New geographical approaches to control of some parasitic zoonoses.

Reviews/Analyses New geographical approaches to control of some parasitic zoonoses K.E. Mott,1 1. Nuttall,2 P. Desjeux,3 & P. Cattand4 The advent of new technology for geographical representation and spatial analysis of databases from different sectors offers a new approach to planning and managing the control of tropical diseases. This article reviews the geographical and intersectoral aspects of the epidemiology and control of African try- panosomiasis, cutaneous and visceral leishmaniasis, Chagas disease, schistosomiasis, and foodborne trematode infections. The focal nature of their transmission, increasing recognition of the importance of animal reservoirs, and the need to understand environmental factors influencing their distribution are common to all these diseases. Geographical information systems (GIS) open a completely new per- spective for intersectoral collaboration in adapting new technology to promote control of these diseases. Introduction Geographical approaches, particularly geographical information systems (GIS), provide common ground for dialogue between zoologists, veterinarians and medical public health workers. As a result, there is now increasing awareness of the role of animals in the epidemiology of human "emerging" infections (5), and the control of tropical diseases is gradually involving other sectors besides traditional public health services. GIS is a computer-based technology for input, storage, analysis and display of spatial data (19). It permits cross-sectional display and analysis of multi- ple databases using real geographical coordinates to a specific scale, i.e., rainfall, soil type, soil humidity, health services infrastructure, disease vectors, and infection or disease in animals and people. This rela- tively new analytical tool in the field of epidemiol- ogy facilitates the collaboration of different sectors 1 Chief, Schistosomiasis Control, Division of Control of Tropical Diseases, World Health Organization, 1211 Geneva 27, Switzer- land. Requests for reprints should be sent to this address. 2 WHO Consultant. 3 Medical Officer, Trypanosomiases and Leishmaniases Control, WHO, Geneva, Switzerland. 4 Scientist, Trypanosomiases and Leishmaniases Control, WHO, Geneva, Switzerland. Reprint No. 5594 and can be flexibly adapted to the needs of the endemic countries. The effective integration of new technology such as GIS is done methodically in a step-by-step fashion. For control of tropical diseases all applications begin at the periphery. If the data generated at the periphery have not been reliably collected and checked, no amount of sophisticated technology at the central level can improve them. Furthermore, GIS analyses to determine causal rela- tions are estimates or approximations to support or refute hypotheses, and their conclusions should be confirmed by field epidemiological studies. Appropriate use of new geographical approaches will depend on (1) proven benefit from older geo- graphical approaches in control, (2) a clear priority for the sequence of introduction of databases, and (3) scientifically sound hypotheses. If these three criteria are not fulfilled, the new geographical approaches may become junkyards or graveyards of databases. This review describes the evolution of geograph- ical approaches and methodology as applied to understanding the epidemiology of parasitic diseases with animal intermediate hosts. It does not include the use of remote sensing and satellite image data in relation to vector-bome diseases, which other authors have recently described (13, 16, 25) and which require sophisticated technologies (both hard- ware and software) to analyse the data input. More- over, remote sensing data are managed and analysed as one of the databases in a GIS. Bulletin of the World Health Organization, 1995, 73 (2): 247-257 © World Health Organization 1995 247 K.E. Mott et al. Geographical approaches This article does not cover the history of geographi- cal approaches for controlling tropical diseases. Their origins have been variously interpreted, e.g., maps of the drainage operations of the Pontine swamps in ancient Rome to eliminate malaria-long before the link between the parasite and the mosqui- to could be imagined, or the map of cholera cases in London in 1833-34 which pinpointed the Broad Street public pump as a point source of infection. In all operational tropical disease programmes the necessity of maintaining a geographical approach is recognized. In the field, simple maps bring out information which can easily be interpreted and used in management of control activities. In the 1960s, for example, sketch maps of malaria endemic areas served as the basis for all operations, surveillance and monitoring in attempts at global malaria eradica- tion. Simple sketch maps of endemic areas have been used so extensively in national control programmes under the Ministry of Health that the entire northeast of Brazil has been literally mapped, by household, on A4 sheets to define malaria or smallpox surveillance routes, or the distributions of Chagas disease, yellow fever, leishmaniasis and schistosomiasis. The advent of technologies for geographical rep- resentation of epidemiological data presents new opportunities and challenges for the health sector. One of the principal challenges is to establish a working dialogue with other sectors. Geography is not the domain of any one sector. Its very nature sur- passes the territoriality of any one field of science. While computer-assisted cartography and geographi- cal information systems have expanded the potential, if not the ease of geographical approaches, we are confronted by constraints such as lack of map boun- dary data and the intrinsic unreliability of other data- bases. The standard atlases of disease distribution have always recognized this fact (9). More recently this has been the cause of a dramatic appeal to cor- rect the absence of adequate map databases (11). African trypanosomiasis The history of the control of African trypanosomiasis stands out as a model on the use of geographical approaches linking veterinary and human public health objectives and operations. During the colonial period in Africa, mapping the distribution of the geo- graphical limits of tsetse flies, animal reservoirs, and human disease preceded the control efforts. Since 1954, FAO/WHO collaboration in this area has confirmed the intrinsic necessity of a geographical approach to control. The parasites causing human African trypanoso- miasis, Trypanosoma brucei gambiense and T.b. rhodesiense, are transmitted by Glossina (tsetse flies). Both parasites infect domestic and wild mam- mals. The disease due to T.b. gambiense occurs in 27 countries of West and Central Africa and is mainly transmitted between infected persons. The role of animals as a reservoir and in the epidemiology of this form of the disease is ill defined. On the other hand, disease due to T.b. rhodesiense occurs in 10 countries of East and South-East Africa and the dis- ease is clearly zoonotic. Wild and domestic animals play an important role in foci for disease mainten- ance and resurgence (32). Transmission of sleeping sickness is focal, even within the limits of an endemic area so that transmis- sion and risk of infection are not homogeneous. Only 10 of the 31 known species and subspecies of Glossi- na are incriminated as vectors in sleeping sickness transmission. Transmission is predicated on the pres- ence, type and behaviour of vectors and is closely linked to the environment (i.e., minor hydrographical features play a major role in determining transmis- sion sites). Peoples' attitudes and habits are also major factors in transmission occurrence and inten- sity. In general, risk areas and populations at risk are poorly defined. In most endemic countries, large geographical areas have been designated as being "high risk" for transmission, while in reality the actual areas at risk of transmission are focal and separated, sometimes by large distances. Within a geographical approach to sleeping sick- ness, the distribution of infected persons, infected animals, tsetse flies (species type as well), and of various environmental factors such as water bodies, forests, roads, villages, plantations, etc. are needed. To implement or improve control approaches, iden- tification of cross-sectional areas where several com- plementary or limiting factors coexist may be useful. Circumscription of the sites to be brought under sur- veillance, where control is needed, is the first step in a long-term control strategy-not unlike the GIS- based methodology proposed in Guatemala for onchocerciasis (24). The operational aspects of control also have to be considered in the application of geographical methodologies. In the course of passive surveillance (fixed post-systematic surveillance), the identifica- tion of the origin of patients with African trypanoso- miasis may play the role of an "early warning device" for the resumption of transmission and the identification of the geographical area at risk of an epidemic. The usefulness of the data will depend on the extent to which they have been accurately recorded and integrated into a GIS-based file and analysed immediately rather than accumulated and analysed in an aggregate fashion. Certainly data from active surveillance should have carefully defined sur- WHO Bulletin OMS. Vol 73 1995248 Geographical approaches to control of parasitic zoonoses vey areas on a GIS map and the database records should include the origin, number of patients, treat- ment and secondary reactions, date of examination, patients' follow-up including date, location, test results, physical condition, and relevant medical, environmental and social information. Vector control is generally considered a crucial component of all operational programmes. The data- bases already available are usually very extensive and detailed. Integration of existing databases can be time-consuming and may not be necessary to address the current programme needs. All data collected and integrated into GIS should optimize the available resources to achieve and sustain control of sleeping sickness. An environmental study in the Vavoua focus in Daloa prefecture, Cote d'Ivoire, has opened a new area of GIS analysis to support the control of African trypanosomiasis.a This study has shown that there is a direct correlation between the risk of Trypanosoma infection and the agricultural cultivation methods of certain "socially open" groups (communities com- prised of different ethnic groups, e.g., exogenous tribes such as Mossi or Dioula and endogenous tribes such as Gouro or Baoule), compared with the agri- cultural methods of "socially closed" communities of monolithic tribes. Landscape contrast due to the dif- ference in types and methods of cultivation can be observed in the field. Correlation between agricultu- ral habits and disease distribution was first noticed by comparing static maps. Studies are now being implemented to combine remote sensing and GIS to allow for the quick assessment of risk for rapid pre- ventive intervention (vector trapping) and for case detection (mobile surveys or in fixed centres). A detailed geographical analysis of the environ- ment-vegetation/vector distribution/human popula- tion and infection in Nola Biolo area of the Nola pre- fecture of the Central Africa Republic-has shown that G. palpalis palpalis tsetse flies were the main vectors in the coffee plantations and G. fuscipes fuscipes in the manioc fields and at bathing sites.b Operationally this was important since there is a short geographical overlap of the ranges of the tsetse flies and the trapping methodology must be adapted for each species. In a large-scale control programme in southem Uganda, the distribution of tsetse fly traps was deter- a Laveissiere C et al. Les communaut6s rurales et la lutte con- tre la maladie du sommeil en foret de COte d'lvoire. Unpublished document, WHO/TRY/94.1, 1994. b Kounda Gboumbi J-C. Les glossines de la region de Nola. Especes et repartition. Relation avec la trypanosomiase humaine et possibilites de lutte. Thesis for diploma in Public Health, Universite Marien N'Gouabi, Bangui, 1990. mined by the geographical origin of sick persons.c Exhaustive trap placement throughout the entire endemic area was not feasible owing to lack of logis- tics and limited financial resources. Thus, detailed maps of the patient's origins were required for per- manent adjustment of planning and implementing control. In northem Uganda, population movements related to civil unrest and refugees were of strategic importance since the majority of infected persons originated from Sudan. Before user-friendly GIS software for the per- sonal computer became readily available, an opera- tional plan was completed to define the peripheral health care system in the Moundou prefecture of Chad.d Subsequently, an attempt was made to inte- grate control of African trypanosomiasis (T. brucei gambiense) by superimposing a series of maps with the health centres having a catchment area of 10 000 persons and accessible within 30 km. Using static maps it was difficult to adjust the catchment areas to correspond to the population density and distance to the health centres. It was found that the major limita- tion in using static maps was the need for constant recalculation of the data, even though the disaggre- gated or discriminant data on habitations, river basin boundaries, and agricultural areas were available and were represented on the static maps. Today a similar type of operational GIS analysis could be imple- mented within weeks and kept up-to-date. Using sequential geographical mapping, Simarro et al. (28) showed that foci in three villages were the origin of transmission of T. gambiense in the larger endemic area of Luba district of the island of Bioco in Equatorial Guinea. Cases on the island occurred mainly in proximity to these three foci, each limited by altitude and surrounding vegetation. From a his- torical review of the geographical distribution of human infection, they concluded that if early preven- tive action had been undertaken in the areas of the appearance of the first cases, the extension of the foci could have been avoided. It is now possible to integrate data derived from health services and spe- cific surveys into a GIS which would establish the potential limits of a focus of transmission. Thus, the preventive control measures could effectively be tar- geted to the population at risk. c C. Lancien, personal communication in report of the National Sleeping Sickness Control Programme, Center of Control of Try- panosomiasis, Vector Control Project in Busoga. d Tobyo AS, Bouchet B. i) Programme de controle de la try- panosomiase humaine en Republique du Tchad. ii) Propositions du Comite-Sant6 pour la mise en place du plan de couverture sanitaire des districts du Logone Occidental. Unpublished docu- ments, Projet de cooperation Franco-Tchadienne, Convention No. 356/C/89, 1991. WHO Bulletin OMS. Vol 73 1995 249 K.E. Mott et al. Chagas disease Chagas disease or American trypanosomiasis affects up to 18 million persons in 17 countries of the West- ern Hemisphere (34). Since its discovery in 1909 by Carlos Chagas, the complex interaction between Try- panosoma cruzi in the arthropod reduviid vector and in wild and domestic animals and man has been a focus of research and been taken into account by all control strategies. The zootic cycle of T. cruzi extends far beyond the endemic areas of human infections; however, in areas where the disease is primarily zootic, the level of socioeconomic deve- lopment has been extensive and contact of people with both the reduviid vector and reservoirs is ran- dom. In Latin America, the domestic cycle of the disease is rapidly disappearing in the face of exten- sive control efforts and public recognition of the danger of the presence of the bug in households (34). Prospective cross-discipline studies to assess the interrelationship between the sylvatic, peridomestic and domestic dynamics of T. cruzi transmission are few. Using isoenzyme techniques, it was possible to identify and describe the distribution of three different zymodemes in one defined geographical area in Cas- tro Alves, Brazil (3). The dynamics and spread of one reduviid vector, Triatoma infestans, from south Brazil into the north along the pilgrimage and migration routes through the interior has been documented (2). The potential for application of new geographi- cal approaches to assist management and control of Chagas disease is now accentuated by the increase in transmission in urban and periurban areas and the extensive transmission through blood products (21). Although the quality of data collection is improving with the commitment of the Southern Cone countries towards eradication of Chagas disease, accurate defi- nition of the geographical distribution and preva- lence are still lacking. Leishmaniasis The complex of diseases caused by Leishmania affect more than 12 million persons in 88 countries. Cutaneous leishmaniasis due to Leishmania major affects more than 500 000 persons in 35 countries, principally in the Eastern Mediterranean region. In developing a geographical perspective of these diseases, it is noteworthy that there is no person-to- person transmission of leishmaniasis in the Western Hemisphere. On the other hand, Leishmania major which causes cutaneous leishmaniasis in the Eastern Mediterranean region involves rodent reservoirs (zoonotic), while L. donovani causing kala-azar in Bangladesh, India, Nepal, and East Africa and L. tropica causing cutaneous leishmaniasis in urban and periurban areas are transmitted from person to person by Phlebotomus (sandfly) species (33). Natural population fluctuations of rodents are directly linked to the epidemics of cutaneous leish- maniasis. Migrations of nonimmune persons into endemic areas are associated with epidemics of cuta- neous leishmaniasis or visceral leishmaniasis (kala- azar) spreading from person to person. Four exam- ples in different geographical areas are presented which give insights into how management and moni- toring of control require a geographical approach. Since 1983 in Tunisia more than 36 000 persons have been diagnosed to have cutaneous leishmaniasis which is now recognized as a national health priority (5). The majority (65%) of these cases have been reported within the governorate of Sidi Bouzid in northern Tunisia and many of those infected come from the urban and periurban area of the capital of the governorate. This urban epidemic is linked to the increasing populations of the sand rat (Psammomys obeus), the main mammalian animal reservoir, and the distribution of Atriplex sp., a chenopode which is the sole food source for Psammomys. The other important animal reservoir is the gerbil (Meriones shawi), whose habitat in this and other endemic areas does not overlap with Psammomys. The control of cutaneous leishmaniasis in Tuni- sia demands a sound geographical approach. The actual and potential habitats of Psammomys are strictly defined by the distribution of the chenopode, which is mechanically destroyed and replaced by planting Acacia cyanophilla. Meriones populations are limited by wheat/oil baits soaked in 3% zinc phosphate. Thus, the geographical extent of animal habitats will define both the zone of risk of human infection and the zone of intervention for control. In Ethiopia, long-standing foci of cutaneous leishmaniasis (locally known as "bolho") due to L. aethiopica are found in the basalt rock hills at an altitude of 1500-2700 metres above sea level. Annu- al rainfall averages 400-600 mm. Each family lives in a self-contained compound with dense banana and coffee plantations. Most of the houses are grass- thatched huts. After careful mapping and census of the relatively isolated village of Ocholo, surrounded by steep cliffs, a house-to-house survey was com- pleted (20). The prevalence of active lesions was 3.9%, and 34.3% of the 3200 persons living in 900 houses had scarring due to leishmaniasis. The high level of transmission was explained by the particu- larly close proximity between the houses, the caves where abundant colonies of hyrax (Procavia sp.), the proven animal reservoir, live together with the prov- en sandfly vector (Phlebotomus pedifer). Within this village, the disease was heterogenous: localized cuta- neous, diffuse cutaneous and mucocutaneous leish- maniasis were observed. In this situation, not unlike most, it is necessary to define the dispersal range of WHO Bulletin OMS. Vol 73 1995250 Geographical approaches to control of parasitic zoonoses the sandfly, i.e., the distance at which the mean num- ber of infected females is maximum, the population dynamics of the mammalian host, the distribution of the sandfly habitats (caves, burrows), and the loca- tion of houses with infected persons to understand the spatial correlates which are useful for interven- tion. In Sudan, visceral leishmaniasis is apparently directly linked to the traditional livestock grazing patterns. Transmission of visceral leishmaniasis occurs seasonally in the acacia forests of the south bordering the rivers. The cattle are brought there in the dry season from the central and western parts of the country. The recent epidemics have been linked to refugee movements due to civil unrest. Utilization of GIS has been proposed as part of the planning and implementation of control in this area.e It is note- worthy that in this area of the Rift Valley, extensive geo-referenced data as well as remote sensing images have been used to analyse earthquake patterns (12). The initial geographical analysis has indicated that the only feasible control approach is the use of individual protection among the migrants. Trials with impregnated bednets are postponed due to war. Visceral leishmaniasis had not been reported from the state of Rio de Janeiro until 1977; by 1990, 59 cases had been reported (17). The peak incidence occurred in 1981. The high-risk areas were defined by simple mapping of cases. The majority of cases were in proximity to the Gericinio and Pedra Branca massifs. To test retrospectively the hypothesis that deforestation was a major factor in the changing epi- demiology, the remote sensing data (Landstat MSS) from the Rio de Janeiro Municipal Secretariat for General Planning and Coordinations were reviewed in an attempt to understand the rates of deforestation in that state. It is striking that in areas over 100 m altitude, the rate of deforestation was over 10% between 1972 and 1978, in spite of extensive legisla- tion against deforestation and promotion of conser- vation of forest cover at above 100 metres. Further- more, within the state, several long-distance electric powerlines were constructed with accompanying deforestation linking the endemic areas of kala-azar. The geographical analyses have not provided causal links between environmental and epidemi- ological observations. On the other hand, they have demonstrated the effectiveness of standard control measures which focused on geographical principals, e Connor SJ, Thomson MC. Towards an understanding of the 'Eco-epidemiology' underlying the current leishmaniasis epidem- ic in southern Sudan: the role of remote sensing and geographi- cal information systems in the decision support. Liverpool School of Tropical Medicine, Consultancy report to MSF-Holland. March 1994. e.g., canine serological surveys, elimination of infected and seropositive dogs, and spraying of all households of infected persons and a 100-metre radius around the households. Currently, the rates of seropositive dogs has decreased significantly (from 4.3-12.7% to 0.4-1.8%) in different areas. With the rapid urbanization in most countries endemic for leishmaniasis, GIS offers a new tool for control. It appears that national authorities are increasingly aware of the usefulness of GIS for urban planning and monitoring growth. Monitoring of disease patterns and intervention can be enhanced by integration of health/disease variables into the urban GIS systems. Schistosomiasis Schistosomiasis affects more than 200 million per- sons in a population of 600 million in 74 developing countries. The epidemiology of schistosomiasis is rapidly changing and new geographical approaches are required to promote prevention and control. Four situations where schistosomiasis transmission is increasingly linked with infection in animals have now been documented: urban/periurban, irrigation systems, delta areas related to dam construction, and civil strife/war (35). Aside from the role of people as the major reser- voirs of infection, future epidemiological studies and control programmes should give attention to the role of other mammals in maintaining transmission. No less than 38 mammals have been found to be natural- ly infected by S. mansoni (26). A review of the situa- tion in Brazil showed that at least 15 species of mammals, mostly rodents, were naturally infected (23). Most importantly, it was noted that the reports of mammalian infection are increasing and it appeared that in urban and periurban areas the epi- demiology of human infection may be linked to that of these rodents in the northeast of Brazil. Since 1962, natural S. mansoni infections have been found in cattle (Bos taurus) in Brazil, Sudan and Venezuela (6). However, these anecdotal reports have never been followed up with detailed studies in endemic areas with dense cattle populations. Al- though experimental infections with S. haematobium have been achieved in large primates, and most recently in mice (14), there is no published evidence of natural infection in other mammals. Likewise, over 30 mammals have been found infected with S. japonicum and recent evidence suggests that the epidemiology of human infection is linked to the dis- tribution and prevalence among coinciding animal populations. The changing trend towards urbanization of schistosomiasis in the northeast of Brazil and in WHO Bulletin OMS. Vol 73 1995 251 K.E. Mott et al. Africa has been reviewed (4, 21). Further reports from Guider in Cameroon (31), Bata in Equatorial Guinea (27), and Brazzaville in Congo (1) confirm that this phenomenon is now widespread and involves all species of human Schistosoma. In three areas of the world a change in the distri- bution of the type of schistosomiasis has occurred in the past 20 years: in the Nile delta of lower Egypt, in the Volta delta of Ghana, and in the deltas of the rivers of southwest Cameroon. In all these areas the predominant type of schistosomiasis has shifted from S. haematobium to S. mansoni, and possibly in the Cameroon also the spread of S. intercalatum. All these areas are downstream from major dams which have altered the ecology of their respective deltas. The most recent dramatic example is in Senegal. The first dam on the Senegal river was constructed in Taouey at Richard Toll in 1948. This retarded the seasonal flood from backflowing into Lake Guiers. It permitted irrigation of 6000 hectares for rice cultiva- tion and was subsequently changed to sugar cane production. Subsequently, the Diama dam, a gate structure dam with an embankment dam on the Sene- gal River at St Louis, was completed in August 1986, blocking salt water intrusion during the dry season and creating a reservoir for irrigation and industrial installations. Surveillance through the laboratory of the health centre in Richard Toll began in May 1987. The first cases of S. mansoni infection were reported 18 months after the dam became oper- ational in early 1988. In the last trimester of 1989, 71.5% of 2086 passive stool examinations were posi- tive (30). The role of increased rodent populations (Arvicanthus niloticus, Mastomys huberti and M. erythroleucus), with focally high levels of infection in this area, may have an important role in changing the pattern of the transmission (10). Strife and unrest have caused the movement of people to an extent never before imagined. The situa- tion is so dynamic and changing that the effect on the introduction, spread, or aggravation of schistoso- miasis is difficult to monitor. In Somalia, S. mansoni is now endemic in the areas where Ethiopian refu- gees have been present. In Laos and Cambodia, the impact of refugee movements on transmission must still be determined, but S. mekongi has been diag- nosed frequently in some Thailand refugee centres. The potential for predictive modelling of schis- tosomiasis is an attractive application of GIS. By layering different maps of soil type, soil humidity and hydrology, Nihei et al. (22) successfully predic- ted the habitats of Oncomelania in Japan. Sub- sequently, remote sensing environmental data were used to classify potential levels of transmission of schistosomiasis in the Philippines (7, 8), but this study was not supported by ground proofing. Foodborne trematode infections Foodbome trematode infections offer new and exci- ting opportunities for intersectoral collaboration through use of new geographical approaches. Global- ly, more than 40 million persons are infected with some type of foodbome trematode; the extent of the problem is indicated in Table 1. These diseases represent a new global challenge for food safety and parasitic disease control programmes. The WHO database is integrated into a GIS using MapInfo (from MapInfo Corporation) with, for the present, first-order (NUTS 1) boundary files for countries (18). As the national programmes develop it is anticipated to have up to fourth-order boundaries at the smallest administrative unit. Aquaculture is rapidly expanding and all esti- mates suggest that this expansion will continue and accelerate into the next century. GIS applications in aquaculture are extensive (19). As yet there is no linkage with the public health problems related to wastewater use and risk of infectious and parasitic diseases. On the other hand, the economic imperative of freshwater fish production will increase internal and foreign markets. It is foreseen that the Micro- computer-based Commodity Analysis and Informa- tion System (MICAS) developed by UNCTADf to f Special UNCTAD Unit of Commodity Analysis and Information System. Commodity Division, UNCTAD, Palais des Nations, 1211 Geneva 10, Switzerland. Table 1: Global distribution of foodborne trematode Infections No. of Population Persons Disease/food item Parasite countries at risk (x106) infected (x106) Clonorchiasis/freshwater fish Clonorchis sinensis 9 290 7 Opisthorchiasis/freshwater fish Opisthorchis felineus 2 14 1.5 Opisthorchis viverrini 2 50 9 Paragonimiasis/freshwater crab Paragonimus species 39 195 21 Fascioliasis/watercress Fasciola hepatical 61 180 2.4 F. gigantica 252 WHO Bulletin OMS. Vol 73 1995 Geographical approaches to control of parasitic zoonoses assist investment and diversification will include fish and fish products incorporating FAO databases in a GIS. The epidemiology of human fascioliasis has undergone dramatic revision in the last five years. Its extent is far greater in the Andean altiplano from Bolivia to Ecuador, in the Nile delta, and in the Cas- pian region of the Islamic Republic of Iran than was previously recognized. The usefulness of GIS in inte- grating remote sensing data has been demonstrated by Malone et al. (16) who assessed the epidemiology of animal fascioliasis in Louisiana (USA). As yet, we do not have applications in areas where human disease has been reported. Since the closure of the Aswan High Dam in 1964, there has been a rapid geological evolution of the Nile delta (29). The geological analysis of these changes has incorporated remote sensing data. In view of the increasing prevalence of human fascioli- asis in the Nile delta, a GIS-based analysis seems opportune. Discussion There are many common geographical issues among the control approaches against parasitic diseases reviewed in this paper. These issues may be classi- fied into two major types: those inherent to the disease (epidemiological, environmental or biomedi- cal), and those relating to control activities (human resources, organization, coordination, etc.). Schistosomiasis, trypanosomiasis and other para- sitic diseases occur in relation to common environ- mental denominators. Understanding of the first group of issues is necessary to plan for control. A single control action, based on appropriate epidemi- ological, environmental and biomedical interpreta- tion, could considerably reduce transmission of two diseases. Moreover, additional actions could totally prevent or even eliminate other diseases under cer- tain conditions. The distribution of health services is of common interest to all control programmes. Control activities will eventually, if not from the outset, be part of the local health services. Thus the spatial relations between disease distribution and health services should be an essential part of any GIS for control of tropical diseases. Control action, mostly managerial, involves mainly the second set of issues. It is pos- sible to pinpoint where changes in staffing and infra- structure can be useful to achieve better delivery of control (diagnosis and treatment, environmental management) and maintenance of control. The focal nature of the distribution, although not unique to parasitic diseases, is a characteristic of all diseases in this review. Thus the extent to which a geographical approach can benefit implementation, monitoring and maintenance of control operations will depend on the level of discrimination of the data collected and their reliability. Ultimately a control operation must reliably determine if it can be stopped without risk of recurrence or reintroduction. For the first time GIS affords a practical way of determining overlap of "focality" and of making this information readily available for programme plan- ning and execution, both centrally and at the periph- ery of endemic countries. Aside from the common issues, new problems will become evident as GIS attracts more attention. Today single disease databases are rare; where they exist they are incomplete. While an enormous amount of information has been and is being collect- ed in the peripheral health services, it is usually not compiled in a format suitable for common databases. Usually the original data are analysed at a peripher- al health unit and synthesized into a report where they become so aggregated that they are difficult to interpret or integrate into a database. Furthermore, in the process of aggregation, any geo-referenced data will become obscured. This methodology is not due to the technical impossibility to obtain the required information or to develop a proper database, but due to the lack of current management guidelines of epi- demiological reporting of tropical diseases in health care systems of endemic countries. Exceptionally, only certain diseases may be specifically reported. In general, the broad classifications of "febrile illness" or "parasitic disease" may mask the actual cause. Many tropical diseases coexist in the same geo- graphical area. Yet the creation of common disease databases continues to be a chimera. From a distance or intuitively such common databases seem feasible, but experience dictates otherwise. For example, the separate organizational and administrative structure in the Ministries of Health of endemic countries limit any initiative to create common databases, let alone to share scarce resources. Before databases are input into GIS, the final output objectives should be discussed and clearly defined. In the early stage of implementation it is not uncommon to create databases without forethought as to the expected outcome. GIS software is capable of storage of numerous and extensive databases. Thus, a GIS can potentially calculate an enormous number of variables which might not all be relevant to planning or monitoring of tropical diseases. GIS does not eliminate misinterpretation of the data, so sound epidemiological understanding and mapping for comprehension are prerequisites. A map prepared with inappropriate variables (i.e., using absolute WHO Bulletin OMS. Vol 73 1995 253 K.E. Mott et al. Table 2: Geographical databases used in the control and monitoring of some parasitic dis- eases African Chagas Foodborne trypanosomiasis disease Leishmaniasis Schistosomiasis trematodes Demography +++a ... +++ Geology + + + + + Hydrology ++ +++ ... Temperature + + + + + Altitude + + ++ + + Health services ... +++ ... ... ... Schools + Roads ... ... ... ... ... Crops/agriculture +++ ++ Dams + Vectors: Tsetse ++ Reduviid ++ Sandfly ++ Snails ++ ++ Hosts: Wild mammals + ++ ++ + Domestic animals ++ + Eating raw freshwater + fish/aquaculture Eating raw freshwater + crabs Watercress (Fasciola only) + a + = useful, if available; ++ = needed; +++ = essential. values without denominators instead of prevalence), or which attempts to display too many or unselected variables, will be detrimental to communicating the intended information. At this stage in time, GIS presents an unforeseen difficulty which may be transitory, but is neverthe- less very real-the need for reliable map boundary files. We believe that electronic map boundary files are available from national or international sources. The problem is where to find them. Map boundary files which have been approved for national govern- ment programmes are usually available through the government agency concerned with maintaining the records of national territory. In principle, we believe that any statement on the lack of map boundary files reflects a limitation of the search rather the absence of such files. Thus, it is not necessary to digitize stat- ic maps de novo except for specific local programme requirements. Various WHO programmes are now exploring GIS applications and the first hurdle, if not obstacle, is the acquisition of boundary files. An inventory of public domain boundary files, which can be easily accessed and used in specific GIS soft- ware, has been prepared which will be revised peri- odically.9 In spite of these current limitations and "growing pains", GIS offers new and attractive opportunities. While the potential uses of GIS for predictive modelling are attractive research objectives, the con- trol of tropical diseases requires simple and robust applications to support current operations. The com- plex epidemiology of parasitic infections involving environment, animals and people must therefore be reduced to simplified data collection and analysis which can be carried out in the endemic areas. The new geographical approaches provide a focus for different sectors to converge as equal part- ners-a good example of intersectoral cooperation. However, the public health sector is both chronologi- cally and technically behind other sectors in the functional application of GIS. Rapid progress in GIS software development and applications in the other sectors can certainly benefit the integration of health data into the ongoing national GIS. 9 Nuttall I. Inventory of administrative boundary files for compu- terized geographical applications. Unpublished document, WHO/ SCHISTO/94.109, 1994. 254 WHO Bulletin OMS. Vol 73 1995 Geographical approaches to control of parasitic zoonoses All development sectors are confronted with decentralization and the question of sustainable development. The global trend of governments and the public as well as private sectors is towards the periphery, whether it be in marketing or health care delivery. This changes the strategy for control of tropical diseases which has traditionally been carried out in campaign fashion under the aegis of national quasi-military public health organizations. The costs of logistics and the limitation of personnel have been enormous-and are unsustainable. As endemic coun- tries struggle to implement control without foreign aid or interference, the existing health systems and staff are being designated to undertake control meas- ures. GIS offers a new and practical methodology to define the spatial relationships between the health care system, disease distribution, and the factors influencing the distribution. Moreover, as its analyti- cal potential is exploited, GIS can provide allocation options for decreasing health budgets and reducing the number of staff, which might not be possible in any other way. At present, most estimates of the distribution, prevalence and incidence of parasitic diseases are based on assumptions of the homogeneity of distri- bution in large administrative units. The current epi- demiological methodology is unable to utilize or synthesize the discriminative data from small geo- graphical units (villages or individual households) in a statistically acceptable manner, which can be done by GIS. If extensively implemented to promote plan- ning and monitoring of control of tropical diseases, GIS could contribute to a complete revision of our current understanding of the distribution and preva- lence of tropical diseases. Acknowledgements The Schistosomiasis Control unit in the WHO Division of Control of Tropical Diseases gratefully acknowledges finan- cial support from the Edna McConnell Clark Foundation; German Pharma Health Fund; International Development Research Centre of Canada; Ministry of Foreign Affairs, Direzione Generale per la Cooperazione allo Sviluppo, Italy; and the U.S. Agency for International Development. Resume Nouvelles approches g6ographiques de la lutte contre certaines zoonoses parasitaires Les approches geographiques, en particulier les systemes d'information geographique (GIS) cons- tituent une base de dialogue entre differents sec- teurs, rassemblant des zoologistes, des vet6ri- naires et des agents de sante publique. La recon- naissance de plus en plus grande du role des animaux dans l'epidemiologie des maladies humaines entre dans le cadre du phenomene des infections "emergentes". Outil analytique relati- vement nouveau dans le domaine de l'epidemio- logie, le GIS consiste en une technologie infor- matique de saisie, de stockage, d'analyse et de visualisation de donnees spatiales. L'integration d'une nouvelle technologie comme le GIS se fait d'une fagon methodique, etape par 6tape. L'utilisation appropriee des nouvelles appro- ches geographiques reposera sur: 1) l'int6ret prou- ve de l'utilisation des anciennes approches g6o- graphiques dans la lutte contre la maladie; 2) une priorit6 clairement etablie pour la sequence d'in- troduction des bases de donnees; et 3) des hypo- theses scientifiquement fondees. L'historique de la lutte contre la trypanosomia- se africaine constitue un remarquable modele de l'utilisation des approches geographiques pour relier les objectifs et les operations de sante pu- blique humaine et v6terinaire. La transmission de la maladie du sommeil est focale, meme a l'inte- rieur d'une zone d'endemie, de sorte que la trans- mission et le risque d'infection ne sont pas homo- genes sur le plan geographique. Le potentiel d'application des nouvelles approches geogra- phiques a la prise en charge de la maladie de Chagas et aux activites de lutte est actuellement renforce par I'augmentation de la transmission dans les zones urbaines et p6riurbaines et par la transmission de plus en plus fr6quente par le biais de produits sanguins. Etant donne l'urbanisation rapide dans la plu- part des pays ou la leishmaniose est end6mique, le GIS constitue un nouvel outil de lutte. Les analyses g6ographiques n'ont pas demontre de lien causal entre les observations environnementales et epide- miologiques sur la leishmaniose. En revanche, 1'effi- cacite de mesures de lutte standard axees sur des considerations geographiques a ete demontr6e. La possibilite d'une modelisation pr6dictive de la schistosomiase constitue une des applications interessantes du GIS. En superposant differentes cartes representant les types de sol, I'humidite du sol et I'hydrologie, il a ete possible de predire les habitats du gast6ropode h6te intermediaire au Japon. Des donnees environnementales obtenues par t6led6tection ont ete utilis6es pour classer les niveaux potentiels de transmission de la schisto- somiase aux Philippines. Les infestations a trema- todes transmises par les aliments offrent de nou- velles occasions interessantes de collaboration intersectorielle grace aux nouvelles approches geographiques. WHO Bulletin OMS. Vol 73 1995 255 K.E. Mott et al. Les aspects g6ographiques communs aux approches de lutte utilisees contre les maladies parasitaires peuvent etre class6s en deux grands types: les questions inherentes a la maladie (epidemiologiques, environnementales ou biome- dicales), et les questions relatives aux activit6s de lutte (ressources humaines, organisation, coordi- nation etc.). Outre ces questions, de nouveaux problemes apparaitront a mesure de l'utilisation du GIS. Actuellement, les bases de donnees s6parees sur les maladies sont rares, et lors- qu'elles existent elles sont incompletes. De nom- breuses maladies tropicales coexistent dans la meme r6gion g6ographique, mais la creation de bases de donn6es communes sur les maladies se heurte a la s6paration des structures organisation- nelles et administratives au sein des ministeres de la sant6 des pays d'endemie. Malgr6 ces limitations actuelles et ces "dou- leurs de croissance", le GIS offre des perspec- tives int6ressantes. Mais tandis que les utilisations potentielles du GIS pour la mod6lisation predic- tive constituent des objectifs de recherche int6res- sants, la lutte contre les maladies tropicales exige des applications simples et robustes a l'appui des op6rations actuelles. Les nouvelles approches g6ographiques constituent le point de convergen- ce des diff6rents secteurs en tant que partenaires 6gaux. La methodologie epidemiologique actuelle n'est pas capable d'utiliser ou de synth6tiser les donnees distinctes provenant de petites unit6s geographiques (villages ou menages individuels) d'une faqon statistiquement acceptable, comme peut le faire le GIS. S'il est mis en ceuvre a gran- de echelle pour promouvoir la planification et la surveillance de la lutte contre les maladies tropi- cales, le GIS pourrait contribuer a une r6vision complete de nos connaissances actuelles sur la r6partition et la prevalence des maladies tropi- cales. References 1. Akoulala JJ et al. Le foyer de schistosomiase uri- naire de Brazzaville: epidemiologie et geographie des sites urbains de transmission. Bulletin de liaison et documentation OCEAC, juillet-decembre 1989, No. 89-90: 11-14. 2. Barrett TV et al. An outbreak of acute Chagas' dis- ease in the Sao Francisco Valley region of Bahia, Brazil: triatomine vectors and animal reservoirs of Trypanosoma cruzi. Transactions of the Royal Soci- ety of Tropical Medicine and Hygiene, 1979, 73: 703-709. 3. Barrett TV et al. Epidemiological aspects of three Trypanosoma cruzi zymodemes in Bahia State, Bra- zil. Transactions of the Royal Society of Tropical Medicine and Hygiene, 1980, 74: 84-90. 4. Barreto ML. Geographical and socioeconomic fac- tors relating to the distribution of Schistosoma man- soni infection in an urban area of northeast Brazil. Bulletin of the World Health Organization, 1991, 69: 93-102. 5. Ben-Ismail R, Ben-Rachid MS. Epidemiologie des leishmanioses en Tunisie. In: Gentilini M, Veins P, eds. Maladies tropicales transmissibles. Paris, Lib- bey, 1989: 73-80. 6. Coelho PMZ, Lima W, Nogueira RHG. Schistoso- ma mansoni: on the possibility of Indian buffalo (Bubalus babalis) being infected experimentally. Revista da Sociedade Brasileira de Medicina Tropi- cal, 1989, 22: 157-158. 7. Cross ER, Bailey RC. Prediction of areas endemic for schistosomiasis through the use of discriminate analysis of environmental data. Military medicine, 1984, 149: 28-30. 8. Cross ER et al. Predicting areas endemic for schis- tosomiasis using weather variables and a LANDSTAT database. Military medicine, 1984, 149: 542-544. 9. Doumenge JP et al. Atlas of the global distribution of schistosomiasis. Bordeaux, Presses universi- taires, 1987 (for the World Health Organization and Centre d'Etudes de Geographie tropicale). 10. Duplantier J et al. (1991) Rongeurs et degats des rongeurs. Dakar, ORSTOM Programme "Eau, Sante et Developpement" Annual Report for 1990. 11. Estes JE, Mooneyhan DW. The mythical map. Pro- ceedings of Pecora XII Symposium, Sioux Falls, SD, USA. United States Geological Service (in press). 12. Girdler RW, McConnell DA. The 1990 to 1991 Sudan earthquake sequence and the extent of the East African rift system. Science, 1994, 264: 67-70. 13. Hugh-Jones M. Application of remote sensing to the identification of the habitats of parasites and dis- ease vectors. Parasitology today, 1989, 5: 244-251. 14. Imbert-Establet D et al. The mouse as a suitable host for an isolate of Schistosoma haematobium from Niger. Journal of helminthology, 1992, 66: 1-5. 15. Lederberg J, Shope RE, Oaks SC. Emerging infec- tions: microbial threats to health in the United States. Institute of Medicine, National Academy Press, Washington DC, 1992. 16. Malone J et al. Use of LANDSTAT MSS imagery and soil type in geographical information system to assess site-specific risk of fascioliasis on Red River Basin Farms in Louisiana. Annals of the New York Academy of Science, 1992, No. 653 (Tropical veteri- nary medicine: current issues and perspectives): 389-397. 17. Marzochi MCA, Marzochi KBF, Carvalho R. Vis- ceral leishmaniasis in Rio de Janeiro. Parasitology today, 1994, 10: 37-40. 18. Maurice J. Is something lurking in your liver? New Scientist, 19 March 1994: 26-31. 19. Meaden GJ, Kapetsky JM. Geographical information systems and remote sensing in inland fisheries and aquaculture. Rome, Food and Agriculture Organiza- tion, 1991. (FAO Fisheries Technical Paper, No. 318). 20. Mengistu G et al. Cutaneous leishmaniasis in south-western Ethiopia: Ocholo revisited. Transac- 256 WHO Bulletin OMS. Vol 73 1995 Geographical approaches to control of parasitic zoonoses tions of the Royal Society of Tropical Medicine and Hygiene, 1992, 86: 149-153. 21. Mott KE et al. Parasitic diseases and urban devel- opment. Bulletin of the World Health Organization, 1990, 68: 691-698. 22. Nihei N, Asami S, Tanaka H. Geographical factors influencing the population numbers and distribution of Oncomelania nosophora and the subsequent effect on the control of schistosomiasis japonica in Japan. Social science and medicine, 1981, 15D: 149-157. 23. Rey L. Non-human vertebrate hosts of Schistosoma mansoni and schistosomiasis transmission in Brazil. Research and reviews in parasitology, 1992, 53: 13-25. 24. Richards FO Jr. Use of geographic information systems in control programs for onchocerciasis in Guatemala. Bulletin of the Pan American Health Organization, 1993, 27: 52-55. 25. Rogers DJ, Randolph SE. Mortality rates and pop- ulation density of tsetse flies correlated with satellite imagery. Nature, 1991, 351: 739-741. 26. Rollinson D, Southgate VR. The genus Schistoso- ma: a taxonomic appraisal. In: Rollinson D, Simpson AJG, eds., The biology of schistosomes: from genes to latrines. London, Academic Press, 1987: 1-49. 27. Simarro PP, Sima FO, Mir M. Epidemiologie de la schistosomiase en Guinee equatoriale. 1. Etude de Schistosoma intercalatum dans la ville de Bata. Bul- letin de liaison et documentation OCEAC, juillet- decembre 1989, No. 89-90: 5-7. 28. Simarro PP et al. La lutte contre la trypanosomiase humaine africaine dans le foyer de Luba en Guinde equatoriale: bilan de trois methodes. Bulletin of the World Health Organization, 1991, 69: 451-457. 29. Stanley DJ, Warne AG. Nile delta: recent geologi- cal evolution and human impact. Science, 1993, 260: 628-634. 30. Talla I et al. Outbreak of intestinal schistosomiasis in the Senegal River Basin. Annales de la Soci6t6 belge de M6dicine tropicale, 1990, 70: 173-180. 31. Toupoury A, Amadou A. Enquete sur la bilharziose urinaire dans l'ecole publique de Sanguere a Guider (Departement du Mayo-Louti). Bulletin de liaison et documentation OCEAC, juillet-d6cembre 1989, No. 89-90: 9-10. 32. Epidemiology and control of African trypanosomia- sis. Report of a WHO Expert Committee. Geneva, World Health Organization, 1986 (WHO Technical Report Series, No. 739). 33. Control of the leishmaniases. Report of a WHO Expert Committee. Geneva, World Health Organiza- tion, 1990, (WHO Technical Report Series, No. 790). 34. The control of Chagas diseases. Report of a WHO Expert Committee. Geneva, World Health Organiza- tion, 1991 (WHO Technical Report Series, No. 81 1). 35. The control of schistosomiasis. Second report of the WHO Expert Committee. Geneva, World Health Organization, 1993 (WHO Technical Report Series, No. 830). WHO Bulletin OMS. Vol 73 1995 257