Volume 42 Supplement 1
Geographical Information System (GIS) as a Tool in Surveillance and Monitoring of Animal Diseases
- Madelaine Norstrøm1
© The Author(s) 2001
Published: 31 March 2001
In the veterinary epidemiology, the advantage of mapping the locations of farms and other facilities with animals is obvious. In an outbreak of a disease it could make the management of the situation easier, and it could also provide a tool to evaluate different strategies to prevent the spread of infectious diseases. This paper aims to describe and give an overview of the possibilities and potential uses of a Geographical Information System (GIS) in the field of surveillance and monitoring of animal diseases. The following areas in which GIS and special GIS-functions could be incorporated are presented: recording and reporting information, epidemic emergency, cluster analysis, modelling disease spread, and planning control strategies. Different sources of data; geographical data, farm locations and disease information, used in the development of the GIS at the National Veterinary Institute in Norway are thoroughly described in the paper. Further, it presents a few examples where the GIS has been applied to studies of epidemiology and surveillance of animal diseases in Norway, which shows the significant value of GIS in these areas. At the same time, the incorporation of GIS in this field shows the scarcity of the data available, which should encourage improvement in the data recording and the quality of the registries.
KeywordsGeographical Information Systems Surveillance Epidemiology
A Geographical Information System (GIS) can be used as a tool for any discipline which handles with data that can be connected with geographical locations, such as countries, regions, communities, or co-ordinates. The systems have been developing rapidly in the past and today there are a number of different software which are more user-friendly than in the past. GIS is about to become tools for everyone.
The need for using this system also in the field of veterinary medicine has been emerging during the last decade. In 1991 Sanson et al. described the systems and possible applications in the field of veterinary medicine. Still, the most used application of GIS is to produce descriptive maps. However, the potential of GIS is much larger. Reviews in the field of environment and human health , and in the field of animal health  have been undertaken. GIS has been included in decision support systems for control of infectious diseases in animals [8, 2].
This paper will attempt to present the technology and possibilities of GIS with regard to surveillance and monitoring of animal diseases, and will discuss some applications of GIS in the field of veterinary epidemiology in Norway.
Geographical information system
The data can be stored in two formats; vector-based and grid-based.
The maps of the vector-based format display models of the real world using points, lines and polygons. Vector digitising captures a point as a x, y co-ordinate, while a line is captured as an ordered string of such co-ordinates. A polygon is a closed line. The grid-based format of data is captured as information of each quadratic cell in a screen and could be looked at as a photo of the area. GIS displays the geo-referenced data as theme layers which can be displayed one at a time or on top of each other, like overheads on a projector. These are stored in a geo-relational database. Each feature has attribute data linked to it which is stored in a table. Attributes can be any item of a feature which relate to the map, without being a part of it. The attribute data of the object with a geographical connection is stored in tables which can be joined with the geographical data through a common identifier (ID). An ID relevant to animal disease data could be a farm or region. Numbers are to prefer as ID as character variables often can be misspelled. The farms can be visualised using points, and regions such as veterinary districts, municipalities or counties are stored as polygons.
Description of GIS-functions useful in the veterinary surveillance
Recording and reporting disease information
GIS can be used to produce maps of disease incidence, prevalence, mortality, morbidity on farm, region, or national levels. The information is more easily understood when visualised on a map. Because information on diseases often tends to be aggregated (from information on each individual herd to municipality or county level) the information loses some of its value. If the information is mapped at the farm level, only small parts of a region can be visualised at the same time.
Another way to describe the incidences of diseases in a defined area can be to create density maps by using the density function. The density function creates a grid with a defined cell size and gives each cell in the area a density value of the infected farms. To adjust for the underlying population, a density map of the whole population at risk is created with the same cell size. The density maps are then divided to provide a map that shows the incidence of the particular disease in each area unit at the time unit chosen. This function can further provide maps which show the spread of the disease by displaying the maps as a movie. The GIS can also be incorporated in a real time outbreak notification, as done in an eradication program of the Aujeszky's disease in North Carolina . Maps displaying the updated situation in a region, together with farm information are important tools for field personnel and can also be incorporated in reports to producers, administrators and the media.
Analysis of clustering of diseases
To analyse whether a disease is clustered in space, time or in time and space other programs still have to be used because this is not yet a standard tool in the available GIS-packages.
The visualisation of the disease rates on digital maps can be misleading because the eye tends to interpret point patterns as clusters more often than what is real. Therefore, a cluster analysis should be carried out for an objective evaluation of the reported disease cases. The results of some of the cluster analyses can, thereafter, be imported into a GIS to visualise the location of clusters or cluster areas.
Model disease spread
Simulation models using programmes packages as @Risk (Palisade Corporation, Newfield, NY, USA) can be integrated within a GIS. Such simulation models can incorporate farm information such as herd size, production type as well as spatial factors like distance to the source of outbreak, population density and climate conditions, vegetation and landscape, all of which have been defined as risk factors for the spread of the modelled disease.  has developed a model of a potential outbreak of foot and mouth disease in New Zealand.
Planning disease control strategies
The neighbourhood analysis function can be used to identify all adjacent farms to an infected farm. It is a function that identifies all adjacent features with a certain criteria to a particular feature. Contact patterns such as common use of grasslands or sources of purchasing etc. could be visualised with a so-called spider diagram. This could provide insight into the possibility of transmission of infectious diseases between herds. In the planning of eradication of diseases, GIS has the possibility to perform overlay analysis to find high or low risk areas for diseases which depend on geographical features or conditions related to the geography. Studies of trypanosomiasis  and theileriosis [7, 3], are just some examples of how to use GIS to plan eradication of diseases depending on habitats of vectors or wild animal population. GIS could also be used to find areas with a low density of other farms [4, 10, 5] or risk areas of diseases as shown by  in case of Echinococcus multilocularis in foxes.
Description of the sources of the data used in the GIS in the field of veterinary epidemiology in Norway
Digital maps of Norway are provided and can be purchased from the National Map Department of Norway. The geographic data consisting of themes of each geographical feature are complete for the whole country in the scales 1:1 Mill. and 1:250 000. There are maps in the scale 1:50 000 for some parts of Norway.
The administrative boundaries of Norway can be divided into regions such as counties, municipalities, and in the veterinary field, veterinary districts which mostly consist of one ore more municipalities. The themes of veterinary districts were manually created and derived from the themes of the municipalities with the use of ArcView 3.1 (ESRI., Redlands, CA, USA).
The farm locations were provided by the Agriculture Property Registry, which is the official database of all information regarding agriculture in Norway. This registry includes all agricultural properties in the country, including properties with as well as without animal production. Animal producers can be found in the Registry of Production Subsidies (RPS), which records all farms which apply for financial support for their production. This registry is updated twice a year. This registry contains the production number, name, address of the applicant and number of animals in each production category at the day of application. The information of the locations of the farms with animal production as well as their production type and herd sizes are collected from these two registries. The disease recording system of the National Veterinary Institute includes the results from all tests of samples tested according to surveillance programs as well as diagnostic purposes of disease investigation. Today, all information about disease status in the counties, municipalities or on each farm can be collected from this database and imported into ArcView 3.1 as text files for joining with a geo-referenced theme such as farm, municipality, veterinary district or region. The GIS can thereby show the summarised information at a specific time or over any desired time period. Alternatively, the information in the database can be accessed through the ODBC interface.
A goal of the introduction of GIS is to have maps continuously displaying the situation for each of the diseases included in the Norwegian Surveillance Program.
Mycoplasma eradication in the swine population
One of the goals of The Norwegian Pig Health Service is to eradicate Mycoplasma hyopneumoniae from the Norwegian swine population. There are several projects going on and the role of the geographical information system in these projects will be to describe and follow the situation over time. The system also provides a tool to plan the eradication of the disease. The swine population in Norway is built up in a breeding pyramid, where the elite herds are at the top, followed by multipliers, conventional herds, and at the bottom of the pyramid; the slaughter pig herds. The strategy of the eradication program is to try to eradicate one level at a time with the starting point at the top. With a GIS the spatial aspects can be included in the eradication program. To avoid re-infection from nearby herds of another level, the program can identify those herds and help in the planning of further eradication of the disease.
In Norway, Paratuberculosis (PTB) has been considered to be a significant problem in the goat population, whereas PTB in cattle hadn't been diagnosed since 1979. Nevertheless, the fact that Norway claimed to have a PTB free status in the cattle population forced the authorities to start with an active surveillance program to test the cattle population systematically. The surveillance has focused on several risk groups, starting with the imported cattle, thereafter cattle in goat herds with PTB, older cattle, cattle in goat herds and finally a random sample dairy and beef cattle. The GIS has been used to identify the location of all goat herds, goat herds positive for PTB goat herds with cattle. During the test period the GIS has been used to identify the location of sero-positive and bacteriological positive cattle herds and to look for spatial relationship between positive cattle herds and positive goat herds.
An outbreak investigation of bovine respiratory syncytial virus in cattle
In a study of the transmission of epidemic respiratory disease between cattle herds, data from an outbreak of acute respiratory disease associated with BRSV (Norström et al. submitted) have been used to map the disease occurrence weekly as well as to provide incidence maps, and most likely clusters. The distances between all herds have been calculated by the use of ArcView 3.1 and will be used in a further study of risk factors, involving spatial factors. It is also planned to create a transmission model of acute respiratory disease and apply it in a GIS.
Discussion and conclusions
A GIS provides significant added value to current routine data that is usually taken into low consideration for either epidemiological or management purposes in veterinary medicine. A GIS considerably increases the efficacy of communication. Management and veterinary service tasks and resources during emergency can be improved with the use of GIS. Description of geographical disease dynamics over time, of risk factors due to spatial relationships as well as the drawing of risk and damage maps become feasible.
The deficiencies in a surveillance system also become more obvious and as a by-product of introduction of GIS, the system of collecting, storing and managing data can be improved. Last but not least, keep in mind: The maps will never be better than the original input data!
- Briggs DJ, Elliott P: The use of geographical information systems in studies on environment and health. World Health Statistics quarterly. 1995, 48 (2): 85-94.Google Scholar
- Laube P, Stärk KDC, Keller H: GIS-based computer program for the region-wide eradication of Enzootic Pneumonia (EP). Proceedings of the 8th International symposium on veterinary epidemiology and economics in Paris in 1997, published in Epidemiologie et sante animale. 1997, 31-32. 12.13.1–12.13.2Google Scholar
- Lessard P, L'Eplattenier RL, Norval RAI, Kundert K, Dolan TT, Croze H, Walker B, Irvin AD, Perry BD: Geographical information systems for studying the epidemiology of cattle diseases caused by Theileria parva. Vet Rec. 1990, 126: 255-262.PubMedGoogle Scholar
- Marsh WE, Damrongwatanapokin T, Larntz K, Morrison RB: The use of geographic information system in an epidemiological study of pseudorabies (Aujeszky's disease) in Minnesota swine herds. Prev Vet Med. 1991, 11: 249-254. 10.1016/S0167-5877(05)80010-8.View ArticleGoogle Scholar
- McGinn TJ, Cowen P, Wray DW: Intergrating a geographic information system with animal health management. Proceedings of the 8th International symposium on veterinary epidemiology and economics in Paris in 1997, published in Epidemiologie et sante animale. 1997, 31-32. 12.C.36Google Scholar
- Rogers DJ: Satellite imagery tsetse and trypanosomiasis in Africa. Prev Vet Med. 1991, 11: 201-220. 10.1016/S0167-5877(05)80005-4.View ArticleGoogle Scholar
- Perry BD, Kruska R, Lessard P, Norval RAI, Kundert K: Estimating the distribution and abundance of Rhipicephalus appendiculatus in Africa. Prev Vet Med. 1991, 11: 261-268. 10.1016/S0167-5877(05)80012-1.View ArticleGoogle Scholar
- Sanson RL, Ster MW, Morris RS: Interspread-A spatial stochastic simulation model of epidemic foot-and-mouth disease. The Kenyan Veterinarian. 1994, 18 (2): 493-495.Google Scholar
- Sharma P: Use of geographic information systems in animal health information programs. ACIAR Proceedings in. 1994, 51: 119-125.Google Scholar
- Staubach C, Teuffert J, Thulke H-H: Risk analysis and local spread mechanisms of classical swine. Proceedings of the 8th International symposium on veterinary epidemiology and economics in Paris in 1997, published in Epidemiologie et sante animale. 1997, 31-32. 6.12.1–6.12.3Google Scholar
- Staubach C, Tackmann K, Löschner U, Mix H, Busse W, Thulke H-H, Territo BM, Conraths FJ: Geographic information system-aided analysis of factors potentially influencing the spatial distribution of Echinococcus multilocularis infections of foxes. Proceedings of a meeting held at the West County Hotel Ennis CoClare on the 25th 26th and 27th of March 1998. Edited by: Trushfield MV, Goodall EA. 1998, 40-47.Google Scholar