Revolutionalising epidemiology with GIS

Revolutionalising epidemiology with GIS

SHARE

Remote sensing and GIS are becoming essential tools in public health, being used not only to map the spatial distribution of disease prevalence but also for disease surveillance, epidemiological research, and providing the guidelines for decision making. Here’s a look at the use of geospatial in vector borne diseases in India

Public health epidemiology is the study of horizontal and vertical structure of disease infection, transmission, diffusion, magnitude, age groups and sex of the diseases with reference to space and time. GIS has been used to updating and mapping disease epidemiological information.  Perhaps, remote sensing and GPS have been integrated under the GIS umbrella for disease surveillance, situation analyse and the spatial modeling of disease transmission. The first application of cartography was used in public health epidemiology for mapping diarrhea disease in London, in 1854 by UK physician Jonson Smith. However, applied GIS and remote sensing are becoming essential tools in mapping epidemiological information, disease surveillance, health monitoring, surveying, sampling design, disease control programmes and predicting disease transmission. It has become a significant decision making tool in public health epidemiology. This article focuses on use of geospatial in vector borne diseases in India.

Introduction
GIS is computer software for data capturing, thematic mapping, updating, retrieving, structured querying and analysing the distribution and the differentiation of various phenomenon including communicable and non-communicable diseases across the world with reference to various time periods. It is “tailor-made maps/ layers of thematic map information.” Satellite data is reliable, offers repetitive coverage and is accurate. It has been used for studying and mapping surrogate information relevant to the environments of disease transmission at particular time periods.  Integrated remote sensing and GPS under the GIS umbrella have also been used for disease surveillance and disease epidemic control.

The study of public health epidemiology contains information on the occurrences of diseases, infection rate, age group, sex, disease transmission, site specification of the patients, host availability of the parasite or virus loads, etc. It was used to state the horizontal and vertical structure of the diseases, history of the disease, etc. with reference to space and times. GIS has been used to map the geographical distribution of prevalence of disease (communicable and non-communicable diseases), the trend of the disease transmission and spatial modeling of environmental aspects of disease occurrence (Bailey TC 1995, Gatrell AC, 1995, Cressie NAC, 1993, and Srividya A, et al, 2002). GIS was also used for spatial analysis, spatial modeling, cause and effect analysis, cognate models, temporal analysis etc.

GIS has in-built facilities of conventional and the scientific knowledge of traditional, fundamental concepts of formal mapping with signs and symbols, variety of colours, shades, lines and poly lines, patterns, etc. GIS has computer-aided designs, symbols and colours for thematic mapping or customised mapping, and perhaps, embedding mapping facilities, overlay analysis, cluster analysis, nearest neighborhood analysis, pattern recognition, temporal analysis, interpolation of point data (Kriging, Co-Kriging, Universal Kriging), spatial correlation, fussy analysis, linear determinant analysis, the probability of minimum and maximum likelihood analysis etc., of geospatial analysis of thematic information. Thus, remote sensing and GIS could be used for mapping and studying and analysing the information relevant to the disease transmission of public health epidemiology with reference to space and time.

GPS for epidemic surveillance

GPS has been used directly on top of a map for site specific location to collect field data in real time, convert and log real-time GPS coordinates. It has been assisting field surveys to collect information continuously and to automatically update the geographic coordinates with minimum 500 points. The latest version of geographic tracker includes a map basic application, which allows “GPS tracking” by showing a real-time GPS derived position directly on top of a map. It has facilities to collect and attribute field data directly into one’s geospatial database engine (GIS software) in real time, an exciting concept that may be called “GPS Geo-coding”. The geographic tracker can process live or simulated GPS message data (“Live GPS Data” or “Simulated GPS Data) on online database connectivity. GPS has been used for disease surveillance in crucial situations like dengue epidemic in India. The dengue vector mosquito’s flight range is between 400 to 600 meters and the Aedes species has the outdoor resting habitats and is a day-biting mosquito. A reconnaissance survey was conducted in the nearest house of closeness to the intersection points of 100 meters grid samples. The available GPS instruments has an inbuilt error of (+) or (–) 100 metres. Therefore, the GPS instrument was found useful to map the dengue vector’s breeding habitats with site specifications including house locations, streets, house type, and locality of the areas with interval 100 metre interval.

GIS for mapping point data and interpolation of contour surface

GIS has been used for mapping epidemiological data and for spatial interpolation of data for data not available/ un-surveyed places (Bailey TC, 1995, Cressie NAC, 1993, and Srividya A, et al, 2002). GPS was used to collect the filariasis epidemiological information of selected villages, based on the GIS-based 25 km x 25 km grid sample procedures. The data pertaining to the mF and disease rate was mapped with graduated point symbol and the interpolation of contour surface was created for predicting the filariasis mF rate in the areas where data was not collected. The mF infection rate of selected sample villages was overlaid on the interpolation of contour surface of the predicted filariasis map of part of Tamil Nadu State in India. The procedures applied in the study was used for mapping the disease infection in the areas where data was not available and it was used for action plan for implementing disease surveillance, management of disease control programmes and for disease management in a vast country like India.

spatial interpolation of filarial diseases and infection

GIS for mapping lineament data using line symbols

The population movements to specialised hospitals located in the cities, floating population of the hospital outpatients and the inpatients, health services, the roads and railway service facilities to the hospitals etc., have been mapped using line symbols and flow maps. The drainages, irrigation canals, rivers, streams etc., of malaria vector mosquitoes (Anopheles genus) and the JE vector mosquitoes (Culex genus) groups breeding habitats were mapped using line symbols. The site specifications of the houses in the streets with dengue vector mosquitoes of Aedes species (Aedes aegypti or Ae. Albopictus) breeding habitats have been mapped with line symbols. The mosquitogenic conditions suitable for profusion of mosquitoes around the rice fields and the lineament features of irrigation canals from the water resources (rivers, streams, and lakes, tanks, dams etc.) with buffer zone of 2.5 km radius of malaria and JE vector mosquitoes flight range have also been mapped with line symbols. A hypothetical model of flow map showing the frequency and the number of population movement from various parts of Tamil Nadu state to Pondicherry for seeking hospital health treatments was mapped with flow maps using graduated line symbols with colour .

Population movements from various parts of Tamil Nadu to Pondicherry (JIPMER Hospital) for treatments

GIS for mapping regional data with polygon / area symbol
GIS has been used for mapping the district level malaria disease prevalence and the epidemiological information with polygon symbol. The traditional method of vector-borne disease control was based on the empirical knowledge; however, it was most crude, laborious, expenditure, erroneous and time consuming, whereas the remote sensing and GIS techniques are scientific, accuracy, speed, reliable. GIS and remote sensing have been used for mapping vector habitats, vectors abundance and density, assessing the risk of vector borne diseases. Perhaps, it was used for finding the source of infection, root cause of disease transmission and diffusion of the diseases (Palaniyandi M, 2012, Sabesan S, et al, 2000 and 2006).  It was also used for assessing the community at risk of disease transmission, and thus it has been epidemiologically important for choosing appropriate controlling methods and priority of the areas for both vector and disease control.

GIS for real time mapping and structured querying

GIS has facilitates for structured querying and decision making process at certain level. The structured spatial queries relevant to demographic features, disease prevalence, environmental aspects and the socio-economic risk factors have been provided the diffusion of disease transmission, and hence, the action plan for the disease control operations was taken to prevent the disease epidemics in the country. The web mapping GIS using API has been readily available to customize the embed mapping of the real time epidemiological disease information to the individual and planners for browsing the information from the public domain of health GIS websites. The web mapping API are becoming important, mainly the embed customized web mapping GIS (ASP, .Net, html, java, python, CSS, PHP, Arc IMS, Geo ext, C, C++, Visual Basic, Arc objects), which has user interface facilities for browsing, querying, and table sorting and drawing the disease epidemiological information in different part of the country.

Spatial interpolation of filariasis for Tamilnadu

Remote sensing and GIS for spatial prediction of disease transmission

The geostatistical analysis of remote sensing and climate, geo-environmental variables, the spatial models have been providing us significant and reliable results and the guidelines of algorithms for predicting the people of community at risk of disease transmission with reference to space and time (Vounatsou P, et al, 2009).  For example, a Geo-Environmental Risk Model (GEFRM) for filariasis transmission was developed using remote sensing and GIS during the period 2000-2003. The GERM model provided us the guidelines for predicting the probability of filariasis transmission risk in Tamil Nadu region with reliable, scientific, accuracy and spatially significant information. The model was customized according to the environmental parameters, encompassing: altitude 0 – 2000 m mean sea level; temperature: 80 – 370 C; rainfall 300mm – 1500mm and relative humidity 40% – 90% for deriving filariasis risk index (FRI). Based on the results obtained in the analysis of FRI, the Geo-environmental filariasis transmission risk map was created on the GIS platform, and further it was stratified into four spatial entities, which are hypothesized as potentially high risk (FRI: 31 – 38), moderate risk (FRI: 23 – 30), low risk (FRI: 15 – 22) and no risk areas (FRI: < 15). The GERM spatial model for filariasis transmission risk was validated in the different geographical regions (plain, plateau, hills, river beds, coastal, and the uplands) with supported ground truth data. The negative value of spatial prediction provides the guidelines for decision making and planning for deciding to pass up the areas for resurveying or to avoid the implementation of disease control programme where (or) the efforts need not be taken / are not required (Sabesan S, et al, 2006), and hence, this model could be assisted to the planners for action plan in the right place and the right ways. This kind of spatial models could be importantly useful for decision making for disease control programme Remote sensing and GIS for mapping vector breeding sources

The visual interpretation of the multi spectral and multi temporal satellite sensors data products derived from the earth observation resource satellites Landsat TM, SPOT, IRS LISS I, LISS-II, LISS III and IRS WiFs, IKONOS, and the meteorological satellites NOVA- AVHRR has been used for mapping the mosquito breeding habitats (Palaniyandi M, 2004, and 2008, Sharma VP, et al, 1996, and Wood BL et al, 1991, and 1992). The range of NDVI values derived from the satellite data were highly significant with the vector abundance and the spatial occurrences of vector borne diseases (Liu J and XP Chen, 2006). The vectors and vector borne diseases have been directly controlled by the geo-climatic variables (altitude, mean annual temperature, mean annual rainfall, potential evapotranspiration, readily available soil moisture, soil types and water logging potential, and terrain slope, and the normalised difference vegetation index (NDVI) of space-borne remote sensing data (Leonardo LR, et al., 2005, Dale PE, et al., 1998, Hay SI, et al., 1998, Lindsay SW, et al, 1998, and Rossi RE, et al, 1992).

GIS for optimum health service coverage
The spatial clustering, nearest neighborhood analysis was performed for easy understanding of the filariasis spatial pattern and disease clustering and the spatial ring buffering was created for the optimum service coverage of the patients. The different distance rule of 0.2KM, 0.3KM, 0.4KM 0.5 KM, 0.6, 0.7 KM, 0.8 and 0.9 KM were created over the disease distribution map, using spatial ring buffering technique at GIS platform. The minimum, maximum and the mean distances of each disease cluster are calculated against to each distance rule/ ring buffering. The lymphodema cases proportionally high in the 56 -75 years of age group. The list of existing PHC/ GH is depicted on the Pondicherry urban boundary map. The study finds that the 0.7 km ring buffering distance has optimum service coverage. The hypothesis of the present study is that aged patients could travel less than 1 km distance from their residence to the health centers for morbidity management. The study area required 15 centers with 0.7 km distance ring buffer or coverage area. Out of 15, 10 centres already existed, and 5 more new centers were required to cover all the patients. We suggest opening up of self-help health service centers with coverage of less than 1-kilometer distance in urban places like Pondicherry for lymphatic filariasis morbidity management (Palaniyandi M, 2008).

Filariasis Transmission Risk in Tamilnadu, based on the GERM model


GIS for mapping, health monitoring, and decision making

GIS has been facilitating integration of remote sensing, database management systems (DBMS), computer cartography, and geo-statistics. It has been used for mapping, monitoring, visualising, monitoring, retrieving, analysing and modeling of geo-referenced data with high accuracy. For example, it was used to map the biodiversity and the ecology of vectors, disease prevalence, disease transmission, spatial diffusion etc. Perhaps, it has been used for monitoring the past, present and the future disease control programmes. The current situation of the disease prevalence in the country, based on the historical data, may cause error in the disease control programme. Therefore, it is mandatory to resurvey the areas for health monitoring. (Example 10: Mapping the priorities of districts for resurveying the filarial antigenaemia detection for implementing the national filariasis disease control programme at national level).

Priority for Filariasis Survey in India

Conclusion
Remote sensing and GIS are becoming essential in public health for mapping the geographical aspects of vector borne diseases, bio-diversity of vectors, viral diseases, parasites, bacterial diseases and studying the geo-environmental risk factors associated with disease occurrences. It is being used not only to map the spatial distribution of disease prevalence but is becoming an essential tool for disease surveillance, epidemiological research, and providing the guidelines for decision making. There is no limit on applications; it can be used in national disease control programmes, disease surveillance, spatial modeling, and disease transmission.

References
1.    Bailey TC, Gatrell AC, 1995. Interactive Spatial Data Analysis, Longman Scientific & Technical, Harlow England

2.    Dale PE, et al., 1998. An overview of remote sensing and GIS for surveillance of mosquito vector habitats and risk assessment. J Vector Ecol, 23: 54-61

3.    Hay SI, et al., 1998. Prediction of malaria seasons in Kenya using multi-temporal meteorological satellite sensor data. Trans. R. Soc. Trop. Med. Hyg. 92, 12–20

4.    Leonardo LR, et al., 2005. A study of the environmental determinants of malaria and schistosomiasis in the Philippines using remote sensing and geographic information systems. Parassitologia 47 (1):105-14

5.    Lindsay SW, L Parson and CJ Thomas, 1998. Mapping the ranges and relative abundance of the two principal African malaria vectors, Anopheles gambiae sensu stricto and An-arabiensis, using climate data. Proceedings of the Royal Society of London Series B – Biological Sciences,  265: 847-854

6.    Liu J and XP Chen, 2006. Relationship of remote sensing normalized differential vegetation index to Anopheles density and malaria incidence rate. Biomed Environ Sci 19 (2):130-2

7.    Palaniyandi M and T Mariappan, (2012), Remote Sensing and GIS for Malaria Control in the Urban Environment – a case study, (Containing the spread of Malaria with geospatial Tech), Geospatial World Weekly  27th February 2012 Vol. 8 Issue 9,PP 1-9

8.    Palaniyandi M (2008), “GIS for lymphatic filariasis morbidity management and control”, Coordinates (e-journal), Vol. V, Issue 5, PP 24-28, May 2008. ISSN: 0973-2136

9.    Srividya A, Michael E, Palaniyandi M, Pani SP. and Das PK (2002), “A Geostatistical Analysis of Lymphatic Filariasis Prevalence in Southern India“, American Journal of Tropical Medicine & Hygiene, Vol.67 (5), 2002, PP. 480-489.ISSN: 0002-9637

10.    Rossi RE, Mulla DJ, Journel AG, Franz EH, (1992). Geostatistical Tools for modeling and interpreting ecological spatial dependence. Ecol Mono 62: 277-314

11.    Sabesan S, M.Palaniyandi, Edwin Michael and P.K. Das, (2000). Mapping of Lymphatic   Filariasis at the district level in India.  Ann Trop. Med & Parasit  Vol. 94 (6) 591-606

12.    Sabesan.S, K.H.K. Raju, A.Srividya, and P.K.Das (2006). Delineation of Lymphatic Filariasis Transmission Risk Areas: A geo-environmental approach. Filaria Journal (online journal) 5(12): 1-6

13.    Sharma VP, et al., 1996. Estimation of larval production in Sanjay Lake and its surrounding ponds in Delhi, India using remote sensing technology, Southeast Asian J Trop Med Public Health, 27: 834-40

14.    Sharma VP, et al., 1996. Study on the feasibility of delineating mosquitogenic conditions in and around Delhi using IRS satellite data. Indian J. Malariology 33: 107–125

15.    Vounatsou P, G Raso, M Tanner, EKN Goran and J Utzinger, (2009), “Bayesian geostatistical modeling for mapping Schistosomiasis transmission, Parasit. Vol. 136(13): 1695-1705

16.    Wood BL, et al., 1991. Distinguishing high and low anopheline-producing rice fields using remote sensing and GIS technologies. Prevent. Vet. Med. 11, 277–288

17.    Wood BL, et al., 1992. Estimating high mosquito-producing rice fields using spectral and spatial data. Int. J. Remote Sensing, 13, 2813–2826

18.    Wood BL. et al., 1991. Spectral and spatial characterization of rice field mosquito habitat. Int. J. Remote Sensing 12: 621–626