the priorities, the problems and provide an analysis and evidence based menu of options for programme implementation. Mapping is also used in participatory learning and evaluation techniques for community health projects. Small or large maps may be drawn or painted by groups or individuals to represent the context in which they are living. These maps, showing the location of the community structures – the houses, the wells, pumps, latrines, roads, PHCs, sub-centres and anganwadi centres give participants a wider view of where they are living. Such maps can help in discussion, assessment, analysis and decision making.

These maps when posted in public places and if updated provide a mirror for review and continuous updation of decisions at the community level. Such efforts have been tried out in different parts of India not only for decisions in the health sector but also for critical decisions on siting of handpumps, sub-centres and even primary schools.

   

Some resources related to GIS in Health
Anyone interested in the geography of disease will need a good understanding of the basics of epidemiology, or at least of health statistics. The World Health Organization website (www.who.int) is a good place to start. The WHO collects mortality and morbidity data from member countries, allowing international and inter-regional comparisons of health and disease. They publish the annual World Health Report, the Weekly Epidemiological Record, WHO Statistical Information System (WHOSIS), and a wealth of other useful publications. The Noncommunicable Disease Division of WHO describes the global epidemiology of diseases such as cardiovascular diseases and diabetes.

Data on the global epidemiology of cancer can be found at CANCERmondial, (www.dep.iarc.fr/dephome.htm) a website of the International Agency for Research on Cancer (IARC).

After the eradication of smallpox, poliomyelitis is slated for eradication. As of end of 1999, only 33 countries including India continue to have cases of Polio being reported. In studying the surveillance of Polio, it is important to determine, which type of polio is occurring in which parts of the world as these have important implications for the disease eradication strategy employed. The map (fig. 4) produced by WHO, thus has very important application. Further, in each of the countries, specific geographic areas are identified for concerted action and additional rounds of polio drops to the children below 5 years. Thus in addition to the NIDs for eradicating polio from India, which is carried out throughout the country, the states of Delhi, Uttar Pradesh, Bihar and West Bengal require special focus and additional attention. These interpretations become possible when GIS is used effectively.

In recent years, there has been a great deal of research on the geography of AIDS diffusion, particularly in the United States.

The AIDS Data Animation Project is a project of the Consortium of Earth Sciences Information Network (CIESIN). The still frames and animations illustrated at this web site document US disease trends for the years 1981 to 1993 using mortality data from the US National Centers for Health Statistics. The US Centers for Disease Control and Prevention (CDC) provide a wealth of information for public health professionals, including this series of slides on the epidemiology of AIDS. For a dramatic illustration of the impact of the AIDS pandemic worldwide, visit the UNAIDS AIDS Clock.

EpiMap – No treatise on GIS in Health is complete without a description of EpiMap.

EpiMap is a programme for IBM compatible microcomputers that displays data using geographic or other maps. Data values may be entered from the keyboard or supplied in Epi Info or dBASE files. The data may be counts, rates or other numeric values. In colour or patterns maps, the values are represented as shading or colour patterns for each geographic entity. In Dot density maps, randomly placed dots proportional in number to the values are placed in each entity. Epi Map also produces cartograms, in which the value for each geographic entity is allowed to control the size of the entity. Available in the public domain, it is freely available for researchers, practitioners from CDC, Atlanta, now for over 10 years. Outline maps are supplied with the software. It is designed to work both independently or as a companion to Epi Info.

HealthMap – It is a joint WHO/UNICEF programme based within the Department of Communicable Diseases of WHO. Created in 1993, to establish a GIS to support management and monitoring of the Guineaworm Eradication programme, it has been expanded to a much wider range of public health applications and now includes the promotion and use of GIS for other disease control programmes and in public health departments of a number of countries. Some of the specific applications of HealthMap include the Programme on Onchocerciasis in Africa, Surveillance and screening for African trypanosomiasis, Severe trachoma in Morocco, Mali and Gambia and Surveillance for Malaria stratification in Ethiopia. The HealthMap provides an excellent means of analysing epidemiological data, revealing trends, dependencies and inter-relationships that would be more difficult to discover in tabular format. It allows policy makers to easily visualise the problems in relation to the resources and more efficiently target resources to those communities in need. Public Health resources, the specific diseases and other health events can be mapped in relation to their surrounding environment and existing health and social infrastructures. Such information when mapped together creates a powerful tool for monitoring and management of disease and other public health programmes.

ChildInfo – ChildInfo is a database initially developed in India as a database on indicators related to nutrition and eventually expanded to include over 100 indicators on women and children. The database developed by UNICEF has simple tools to link with features that allow users to easily make tables, graphs and maps based on the data, without having to learn any mapping technology. Graphs and maps on a variety of indicators can be developed by users with average computer literacy in minutes and can be directly imported into documents, reports and presentations. The database for India has indicators and mapping facility down to the district level. ChildInfo is currently available from UNICEF as a LAN version and will be developed and available in the public domain as an internet version early next year. The tool has already expanded to cover not only all countries of the SAARC region of the sub-continent, but has also expanded to countries of East Asia, Pacific, Southern and Eastern Africa. Experience with ChildInfo has led the UN community in India to conceive a common UN Database called DevInfo, on indicators derived from various UN conferences and World Summits. The DevInfo when fully developed and released will have mapping facility right upto the block level for India.

Discussion groups – There are many email discussion groups related to use of GIS in health. One of the such popular groups is the Health-GIS, an electronic mailing service provided by HealthMap. To subscribe, a user has to send a message to [email protected]  with a message “subscribe health-gis” in the body of the mail.