356 Am. J. Trop. Med. Hyg., 61(3), 1999, pp. 356–360 Copyright  1999 by The American Society of Tropical Medicine and Hygiene APPLICATION OF RISK ANALYSIS AND GEOGRAPHIC INFORMATION SYSTEM TECHNOLOGIES TO THE PREVENTION OF DIARRHEAL DISEASES IN NIGERIA PHILIP C. NJEMANZE, JOSEPHINE ANOZIE, JACINTHA O. IHENACHO, MARCIA J. RUSSELL, AND AMARACHUKWU B. UWAEZIOZI Institute of Space Medicine and Terrestrial Science, and Institute of Non-invasive Imaging for Parasitology, International Institutes of Advanced Research and Training, Chidicon Medical Center Owerri, Imo State, Nigeria; Imo State Water Corporation, Imo State Government, Owerri, Imo State, Nigeria Abstract. Among the poor in developing countries, up to 20% of an infant’s life experience may include diarrhea. This problem is spatially related to the lack of potable water at different sites. This project used risk analysis (RA) methods and geographic information system (GIS) technologies to evaluate the health impact of water source. Maps of Imo State, Nigeria were converted into digital form using ARC/INFO GIS software, and the resulting coverages included geology, hydrology, towns, and villages. A total of 11,537 diarrheal cases were reported. Thirty-nine water sources were evaluated. A computer modeling approach called probabilistic layer analysis (PLA) spatially displayed the water source at layers of geology, hydrology, population, environmental pollution, and electricity according to a color-coded five-point ranking. The water sources were categorized into A, B, and C based on the cumulative scores  10 for A, 10–19 for B, and  19 for C. T-test showed revealed significant differences in diarrheal disease incidence between categories A, B, and C with mean  SEM values of 1.612  0.325, 6.257  0.408, and 15.608  2.151, respectively. The differences were significant between categories A and B (P = 0.0000022), A and C (P = 0.0000188), and B and C (P = 0.0011348). The PLA enabled estimation of the probability of the risk of diarrheal diseases occurring at each layer and solutions to eliminate these risks. Diarrhea is an increase in daily stool weight above 200 g per day and an increase in stool liquidity and frequency. Watery diarrhea is caused by several microorganisms; the most prevalent causative agents include Vibrio cholerae, Escherichia coli, Giardia lamblia, and Salmonella typhi. In Africa, Asia, and South America, acute diarrheal illnesses are not only a leading cause of morbidity in children, pro- ducing an estimated 1 billion cases per year, but also are responsible for 4–6 million deaths per year, or 12,600 deaths per day. In some areas in sub-Saharan Africa, more than 50% of the deaths of children are directly attributable to acute diarrheal illnesses. 1,2 This problem is most evident in Nigeria, Africa’s most populous country. 3 The objective of the present project was to devise a model that will eventually lead to reduction in the incidence of diarrheal diseases related to water-borne infections. This project used a new approach of risk analysis and geographic information system (RA/GIS) technologies to perform health and environmental impact assessment of water source (treat- ed surface water and tube wells or bore holes). The GIS provides computerized capture, storage, manage- ment, analysis, retrieval, and display of spatial and descrip- tive data that are geographically referenced to a common coordinate system. It is composed of a database (spatial and attribute), data input (digitized scan), cartographic display system, database management, and geographic analysis sys- tems (e.g., overlay process and buffer zone creation). The spatial database contains information held in the form of digital coordinates, which describe spatial features. 4 Risk analysis includes risk assessment, risk management, and risk communication. Risk assessment is the process of identifying a hazard and evaluating the risk of a specific hazard, either in absolute or relative terms. Risk assessment provides estimates of the probability of occurrence of un- wanted future events. This information is used by the risk management teams to choose the best course of action. This approach has been previously applied in environmental as- sessment. 5 MATERIALS AND METHODS The population of the area studied was 2,311,675 persons 0–85 years of age in 18 local government areas of Imo State, Nigeria. The populations in the various areas did not differ significantly with regard to socioeconomic status, education- al level and access to health care based on a self-reported questionnaire. The reported diarrheal diseases were obtained from the annual disease surveillance report from the various health centers in the local government areas studied. Digital coverages were created using ARC/INFO GIS software (En- vironmental Systems Research Institute, Inc., Redlands, CA). These coverages were derived from existing 1:350,000 scale maps of Imo State. The resulting GIS layers included geology, hydrologic features, towns, and villages. All cov- erages were converted from state plane coordinates to the universal transverse mercator coordinate system and projec- tion. The digitized geology was color-coded according to soil types. The true spatial relationship of the geology, hydrolo- gy, towns, and villages was provided using an overlay of the layers in the GIS database. The geology coverage was used to determine suitability of the various areas for placement of bore holes. The hydrologic features were used initially to determine the proximity of rivers and streams to towns and villages studied. The use of the water source by inhabitants of the towns and villages was latter confirmed by field data. The means of distribution of water was considered reticu- lated when water distribution was by an extensive pipe- borne water network or satellite when water was collected for use from the site of a bore hole. The latter entailed walk- ing short distances to the site of the bore hole by users. To examine the contribution of the various layers of the database to the overall probability (P) of a hazard occurring, each data layer was sequentially considered by what we here refer to as probabilistic layer analysis (PLA). The PLA is a method of risk assessment that begins with identification of a hazard endpoint and sequentially weighing the factors con- tributing to this endpoint or its probability at each layer of