vectorborne diseases in west africa: geographic ... · at disease control, research and advocacy.16...
TRANSCRIPT
Vectorborne diseases in West Africa geographic distribution andgeospatial characteristics
Pavel Ratmanov Oleg Mediannikov and Didier Raoult
Aix Marseille Universite URMITE UMR CNRS 7278 IRD 198 INSERM 1095 27 Boulevard Jean Moulin 13385 Marseille cedex 05 France
Corresponding author Tel +33 4 91 32 43 75 Fax +33 4 91 83 03 90 E-mail didierraoultgmailcom
Received 2 August 2012 revised 8 November 2012 accepted 17 December 2012
This paper provides an overview of the methods in which geographic information systems (GIS) and remotesensing (RS) technology have been used to visualise and analyse data related to vectorborne diseases (VBD)in West Africa and to discuss the potential for these approaches to be routinely included in future studies ofVBDs GISRS studies of diseases that are associated with a specific geographic landscape were reviewed in-cluding malaria human African trypanosomiasis leishmaniasis lymphatic filariasis Loa loa filariasis onchocer-ciasis Rift Valley fever dengue yellow fever borreliosis rickettsioses Buruli ulcer and Q fever RS data andpowerful spatial modelling methods improve our understanding of how environmental factors affect thevectors and transmission of VBDs There is great potential for the use of GISRS technologies in the surveillanceprevention and control of vectorborne and other infectious diseases in West Africa
Keywords Epidemiology Vectorborne diseases Spatial analysis Geographic information systems Remote sensing West Africa
IntroductionVectorborne diseases (VBD) place a terrible and unacceptablepublic health burden on developing countries Specifically 7 ofthe 10 diseases targeted by the WHO Special Programme for Re-search and Training in Tropical Diseases and 7 of the 17 diseasesclassified as neglected tropical diseases (NTD) by WHO aretransmitted by arthropods12 VBDs play a particularly importantrole in West Africa because many of them are endemic to theregion and the burden of VBDs continues to be very heavy
This review addresses a broad range of VBDs (Table 1) andvectors (Table 2) that are prevalent in West Africa (BeninBurkina Faso Cape Verde Cote drsquoIvoire The Gambia GhanaGuinea Guinea-Bissau Liberia Mali Niger Nigeria SenegalSierra Leone Togo) The data for this review were collected bysearching the National Center for Biotechnology Information(NCBI) PubMed database and the reference lists of relevant arti-cles The criteria for inclusion in the study were relaxed for thereferences related to geographic information systems (GIS) andremote sensing (RS) applications used to study VBDs in WestAfrica (see the review of selected studies in SupplementaryTable 1)
GIS and RS technology have opened new avenues for evaluat-ing digital map data generated by earth-observing satellitesensors and for conducting analyses of spatial and temporalenvironments3 It is well known that the distribution of tropicalVBDs is particularly sensitive to climatic and environmentalfactors because of the vulnerability of vectors intermediate
hosts and free-living stages The concept of spatial focality ofVBDs has been observed and discussed for many years4
Numerous reviews have broadly addressed the use of GISRStechnologies and spatial and space-time modelling approachesin the field of VBDs56 However the potential for such technolo-gies and methodologies to be used for the prevention surveil-lance and control of tropical VBDs is a critically important issuethat has not yet received the attention it deserves Adaptingmapping and modelling techniques for resource-constraineddisease-endemic environments must play a role in the next fron-tier of research on VBDs7
Our team specifically studies VBDs in West Africa89 and theaims of this review were to study the geographical distributionof VBDs in West Africa to provide an overview of the methodsin which mapping and spatial and space-time modellingapproaches have been used to visualise and analyse vectorand epidemiological data and to discuss the potential forthese approaches to be further developed in future studies
Geographic distribution of vectorbornediseases in West Africa
Malaria
Most of the papers on VBDs in West Africa focus on malaria (1540 Supplementary Table 1) Malaria is an infectious diseasecaused by parasites of the genus Plasmodium transmitted tohumans through the bites of infected female mosquitoes ofthe genus Anopheles Malaria is a major public health problem
Royal Society of Tropical Medicine and Hygiene 2013 All rights reserved For Permissions please email journalspermissionsoupcom
REV
IEW
AR
TIC
LE
Trans R Soc Trop Med Hygdoi101093trstmhtrt020
1 of 12
Transactions of the Royal Society of Tropical Medicine and Hygiene Advance Access published March 10 2013 at U
niversity of Liverpool on M
ay 9 2013httptrstm
hoxfordjournalsorgD
ownloaded from
with more than 200 million cases and causing up to 1 milliondeaths each year People in endemic areas with symptomaticand asymptomatic malaria are reservoirs for the infectionMalaria is endemic in all countries in West Africa but theburden it imposes depends upon the scale of malaria controlefforts in the country (Figure 1)10
Many efforts have been made to collect and centralise existingentomological parasitological and epidemiological data in AfricaNevertheless a high degree of uncertainty still exists regarding theannual number of malaria cases and their geographic distribu-tion11 A summary of the environmental satellite data that areavailable for studying malaria can be found in the review ofMachault et al12 The Mapping Malaria Risk in Africa (MARAARMA) project was established for developing malaria risk mapson the scale of the entire continent13 More recently the Malaria
Atlas Project (MAP) was born with the objective of gathering world-wide parasite prevalence data and making them freely availableon the internet (httpwwwmapoxacuk)14
Human African trypanosomiasis
The aetiological agent of human African trypanosomiasis (HAT)also known as sleeping sickness is protozoa of the species Trypa-nosoma brucei gambiense that are transmitted by tsetse flies(Glossina spp) In West Africa humans are the main reservoir forTb gambiense Animals play a less important role but pigs andsome wild animal species have been reported as reservoirs15 InWest Africa HAT has been reported in Guinea Cote drsquoIvoire andNigeria and is also endemic to several other countries (Figure 2)Control of sleeping sickness has always been closely related to
Table 1 Selected vectorborne diseases in West Africa
Disease Pathogen Reservoir Vector
ProtozoalMalaria Plasmodium falciparum P vivax
P ovale P malariaeHumans Mosquitoes (Anopheles gambiae sl
An funestus)Human Africantrypanosomiasis
Trypanosoma brucei gambiense Humans some wild anddomestic mammals
Tsetse flies (Glossina palpalis slG tachinoides)
Leishmaniasis Leishmania genus Mammals Phlebotomus (Phlebotomus spp)Sergentomyia (Spelaeomyia) darlingi
HelminthicLymphatic filariasis Wuchereria bancrofti Humans Mosquitoes (Anopheles spp)Loa loa filariasis (loiasis) Loa loa Humans wild mammals Tabanid flies (Chrysops spp)Onchocerciasis (lsquoriverblindnessrsquo)
Onchocerca volvulus Humans Black flies (Simulium genus)
ViralRift Valley fever Rift Valley fever virus Wild and domestic
mammalsMosquitoes (Aedes spp Culex spp)
Dengue Dengue virus Humans Mosquitoes (Aedes spp)Chikungunya fever Chikungunya virus Primates humans Mosquitoes (Aedes spp)CrimeanndashCongohaemorrhagic fever
CrimeanndashCongo haemorrhagicfever virus
Domestic and wild animals Ticks (Hyalomma genus Amblyommavariegatum)
Yellow fever Yellow fever virus Primates humans Mosquitoes (Aedes spp)Bacterial
Mediterranean spottedfever
Rickettsia conorii Wild and domesticmammals
Ticks (Rhipicephalus evertsi)
African tick-bite fever Rickettsia africae Wild and domesticmammals
Ticks (Amblyomma variegatumRhipicephalus evertsi)
Tickborne relapsing fever Borrelia crocidurae Rodents Ticks (Ornithodoros sonrai)Rickettsia felis infection Rickettsia felis Unknown Fleas (Ctenocephalides felis and others)Trench fever Bartonella quintana Humans Lice (Pediculus humanus humanus)Louseborne relapsingfever
Borrelia recurrentis Humans Lice (Pediculus humanus humanus)
Epidemic typhus Rickettsia prowazekii Humans Lice (Pediculus humanus humanus)Buruli ulcer Mycobacterium ulcerans UnknownQ fever Coxiella burnetii Mammals birds and
arthropodsTicksa
aPossibility of transmission of Q fever by tick was not studied
P Ratmanov et al
2 of 12
at University of L
iverpool on May 9 2013
httptrstmhoxfordjournalsorg
Dow
nloaded from
disease mapping The Atlas of HAT is the most prominent attemptat disease control research and advocacy16
Leishmaniasis
Leishmaniasis refers to a group of VBDs that are caused by morethan 20 species of the protozoan genus Leishmania rangingfrom localized skin ulcers to lethal systemic disease Leishmania-sis is considered one of the lsquomost neglected diseasesrsquo because
limited resources are invested in its diagnosis treatment andcontrol and it is strongly associated with poverty17 Humansare infected via the bite of phlebotomine sandflies (Phlebotomusspp) The leishmaniases can be classified into two epidemio-logical entities according to the type of transmission anthropo-notic if humans are the sole reservoir involved in transmission(and the sole source for vector infection) or zoonotic if atleast one mammalian reservoir is involved18 Although thereare no major geographic foci of leishmaniasis in West Africathe cutaneous forms of leishmaniasis have been reported in 11of 15 countries in the region (Figure 3) In recent studies theseroprevalence of specific antibodies against L infantum (theagent of visceral leishmaniasis) in the human population wasdetermined in Senegal Larger-scale studies with application ofGISRS technologies are now required to define the distributionof L infantum in West Africa and an identification of its potentialvector19
Filariasis
Lymphatic filariasis (LF) is a vectorborne parasitic infectiousdisease caused by Wuchereria bancrofti which is endemic inthe tropics including in sub-Saharan Africa It is transmitted tohumans by infected mosquitoes In West Africa the mosquitoAn gambiae sl is a vector of LF and malaria caused by P falcip-arum Humans are the only reservoir host of the LF parasite inAfrica20 Onchocerciasis (or lsquoriver blindnessrsquo) is a parasiticdisease caused by the filarial parasitic nematode Onchocerca vol-vulus It is transmitted through the bites of infected Simulium(black fly) vectors which breed in fast-flowing streams andrivers21 Control of onchocerciasis in sub-Saharan Africa is over-seen by the African Programme for Onchocerciasis Control(APOC)22 Loa loa filariasis (loiasis) is a NTD caused by the filarialparasite Loa loa It is an African disease restricted to the equator-ial rainforest regions of Central and West Africa23 The insectvectors of L loa are flies of the genus Chrysops Humans arethe primary reservoir for L loa LF loiasis and onchocerciasisare endemic in many countries of West Africa (Figure 4)
Rift Valley fever
Rift Valley fever (RVF) is an arthropodborne viral disease that pri-marily causes epizootics of abortion and high mortality rates indomestic animals but it can also infect humans The RVF virusis mainly transmitted by mosquitoes of the Aedes and Culexgenera to a wide range of animals from rodents to camelswhich are the natural reservoirs for RVF Although mosquitoesmay transmit the RVF virus to humans human infections resultfrom contact with infected animals24 A large RVF epidemic oc-curred in 1987 in southern Mauritania and northern SenegalHowever all of the countries in West Africa are at risk of RVF(Figure 5)25
Dengue and yellow fever
Dengue fever is a viral infectious tropical disease The mosquitoAe aegypti is the primary vector of the dengue virus andhumans are the most common reservoir Dengue is a widespreaddisease in the subtropics and tropics but in Africa the burden ofthe disease is poorly understood7 In West Africa sylvatic
Table 2 Selected vectors of the human vectorborne diseasesin West Africa
Vector Disease
MosquitoesAnopheles spp Lymphatic filariasisa
An gambiae sl MalariaAedes spp Rift Valley fever
DengueYellow feverChikungunya fever
Culex spp Rift Valley feverPhlebotomus
Phlebotomus spp LeishmaniasisSergentomyia (Spelaeomyia)darlingi
Leishmaniasis
Tsetse fliesGlossina palpalis sl Human African
trypanosomiasisG tachinoides Human African
trypanosomiasisTabanid flies
Chrysops spp Loa loa filariasis (loiasis)Black flies
Simulium spp Onchocerciasis (lsquoriverblindnessrsquo)
TicksRhipicephalus evertsi Mediterranean spotted fever
African tick-bite feverAmblyomma variegatum African tick-bite fever
CrimeanndashCongo haemorrhagicfever
Ornithodoros sonrai Tickborne relapsing feverOrnithodoros moubata Tickborne relapsing feverHyalomma spp CrimeanndashCongo haemorrhagic
feverLice
Pediculus humanus humanus Trench feverLouseborne relapsing feverEpidemic typhus
FleasCtenocephalides felis andothers
Rickettsia felis infectionb
aAnopheles is a primary vector of lymphatic filariasis in AfricabFleas are considered as vectors of R felis worldwide
Transactions of the Royal Society of Tropical Medicine and Hygiene
3 of 12
at University of L
iverpool on May 9 2013
httptrstmhoxfordjournalsorg
Dow
nloaded from
circulation of the dengue virus is the predominant form of circu-lation (with lower primates as the main reservoir)26 Evidenceregarding circulation of the dengue virus was obtained from 11countries in the region although the mosquito vectors arepresent throughout West Africa (Figure 6)
Yellow fever (YF) is an acute viral haemorrhagic disease trans-mitted in West Africa by infected Aedes spp mosquitoes Up to50 of severely affected persons who do not receive treatmentdie from YF and there is no cure The YF virus circulates both inurban and sylvatic settings involving several vertebrate speciesIn the sylvatic cycle mosquitoes act as the main vectors andmonkeys act as the primary hosts In the urban cycle the virusis transmitted between human beings and mosquitoes (predom-inately Ae aegypti) In Africa transmission of the virus can alsooccur in an intermediate cycle between human beings or non-human primates and Aedes spp mosquitoes that breed in treeholes on the savannah27 Vertical transmission also occurswithin the mosquito population and may play an importantrole in maintaining the sylvatic cycle28 In West Africa YF isholoendemic in all countries except in the Sahara Desertregions of Mali and Niger (Figure 7)27
CrimeanndashCongo haemorrhagic fever
CrimeanndashCongo haemorrhagic fever is a viral tickborne diseaseThe virus causes severe illness throughout the world includingWest Africa Its distribution closely matches that of its mainarthropod vector ixodid ticks belonging to the genus Hyalomma
Human infection occurs through tick bites contact with infectedlivestock or nosocomial transmission29 In West Africa enzooticcirculation of the CrimeanndashCongo haemorrhagic fever virus hasbeen shown in serological surveys of cattle but only sporadichuman cases have been reported (Figure 8)30
Borrelioses
In West Africa tickborne relapsing fever (TBRF) is caused by Bor-relia crocidurae Studies in Senegal indicate that TBRF is aftermalaria the most common cause of outpatient visits to a ruraldispensary Investigations found that the vector tick (Ornitho-doros sonrai) was present in villages in Senegal Mauritania andMali with high infection rates Rodents and insectivores are thereservoirs of B crocidurae in West Africa31 Unfortunately noattempts have been made to perform a spatial analysis ofTBRF prevalence in West Africa
Rickettsioses
Following malaria tickborne rickettsioses are one of the mostcommon causes of systemic febrile illness among travellersfrom developed countries but little is known about their preva-lence in indigenous populations especially in West Africa32
There is a high incidence (44) of Rickettsia felis in Senegal33
A case of Mediterranean spotted fever was described inSenegal and its agent R conorii was found in the ticks Rhipice-phalus evertsi evertsi8 There is a plethora of serological studies
Figure 1 Cases of malaria per 100 000 inhabitants in West African countries in 2010 or latest available year (data from the WHO retrieved fromhttpwwwwhoint)
P Ratmanov et al
4 of 12
at University of L
iverpool on May 9 2013
httptrstmhoxfordjournalsorg
Dow
nloaded from
of the prevalence of rickettsioses and entomological studies ofticks for Rickettsia spp Although attempts to map the geographicdistribution of rickettsioses in Africa have been made they simplydisplayed infected sites or counties with reported cases as pointswithout conducting any geostatistical analyses834 Thus thespatial aspects of borrelial and rickettsial VBDs in West Africahave not received the attention they deserve
Buruli ulcer
Buruli ulcer (BU) disease caused by Mycobacterium ulcerans isan emerging infectious disease in many tropical and subtropicalcountries The exact mode of transmission of this diseaseremains unclear Although vectors and modes of transmissionremain unknown it has been hypothesized that transmissionof BU disease is associated with human activities in andaround aquatic environments (the role of water bugs has beendiscussed) and that characteristics of the landscape play a rolein the spread of BU disease35 In West Africa Cote drsquoIvoireGhana and Benin are the countries that are particularly affectedby BU disease (Figure 9)
Q fever
In West Africa Q fever has a wide distribution which has beenshown repeatedly in human serological studies and studies ofreservoirs (domestic animals)36 It was reported that the main
vector of TBRF the soft tick O sonrai also harboured Coxiella bur-netii37 The possibility that this tick transmits Q fever was notstudied but some species of tick may play a role in the transmis-sion of Q fever
Geographic information systems and remotesensing for spatial analysis of vectordistributionThe geographic distribution of VBDs depends on the distributionof their vectors environmental factors and climatic factorsVectors depend on specific abiotic conditions and advances inRS and mapping of variations in abiotic conditions have stimu-lated efforts to create risk maps for VBDs538 This spatial con-cordance of environmental variables and vector distribution isoften used to estimate the current distribution of vectors in un-studied areas
There have been many studies of the entomological inocula-tion rate and parasite prevalence vector density and breedingsites and risk mapping and modelling in West Africa related tothe Anopheles mosquito The distribution of Anopheles sppwhich are the major vectors of LF and malaria in West Africawas studied in relation to normalised difference vegetationindex (NDVI) values39 Several studies were dedicated to investi-gating the risk factors for malaria and identifying the potentialAnopheles mosquito breeding sites in urban environments inCote drsquoIvoire and Senegal4041 In Burkina Faso and The
Figure 2 Geographic distribution of human African trypanosomiasis (Trypanosoma brucei gambiense) in West African countries in 2010 (data from theWHO retrieved from httpwwwwhoint)
Transactions of the Royal Society of Tropical Medicine and Hygiene
5 of 12
at University of L
iverpool on May 9 2013
httptrstmhoxfordjournalsorg
Dow
nloaded from
Gambia potential Anopheles breeding sites were mapped at thevillage level using satellite imagery4243
A critically important review regarding the potential for GISRStechnologies and methodologies to be used for the preventionsurveillance and control of the mosquito Ae aegypti thedengue virus vector can be found in an article by Eisen et al7
In Senegal a clear association between the amount of rainfallthe abundance of vectors and the prevalence of RVF has beendemonstrated44 In another study in Senegal the average totalmonthly rainfall from December to February was the most im-portant spatial predictor of the risk of RVF45
GIS and RS techniques are promising and powerful tools for de-scribing tsetse distribution because thermal data appeared to bethe most useful predictor variable followed by indices of vegeta-tion and rainfall46 Recently a study demonstrated the potentialof high-resolution images for mapping the habitat of Glossina ona local scale as well as in larger areas47
Applications of GISRS have been summarised with examplesof studies on various vectors such as the malaria vector Anoph-eles the arbovirus vector culicine mosquitoes (Aedes spp andCulex spp) the leishmaniasis vector Phlebotomus sandflies thetrypanosomiasis vector tsetse and the loiasis vector Chrysopsin a review by Thomson and Connor48
Previous studies have used multivariate analyses to estimatethe spatial prevalence of particular tick species and to makeinferences about different environmental variables that deter-mine their distribution in Africa4950 Another approach to predict-ing tick distribution was established using a set of methods in
which the author concluded that on average climatic variablesare better predictors of tick distribution than vegetation-relatedvariables and the key to describing tick distribution is the covari-ance of temperature and rainfall51
It should be taken into account that risk mapping based onvectors has serious limitations Primarily disease risk or incidenceis most closely correlated with the abundance of pathogen-infected vectors rather than simply the presence of vectors orthe total abundance of vectors52
Perspectives of application of geographicinformation systemsremote sensingtechnologies for studying vectorbornediseases remaining challenges and conclusionGIS facilitate comparisons between disease patterns and envir-onmental data while RS technologies can use high-resolutionsatellite data to provide estimates of variables such as tempera-ture vegetation and humidity53 During the past two decadessatellite RS technology has shown promising results in assessingthe risk of various VBDs at different spatial scales Satellite-basedimagery is characterised by its spatial spectral and temporalresolution Despite some limited attempts to apply RS to epi-demiology the methods have not yet demonstrated theirexpected potential GISRS technologies are undoubtedly a valu-able source of information for epidemiologists GISRS technolo-gies should not be overestimated and it is necessary to take all of
Figure 3 Geographic distribution of cutaneous leishmaniasis in West African countries in 2010 (data from the WHO retrieved from httpwwwwhoint)
P Ratmanov et al
6 of 12
at University of L
iverpool on May 9 2013
httptrstmhoxfordjournalsorg
Dow
nloaded from
Figure 4 Geographic distribution of lymphatic filariasis Loa loa filariasis (loiasis) and onchocerciasis in West African countries (data from the WHOretrieved from httpwwwwhoint)
Figure 5 Geographic distribution of Rift Valley fever (RVF) in West African countries (data from the WHO retrieved from httpwwwwhoint)
Transactions of the Royal Society of Tropical Medicine and Hygiene
7 of 12
at University of L
iverpool on May 9 2013
httptrstmhoxfordjournalsorg
Dow
nloaded from
Figure 6 Geographic distribution of dengue in West African countries in 2011 (data from the WHO retrieved from httpwwwwhoint)
Figure 7 Geographic distribution of yellow fever in West African countries in 2011 (data from the WHO retrieved from httpwwwwhoint)
P Ratmanov et al
8 of 12
at University of L
iverpool on May 9 2013
httptrstmhoxfordjournalsorg
Dow
nloaded from
Figure 8 Geographic distribution of CrimeanndashCongo haemorrhagic fever in West African countries (data from the WHO retrieved from httpwwwwhoint)
Figure 9 Geographic distribution of Buruli ulcer in 2010 in West African countries (data from the WHO retrieved from httpwwwwhoint)
Transactions of the Royal Society of Tropical Medicine and Hygiene
9 of 12
at University of L
iverpool on May 9 2013
httptrstmhoxfordjournalsorg
Dow
nloaded from
their advantages and limitations into account Moreover VBDinterventions have to be identified and recognised as they canconfound the relationship between the environment anddisease In the recent malaria indicator surveys in Zambia andAngola for example no relationship was observed between re-motely sensed data and malaria risk5455
The typical modelling approach investigates associationsbetween multivariate environmental data and patterns ofvector presence or absence for mapping vectors and VBDs Atthe first step simple statistical models could be a good startingpoint for linking the limited number of environmental variablesthat can be derived from satellite data Simple statisticalmodels are restricted because they often fit linear functionsbetween environmental variables and presenceabsence datawhen it is most likely that such associations are highlycomplex and non-linear20 At the second step sophisticatedprocess-based models that rely on vector biology as predictorsof diseases and their risk should be developed56
Regression models have been widely applied in landscape epi-demiology The type of regression model (logistic Poisson linearetc) is determined by the type of outcome variable to be pre-dicted (eg binary count continuous) and environmental vari-ables measured at sampled locations are entered ascovariates The resultant model is then either used to predictthe outcome variable at non-sampled locations based onobserved values of the covariates at the prediction locations orto explain observed patterns of disease on the basis of themodel covariates57
Besides regression modelling there are many otherapproaches in spatial epidemiology spatial clustering discrimin-ant analysis generalised linear models generalised additivemodels Bayesian estimation methods and others Howeversuch traditional models require both disease presence anddisease absence data At the same time there are differentlsquopresence-onlyrsquo models that could be used when no lsquoabsencedatarsquo are available One of these presence-only machine learning(rule-based) algorithms is an ecological niche modelling algo-rithm based on maximum entropy (Maxent) This method canbe used successfully with very small sample sizes and does notrequire independence of covariates58 Moreover these advan-tages could be crucial in Africa where complete surveillancedata from all regions are not available
Basic spatial modelling approaches include interpolationbased on spatial dependence in vector or VBD data and extrapo-lation based on associations between vector or VBD data and en-vironmental or socioeconomic predictor variables The latterapproach can be a powerful tool to gain insights into levels ofrisk within areas where surveillance data are lacking or unreli-able However it should be noted that model extrapolation isrestricted to areas with ecological and climatic characteristicssimilar to those of the model development area1
Global strategies for controlling infections in sub-SaharanAfrica have focused on the lsquobig threersquo diseases namely AIDSTB and malaria Other causes of fever and infection fall intothe category of NTDs In the twenty-first century it is importantto compile a comprehensive list of the infections in sub-SaharanAfrica and to study their epidemiology59
Few investigations have addressed the unknown causes offever in West Africa The majority of fevers have long been con-sidered to be related to malaria Recent studies have shown that
NTDs are the most frequent causes of fever in both indigenes andtravellers In particular TBRF due to B crocidurae has been redis-covered931 and the rickettsiae including R felis have beenreported33 Overall new pathogens among the rickettsiae wereidentified in the environment8
The geographic distribution of ticksmdashvectors of diseases inWest Africamdashis a neglected area of study However applyingGISRS technologies to the study of Ixodes ticks in North Americahas yielded significant results For example the use of geospatialmodelling has revealed that high concentrations of I pacificusand I scapularis which are the key tick vectors of Lyme diseasein North America can be predicted by GISRS-based environmen-tal factors related to elevation slope of the landscape vegetationtype soil type temperature and moisture60 The same approachescould be used in epidemiological studies of R felis infection anemerging disease in West Africa
Today epidemiologists often use new GISRS techniques tostudy a variety of VBDs The associations between satellite-derived environmental variables (such as temperature humidityelevation vegetation rainfall surface water land use land covertype and soil moisture) and vector density are used to identifyand characterise vector habitats A wide variety of RS datapowerful GIS and statistical software packages are available fordesktop computing environments making it affordable and feas-ible for epidemiologists to experiment with new spatial analysistechniques56
Maps that show the seasonal risk of VBDs will be necessary tomonitor the impacts of changes on vector ecology Using GISRSadvanced analytical tools and a landscape ecology approachthese risk maps could play a major role in defining research ques-tions and surveillance needs and in guiding control efforts andfield studies38
From the other side this approach is based on ecological ana-lysis and requires assembling all epidemiologic available dataand moreover it should be supported by fieldwork The most suc-cessful GISRS applications to study VBDs in West Africa areeither local (country)-level studies using community-based sur-veillance data or continental (sub-continental)-level studiesusing published scientific literature data Cooperation of the epi-demiological community could allow enhancing studies of VBDsin West Africa The most important goal of the application of GISRS technologies to the study of VBDs is to reduce the burden ofdisease by generating information that empowers the public totake protective action and helps public health agencies to effect-ively allocate their limited prevention surveillance and controlresources There is great potential to use GISRS technologiesto improve the surveillance prevention and control of vector-borne infectious diseases in West Africa
Supplementary dataSupplementary data are available at Transactions Online(httptrstmhoxfordjournalsorg)
Authorsrsquo contributions All authors contributed equally to themanuscript DR conceived the review PR and OM researched the dataand literature PR OM and DR analysed and interpreted the data for
P Ratmanov et al
10 of 12
at University of L
iverpool on May 9 2013
httptrstmhoxfordjournalsorg
Dow
nloaded from
the maps and discussed the content of the review PR and DR drafted themanuscript All authors critically revised the manuscript for intellectualcontent and read and approved the final version DR is guarantor ofthe paper
Funding None
Competing interests None declared
Ethical approval Not required
References1 Eisen L Eisen RJ Using geographic information systems and decision
support systems for the prediction prevention and control ofvector-borne diseases Annu Rev Entomol 20115641ndash61
2 WHO Neglected Tropical Diseases Geneva World HealthOrganization 2012 httpwwwwhointneglected_diseasesdiseasesen [accessed 8 January 2013]
3 Hay SI An overview of remote sensing and geodesy for epidemiologyand public health application Adv Parasitol 2000471ndash35
4 Pavlovsky EN Natural Nidality of Transmissible Diseases with SpecialReference to the Landscape Epidemiology of ZooanthroponosesUrbana IL University of Illinois Press 1966
5 Kitron U Landscape ecology and epidemiology of vector-bornediseases tools for spatial analysis J Med Entomol 199835435ndash45
6 Rogers DJ Randolph SE Studying the global distribution of infectiousdiseases using GIS and RS Nat Rev Microbiol 20031231ndash7
7 Eisen L Lozano-Fuentes S Use of mapping and spatial andspace-time modeling approaches in operational control of Aedesaegypti and dengue PLoS Negl Trop Dis 20093e411
8 Mediannikov O Diatta G Fenollar F et al Tick-borne rickettsiosesneglected emerging diseases in rural Senegal PLoS Negl Trop Dis20104e821
9 Parola P Diatta G Socolovschi C et al Tick-borne relapsing feverborreliosis rural Senegal Emerg Infect Dis 201117883ndash5
10 WHO World Malaria Report 2010 Geneva World HealthOrganization 2010
11 Sullivan D Uncertainty in mapping malaria epidemiologyimplications for control Epidemiol Rev 201032175ndash87
12 Machault V Vignolles C Borchi F et al The use of remotely sensedenvironmental data in the study of malaria Geospat Health20115151ndash68
13 Snow RW Marsh K le Sueur D The need for maps of transmissionintensity to guide malaria control in Africa Parasitol Today199612455ndash7
14 Guerra CA Hay SI Lucioparedes LS et al Assembling a globaldatabase of malaria parasite prevalence for the Malaria AtlasProject Malar J 2007617
15 Brun R Blum J Chappuis F Burri C Human African trypanosomiasisLancet 2010375148ndash59
16 Simarro PP Cecchi G Paone M et al The Atlas of human Africantrypanosomiasis a contribution to global mapping of neglectedtropical diseases Int J Health Geogr 2010957
17 Bern C Maguire JH Alvar J Complexities of assessing the diseaseburden attributable to leishmaniasis PLoS Negl Trop Dis 20082e313
18 WHO Report of the Scientific Working Group Meeting onLeishmaniasis (Geneva 2ndash4 February 2004) Geneva World HealthOrganization 2004
19 Faye B Bucheton B Banuls AL et al Seroprevalence of Leishmaniainfantum in a rural area of Senegal analysis of risk factors involvedin transmission to humans Trans R Soc Trop Med Hyg2011105333ndash40
20 Slater H Michael E Predicting the current and future potentialdistributions of lymphatic filariasis in Africa using maximumentropy ecological niche modelling PLoS One 20127e32202
21 Basanez MG Pion SD Churcher TS et al River blindness a successstory under threat PLoS Med 20063e371
22 Noma M Nwoke BE Nutall I et al Rapid epidemiological mapping ofonchocerciasis (REMO) its application by the African Programme forOnchocerciasis Control (APOC) Ann Trop Med Parasitol 200296(Suppl 1)S29ndash39
23 Zoure HG Wanji S Noma M et al The geographic distribution of Loaloa in Africa results of large-scale implementation of the RapidAssessment Procedure for Loiasis (RAPLOA) PLoS Negl Trop Dis20115e1210
24 Favier C Chalvet-Monfray K Sabatier P et al Rift Valley fever in WestAfrica the role of space in endemicity Trop Med Int Health2006111878ndash88
25 Jouan A Coulibaly I Adam F et al Analytical study of a Rift Valleyfever epidemic Res Virol 1989140175ndash86
26 Diallo M Ba Y Sall AA et al Amplification of the sylvatic cycle ofdengue virus type 2 Senegal 1999ndash2000 entomologicfindings and epidemiologic considerations Emerg Infect Dis20039362ndash7
27 Jentes ES Poumerol G Gershman MD et al The revised global yellowfever risk map and recommendations for vaccination 2010consensus of the Informal WHO Working Group on Geographic Riskfor Yellow Fever Lancet Infect Dis 201111622ndash32
28 Rogers DJ Wilson AJ Hay SI Graham AJ The global distribution ofyellow fever and dengue Adv Parasitol 200662181ndash220
29 Grard G Drexler JF Fair J et al Re-emergence of CrimeanndashCongohemorrhagic fever virus in Central Africa PLoS Negl Trop Dis20115e1350
30 Nabeth P Thior M Faye O Simon F Human CrimeanndashCongohemorrhagic fever Senegal Emerg Infect Dis 2004101881ndash2
31 Vial L Diatta G Tall A et al Incidence of tick-borne relapsing fever inWest Africa longitudinal study Lancet 200636837ndash43
32 Jensenius M Davis X von Sonnenburg F et al Multicenter GeoSentinelanalysis of rickettsial diseases in international travelers 1996ndash2008Emerg Infect Dis 2009151791ndash8
33 Socolovschi C Mediannikov O Sokhna C et al Rickettsiafelis-associated uneruptive fever Senegal Emerg Infect Dis2010161140ndash2
34 Cazorla C Socolovschi C Jensenius M Parola P Tick-borne diseasestick-borne spotted fever rickettsioses in Africa Infect Dis Clin NorthAm 200822p531ndash44ixndashx
35 Merritt RW Walker ED Small PL et al Ecology and transmission ofBuruli ulcer disease a systematic review PLoS Negl Trop Dis20104e911
36 Tissot-Dupont H Brouqui P Faugere B Raoult D Prevalence ofantibodies to Coxiella burnetti Rickettsia conorii and Rickettsia typhiin seven African countries Clin Infect Dis 1995211126ndash33
37 Mediannikov O Fenollar F Socolovschi C et al Coxiella burnetii inhumans and ticks in rural Senegal PLoS Negl Trop Dis 20104e654
38 Kitron U Risk maps transmission and burden of vector-bornediseases Parasitol Today 200016324ndash5
Transactions of the Royal Society of Tropical Medicine and Hygiene
11 of 12
at University of L
iverpool on May 9 2013
httptrstmhoxfordjournalsorg
Dow
nloaded from
39 Thomson MC Connor SJ Milligan P Flasse SP Mapping malaria risk inAfrica what can satellite data contribute Parasitol Today199713313ndash8
40 Matthys B Koudou BG NrsquoGoran EK et al Spatial dispersion andcharacterisation of mosquito breeding habitats in urbanvegetable-production areas of Abidjan Cote drsquoIvoire Ann Trop MedParasitol 2010104649ndash66
41 Machault V Vignolles C Pages F et al Spatial heterogeneity andtemporal evolution of malaria transmission risk in Dakar Senegalaccording to remotely sensed environmental data Malar J 20109252
42 Dambach P Sie A Lacaux JP et al Using high spatial resolutionremote sensing for risk mapping of malaria occurrence in theNouna District Burkina Faso Glob Health Action 20092 doi103402ghav2io2094
43 Bogh C Lindsay SW Clarke SE et al High spatial resolution mappingof malaria transmission risk in The Gambia West Africa usingLANDSAT TM satellite imagery Am J Trop Med Hyg 200776875ndash81
44 Bicout DJ Sabatier P Mapping Rift Valley fever vectors and prevalenceusing rainfall variations Vector Borne Zoonotic Dis 2004433ndash42
45 Clements AC Pfeiffer DU Martin V et al Spatial risk assessment of RiftValley fever in Senegal Vector Borne Zoonotic Dis 20077203ndash16
46 Rogers DJ Hay SI Packer MJ Predicting the distribution of tsetse fliesin West Africa using temporal Fourier processed meteorologicalsatellite data Ann Trop Med Parasitol 199690225ndash41
47 Cecchi G Mattioli RC Slingenbergh J de la Rocque S Land cover andtsetse fly distributions in sub-Saharan Africa Med Vet Entomol200822364ndash73
48 Thomson MC Connor SJ Environmental information systems for thecontrol of arthropod vectors of disease Med Vet Entomol200014227ndash44
49 Hay SI Tucker CJ Rogers DJ Packer MJ Remotely sensed surrogatesof meteorological data for the study of the distribution andabundance of arthropod vectors of disease Ann Trop Med Parasitol1996901ndash19
50 Randolph SE Rogers DJ A generic population model for theAfrican tick Rhipicephalus appendiculatus Parasitology 1997115265ndash79
51 Cumming GS Comparing climate and vegetation as limiting factorsfor species ranges of African ticks Ecology 200283255ndash68
52 Ostfeld RS Glass GE Keesing F Spatial epidemiology an emerging (orre-emerging) discipline Trends Ecol Evol 200520328ndash36
53 Hay SI Tatem AJ Graham AJ et al Global environmental data formapping infectious disease distribution Adv Parasitol 20066237ndash77
54 Riedel N Vounatsou P Miller JM et al Geographical patterns andpredictors of malaria risk in Zambia Bayesian geostatisticalmodelling of the 2006 Zambia national malaria indicator survey(ZMIS) Malar J 2010937
55 Gosoniu L Veta AM Vounatsou P Bayesian geostatisticalmodeling of Malaria Indicator Survey data in Angola PLoS One20105e9322
56 Kalluri S Gilruth P Rogers D Szczur M Surveillance of arthropodvector-borne infectious diseases using remote sensing techniquesa review PLoS Pathog 200731361ndash71
57 Clements AC Pfeiffer DU Emerging viral zoonoses frameworks forspatial and spatiotemporal risk assessment and resource planningVet J 200918221ndash30
58 Phillips SJ Anderson RP Schapire RE Maximum entropy modelingof species geographic distributions Ecological Modelling 2006190231ndash59
59 Fenollar F Neglected and emerging diseases in sub-Saharan AfricaClin Microbiol Infect 201117957ndash8
60 Eisen RJ Eisen L Lane RS Predicting density of Ixodes pacificusnymphs in dense woodlands in Mendocino County Californiabased on geographic information systems and remotesensing versus field-derived data Am J Trop Med Hyg 200674632ndash40
P Ratmanov et al
12 of 12
at University of L
iverpool on May 9 2013
httptrstmhoxfordjournalsorg
Dow
nloaded from
with more than 200 million cases and causing up to 1 milliondeaths each year People in endemic areas with symptomaticand asymptomatic malaria are reservoirs for the infectionMalaria is endemic in all countries in West Africa but theburden it imposes depends upon the scale of malaria controlefforts in the country (Figure 1)10
Many efforts have been made to collect and centralise existingentomological parasitological and epidemiological data in AfricaNevertheless a high degree of uncertainty still exists regarding theannual number of malaria cases and their geographic distribu-tion11 A summary of the environmental satellite data that areavailable for studying malaria can be found in the review ofMachault et al12 The Mapping Malaria Risk in Africa (MARAARMA) project was established for developing malaria risk mapson the scale of the entire continent13 More recently the Malaria
Atlas Project (MAP) was born with the objective of gathering world-wide parasite prevalence data and making them freely availableon the internet (httpwwwmapoxacuk)14
Human African trypanosomiasis
The aetiological agent of human African trypanosomiasis (HAT)also known as sleeping sickness is protozoa of the species Trypa-nosoma brucei gambiense that are transmitted by tsetse flies(Glossina spp) In West Africa humans are the main reservoir forTb gambiense Animals play a less important role but pigs andsome wild animal species have been reported as reservoirs15 InWest Africa HAT has been reported in Guinea Cote drsquoIvoire andNigeria and is also endemic to several other countries (Figure 2)Control of sleeping sickness has always been closely related to
Table 1 Selected vectorborne diseases in West Africa
Disease Pathogen Reservoir Vector
ProtozoalMalaria Plasmodium falciparum P vivax
P ovale P malariaeHumans Mosquitoes (Anopheles gambiae sl
An funestus)Human Africantrypanosomiasis
Trypanosoma brucei gambiense Humans some wild anddomestic mammals
Tsetse flies (Glossina palpalis slG tachinoides)
Leishmaniasis Leishmania genus Mammals Phlebotomus (Phlebotomus spp)Sergentomyia (Spelaeomyia) darlingi
HelminthicLymphatic filariasis Wuchereria bancrofti Humans Mosquitoes (Anopheles spp)Loa loa filariasis (loiasis) Loa loa Humans wild mammals Tabanid flies (Chrysops spp)Onchocerciasis (lsquoriverblindnessrsquo)
Onchocerca volvulus Humans Black flies (Simulium genus)
ViralRift Valley fever Rift Valley fever virus Wild and domestic
mammalsMosquitoes (Aedes spp Culex spp)
Dengue Dengue virus Humans Mosquitoes (Aedes spp)Chikungunya fever Chikungunya virus Primates humans Mosquitoes (Aedes spp)CrimeanndashCongohaemorrhagic fever
CrimeanndashCongo haemorrhagicfever virus
Domestic and wild animals Ticks (Hyalomma genus Amblyommavariegatum)
Yellow fever Yellow fever virus Primates humans Mosquitoes (Aedes spp)Bacterial
Mediterranean spottedfever
Rickettsia conorii Wild and domesticmammals
Ticks (Rhipicephalus evertsi)
African tick-bite fever Rickettsia africae Wild and domesticmammals
Ticks (Amblyomma variegatumRhipicephalus evertsi)
Tickborne relapsing fever Borrelia crocidurae Rodents Ticks (Ornithodoros sonrai)Rickettsia felis infection Rickettsia felis Unknown Fleas (Ctenocephalides felis and others)Trench fever Bartonella quintana Humans Lice (Pediculus humanus humanus)Louseborne relapsingfever
Borrelia recurrentis Humans Lice (Pediculus humanus humanus)
Epidemic typhus Rickettsia prowazekii Humans Lice (Pediculus humanus humanus)Buruli ulcer Mycobacterium ulcerans UnknownQ fever Coxiella burnetii Mammals birds and
arthropodsTicksa
aPossibility of transmission of Q fever by tick was not studied
P Ratmanov et al
2 of 12
at University of L
iverpool on May 9 2013
httptrstmhoxfordjournalsorg
Dow
nloaded from
disease mapping The Atlas of HAT is the most prominent attemptat disease control research and advocacy16
Leishmaniasis
Leishmaniasis refers to a group of VBDs that are caused by morethan 20 species of the protozoan genus Leishmania rangingfrom localized skin ulcers to lethal systemic disease Leishmania-sis is considered one of the lsquomost neglected diseasesrsquo because
limited resources are invested in its diagnosis treatment andcontrol and it is strongly associated with poverty17 Humansare infected via the bite of phlebotomine sandflies (Phlebotomusspp) The leishmaniases can be classified into two epidemio-logical entities according to the type of transmission anthropo-notic if humans are the sole reservoir involved in transmission(and the sole source for vector infection) or zoonotic if atleast one mammalian reservoir is involved18 Although thereare no major geographic foci of leishmaniasis in West Africathe cutaneous forms of leishmaniasis have been reported in 11of 15 countries in the region (Figure 3) In recent studies theseroprevalence of specific antibodies against L infantum (theagent of visceral leishmaniasis) in the human population wasdetermined in Senegal Larger-scale studies with application ofGISRS technologies are now required to define the distributionof L infantum in West Africa and an identification of its potentialvector19
Filariasis
Lymphatic filariasis (LF) is a vectorborne parasitic infectiousdisease caused by Wuchereria bancrofti which is endemic inthe tropics including in sub-Saharan Africa It is transmitted tohumans by infected mosquitoes In West Africa the mosquitoAn gambiae sl is a vector of LF and malaria caused by P falcip-arum Humans are the only reservoir host of the LF parasite inAfrica20 Onchocerciasis (or lsquoriver blindnessrsquo) is a parasiticdisease caused by the filarial parasitic nematode Onchocerca vol-vulus It is transmitted through the bites of infected Simulium(black fly) vectors which breed in fast-flowing streams andrivers21 Control of onchocerciasis in sub-Saharan Africa is over-seen by the African Programme for Onchocerciasis Control(APOC)22 Loa loa filariasis (loiasis) is a NTD caused by the filarialparasite Loa loa It is an African disease restricted to the equator-ial rainforest regions of Central and West Africa23 The insectvectors of L loa are flies of the genus Chrysops Humans arethe primary reservoir for L loa LF loiasis and onchocerciasisare endemic in many countries of West Africa (Figure 4)
Rift Valley fever
Rift Valley fever (RVF) is an arthropodborne viral disease that pri-marily causes epizootics of abortion and high mortality rates indomestic animals but it can also infect humans The RVF virusis mainly transmitted by mosquitoes of the Aedes and Culexgenera to a wide range of animals from rodents to camelswhich are the natural reservoirs for RVF Although mosquitoesmay transmit the RVF virus to humans human infections resultfrom contact with infected animals24 A large RVF epidemic oc-curred in 1987 in southern Mauritania and northern SenegalHowever all of the countries in West Africa are at risk of RVF(Figure 5)25
Dengue and yellow fever
Dengue fever is a viral infectious tropical disease The mosquitoAe aegypti is the primary vector of the dengue virus andhumans are the most common reservoir Dengue is a widespreaddisease in the subtropics and tropics but in Africa the burden ofthe disease is poorly understood7 In West Africa sylvatic
Table 2 Selected vectors of the human vectorborne diseasesin West Africa
Vector Disease
MosquitoesAnopheles spp Lymphatic filariasisa
An gambiae sl MalariaAedes spp Rift Valley fever
DengueYellow feverChikungunya fever
Culex spp Rift Valley feverPhlebotomus
Phlebotomus spp LeishmaniasisSergentomyia (Spelaeomyia)darlingi
Leishmaniasis
Tsetse fliesGlossina palpalis sl Human African
trypanosomiasisG tachinoides Human African
trypanosomiasisTabanid flies
Chrysops spp Loa loa filariasis (loiasis)Black flies
Simulium spp Onchocerciasis (lsquoriverblindnessrsquo)
TicksRhipicephalus evertsi Mediterranean spotted fever
African tick-bite feverAmblyomma variegatum African tick-bite fever
CrimeanndashCongo haemorrhagicfever
Ornithodoros sonrai Tickborne relapsing feverOrnithodoros moubata Tickborne relapsing feverHyalomma spp CrimeanndashCongo haemorrhagic
feverLice
Pediculus humanus humanus Trench feverLouseborne relapsing feverEpidemic typhus
FleasCtenocephalides felis andothers
Rickettsia felis infectionb
aAnopheles is a primary vector of lymphatic filariasis in AfricabFleas are considered as vectors of R felis worldwide
Transactions of the Royal Society of Tropical Medicine and Hygiene
3 of 12
at University of L
iverpool on May 9 2013
httptrstmhoxfordjournalsorg
Dow
nloaded from
circulation of the dengue virus is the predominant form of circu-lation (with lower primates as the main reservoir)26 Evidenceregarding circulation of the dengue virus was obtained from 11countries in the region although the mosquito vectors arepresent throughout West Africa (Figure 6)
Yellow fever (YF) is an acute viral haemorrhagic disease trans-mitted in West Africa by infected Aedes spp mosquitoes Up to50 of severely affected persons who do not receive treatmentdie from YF and there is no cure The YF virus circulates both inurban and sylvatic settings involving several vertebrate speciesIn the sylvatic cycle mosquitoes act as the main vectors andmonkeys act as the primary hosts In the urban cycle the virusis transmitted between human beings and mosquitoes (predom-inately Ae aegypti) In Africa transmission of the virus can alsooccur in an intermediate cycle between human beings or non-human primates and Aedes spp mosquitoes that breed in treeholes on the savannah27 Vertical transmission also occurswithin the mosquito population and may play an importantrole in maintaining the sylvatic cycle28 In West Africa YF isholoendemic in all countries except in the Sahara Desertregions of Mali and Niger (Figure 7)27
CrimeanndashCongo haemorrhagic fever
CrimeanndashCongo haemorrhagic fever is a viral tickborne diseaseThe virus causes severe illness throughout the world includingWest Africa Its distribution closely matches that of its mainarthropod vector ixodid ticks belonging to the genus Hyalomma
Human infection occurs through tick bites contact with infectedlivestock or nosocomial transmission29 In West Africa enzooticcirculation of the CrimeanndashCongo haemorrhagic fever virus hasbeen shown in serological surveys of cattle but only sporadichuman cases have been reported (Figure 8)30
Borrelioses
In West Africa tickborne relapsing fever (TBRF) is caused by Bor-relia crocidurae Studies in Senegal indicate that TBRF is aftermalaria the most common cause of outpatient visits to a ruraldispensary Investigations found that the vector tick (Ornitho-doros sonrai) was present in villages in Senegal Mauritania andMali with high infection rates Rodents and insectivores are thereservoirs of B crocidurae in West Africa31 Unfortunately noattempts have been made to perform a spatial analysis ofTBRF prevalence in West Africa
Rickettsioses
Following malaria tickborne rickettsioses are one of the mostcommon causes of systemic febrile illness among travellersfrom developed countries but little is known about their preva-lence in indigenous populations especially in West Africa32
There is a high incidence (44) of Rickettsia felis in Senegal33
A case of Mediterranean spotted fever was described inSenegal and its agent R conorii was found in the ticks Rhipice-phalus evertsi evertsi8 There is a plethora of serological studies
Figure 1 Cases of malaria per 100 000 inhabitants in West African countries in 2010 or latest available year (data from the WHO retrieved fromhttpwwwwhoint)
P Ratmanov et al
4 of 12
at University of L
iverpool on May 9 2013
httptrstmhoxfordjournalsorg
Dow
nloaded from
of the prevalence of rickettsioses and entomological studies ofticks for Rickettsia spp Although attempts to map the geographicdistribution of rickettsioses in Africa have been made they simplydisplayed infected sites or counties with reported cases as pointswithout conducting any geostatistical analyses834 Thus thespatial aspects of borrelial and rickettsial VBDs in West Africahave not received the attention they deserve
Buruli ulcer
Buruli ulcer (BU) disease caused by Mycobacterium ulcerans isan emerging infectious disease in many tropical and subtropicalcountries The exact mode of transmission of this diseaseremains unclear Although vectors and modes of transmissionremain unknown it has been hypothesized that transmissionof BU disease is associated with human activities in andaround aquatic environments (the role of water bugs has beendiscussed) and that characteristics of the landscape play a rolein the spread of BU disease35 In West Africa Cote drsquoIvoireGhana and Benin are the countries that are particularly affectedby BU disease (Figure 9)
Q fever
In West Africa Q fever has a wide distribution which has beenshown repeatedly in human serological studies and studies ofreservoirs (domestic animals)36 It was reported that the main
vector of TBRF the soft tick O sonrai also harboured Coxiella bur-netii37 The possibility that this tick transmits Q fever was notstudied but some species of tick may play a role in the transmis-sion of Q fever
Geographic information systems and remotesensing for spatial analysis of vectordistributionThe geographic distribution of VBDs depends on the distributionof their vectors environmental factors and climatic factorsVectors depend on specific abiotic conditions and advances inRS and mapping of variations in abiotic conditions have stimu-lated efforts to create risk maps for VBDs538 This spatial con-cordance of environmental variables and vector distribution isoften used to estimate the current distribution of vectors in un-studied areas
There have been many studies of the entomological inocula-tion rate and parasite prevalence vector density and breedingsites and risk mapping and modelling in West Africa related tothe Anopheles mosquito The distribution of Anopheles sppwhich are the major vectors of LF and malaria in West Africawas studied in relation to normalised difference vegetationindex (NDVI) values39 Several studies were dedicated to investi-gating the risk factors for malaria and identifying the potentialAnopheles mosquito breeding sites in urban environments inCote drsquoIvoire and Senegal4041 In Burkina Faso and The
Figure 2 Geographic distribution of human African trypanosomiasis (Trypanosoma brucei gambiense) in West African countries in 2010 (data from theWHO retrieved from httpwwwwhoint)
Transactions of the Royal Society of Tropical Medicine and Hygiene
5 of 12
at University of L
iverpool on May 9 2013
httptrstmhoxfordjournalsorg
Dow
nloaded from
Gambia potential Anopheles breeding sites were mapped at thevillage level using satellite imagery4243
A critically important review regarding the potential for GISRStechnologies and methodologies to be used for the preventionsurveillance and control of the mosquito Ae aegypti thedengue virus vector can be found in an article by Eisen et al7
In Senegal a clear association between the amount of rainfallthe abundance of vectors and the prevalence of RVF has beendemonstrated44 In another study in Senegal the average totalmonthly rainfall from December to February was the most im-portant spatial predictor of the risk of RVF45
GIS and RS techniques are promising and powerful tools for de-scribing tsetse distribution because thermal data appeared to bethe most useful predictor variable followed by indices of vegeta-tion and rainfall46 Recently a study demonstrated the potentialof high-resolution images for mapping the habitat of Glossina ona local scale as well as in larger areas47
Applications of GISRS have been summarised with examplesof studies on various vectors such as the malaria vector Anoph-eles the arbovirus vector culicine mosquitoes (Aedes spp andCulex spp) the leishmaniasis vector Phlebotomus sandflies thetrypanosomiasis vector tsetse and the loiasis vector Chrysopsin a review by Thomson and Connor48
Previous studies have used multivariate analyses to estimatethe spatial prevalence of particular tick species and to makeinferences about different environmental variables that deter-mine their distribution in Africa4950 Another approach to predict-ing tick distribution was established using a set of methods in
which the author concluded that on average climatic variablesare better predictors of tick distribution than vegetation-relatedvariables and the key to describing tick distribution is the covari-ance of temperature and rainfall51
It should be taken into account that risk mapping based onvectors has serious limitations Primarily disease risk or incidenceis most closely correlated with the abundance of pathogen-infected vectors rather than simply the presence of vectors orthe total abundance of vectors52
Perspectives of application of geographicinformation systemsremote sensingtechnologies for studying vectorbornediseases remaining challenges and conclusionGIS facilitate comparisons between disease patterns and envir-onmental data while RS technologies can use high-resolutionsatellite data to provide estimates of variables such as tempera-ture vegetation and humidity53 During the past two decadessatellite RS technology has shown promising results in assessingthe risk of various VBDs at different spatial scales Satellite-basedimagery is characterised by its spatial spectral and temporalresolution Despite some limited attempts to apply RS to epi-demiology the methods have not yet demonstrated theirexpected potential GISRS technologies are undoubtedly a valu-able source of information for epidemiologists GISRS technolo-gies should not be overestimated and it is necessary to take all of
Figure 3 Geographic distribution of cutaneous leishmaniasis in West African countries in 2010 (data from the WHO retrieved from httpwwwwhoint)
P Ratmanov et al
6 of 12
at University of L
iverpool on May 9 2013
httptrstmhoxfordjournalsorg
Dow
nloaded from
Figure 4 Geographic distribution of lymphatic filariasis Loa loa filariasis (loiasis) and onchocerciasis in West African countries (data from the WHOretrieved from httpwwwwhoint)
Figure 5 Geographic distribution of Rift Valley fever (RVF) in West African countries (data from the WHO retrieved from httpwwwwhoint)
Transactions of the Royal Society of Tropical Medicine and Hygiene
7 of 12
at University of L
iverpool on May 9 2013
httptrstmhoxfordjournalsorg
Dow
nloaded from
Figure 6 Geographic distribution of dengue in West African countries in 2011 (data from the WHO retrieved from httpwwwwhoint)
Figure 7 Geographic distribution of yellow fever in West African countries in 2011 (data from the WHO retrieved from httpwwwwhoint)
P Ratmanov et al
8 of 12
at University of L
iverpool on May 9 2013
httptrstmhoxfordjournalsorg
Dow
nloaded from
Figure 8 Geographic distribution of CrimeanndashCongo haemorrhagic fever in West African countries (data from the WHO retrieved from httpwwwwhoint)
Figure 9 Geographic distribution of Buruli ulcer in 2010 in West African countries (data from the WHO retrieved from httpwwwwhoint)
Transactions of the Royal Society of Tropical Medicine and Hygiene
9 of 12
at University of L
iverpool on May 9 2013
httptrstmhoxfordjournalsorg
Dow
nloaded from
their advantages and limitations into account Moreover VBDinterventions have to be identified and recognised as they canconfound the relationship between the environment anddisease In the recent malaria indicator surveys in Zambia andAngola for example no relationship was observed between re-motely sensed data and malaria risk5455
The typical modelling approach investigates associationsbetween multivariate environmental data and patterns ofvector presence or absence for mapping vectors and VBDs Atthe first step simple statistical models could be a good startingpoint for linking the limited number of environmental variablesthat can be derived from satellite data Simple statisticalmodels are restricted because they often fit linear functionsbetween environmental variables and presenceabsence datawhen it is most likely that such associations are highlycomplex and non-linear20 At the second step sophisticatedprocess-based models that rely on vector biology as predictorsof diseases and their risk should be developed56
Regression models have been widely applied in landscape epi-demiology The type of regression model (logistic Poisson linearetc) is determined by the type of outcome variable to be pre-dicted (eg binary count continuous) and environmental vari-ables measured at sampled locations are entered ascovariates The resultant model is then either used to predictthe outcome variable at non-sampled locations based onobserved values of the covariates at the prediction locations orto explain observed patterns of disease on the basis of themodel covariates57
Besides regression modelling there are many otherapproaches in spatial epidemiology spatial clustering discrimin-ant analysis generalised linear models generalised additivemodels Bayesian estimation methods and others Howeversuch traditional models require both disease presence anddisease absence data At the same time there are differentlsquopresence-onlyrsquo models that could be used when no lsquoabsencedatarsquo are available One of these presence-only machine learning(rule-based) algorithms is an ecological niche modelling algo-rithm based on maximum entropy (Maxent) This method canbe used successfully with very small sample sizes and does notrequire independence of covariates58 Moreover these advan-tages could be crucial in Africa where complete surveillancedata from all regions are not available
Basic spatial modelling approaches include interpolationbased on spatial dependence in vector or VBD data and extrapo-lation based on associations between vector or VBD data and en-vironmental or socioeconomic predictor variables The latterapproach can be a powerful tool to gain insights into levels ofrisk within areas where surveillance data are lacking or unreli-able However it should be noted that model extrapolation isrestricted to areas with ecological and climatic characteristicssimilar to those of the model development area1
Global strategies for controlling infections in sub-SaharanAfrica have focused on the lsquobig threersquo diseases namely AIDSTB and malaria Other causes of fever and infection fall intothe category of NTDs In the twenty-first century it is importantto compile a comprehensive list of the infections in sub-SaharanAfrica and to study their epidemiology59
Few investigations have addressed the unknown causes offever in West Africa The majority of fevers have long been con-sidered to be related to malaria Recent studies have shown that
NTDs are the most frequent causes of fever in both indigenes andtravellers In particular TBRF due to B crocidurae has been redis-covered931 and the rickettsiae including R felis have beenreported33 Overall new pathogens among the rickettsiae wereidentified in the environment8
The geographic distribution of ticksmdashvectors of diseases inWest Africamdashis a neglected area of study However applyingGISRS technologies to the study of Ixodes ticks in North Americahas yielded significant results For example the use of geospatialmodelling has revealed that high concentrations of I pacificusand I scapularis which are the key tick vectors of Lyme diseasein North America can be predicted by GISRS-based environmen-tal factors related to elevation slope of the landscape vegetationtype soil type temperature and moisture60 The same approachescould be used in epidemiological studies of R felis infection anemerging disease in West Africa
Today epidemiologists often use new GISRS techniques tostudy a variety of VBDs The associations between satellite-derived environmental variables (such as temperature humidityelevation vegetation rainfall surface water land use land covertype and soil moisture) and vector density are used to identifyand characterise vector habitats A wide variety of RS datapowerful GIS and statistical software packages are available fordesktop computing environments making it affordable and feas-ible for epidemiologists to experiment with new spatial analysistechniques56
Maps that show the seasonal risk of VBDs will be necessary tomonitor the impacts of changes on vector ecology Using GISRSadvanced analytical tools and a landscape ecology approachthese risk maps could play a major role in defining research ques-tions and surveillance needs and in guiding control efforts andfield studies38
From the other side this approach is based on ecological ana-lysis and requires assembling all epidemiologic available dataand moreover it should be supported by fieldwork The most suc-cessful GISRS applications to study VBDs in West Africa areeither local (country)-level studies using community-based sur-veillance data or continental (sub-continental)-level studiesusing published scientific literature data Cooperation of the epi-demiological community could allow enhancing studies of VBDsin West Africa The most important goal of the application of GISRS technologies to the study of VBDs is to reduce the burden ofdisease by generating information that empowers the public totake protective action and helps public health agencies to effect-ively allocate their limited prevention surveillance and controlresources There is great potential to use GISRS technologiesto improve the surveillance prevention and control of vector-borne infectious diseases in West Africa
Supplementary dataSupplementary data are available at Transactions Online(httptrstmhoxfordjournalsorg)
Authorsrsquo contributions All authors contributed equally to themanuscript DR conceived the review PR and OM researched the dataand literature PR OM and DR analysed and interpreted the data for
P Ratmanov et al
10 of 12
at University of L
iverpool on May 9 2013
httptrstmhoxfordjournalsorg
Dow
nloaded from
the maps and discussed the content of the review PR and DR drafted themanuscript All authors critically revised the manuscript for intellectualcontent and read and approved the final version DR is guarantor ofthe paper
Funding None
Competing interests None declared
Ethical approval Not required
References1 Eisen L Eisen RJ Using geographic information systems and decision
support systems for the prediction prevention and control ofvector-borne diseases Annu Rev Entomol 20115641ndash61
2 WHO Neglected Tropical Diseases Geneva World HealthOrganization 2012 httpwwwwhointneglected_diseasesdiseasesen [accessed 8 January 2013]
3 Hay SI An overview of remote sensing and geodesy for epidemiologyand public health application Adv Parasitol 2000471ndash35
4 Pavlovsky EN Natural Nidality of Transmissible Diseases with SpecialReference to the Landscape Epidemiology of ZooanthroponosesUrbana IL University of Illinois Press 1966
5 Kitron U Landscape ecology and epidemiology of vector-bornediseases tools for spatial analysis J Med Entomol 199835435ndash45
6 Rogers DJ Randolph SE Studying the global distribution of infectiousdiseases using GIS and RS Nat Rev Microbiol 20031231ndash7
7 Eisen L Lozano-Fuentes S Use of mapping and spatial andspace-time modeling approaches in operational control of Aedesaegypti and dengue PLoS Negl Trop Dis 20093e411
8 Mediannikov O Diatta G Fenollar F et al Tick-borne rickettsiosesneglected emerging diseases in rural Senegal PLoS Negl Trop Dis20104e821
9 Parola P Diatta G Socolovschi C et al Tick-borne relapsing feverborreliosis rural Senegal Emerg Infect Dis 201117883ndash5
10 WHO World Malaria Report 2010 Geneva World HealthOrganization 2010
11 Sullivan D Uncertainty in mapping malaria epidemiologyimplications for control Epidemiol Rev 201032175ndash87
12 Machault V Vignolles C Borchi F et al The use of remotely sensedenvironmental data in the study of malaria Geospat Health20115151ndash68
13 Snow RW Marsh K le Sueur D The need for maps of transmissionintensity to guide malaria control in Africa Parasitol Today199612455ndash7
14 Guerra CA Hay SI Lucioparedes LS et al Assembling a globaldatabase of malaria parasite prevalence for the Malaria AtlasProject Malar J 2007617
15 Brun R Blum J Chappuis F Burri C Human African trypanosomiasisLancet 2010375148ndash59
16 Simarro PP Cecchi G Paone M et al The Atlas of human Africantrypanosomiasis a contribution to global mapping of neglectedtropical diseases Int J Health Geogr 2010957
17 Bern C Maguire JH Alvar J Complexities of assessing the diseaseburden attributable to leishmaniasis PLoS Negl Trop Dis 20082e313
18 WHO Report of the Scientific Working Group Meeting onLeishmaniasis (Geneva 2ndash4 February 2004) Geneva World HealthOrganization 2004
19 Faye B Bucheton B Banuls AL et al Seroprevalence of Leishmaniainfantum in a rural area of Senegal analysis of risk factors involvedin transmission to humans Trans R Soc Trop Med Hyg2011105333ndash40
20 Slater H Michael E Predicting the current and future potentialdistributions of lymphatic filariasis in Africa using maximumentropy ecological niche modelling PLoS One 20127e32202
21 Basanez MG Pion SD Churcher TS et al River blindness a successstory under threat PLoS Med 20063e371
22 Noma M Nwoke BE Nutall I et al Rapid epidemiological mapping ofonchocerciasis (REMO) its application by the African Programme forOnchocerciasis Control (APOC) Ann Trop Med Parasitol 200296(Suppl 1)S29ndash39
23 Zoure HG Wanji S Noma M et al The geographic distribution of Loaloa in Africa results of large-scale implementation of the RapidAssessment Procedure for Loiasis (RAPLOA) PLoS Negl Trop Dis20115e1210
24 Favier C Chalvet-Monfray K Sabatier P et al Rift Valley fever in WestAfrica the role of space in endemicity Trop Med Int Health2006111878ndash88
25 Jouan A Coulibaly I Adam F et al Analytical study of a Rift Valleyfever epidemic Res Virol 1989140175ndash86
26 Diallo M Ba Y Sall AA et al Amplification of the sylvatic cycle ofdengue virus type 2 Senegal 1999ndash2000 entomologicfindings and epidemiologic considerations Emerg Infect Dis20039362ndash7
27 Jentes ES Poumerol G Gershman MD et al The revised global yellowfever risk map and recommendations for vaccination 2010consensus of the Informal WHO Working Group on Geographic Riskfor Yellow Fever Lancet Infect Dis 201111622ndash32
28 Rogers DJ Wilson AJ Hay SI Graham AJ The global distribution ofyellow fever and dengue Adv Parasitol 200662181ndash220
29 Grard G Drexler JF Fair J et al Re-emergence of CrimeanndashCongohemorrhagic fever virus in Central Africa PLoS Negl Trop Dis20115e1350
30 Nabeth P Thior M Faye O Simon F Human CrimeanndashCongohemorrhagic fever Senegal Emerg Infect Dis 2004101881ndash2
31 Vial L Diatta G Tall A et al Incidence of tick-borne relapsing fever inWest Africa longitudinal study Lancet 200636837ndash43
32 Jensenius M Davis X von Sonnenburg F et al Multicenter GeoSentinelanalysis of rickettsial diseases in international travelers 1996ndash2008Emerg Infect Dis 2009151791ndash8
33 Socolovschi C Mediannikov O Sokhna C et al Rickettsiafelis-associated uneruptive fever Senegal Emerg Infect Dis2010161140ndash2
34 Cazorla C Socolovschi C Jensenius M Parola P Tick-borne diseasestick-borne spotted fever rickettsioses in Africa Infect Dis Clin NorthAm 200822p531ndash44ixndashx
35 Merritt RW Walker ED Small PL et al Ecology and transmission ofBuruli ulcer disease a systematic review PLoS Negl Trop Dis20104e911
36 Tissot-Dupont H Brouqui P Faugere B Raoult D Prevalence ofantibodies to Coxiella burnetti Rickettsia conorii and Rickettsia typhiin seven African countries Clin Infect Dis 1995211126ndash33
37 Mediannikov O Fenollar F Socolovschi C et al Coxiella burnetii inhumans and ticks in rural Senegal PLoS Negl Trop Dis 20104e654
38 Kitron U Risk maps transmission and burden of vector-bornediseases Parasitol Today 200016324ndash5
Transactions of the Royal Society of Tropical Medicine and Hygiene
11 of 12
at University of L
iverpool on May 9 2013
httptrstmhoxfordjournalsorg
Dow
nloaded from
39 Thomson MC Connor SJ Milligan P Flasse SP Mapping malaria risk inAfrica what can satellite data contribute Parasitol Today199713313ndash8
40 Matthys B Koudou BG NrsquoGoran EK et al Spatial dispersion andcharacterisation of mosquito breeding habitats in urbanvegetable-production areas of Abidjan Cote drsquoIvoire Ann Trop MedParasitol 2010104649ndash66
41 Machault V Vignolles C Pages F et al Spatial heterogeneity andtemporal evolution of malaria transmission risk in Dakar Senegalaccording to remotely sensed environmental data Malar J 20109252
42 Dambach P Sie A Lacaux JP et al Using high spatial resolutionremote sensing for risk mapping of malaria occurrence in theNouna District Burkina Faso Glob Health Action 20092 doi103402ghav2io2094
43 Bogh C Lindsay SW Clarke SE et al High spatial resolution mappingof malaria transmission risk in The Gambia West Africa usingLANDSAT TM satellite imagery Am J Trop Med Hyg 200776875ndash81
44 Bicout DJ Sabatier P Mapping Rift Valley fever vectors and prevalenceusing rainfall variations Vector Borne Zoonotic Dis 2004433ndash42
45 Clements AC Pfeiffer DU Martin V et al Spatial risk assessment of RiftValley fever in Senegal Vector Borne Zoonotic Dis 20077203ndash16
46 Rogers DJ Hay SI Packer MJ Predicting the distribution of tsetse fliesin West Africa using temporal Fourier processed meteorologicalsatellite data Ann Trop Med Parasitol 199690225ndash41
47 Cecchi G Mattioli RC Slingenbergh J de la Rocque S Land cover andtsetse fly distributions in sub-Saharan Africa Med Vet Entomol200822364ndash73
48 Thomson MC Connor SJ Environmental information systems for thecontrol of arthropod vectors of disease Med Vet Entomol200014227ndash44
49 Hay SI Tucker CJ Rogers DJ Packer MJ Remotely sensed surrogatesof meteorological data for the study of the distribution andabundance of arthropod vectors of disease Ann Trop Med Parasitol1996901ndash19
50 Randolph SE Rogers DJ A generic population model for theAfrican tick Rhipicephalus appendiculatus Parasitology 1997115265ndash79
51 Cumming GS Comparing climate and vegetation as limiting factorsfor species ranges of African ticks Ecology 200283255ndash68
52 Ostfeld RS Glass GE Keesing F Spatial epidemiology an emerging (orre-emerging) discipline Trends Ecol Evol 200520328ndash36
53 Hay SI Tatem AJ Graham AJ et al Global environmental data formapping infectious disease distribution Adv Parasitol 20066237ndash77
54 Riedel N Vounatsou P Miller JM et al Geographical patterns andpredictors of malaria risk in Zambia Bayesian geostatisticalmodelling of the 2006 Zambia national malaria indicator survey(ZMIS) Malar J 2010937
55 Gosoniu L Veta AM Vounatsou P Bayesian geostatisticalmodeling of Malaria Indicator Survey data in Angola PLoS One20105e9322
56 Kalluri S Gilruth P Rogers D Szczur M Surveillance of arthropodvector-borne infectious diseases using remote sensing techniquesa review PLoS Pathog 200731361ndash71
57 Clements AC Pfeiffer DU Emerging viral zoonoses frameworks forspatial and spatiotemporal risk assessment and resource planningVet J 200918221ndash30
58 Phillips SJ Anderson RP Schapire RE Maximum entropy modelingof species geographic distributions Ecological Modelling 2006190231ndash59
59 Fenollar F Neglected and emerging diseases in sub-Saharan AfricaClin Microbiol Infect 201117957ndash8
60 Eisen RJ Eisen L Lane RS Predicting density of Ixodes pacificusnymphs in dense woodlands in Mendocino County Californiabased on geographic information systems and remotesensing versus field-derived data Am J Trop Med Hyg 200674632ndash40
P Ratmanov et al
12 of 12
at University of L
iverpool on May 9 2013
httptrstmhoxfordjournalsorg
Dow
nloaded from
disease mapping The Atlas of HAT is the most prominent attemptat disease control research and advocacy16
Leishmaniasis
Leishmaniasis refers to a group of VBDs that are caused by morethan 20 species of the protozoan genus Leishmania rangingfrom localized skin ulcers to lethal systemic disease Leishmania-sis is considered one of the lsquomost neglected diseasesrsquo because
limited resources are invested in its diagnosis treatment andcontrol and it is strongly associated with poverty17 Humansare infected via the bite of phlebotomine sandflies (Phlebotomusspp) The leishmaniases can be classified into two epidemio-logical entities according to the type of transmission anthropo-notic if humans are the sole reservoir involved in transmission(and the sole source for vector infection) or zoonotic if atleast one mammalian reservoir is involved18 Although thereare no major geographic foci of leishmaniasis in West Africathe cutaneous forms of leishmaniasis have been reported in 11of 15 countries in the region (Figure 3) In recent studies theseroprevalence of specific antibodies against L infantum (theagent of visceral leishmaniasis) in the human population wasdetermined in Senegal Larger-scale studies with application ofGISRS technologies are now required to define the distributionof L infantum in West Africa and an identification of its potentialvector19
Filariasis
Lymphatic filariasis (LF) is a vectorborne parasitic infectiousdisease caused by Wuchereria bancrofti which is endemic inthe tropics including in sub-Saharan Africa It is transmitted tohumans by infected mosquitoes In West Africa the mosquitoAn gambiae sl is a vector of LF and malaria caused by P falcip-arum Humans are the only reservoir host of the LF parasite inAfrica20 Onchocerciasis (or lsquoriver blindnessrsquo) is a parasiticdisease caused by the filarial parasitic nematode Onchocerca vol-vulus It is transmitted through the bites of infected Simulium(black fly) vectors which breed in fast-flowing streams andrivers21 Control of onchocerciasis in sub-Saharan Africa is over-seen by the African Programme for Onchocerciasis Control(APOC)22 Loa loa filariasis (loiasis) is a NTD caused by the filarialparasite Loa loa It is an African disease restricted to the equator-ial rainforest regions of Central and West Africa23 The insectvectors of L loa are flies of the genus Chrysops Humans arethe primary reservoir for L loa LF loiasis and onchocerciasisare endemic in many countries of West Africa (Figure 4)
Rift Valley fever
Rift Valley fever (RVF) is an arthropodborne viral disease that pri-marily causes epizootics of abortion and high mortality rates indomestic animals but it can also infect humans The RVF virusis mainly transmitted by mosquitoes of the Aedes and Culexgenera to a wide range of animals from rodents to camelswhich are the natural reservoirs for RVF Although mosquitoesmay transmit the RVF virus to humans human infections resultfrom contact with infected animals24 A large RVF epidemic oc-curred in 1987 in southern Mauritania and northern SenegalHowever all of the countries in West Africa are at risk of RVF(Figure 5)25
Dengue and yellow fever
Dengue fever is a viral infectious tropical disease The mosquitoAe aegypti is the primary vector of the dengue virus andhumans are the most common reservoir Dengue is a widespreaddisease in the subtropics and tropics but in Africa the burden ofthe disease is poorly understood7 In West Africa sylvatic
Table 2 Selected vectors of the human vectorborne diseasesin West Africa
Vector Disease
MosquitoesAnopheles spp Lymphatic filariasisa
An gambiae sl MalariaAedes spp Rift Valley fever
DengueYellow feverChikungunya fever
Culex spp Rift Valley feverPhlebotomus
Phlebotomus spp LeishmaniasisSergentomyia (Spelaeomyia)darlingi
Leishmaniasis
Tsetse fliesGlossina palpalis sl Human African
trypanosomiasisG tachinoides Human African
trypanosomiasisTabanid flies
Chrysops spp Loa loa filariasis (loiasis)Black flies
Simulium spp Onchocerciasis (lsquoriverblindnessrsquo)
TicksRhipicephalus evertsi Mediterranean spotted fever
African tick-bite feverAmblyomma variegatum African tick-bite fever
CrimeanndashCongo haemorrhagicfever
Ornithodoros sonrai Tickborne relapsing feverOrnithodoros moubata Tickborne relapsing feverHyalomma spp CrimeanndashCongo haemorrhagic
feverLice
Pediculus humanus humanus Trench feverLouseborne relapsing feverEpidemic typhus
FleasCtenocephalides felis andothers
Rickettsia felis infectionb
aAnopheles is a primary vector of lymphatic filariasis in AfricabFleas are considered as vectors of R felis worldwide
Transactions of the Royal Society of Tropical Medicine and Hygiene
3 of 12
at University of L
iverpool on May 9 2013
httptrstmhoxfordjournalsorg
Dow
nloaded from
circulation of the dengue virus is the predominant form of circu-lation (with lower primates as the main reservoir)26 Evidenceregarding circulation of the dengue virus was obtained from 11countries in the region although the mosquito vectors arepresent throughout West Africa (Figure 6)
Yellow fever (YF) is an acute viral haemorrhagic disease trans-mitted in West Africa by infected Aedes spp mosquitoes Up to50 of severely affected persons who do not receive treatmentdie from YF and there is no cure The YF virus circulates both inurban and sylvatic settings involving several vertebrate speciesIn the sylvatic cycle mosquitoes act as the main vectors andmonkeys act as the primary hosts In the urban cycle the virusis transmitted between human beings and mosquitoes (predom-inately Ae aegypti) In Africa transmission of the virus can alsooccur in an intermediate cycle between human beings or non-human primates and Aedes spp mosquitoes that breed in treeholes on the savannah27 Vertical transmission also occurswithin the mosquito population and may play an importantrole in maintaining the sylvatic cycle28 In West Africa YF isholoendemic in all countries except in the Sahara Desertregions of Mali and Niger (Figure 7)27
CrimeanndashCongo haemorrhagic fever
CrimeanndashCongo haemorrhagic fever is a viral tickborne diseaseThe virus causes severe illness throughout the world includingWest Africa Its distribution closely matches that of its mainarthropod vector ixodid ticks belonging to the genus Hyalomma
Human infection occurs through tick bites contact with infectedlivestock or nosocomial transmission29 In West Africa enzooticcirculation of the CrimeanndashCongo haemorrhagic fever virus hasbeen shown in serological surveys of cattle but only sporadichuman cases have been reported (Figure 8)30
Borrelioses
In West Africa tickborne relapsing fever (TBRF) is caused by Bor-relia crocidurae Studies in Senegal indicate that TBRF is aftermalaria the most common cause of outpatient visits to a ruraldispensary Investigations found that the vector tick (Ornitho-doros sonrai) was present in villages in Senegal Mauritania andMali with high infection rates Rodents and insectivores are thereservoirs of B crocidurae in West Africa31 Unfortunately noattempts have been made to perform a spatial analysis ofTBRF prevalence in West Africa
Rickettsioses
Following malaria tickborne rickettsioses are one of the mostcommon causes of systemic febrile illness among travellersfrom developed countries but little is known about their preva-lence in indigenous populations especially in West Africa32
There is a high incidence (44) of Rickettsia felis in Senegal33
A case of Mediterranean spotted fever was described inSenegal and its agent R conorii was found in the ticks Rhipice-phalus evertsi evertsi8 There is a plethora of serological studies
Figure 1 Cases of malaria per 100 000 inhabitants in West African countries in 2010 or latest available year (data from the WHO retrieved fromhttpwwwwhoint)
P Ratmanov et al
4 of 12
at University of L
iverpool on May 9 2013
httptrstmhoxfordjournalsorg
Dow
nloaded from
of the prevalence of rickettsioses and entomological studies ofticks for Rickettsia spp Although attempts to map the geographicdistribution of rickettsioses in Africa have been made they simplydisplayed infected sites or counties with reported cases as pointswithout conducting any geostatistical analyses834 Thus thespatial aspects of borrelial and rickettsial VBDs in West Africahave not received the attention they deserve
Buruli ulcer
Buruli ulcer (BU) disease caused by Mycobacterium ulcerans isan emerging infectious disease in many tropical and subtropicalcountries The exact mode of transmission of this diseaseremains unclear Although vectors and modes of transmissionremain unknown it has been hypothesized that transmissionof BU disease is associated with human activities in andaround aquatic environments (the role of water bugs has beendiscussed) and that characteristics of the landscape play a rolein the spread of BU disease35 In West Africa Cote drsquoIvoireGhana and Benin are the countries that are particularly affectedby BU disease (Figure 9)
Q fever
In West Africa Q fever has a wide distribution which has beenshown repeatedly in human serological studies and studies ofreservoirs (domestic animals)36 It was reported that the main
vector of TBRF the soft tick O sonrai also harboured Coxiella bur-netii37 The possibility that this tick transmits Q fever was notstudied but some species of tick may play a role in the transmis-sion of Q fever
Geographic information systems and remotesensing for spatial analysis of vectordistributionThe geographic distribution of VBDs depends on the distributionof their vectors environmental factors and climatic factorsVectors depend on specific abiotic conditions and advances inRS and mapping of variations in abiotic conditions have stimu-lated efforts to create risk maps for VBDs538 This spatial con-cordance of environmental variables and vector distribution isoften used to estimate the current distribution of vectors in un-studied areas
There have been many studies of the entomological inocula-tion rate and parasite prevalence vector density and breedingsites and risk mapping and modelling in West Africa related tothe Anopheles mosquito The distribution of Anopheles sppwhich are the major vectors of LF and malaria in West Africawas studied in relation to normalised difference vegetationindex (NDVI) values39 Several studies were dedicated to investi-gating the risk factors for malaria and identifying the potentialAnopheles mosquito breeding sites in urban environments inCote drsquoIvoire and Senegal4041 In Burkina Faso and The
Figure 2 Geographic distribution of human African trypanosomiasis (Trypanosoma brucei gambiense) in West African countries in 2010 (data from theWHO retrieved from httpwwwwhoint)
Transactions of the Royal Society of Tropical Medicine and Hygiene
5 of 12
at University of L
iverpool on May 9 2013
httptrstmhoxfordjournalsorg
Dow
nloaded from
Gambia potential Anopheles breeding sites were mapped at thevillage level using satellite imagery4243
A critically important review regarding the potential for GISRStechnologies and methodologies to be used for the preventionsurveillance and control of the mosquito Ae aegypti thedengue virus vector can be found in an article by Eisen et al7
In Senegal a clear association between the amount of rainfallthe abundance of vectors and the prevalence of RVF has beendemonstrated44 In another study in Senegal the average totalmonthly rainfall from December to February was the most im-portant spatial predictor of the risk of RVF45
GIS and RS techniques are promising and powerful tools for de-scribing tsetse distribution because thermal data appeared to bethe most useful predictor variable followed by indices of vegeta-tion and rainfall46 Recently a study demonstrated the potentialof high-resolution images for mapping the habitat of Glossina ona local scale as well as in larger areas47
Applications of GISRS have been summarised with examplesof studies on various vectors such as the malaria vector Anoph-eles the arbovirus vector culicine mosquitoes (Aedes spp andCulex spp) the leishmaniasis vector Phlebotomus sandflies thetrypanosomiasis vector tsetse and the loiasis vector Chrysopsin a review by Thomson and Connor48
Previous studies have used multivariate analyses to estimatethe spatial prevalence of particular tick species and to makeinferences about different environmental variables that deter-mine their distribution in Africa4950 Another approach to predict-ing tick distribution was established using a set of methods in
which the author concluded that on average climatic variablesare better predictors of tick distribution than vegetation-relatedvariables and the key to describing tick distribution is the covari-ance of temperature and rainfall51
It should be taken into account that risk mapping based onvectors has serious limitations Primarily disease risk or incidenceis most closely correlated with the abundance of pathogen-infected vectors rather than simply the presence of vectors orthe total abundance of vectors52
Perspectives of application of geographicinformation systemsremote sensingtechnologies for studying vectorbornediseases remaining challenges and conclusionGIS facilitate comparisons between disease patterns and envir-onmental data while RS technologies can use high-resolutionsatellite data to provide estimates of variables such as tempera-ture vegetation and humidity53 During the past two decadessatellite RS technology has shown promising results in assessingthe risk of various VBDs at different spatial scales Satellite-basedimagery is characterised by its spatial spectral and temporalresolution Despite some limited attempts to apply RS to epi-demiology the methods have not yet demonstrated theirexpected potential GISRS technologies are undoubtedly a valu-able source of information for epidemiologists GISRS technolo-gies should not be overestimated and it is necessary to take all of
Figure 3 Geographic distribution of cutaneous leishmaniasis in West African countries in 2010 (data from the WHO retrieved from httpwwwwhoint)
P Ratmanov et al
6 of 12
at University of L
iverpool on May 9 2013
httptrstmhoxfordjournalsorg
Dow
nloaded from
Figure 4 Geographic distribution of lymphatic filariasis Loa loa filariasis (loiasis) and onchocerciasis in West African countries (data from the WHOretrieved from httpwwwwhoint)
Figure 5 Geographic distribution of Rift Valley fever (RVF) in West African countries (data from the WHO retrieved from httpwwwwhoint)
Transactions of the Royal Society of Tropical Medicine and Hygiene
7 of 12
at University of L
iverpool on May 9 2013
httptrstmhoxfordjournalsorg
Dow
nloaded from
Figure 6 Geographic distribution of dengue in West African countries in 2011 (data from the WHO retrieved from httpwwwwhoint)
Figure 7 Geographic distribution of yellow fever in West African countries in 2011 (data from the WHO retrieved from httpwwwwhoint)
P Ratmanov et al
8 of 12
at University of L
iverpool on May 9 2013
httptrstmhoxfordjournalsorg
Dow
nloaded from
Figure 8 Geographic distribution of CrimeanndashCongo haemorrhagic fever in West African countries (data from the WHO retrieved from httpwwwwhoint)
Figure 9 Geographic distribution of Buruli ulcer in 2010 in West African countries (data from the WHO retrieved from httpwwwwhoint)
Transactions of the Royal Society of Tropical Medicine and Hygiene
9 of 12
at University of L
iverpool on May 9 2013
httptrstmhoxfordjournalsorg
Dow
nloaded from
their advantages and limitations into account Moreover VBDinterventions have to be identified and recognised as they canconfound the relationship between the environment anddisease In the recent malaria indicator surveys in Zambia andAngola for example no relationship was observed between re-motely sensed data and malaria risk5455
The typical modelling approach investigates associationsbetween multivariate environmental data and patterns ofvector presence or absence for mapping vectors and VBDs Atthe first step simple statistical models could be a good startingpoint for linking the limited number of environmental variablesthat can be derived from satellite data Simple statisticalmodels are restricted because they often fit linear functionsbetween environmental variables and presenceabsence datawhen it is most likely that such associations are highlycomplex and non-linear20 At the second step sophisticatedprocess-based models that rely on vector biology as predictorsof diseases and their risk should be developed56
Regression models have been widely applied in landscape epi-demiology The type of regression model (logistic Poisson linearetc) is determined by the type of outcome variable to be pre-dicted (eg binary count continuous) and environmental vari-ables measured at sampled locations are entered ascovariates The resultant model is then either used to predictthe outcome variable at non-sampled locations based onobserved values of the covariates at the prediction locations orto explain observed patterns of disease on the basis of themodel covariates57
Besides regression modelling there are many otherapproaches in spatial epidemiology spatial clustering discrimin-ant analysis generalised linear models generalised additivemodels Bayesian estimation methods and others Howeversuch traditional models require both disease presence anddisease absence data At the same time there are differentlsquopresence-onlyrsquo models that could be used when no lsquoabsencedatarsquo are available One of these presence-only machine learning(rule-based) algorithms is an ecological niche modelling algo-rithm based on maximum entropy (Maxent) This method canbe used successfully with very small sample sizes and does notrequire independence of covariates58 Moreover these advan-tages could be crucial in Africa where complete surveillancedata from all regions are not available
Basic spatial modelling approaches include interpolationbased on spatial dependence in vector or VBD data and extrapo-lation based on associations between vector or VBD data and en-vironmental or socioeconomic predictor variables The latterapproach can be a powerful tool to gain insights into levels ofrisk within areas where surveillance data are lacking or unreli-able However it should be noted that model extrapolation isrestricted to areas with ecological and climatic characteristicssimilar to those of the model development area1
Global strategies for controlling infections in sub-SaharanAfrica have focused on the lsquobig threersquo diseases namely AIDSTB and malaria Other causes of fever and infection fall intothe category of NTDs In the twenty-first century it is importantto compile a comprehensive list of the infections in sub-SaharanAfrica and to study their epidemiology59
Few investigations have addressed the unknown causes offever in West Africa The majority of fevers have long been con-sidered to be related to malaria Recent studies have shown that
NTDs are the most frequent causes of fever in both indigenes andtravellers In particular TBRF due to B crocidurae has been redis-covered931 and the rickettsiae including R felis have beenreported33 Overall new pathogens among the rickettsiae wereidentified in the environment8
The geographic distribution of ticksmdashvectors of diseases inWest Africamdashis a neglected area of study However applyingGISRS technologies to the study of Ixodes ticks in North Americahas yielded significant results For example the use of geospatialmodelling has revealed that high concentrations of I pacificusand I scapularis which are the key tick vectors of Lyme diseasein North America can be predicted by GISRS-based environmen-tal factors related to elevation slope of the landscape vegetationtype soil type temperature and moisture60 The same approachescould be used in epidemiological studies of R felis infection anemerging disease in West Africa
Today epidemiologists often use new GISRS techniques tostudy a variety of VBDs The associations between satellite-derived environmental variables (such as temperature humidityelevation vegetation rainfall surface water land use land covertype and soil moisture) and vector density are used to identifyand characterise vector habitats A wide variety of RS datapowerful GIS and statistical software packages are available fordesktop computing environments making it affordable and feas-ible for epidemiologists to experiment with new spatial analysistechniques56
Maps that show the seasonal risk of VBDs will be necessary tomonitor the impacts of changes on vector ecology Using GISRSadvanced analytical tools and a landscape ecology approachthese risk maps could play a major role in defining research ques-tions and surveillance needs and in guiding control efforts andfield studies38
From the other side this approach is based on ecological ana-lysis and requires assembling all epidemiologic available dataand moreover it should be supported by fieldwork The most suc-cessful GISRS applications to study VBDs in West Africa areeither local (country)-level studies using community-based sur-veillance data or continental (sub-continental)-level studiesusing published scientific literature data Cooperation of the epi-demiological community could allow enhancing studies of VBDsin West Africa The most important goal of the application of GISRS technologies to the study of VBDs is to reduce the burden ofdisease by generating information that empowers the public totake protective action and helps public health agencies to effect-ively allocate their limited prevention surveillance and controlresources There is great potential to use GISRS technologiesto improve the surveillance prevention and control of vector-borne infectious diseases in West Africa
Supplementary dataSupplementary data are available at Transactions Online(httptrstmhoxfordjournalsorg)
Authorsrsquo contributions All authors contributed equally to themanuscript DR conceived the review PR and OM researched the dataand literature PR OM and DR analysed and interpreted the data for
P Ratmanov et al
10 of 12
at University of L
iverpool on May 9 2013
httptrstmhoxfordjournalsorg
Dow
nloaded from
the maps and discussed the content of the review PR and DR drafted themanuscript All authors critically revised the manuscript for intellectualcontent and read and approved the final version DR is guarantor ofthe paper
Funding None
Competing interests None declared
Ethical approval Not required
References1 Eisen L Eisen RJ Using geographic information systems and decision
support systems for the prediction prevention and control ofvector-borne diseases Annu Rev Entomol 20115641ndash61
2 WHO Neglected Tropical Diseases Geneva World HealthOrganization 2012 httpwwwwhointneglected_diseasesdiseasesen [accessed 8 January 2013]
3 Hay SI An overview of remote sensing and geodesy for epidemiologyand public health application Adv Parasitol 2000471ndash35
4 Pavlovsky EN Natural Nidality of Transmissible Diseases with SpecialReference to the Landscape Epidemiology of ZooanthroponosesUrbana IL University of Illinois Press 1966
5 Kitron U Landscape ecology and epidemiology of vector-bornediseases tools for spatial analysis J Med Entomol 199835435ndash45
6 Rogers DJ Randolph SE Studying the global distribution of infectiousdiseases using GIS and RS Nat Rev Microbiol 20031231ndash7
7 Eisen L Lozano-Fuentes S Use of mapping and spatial andspace-time modeling approaches in operational control of Aedesaegypti and dengue PLoS Negl Trop Dis 20093e411
8 Mediannikov O Diatta G Fenollar F et al Tick-borne rickettsiosesneglected emerging diseases in rural Senegal PLoS Negl Trop Dis20104e821
9 Parola P Diatta G Socolovschi C et al Tick-borne relapsing feverborreliosis rural Senegal Emerg Infect Dis 201117883ndash5
10 WHO World Malaria Report 2010 Geneva World HealthOrganization 2010
11 Sullivan D Uncertainty in mapping malaria epidemiologyimplications for control Epidemiol Rev 201032175ndash87
12 Machault V Vignolles C Borchi F et al The use of remotely sensedenvironmental data in the study of malaria Geospat Health20115151ndash68
13 Snow RW Marsh K le Sueur D The need for maps of transmissionintensity to guide malaria control in Africa Parasitol Today199612455ndash7
14 Guerra CA Hay SI Lucioparedes LS et al Assembling a globaldatabase of malaria parasite prevalence for the Malaria AtlasProject Malar J 2007617
15 Brun R Blum J Chappuis F Burri C Human African trypanosomiasisLancet 2010375148ndash59
16 Simarro PP Cecchi G Paone M et al The Atlas of human Africantrypanosomiasis a contribution to global mapping of neglectedtropical diseases Int J Health Geogr 2010957
17 Bern C Maguire JH Alvar J Complexities of assessing the diseaseburden attributable to leishmaniasis PLoS Negl Trop Dis 20082e313
18 WHO Report of the Scientific Working Group Meeting onLeishmaniasis (Geneva 2ndash4 February 2004) Geneva World HealthOrganization 2004
19 Faye B Bucheton B Banuls AL et al Seroprevalence of Leishmaniainfantum in a rural area of Senegal analysis of risk factors involvedin transmission to humans Trans R Soc Trop Med Hyg2011105333ndash40
20 Slater H Michael E Predicting the current and future potentialdistributions of lymphatic filariasis in Africa using maximumentropy ecological niche modelling PLoS One 20127e32202
21 Basanez MG Pion SD Churcher TS et al River blindness a successstory under threat PLoS Med 20063e371
22 Noma M Nwoke BE Nutall I et al Rapid epidemiological mapping ofonchocerciasis (REMO) its application by the African Programme forOnchocerciasis Control (APOC) Ann Trop Med Parasitol 200296(Suppl 1)S29ndash39
23 Zoure HG Wanji S Noma M et al The geographic distribution of Loaloa in Africa results of large-scale implementation of the RapidAssessment Procedure for Loiasis (RAPLOA) PLoS Negl Trop Dis20115e1210
24 Favier C Chalvet-Monfray K Sabatier P et al Rift Valley fever in WestAfrica the role of space in endemicity Trop Med Int Health2006111878ndash88
25 Jouan A Coulibaly I Adam F et al Analytical study of a Rift Valleyfever epidemic Res Virol 1989140175ndash86
26 Diallo M Ba Y Sall AA et al Amplification of the sylvatic cycle ofdengue virus type 2 Senegal 1999ndash2000 entomologicfindings and epidemiologic considerations Emerg Infect Dis20039362ndash7
27 Jentes ES Poumerol G Gershman MD et al The revised global yellowfever risk map and recommendations for vaccination 2010consensus of the Informal WHO Working Group on Geographic Riskfor Yellow Fever Lancet Infect Dis 201111622ndash32
28 Rogers DJ Wilson AJ Hay SI Graham AJ The global distribution ofyellow fever and dengue Adv Parasitol 200662181ndash220
29 Grard G Drexler JF Fair J et al Re-emergence of CrimeanndashCongohemorrhagic fever virus in Central Africa PLoS Negl Trop Dis20115e1350
30 Nabeth P Thior M Faye O Simon F Human CrimeanndashCongohemorrhagic fever Senegal Emerg Infect Dis 2004101881ndash2
31 Vial L Diatta G Tall A et al Incidence of tick-borne relapsing fever inWest Africa longitudinal study Lancet 200636837ndash43
32 Jensenius M Davis X von Sonnenburg F et al Multicenter GeoSentinelanalysis of rickettsial diseases in international travelers 1996ndash2008Emerg Infect Dis 2009151791ndash8
33 Socolovschi C Mediannikov O Sokhna C et al Rickettsiafelis-associated uneruptive fever Senegal Emerg Infect Dis2010161140ndash2
34 Cazorla C Socolovschi C Jensenius M Parola P Tick-borne diseasestick-borne spotted fever rickettsioses in Africa Infect Dis Clin NorthAm 200822p531ndash44ixndashx
35 Merritt RW Walker ED Small PL et al Ecology and transmission ofBuruli ulcer disease a systematic review PLoS Negl Trop Dis20104e911
36 Tissot-Dupont H Brouqui P Faugere B Raoult D Prevalence ofantibodies to Coxiella burnetti Rickettsia conorii and Rickettsia typhiin seven African countries Clin Infect Dis 1995211126ndash33
37 Mediannikov O Fenollar F Socolovschi C et al Coxiella burnetii inhumans and ticks in rural Senegal PLoS Negl Trop Dis 20104e654
38 Kitron U Risk maps transmission and burden of vector-bornediseases Parasitol Today 200016324ndash5
Transactions of the Royal Society of Tropical Medicine and Hygiene
11 of 12
at University of L
iverpool on May 9 2013
httptrstmhoxfordjournalsorg
Dow
nloaded from
39 Thomson MC Connor SJ Milligan P Flasse SP Mapping malaria risk inAfrica what can satellite data contribute Parasitol Today199713313ndash8
40 Matthys B Koudou BG NrsquoGoran EK et al Spatial dispersion andcharacterisation of mosquito breeding habitats in urbanvegetable-production areas of Abidjan Cote drsquoIvoire Ann Trop MedParasitol 2010104649ndash66
41 Machault V Vignolles C Pages F et al Spatial heterogeneity andtemporal evolution of malaria transmission risk in Dakar Senegalaccording to remotely sensed environmental data Malar J 20109252
42 Dambach P Sie A Lacaux JP et al Using high spatial resolutionremote sensing for risk mapping of malaria occurrence in theNouna District Burkina Faso Glob Health Action 20092 doi103402ghav2io2094
43 Bogh C Lindsay SW Clarke SE et al High spatial resolution mappingof malaria transmission risk in The Gambia West Africa usingLANDSAT TM satellite imagery Am J Trop Med Hyg 200776875ndash81
44 Bicout DJ Sabatier P Mapping Rift Valley fever vectors and prevalenceusing rainfall variations Vector Borne Zoonotic Dis 2004433ndash42
45 Clements AC Pfeiffer DU Martin V et al Spatial risk assessment of RiftValley fever in Senegal Vector Borne Zoonotic Dis 20077203ndash16
46 Rogers DJ Hay SI Packer MJ Predicting the distribution of tsetse fliesin West Africa using temporal Fourier processed meteorologicalsatellite data Ann Trop Med Parasitol 199690225ndash41
47 Cecchi G Mattioli RC Slingenbergh J de la Rocque S Land cover andtsetse fly distributions in sub-Saharan Africa Med Vet Entomol200822364ndash73
48 Thomson MC Connor SJ Environmental information systems for thecontrol of arthropod vectors of disease Med Vet Entomol200014227ndash44
49 Hay SI Tucker CJ Rogers DJ Packer MJ Remotely sensed surrogatesof meteorological data for the study of the distribution andabundance of arthropod vectors of disease Ann Trop Med Parasitol1996901ndash19
50 Randolph SE Rogers DJ A generic population model for theAfrican tick Rhipicephalus appendiculatus Parasitology 1997115265ndash79
51 Cumming GS Comparing climate and vegetation as limiting factorsfor species ranges of African ticks Ecology 200283255ndash68
52 Ostfeld RS Glass GE Keesing F Spatial epidemiology an emerging (orre-emerging) discipline Trends Ecol Evol 200520328ndash36
53 Hay SI Tatem AJ Graham AJ et al Global environmental data formapping infectious disease distribution Adv Parasitol 20066237ndash77
54 Riedel N Vounatsou P Miller JM et al Geographical patterns andpredictors of malaria risk in Zambia Bayesian geostatisticalmodelling of the 2006 Zambia national malaria indicator survey(ZMIS) Malar J 2010937
55 Gosoniu L Veta AM Vounatsou P Bayesian geostatisticalmodeling of Malaria Indicator Survey data in Angola PLoS One20105e9322
56 Kalluri S Gilruth P Rogers D Szczur M Surveillance of arthropodvector-borne infectious diseases using remote sensing techniquesa review PLoS Pathog 200731361ndash71
57 Clements AC Pfeiffer DU Emerging viral zoonoses frameworks forspatial and spatiotemporal risk assessment and resource planningVet J 200918221ndash30
58 Phillips SJ Anderson RP Schapire RE Maximum entropy modelingof species geographic distributions Ecological Modelling 2006190231ndash59
59 Fenollar F Neglected and emerging diseases in sub-Saharan AfricaClin Microbiol Infect 201117957ndash8
60 Eisen RJ Eisen L Lane RS Predicting density of Ixodes pacificusnymphs in dense woodlands in Mendocino County Californiabased on geographic information systems and remotesensing versus field-derived data Am J Trop Med Hyg 200674632ndash40
P Ratmanov et al
12 of 12
at University of L
iverpool on May 9 2013
httptrstmhoxfordjournalsorg
Dow
nloaded from
circulation of the dengue virus is the predominant form of circu-lation (with lower primates as the main reservoir)26 Evidenceregarding circulation of the dengue virus was obtained from 11countries in the region although the mosquito vectors arepresent throughout West Africa (Figure 6)
Yellow fever (YF) is an acute viral haemorrhagic disease trans-mitted in West Africa by infected Aedes spp mosquitoes Up to50 of severely affected persons who do not receive treatmentdie from YF and there is no cure The YF virus circulates both inurban and sylvatic settings involving several vertebrate speciesIn the sylvatic cycle mosquitoes act as the main vectors andmonkeys act as the primary hosts In the urban cycle the virusis transmitted between human beings and mosquitoes (predom-inately Ae aegypti) In Africa transmission of the virus can alsooccur in an intermediate cycle between human beings or non-human primates and Aedes spp mosquitoes that breed in treeholes on the savannah27 Vertical transmission also occurswithin the mosquito population and may play an importantrole in maintaining the sylvatic cycle28 In West Africa YF isholoendemic in all countries except in the Sahara Desertregions of Mali and Niger (Figure 7)27
CrimeanndashCongo haemorrhagic fever
CrimeanndashCongo haemorrhagic fever is a viral tickborne diseaseThe virus causes severe illness throughout the world includingWest Africa Its distribution closely matches that of its mainarthropod vector ixodid ticks belonging to the genus Hyalomma
Human infection occurs through tick bites contact with infectedlivestock or nosocomial transmission29 In West Africa enzooticcirculation of the CrimeanndashCongo haemorrhagic fever virus hasbeen shown in serological surveys of cattle but only sporadichuman cases have been reported (Figure 8)30
Borrelioses
In West Africa tickborne relapsing fever (TBRF) is caused by Bor-relia crocidurae Studies in Senegal indicate that TBRF is aftermalaria the most common cause of outpatient visits to a ruraldispensary Investigations found that the vector tick (Ornitho-doros sonrai) was present in villages in Senegal Mauritania andMali with high infection rates Rodents and insectivores are thereservoirs of B crocidurae in West Africa31 Unfortunately noattempts have been made to perform a spatial analysis ofTBRF prevalence in West Africa
Rickettsioses
Following malaria tickborne rickettsioses are one of the mostcommon causes of systemic febrile illness among travellersfrom developed countries but little is known about their preva-lence in indigenous populations especially in West Africa32
There is a high incidence (44) of Rickettsia felis in Senegal33
A case of Mediterranean spotted fever was described inSenegal and its agent R conorii was found in the ticks Rhipice-phalus evertsi evertsi8 There is a plethora of serological studies
Figure 1 Cases of malaria per 100 000 inhabitants in West African countries in 2010 or latest available year (data from the WHO retrieved fromhttpwwwwhoint)
P Ratmanov et al
4 of 12
at University of L
iverpool on May 9 2013
httptrstmhoxfordjournalsorg
Dow
nloaded from
of the prevalence of rickettsioses and entomological studies ofticks for Rickettsia spp Although attempts to map the geographicdistribution of rickettsioses in Africa have been made they simplydisplayed infected sites or counties with reported cases as pointswithout conducting any geostatistical analyses834 Thus thespatial aspects of borrelial and rickettsial VBDs in West Africahave not received the attention they deserve
Buruli ulcer
Buruli ulcer (BU) disease caused by Mycobacterium ulcerans isan emerging infectious disease in many tropical and subtropicalcountries The exact mode of transmission of this diseaseremains unclear Although vectors and modes of transmissionremain unknown it has been hypothesized that transmissionof BU disease is associated with human activities in andaround aquatic environments (the role of water bugs has beendiscussed) and that characteristics of the landscape play a rolein the spread of BU disease35 In West Africa Cote drsquoIvoireGhana and Benin are the countries that are particularly affectedby BU disease (Figure 9)
Q fever
In West Africa Q fever has a wide distribution which has beenshown repeatedly in human serological studies and studies ofreservoirs (domestic animals)36 It was reported that the main
vector of TBRF the soft tick O sonrai also harboured Coxiella bur-netii37 The possibility that this tick transmits Q fever was notstudied but some species of tick may play a role in the transmis-sion of Q fever
Geographic information systems and remotesensing for spatial analysis of vectordistributionThe geographic distribution of VBDs depends on the distributionof their vectors environmental factors and climatic factorsVectors depend on specific abiotic conditions and advances inRS and mapping of variations in abiotic conditions have stimu-lated efforts to create risk maps for VBDs538 This spatial con-cordance of environmental variables and vector distribution isoften used to estimate the current distribution of vectors in un-studied areas
There have been many studies of the entomological inocula-tion rate and parasite prevalence vector density and breedingsites and risk mapping and modelling in West Africa related tothe Anopheles mosquito The distribution of Anopheles sppwhich are the major vectors of LF and malaria in West Africawas studied in relation to normalised difference vegetationindex (NDVI) values39 Several studies were dedicated to investi-gating the risk factors for malaria and identifying the potentialAnopheles mosquito breeding sites in urban environments inCote drsquoIvoire and Senegal4041 In Burkina Faso and The
Figure 2 Geographic distribution of human African trypanosomiasis (Trypanosoma brucei gambiense) in West African countries in 2010 (data from theWHO retrieved from httpwwwwhoint)
Transactions of the Royal Society of Tropical Medicine and Hygiene
5 of 12
at University of L
iverpool on May 9 2013
httptrstmhoxfordjournalsorg
Dow
nloaded from
Gambia potential Anopheles breeding sites were mapped at thevillage level using satellite imagery4243
A critically important review regarding the potential for GISRStechnologies and methodologies to be used for the preventionsurveillance and control of the mosquito Ae aegypti thedengue virus vector can be found in an article by Eisen et al7
In Senegal a clear association between the amount of rainfallthe abundance of vectors and the prevalence of RVF has beendemonstrated44 In another study in Senegal the average totalmonthly rainfall from December to February was the most im-portant spatial predictor of the risk of RVF45
GIS and RS techniques are promising and powerful tools for de-scribing tsetse distribution because thermal data appeared to bethe most useful predictor variable followed by indices of vegeta-tion and rainfall46 Recently a study demonstrated the potentialof high-resolution images for mapping the habitat of Glossina ona local scale as well as in larger areas47
Applications of GISRS have been summarised with examplesof studies on various vectors such as the malaria vector Anoph-eles the arbovirus vector culicine mosquitoes (Aedes spp andCulex spp) the leishmaniasis vector Phlebotomus sandflies thetrypanosomiasis vector tsetse and the loiasis vector Chrysopsin a review by Thomson and Connor48
Previous studies have used multivariate analyses to estimatethe spatial prevalence of particular tick species and to makeinferences about different environmental variables that deter-mine their distribution in Africa4950 Another approach to predict-ing tick distribution was established using a set of methods in
which the author concluded that on average climatic variablesare better predictors of tick distribution than vegetation-relatedvariables and the key to describing tick distribution is the covari-ance of temperature and rainfall51
It should be taken into account that risk mapping based onvectors has serious limitations Primarily disease risk or incidenceis most closely correlated with the abundance of pathogen-infected vectors rather than simply the presence of vectors orthe total abundance of vectors52
Perspectives of application of geographicinformation systemsremote sensingtechnologies for studying vectorbornediseases remaining challenges and conclusionGIS facilitate comparisons between disease patterns and envir-onmental data while RS technologies can use high-resolutionsatellite data to provide estimates of variables such as tempera-ture vegetation and humidity53 During the past two decadessatellite RS technology has shown promising results in assessingthe risk of various VBDs at different spatial scales Satellite-basedimagery is characterised by its spatial spectral and temporalresolution Despite some limited attempts to apply RS to epi-demiology the methods have not yet demonstrated theirexpected potential GISRS technologies are undoubtedly a valu-able source of information for epidemiologists GISRS technolo-gies should not be overestimated and it is necessary to take all of
Figure 3 Geographic distribution of cutaneous leishmaniasis in West African countries in 2010 (data from the WHO retrieved from httpwwwwhoint)
P Ratmanov et al
6 of 12
at University of L
iverpool on May 9 2013
httptrstmhoxfordjournalsorg
Dow
nloaded from
Figure 4 Geographic distribution of lymphatic filariasis Loa loa filariasis (loiasis) and onchocerciasis in West African countries (data from the WHOretrieved from httpwwwwhoint)
Figure 5 Geographic distribution of Rift Valley fever (RVF) in West African countries (data from the WHO retrieved from httpwwwwhoint)
Transactions of the Royal Society of Tropical Medicine and Hygiene
7 of 12
at University of L
iverpool on May 9 2013
httptrstmhoxfordjournalsorg
Dow
nloaded from
Figure 6 Geographic distribution of dengue in West African countries in 2011 (data from the WHO retrieved from httpwwwwhoint)
Figure 7 Geographic distribution of yellow fever in West African countries in 2011 (data from the WHO retrieved from httpwwwwhoint)
P Ratmanov et al
8 of 12
at University of L
iverpool on May 9 2013
httptrstmhoxfordjournalsorg
Dow
nloaded from
Figure 8 Geographic distribution of CrimeanndashCongo haemorrhagic fever in West African countries (data from the WHO retrieved from httpwwwwhoint)
Figure 9 Geographic distribution of Buruli ulcer in 2010 in West African countries (data from the WHO retrieved from httpwwwwhoint)
Transactions of the Royal Society of Tropical Medicine and Hygiene
9 of 12
at University of L
iverpool on May 9 2013
httptrstmhoxfordjournalsorg
Dow
nloaded from
their advantages and limitations into account Moreover VBDinterventions have to be identified and recognised as they canconfound the relationship between the environment anddisease In the recent malaria indicator surveys in Zambia andAngola for example no relationship was observed between re-motely sensed data and malaria risk5455
The typical modelling approach investigates associationsbetween multivariate environmental data and patterns ofvector presence or absence for mapping vectors and VBDs Atthe first step simple statistical models could be a good startingpoint for linking the limited number of environmental variablesthat can be derived from satellite data Simple statisticalmodels are restricted because they often fit linear functionsbetween environmental variables and presenceabsence datawhen it is most likely that such associations are highlycomplex and non-linear20 At the second step sophisticatedprocess-based models that rely on vector biology as predictorsof diseases and their risk should be developed56
Regression models have been widely applied in landscape epi-demiology The type of regression model (logistic Poisson linearetc) is determined by the type of outcome variable to be pre-dicted (eg binary count continuous) and environmental vari-ables measured at sampled locations are entered ascovariates The resultant model is then either used to predictthe outcome variable at non-sampled locations based onobserved values of the covariates at the prediction locations orto explain observed patterns of disease on the basis of themodel covariates57
Besides regression modelling there are many otherapproaches in spatial epidemiology spatial clustering discrimin-ant analysis generalised linear models generalised additivemodels Bayesian estimation methods and others Howeversuch traditional models require both disease presence anddisease absence data At the same time there are differentlsquopresence-onlyrsquo models that could be used when no lsquoabsencedatarsquo are available One of these presence-only machine learning(rule-based) algorithms is an ecological niche modelling algo-rithm based on maximum entropy (Maxent) This method canbe used successfully with very small sample sizes and does notrequire independence of covariates58 Moreover these advan-tages could be crucial in Africa where complete surveillancedata from all regions are not available
Basic spatial modelling approaches include interpolationbased on spatial dependence in vector or VBD data and extrapo-lation based on associations between vector or VBD data and en-vironmental or socioeconomic predictor variables The latterapproach can be a powerful tool to gain insights into levels ofrisk within areas where surveillance data are lacking or unreli-able However it should be noted that model extrapolation isrestricted to areas with ecological and climatic characteristicssimilar to those of the model development area1
Global strategies for controlling infections in sub-SaharanAfrica have focused on the lsquobig threersquo diseases namely AIDSTB and malaria Other causes of fever and infection fall intothe category of NTDs In the twenty-first century it is importantto compile a comprehensive list of the infections in sub-SaharanAfrica and to study their epidemiology59
Few investigations have addressed the unknown causes offever in West Africa The majority of fevers have long been con-sidered to be related to malaria Recent studies have shown that
NTDs are the most frequent causes of fever in both indigenes andtravellers In particular TBRF due to B crocidurae has been redis-covered931 and the rickettsiae including R felis have beenreported33 Overall new pathogens among the rickettsiae wereidentified in the environment8
The geographic distribution of ticksmdashvectors of diseases inWest Africamdashis a neglected area of study However applyingGISRS technologies to the study of Ixodes ticks in North Americahas yielded significant results For example the use of geospatialmodelling has revealed that high concentrations of I pacificusand I scapularis which are the key tick vectors of Lyme diseasein North America can be predicted by GISRS-based environmen-tal factors related to elevation slope of the landscape vegetationtype soil type temperature and moisture60 The same approachescould be used in epidemiological studies of R felis infection anemerging disease in West Africa
Today epidemiologists often use new GISRS techniques tostudy a variety of VBDs The associations between satellite-derived environmental variables (such as temperature humidityelevation vegetation rainfall surface water land use land covertype and soil moisture) and vector density are used to identifyand characterise vector habitats A wide variety of RS datapowerful GIS and statistical software packages are available fordesktop computing environments making it affordable and feas-ible for epidemiologists to experiment with new spatial analysistechniques56
Maps that show the seasonal risk of VBDs will be necessary tomonitor the impacts of changes on vector ecology Using GISRSadvanced analytical tools and a landscape ecology approachthese risk maps could play a major role in defining research ques-tions and surveillance needs and in guiding control efforts andfield studies38
From the other side this approach is based on ecological ana-lysis and requires assembling all epidemiologic available dataand moreover it should be supported by fieldwork The most suc-cessful GISRS applications to study VBDs in West Africa areeither local (country)-level studies using community-based sur-veillance data or continental (sub-continental)-level studiesusing published scientific literature data Cooperation of the epi-demiological community could allow enhancing studies of VBDsin West Africa The most important goal of the application of GISRS technologies to the study of VBDs is to reduce the burden ofdisease by generating information that empowers the public totake protective action and helps public health agencies to effect-ively allocate their limited prevention surveillance and controlresources There is great potential to use GISRS technologiesto improve the surveillance prevention and control of vector-borne infectious diseases in West Africa
Supplementary dataSupplementary data are available at Transactions Online(httptrstmhoxfordjournalsorg)
Authorsrsquo contributions All authors contributed equally to themanuscript DR conceived the review PR and OM researched the dataand literature PR OM and DR analysed and interpreted the data for
P Ratmanov et al
10 of 12
at University of L
iverpool on May 9 2013
httptrstmhoxfordjournalsorg
Dow
nloaded from
the maps and discussed the content of the review PR and DR drafted themanuscript All authors critically revised the manuscript for intellectualcontent and read and approved the final version DR is guarantor ofthe paper
Funding None
Competing interests None declared
Ethical approval Not required
References1 Eisen L Eisen RJ Using geographic information systems and decision
support systems for the prediction prevention and control ofvector-borne diseases Annu Rev Entomol 20115641ndash61
2 WHO Neglected Tropical Diseases Geneva World HealthOrganization 2012 httpwwwwhointneglected_diseasesdiseasesen [accessed 8 January 2013]
3 Hay SI An overview of remote sensing and geodesy for epidemiologyand public health application Adv Parasitol 2000471ndash35
4 Pavlovsky EN Natural Nidality of Transmissible Diseases with SpecialReference to the Landscape Epidemiology of ZooanthroponosesUrbana IL University of Illinois Press 1966
5 Kitron U Landscape ecology and epidemiology of vector-bornediseases tools for spatial analysis J Med Entomol 199835435ndash45
6 Rogers DJ Randolph SE Studying the global distribution of infectiousdiseases using GIS and RS Nat Rev Microbiol 20031231ndash7
7 Eisen L Lozano-Fuentes S Use of mapping and spatial andspace-time modeling approaches in operational control of Aedesaegypti and dengue PLoS Negl Trop Dis 20093e411
8 Mediannikov O Diatta G Fenollar F et al Tick-borne rickettsiosesneglected emerging diseases in rural Senegal PLoS Negl Trop Dis20104e821
9 Parola P Diatta G Socolovschi C et al Tick-borne relapsing feverborreliosis rural Senegal Emerg Infect Dis 201117883ndash5
10 WHO World Malaria Report 2010 Geneva World HealthOrganization 2010
11 Sullivan D Uncertainty in mapping malaria epidemiologyimplications for control Epidemiol Rev 201032175ndash87
12 Machault V Vignolles C Borchi F et al The use of remotely sensedenvironmental data in the study of malaria Geospat Health20115151ndash68
13 Snow RW Marsh K le Sueur D The need for maps of transmissionintensity to guide malaria control in Africa Parasitol Today199612455ndash7
14 Guerra CA Hay SI Lucioparedes LS et al Assembling a globaldatabase of malaria parasite prevalence for the Malaria AtlasProject Malar J 2007617
15 Brun R Blum J Chappuis F Burri C Human African trypanosomiasisLancet 2010375148ndash59
16 Simarro PP Cecchi G Paone M et al The Atlas of human Africantrypanosomiasis a contribution to global mapping of neglectedtropical diseases Int J Health Geogr 2010957
17 Bern C Maguire JH Alvar J Complexities of assessing the diseaseburden attributable to leishmaniasis PLoS Negl Trop Dis 20082e313
18 WHO Report of the Scientific Working Group Meeting onLeishmaniasis (Geneva 2ndash4 February 2004) Geneva World HealthOrganization 2004
19 Faye B Bucheton B Banuls AL et al Seroprevalence of Leishmaniainfantum in a rural area of Senegal analysis of risk factors involvedin transmission to humans Trans R Soc Trop Med Hyg2011105333ndash40
20 Slater H Michael E Predicting the current and future potentialdistributions of lymphatic filariasis in Africa using maximumentropy ecological niche modelling PLoS One 20127e32202
21 Basanez MG Pion SD Churcher TS et al River blindness a successstory under threat PLoS Med 20063e371
22 Noma M Nwoke BE Nutall I et al Rapid epidemiological mapping ofonchocerciasis (REMO) its application by the African Programme forOnchocerciasis Control (APOC) Ann Trop Med Parasitol 200296(Suppl 1)S29ndash39
23 Zoure HG Wanji S Noma M et al The geographic distribution of Loaloa in Africa results of large-scale implementation of the RapidAssessment Procedure for Loiasis (RAPLOA) PLoS Negl Trop Dis20115e1210
24 Favier C Chalvet-Monfray K Sabatier P et al Rift Valley fever in WestAfrica the role of space in endemicity Trop Med Int Health2006111878ndash88
25 Jouan A Coulibaly I Adam F et al Analytical study of a Rift Valleyfever epidemic Res Virol 1989140175ndash86
26 Diallo M Ba Y Sall AA et al Amplification of the sylvatic cycle ofdengue virus type 2 Senegal 1999ndash2000 entomologicfindings and epidemiologic considerations Emerg Infect Dis20039362ndash7
27 Jentes ES Poumerol G Gershman MD et al The revised global yellowfever risk map and recommendations for vaccination 2010consensus of the Informal WHO Working Group on Geographic Riskfor Yellow Fever Lancet Infect Dis 201111622ndash32
28 Rogers DJ Wilson AJ Hay SI Graham AJ The global distribution ofyellow fever and dengue Adv Parasitol 200662181ndash220
29 Grard G Drexler JF Fair J et al Re-emergence of CrimeanndashCongohemorrhagic fever virus in Central Africa PLoS Negl Trop Dis20115e1350
30 Nabeth P Thior M Faye O Simon F Human CrimeanndashCongohemorrhagic fever Senegal Emerg Infect Dis 2004101881ndash2
31 Vial L Diatta G Tall A et al Incidence of tick-borne relapsing fever inWest Africa longitudinal study Lancet 200636837ndash43
32 Jensenius M Davis X von Sonnenburg F et al Multicenter GeoSentinelanalysis of rickettsial diseases in international travelers 1996ndash2008Emerg Infect Dis 2009151791ndash8
33 Socolovschi C Mediannikov O Sokhna C et al Rickettsiafelis-associated uneruptive fever Senegal Emerg Infect Dis2010161140ndash2
34 Cazorla C Socolovschi C Jensenius M Parola P Tick-borne diseasestick-borne spotted fever rickettsioses in Africa Infect Dis Clin NorthAm 200822p531ndash44ixndashx
35 Merritt RW Walker ED Small PL et al Ecology and transmission ofBuruli ulcer disease a systematic review PLoS Negl Trop Dis20104e911
36 Tissot-Dupont H Brouqui P Faugere B Raoult D Prevalence ofantibodies to Coxiella burnetti Rickettsia conorii and Rickettsia typhiin seven African countries Clin Infect Dis 1995211126ndash33
37 Mediannikov O Fenollar F Socolovschi C et al Coxiella burnetii inhumans and ticks in rural Senegal PLoS Negl Trop Dis 20104e654
38 Kitron U Risk maps transmission and burden of vector-bornediseases Parasitol Today 200016324ndash5
Transactions of the Royal Society of Tropical Medicine and Hygiene
11 of 12
at University of L
iverpool on May 9 2013
httptrstmhoxfordjournalsorg
Dow
nloaded from
39 Thomson MC Connor SJ Milligan P Flasse SP Mapping malaria risk inAfrica what can satellite data contribute Parasitol Today199713313ndash8
40 Matthys B Koudou BG NrsquoGoran EK et al Spatial dispersion andcharacterisation of mosquito breeding habitats in urbanvegetable-production areas of Abidjan Cote drsquoIvoire Ann Trop MedParasitol 2010104649ndash66
41 Machault V Vignolles C Pages F et al Spatial heterogeneity andtemporal evolution of malaria transmission risk in Dakar Senegalaccording to remotely sensed environmental data Malar J 20109252
42 Dambach P Sie A Lacaux JP et al Using high spatial resolutionremote sensing for risk mapping of malaria occurrence in theNouna District Burkina Faso Glob Health Action 20092 doi103402ghav2io2094
43 Bogh C Lindsay SW Clarke SE et al High spatial resolution mappingof malaria transmission risk in The Gambia West Africa usingLANDSAT TM satellite imagery Am J Trop Med Hyg 200776875ndash81
44 Bicout DJ Sabatier P Mapping Rift Valley fever vectors and prevalenceusing rainfall variations Vector Borne Zoonotic Dis 2004433ndash42
45 Clements AC Pfeiffer DU Martin V et al Spatial risk assessment of RiftValley fever in Senegal Vector Borne Zoonotic Dis 20077203ndash16
46 Rogers DJ Hay SI Packer MJ Predicting the distribution of tsetse fliesin West Africa using temporal Fourier processed meteorologicalsatellite data Ann Trop Med Parasitol 199690225ndash41
47 Cecchi G Mattioli RC Slingenbergh J de la Rocque S Land cover andtsetse fly distributions in sub-Saharan Africa Med Vet Entomol200822364ndash73
48 Thomson MC Connor SJ Environmental information systems for thecontrol of arthropod vectors of disease Med Vet Entomol200014227ndash44
49 Hay SI Tucker CJ Rogers DJ Packer MJ Remotely sensed surrogatesof meteorological data for the study of the distribution andabundance of arthropod vectors of disease Ann Trop Med Parasitol1996901ndash19
50 Randolph SE Rogers DJ A generic population model for theAfrican tick Rhipicephalus appendiculatus Parasitology 1997115265ndash79
51 Cumming GS Comparing climate and vegetation as limiting factorsfor species ranges of African ticks Ecology 200283255ndash68
52 Ostfeld RS Glass GE Keesing F Spatial epidemiology an emerging (orre-emerging) discipline Trends Ecol Evol 200520328ndash36
53 Hay SI Tatem AJ Graham AJ et al Global environmental data formapping infectious disease distribution Adv Parasitol 20066237ndash77
54 Riedel N Vounatsou P Miller JM et al Geographical patterns andpredictors of malaria risk in Zambia Bayesian geostatisticalmodelling of the 2006 Zambia national malaria indicator survey(ZMIS) Malar J 2010937
55 Gosoniu L Veta AM Vounatsou P Bayesian geostatisticalmodeling of Malaria Indicator Survey data in Angola PLoS One20105e9322
56 Kalluri S Gilruth P Rogers D Szczur M Surveillance of arthropodvector-borne infectious diseases using remote sensing techniquesa review PLoS Pathog 200731361ndash71
57 Clements AC Pfeiffer DU Emerging viral zoonoses frameworks forspatial and spatiotemporal risk assessment and resource planningVet J 200918221ndash30
58 Phillips SJ Anderson RP Schapire RE Maximum entropy modelingof species geographic distributions Ecological Modelling 2006190231ndash59
59 Fenollar F Neglected and emerging diseases in sub-Saharan AfricaClin Microbiol Infect 201117957ndash8
60 Eisen RJ Eisen L Lane RS Predicting density of Ixodes pacificusnymphs in dense woodlands in Mendocino County Californiabased on geographic information systems and remotesensing versus field-derived data Am J Trop Med Hyg 200674632ndash40
P Ratmanov et al
12 of 12
at University of L
iverpool on May 9 2013
httptrstmhoxfordjournalsorg
Dow
nloaded from
of the prevalence of rickettsioses and entomological studies ofticks for Rickettsia spp Although attempts to map the geographicdistribution of rickettsioses in Africa have been made they simplydisplayed infected sites or counties with reported cases as pointswithout conducting any geostatistical analyses834 Thus thespatial aspects of borrelial and rickettsial VBDs in West Africahave not received the attention they deserve
Buruli ulcer
Buruli ulcer (BU) disease caused by Mycobacterium ulcerans isan emerging infectious disease in many tropical and subtropicalcountries The exact mode of transmission of this diseaseremains unclear Although vectors and modes of transmissionremain unknown it has been hypothesized that transmissionof BU disease is associated with human activities in andaround aquatic environments (the role of water bugs has beendiscussed) and that characteristics of the landscape play a rolein the spread of BU disease35 In West Africa Cote drsquoIvoireGhana and Benin are the countries that are particularly affectedby BU disease (Figure 9)
Q fever
In West Africa Q fever has a wide distribution which has beenshown repeatedly in human serological studies and studies ofreservoirs (domestic animals)36 It was reported that the main
vector of TBRF the soft tick O sonrai also harboured Coxiella bur-netii37 The possibility that this tick transmits Q fever was notstudied but some species of tick may play a role in the transmis-sion of Q fever
Geographic information systems and remotesensing for spatial analysis of vectordistributionThe geographic distribution of VBDs depends on the distributionof their vectors environmental factors and climatic factorsVectors depend on specific abiotic conditions and advances inRS and mapping of variations in abiotic conditions have stimu-lated efforts to create risk maps for VBDs538 This spatial con-cordance of environmental variables and vector distribution isoften used to estimate the current distribution of vectors in un-studied areas
There have been many studies of the entomological inocula-tion rate and parasite prevalence vector density and breedingsites and risk mapping and modelling in West Africa related tothe Anopheles mosquito The distribution of Anopheles sppwhich are the major vectors of LF and malaria in West Africawas studied in relation to normalised difference vegetationindex (NDVI) values39 Several studies were dedicated to investi-gating the risk factors for malaria and identifying the potentialAnopheles mosquito breeding sites in urban environments inCote drsquoIvoire and Senegal4041 In Burkina Faso and The
Figure 2 Geographic distribution of human African trypanosomiasis (Trypanosoma brucei gambiense) in West African countries in 2010 (data from theWHO retrieved from httpwwwwhoint)
Transactions of the Royal Society of Tropical Medicine and Hygiene
5 of 12
at University of L
iverpool on May 9 2013
httptrstmhoxfordjournalsorg
Dow
nloaded from
Gambia potential Anopheles breeding sites were mapped at thevillage level using satellite imagery4243
A critically important review regarding the potential for GISRStechnologies and methodologies to be used for the preventionsurveillance and control of the mosquito Ae aegypti thedengue virus vector can be found in an article by Eisen et al7
In Senegal a clear association between the amount of rainfallthe abundance of vectors and the prevalence of RVF has beendemonstrated44 In another study in Senegal the average totalmonthly rainfall from December to February was the most im-portant spatial predictor of the risk of RVF45
GIS and RS techniques are promising and powerful tools for de-scribing tsetse distribution because thermal data appeared to bethe most useful predictor variable followed by indices of vegeta-tion and rainfall46 Recently a study demonstrated the potentialof high-resolution images for mapping the habitat of Glossina ona local scale as well as in larger areas47
Applications of GISRS have been summarised with examplesof studies on various vectors such as the malaria vector Anoph-eles the arbovirus vector culicine mosquitoes (Aedes spp andCulex spp) the leishmaniasis vector Phlebotomus sandflies thetrypanosomiasis vector tsetse and the loiasis vector Chrysopsin a review by Thomson and Connor48
Previous studies have used multivariate analyses to estimatethe spatial prevalence of particular tick species and to makeinferences about different environmental variables that deter-mine their distribution in Africa4950 Another approach to predict-ing tick distribution was established using a set of methods in
which the author concluded that on average climatic variablesare better predictors of tick distribution than vegetation-relatedvariables and the key to describing tick distribution is the covari-ance of temperature and rainfall51
It should be taken into account that risk mapping based onvectors has serious limitations Primarily disease risk or incidenceis most closely correlated with the abundance of pathogen-infected vectors rather than simply the presence of vectors orthe total abundance of vectors52
Perspectives of application of geographicinformation systemsremote sensingtechnologies for studying vectorbornediseases remaining challenges and conclusionGIS facilitate comparisons between disease patterns and envir-onmental data while RS technologies can use high-resolutionsatellite data to provide estimates of variables such as tempera-ture vegetation and humidity53 During the past two decadessatellite RS technology has shown promising results in assessingthe risk of various VBDs at different spatial scales Satellite-basedimagery is characterised by its spatial spectral and temporalresolution Despite some limited attempts to apply RS to epi-demiology the methods have not yet demonstrated theirexpected potential GISRS technologies are undoubtedly a valu-able source of information for epidemiologists GISRS technolo-gies should not be overestimated and it is necessary to take all of
Figure 3 Geographic distribution of cutaneous leishmaniasis in West African countries in 2010 (data from the WHO retrieved from httpwwwwhoint)
P Ratmanov et al
6 of 12
at University of L
iverpool on May 9 2013
httptrstmhoxfordjournalsorg
Dow
nloaded from
Figure 4 Geographic distribution of lymphatic filariasis Loa loa filariasis (loiasis) and onchocerciasis in West African countries (data from the WHOretrieved from httpwwwwhoint)
Figure 5 Geographic distribution of Rift Valley fever (RVF) in West African countries (data from the WHO retrieved from httpwwwwhoint)
Transactions of the Royal Society of Tropical Medicine and Hygiene
7 of 12
at University of L
iverpool on May 9 2013
httptrstmhoxfordjournalsorg
Dow
nloaded from
Figure 6 Geographic distribution of dengue in West African countries in 2011 (data from the WHO retrieved from httpwwwwhoint)
Figure 7 Geographic distribution of yellow fever in West African countries in 2011 (data from the WHO retrieved from httpwwwwhoint)
P Ratmanov et al
8 of 12
at University of L
iverpool on May 9 2013
httptrstmhoxfordjournalsorg
Dow
nloaded from
Figure 8 Geographic distribution of CrimeanndashCongo haemorrhagic fever in West African countries (data from the WHO retrieved from httpwwwwhoint)
Figure 9 Geographic distribution of Buruli ulcer in 2010 in West African countries (data from the WHO retrieved from httpwwwwhoint)
Transactions of the Royal Society of Tropical Medicine and Hygiene
9 of 12
at University of L
iverpool on May 9 2013
httptrstmhoxfordjournalsorg
Dow
nloaded from
their advantages and limitations into account Moreover VBDinterventions have to be identified and recognised as they canconfound the relationship between the environment anddisease In the recent malaria indicator surveys in Zambia andAngola for example no relationship was observed between re-motely sensed data and malaria risk5455
The typical modelling approach investigates associationsbetween multivariate environmental data and patterns ofvector presence or absence for mapping vectors and VBDs Atthe first step simple statistical models could be a good startingpoint for linking the limited number of environmental variablesthat can be derived from satellite data Simple statisticalmodels are restricted because they often fit linear functionsbetween environmental variables and presenceabsence datawhen it is most likely that such associations are highlycomplex and non-linear20 At the second step sophisticatedprocess-based models that rely on vector biology as predictorsof diseases and their risk should be developed56
Regression models have been widely applied in landscape epi-demiology The type of regression model (logistic Poisson linearetc) is determined by the type of outcome variable to be pre-dicted (eg binary count continuous) and environmental vari-ables measured at sampled locations are entered ascovariates The resultant model is then either used to predictthe outcome variable at non-sampled locations based onobserved values of the covariates at the prediction locations orto explain observed patterns of disease on the basis of themodel covariates57
Besides regression modelling there are many otherapproaches in spatial epidemiology spatial clustering discrimin-ant analysis generalised linear models generalised additivemodels Bayesian estimation methods and others Howeversuch traditional models require both disease presence anddisease absence data At the same time there are differentlsquopresence-onlyrsquo models that could be used when no lsquoabsencedatarsquo are available One of these presence-only machine learning(rule-based) algorithms is an ecological niche modelling algo-rithm based on maximum entropy (Maxent) This method canbe used successfully with very small sample sizes and does notrequire independence of covariates58 Moreover these advan-tages could be crucial in Africa where complete surveillancedata from all regions are not available
Basic spatial modelling approaches include interpolationbased on spatial dependence in vector or VBD data and extrapo-lation based on associations between vector or VBD data and en-vironmental or socioeconomic predictor variables The latterapproach can be a powerful tool to gain insights into levels ofrisk within areas where surveillance data are lacking or unreli-able However it should be noted that model extrapolation isrestricted to areas with ecological and climatic characteristicssimilar to those of the model development area1
Global strategies for controlling infections in sub-SaharanAfrica have focused on the lsquobig threersquo diseases namely AIDSTB and malaria Other causes of fever and infection fall intothe category of NTDs In the twenty-first century it is importantto compile a comprehensive list of the infections in sub-SaharanAfrica and to study their epidemiology59
Few investigations have addressed the unknown causes offever in West Africa The majority of fevers have long been con-sidered to be related to malaria Recent studies have shown that
NTDs are the most frequent causes of fever in both indigenes andtravellers In particular TBRF due to B crocidurae has been redis-covered931 and the rickettsiae including R felis have beenreported33 Overall new pathogens among the rickettsiae wereidentified in the environment8
The geographic distribution of ticksmdashvectors of diseases inWest Africamdashis a neglected area of study However applyingGISRS technologies to the study of Ixodes ticks in North Americahas yielded significant results For example the use of geospatialmodelling has revealed that high concentrations of I pacificusand I scapularis which are the key tick vectors of Lyme diseasein North America can be predicted by GISRS-based environmen-tal factors related to elevation slope of the landscape vegetationtype soil type temperature and moisture60 The same approachescould be used in epidemiological studies of R felis infection anemerging disease in West Africa
Today epidemiologists often use new GISRS techniques tostudy a variety of VBDs The associations between satellite-derived environmental variables (such as temperature humidityelevation vegetation rainfall surface water land use land covertype and soil moisture) and vector density are used to identifyand characterise vector habitats A wide variety of RS datapowerful GIS and statistical software packages are available fordesktop computing environments making it affordable and feas-ible for epidemiologists to experiment with new spatial analysistechniques56
Maps that show the seasonal risk of VBDs will be necessary tomonitor the impacts of changes on vector ecology Using GISRSadvanced analytical tools and a landscape ecology approachthese risk maps could play a major role in defining research ques-tions and surveillance needs and in guiding control efforts andfield studies38
From the other side this approach is based on ecological ana-lysis and requires assembling all epidemiologic available dataand moreover it should be supported by fieldwork The most suc-cessful GISRS applications to study VBDs in West Africa areeither local (country)-level studies using community-based sur-veillance data or continental (sub-continental)-level studiesusing published scientific literature data Cooperation of the epi-demiological community could allow enhancing studies of VBDsin West Africa The most important goal of the application of GISRS technologies to the study of VBDs is to reduce the burden ofdisease by generating information that empowers the public totake protective action and helps public health agencies to effect-ively allocate their limited prevention surveillance and controlresources There is great potential to use GISRS technologiesto improve the surveillance prevention and control of vector-borne infectious diseases in West Africa
Supplementary dataSupplementary data are available at Transactions Online(httptrstmhoxfordjournalsorg)
Authorsrsquo contributions All authors contributed equally to themanuscript DR conceived the review PR and OM researched the dataand literature PR OM and DR analysed and interpreted the data for
P Ratmanov et al
10 of 12
at University of L
iverpool on May 9 2013
httptrstmhoxfordjournalsorg
Dow
nloaded from
the maps and discussed the content of the review PR and DR drafted themanuscript All authors critically revised the manuscript for intellectualcontent and read and approved the final version DR is guarantor ofthe paper
Funding None
Competing interests None declared
Ethical approval Not required
References1 Eisen L Eisen RJ Using geographic information systems and decision
support systems for the prediction prevention and control ofvector-borne diseases Annu Rev Entomol 20115641ndash61
2 WHO Neglected Tropical Diseases Geneva World HealthOrganization 2012 httpwwwwhointneglected_diseasesdiseasesen [accessed 8 January 2013]
3 Hay SI An overview of remote sensing and geodesy for epidemiologyand public health application Adv Parasitol 2000471ndash35
4 Pavlovsky EN Natural Nidality of Transmissible Diseases with SpecialReference to the Landscape Epidemiology of ZooanthroponosesUrbana IL University of Illinois Press 1966
5 Kitron U Landscape ecology and epidemiology of vector-bornediseases tools for spatial analysis J Med Entomol 199835435ndash45
6 Rogers DJ Randolph SE Studying the global distribution of infectiousdiseases using GIS and RS Nat Rev Microbiol 20031231ndash7
7 Eisen L Lozano-Fuentes S Use of mapping and spatial andspace-time modeling approaches in operational control of Aedesaegypti and dengue PLoS Negl Trop Dis 20093e411
8 Mediannikov O Diatta G Fenollar F et al Tick-borne rickettsiosesneglected emerging diseases in rural Senegal PLoS Negl Trop Dis20104e821
9 Parola P Diatta G Socolovschi C et al Tick-borne relapsing feverborreliosis rural Senegal Emerg Infect Dis 201117883ndash5
10 WHO World Malaria Report 2010 Geneva World HealthOrganization 2010
11 Sullivan D Uncertainty in mapping malaria epidemiologyimplications for control Epidemiol Rev 201032175ndash87
12 Machault V Vignolles C Borchi F et al The use of remotely sensedenvironmental data in the study of malaria Geospat Health20115151ndash68
13 Snow RW Marsh K le Sueur D The need for maps of transmissionintensity to guide malaria control in Africa Parasitol Today199612455ndash7
14 Guerra CA Hay SI Lucioparedes LS et al Assembling a globaldatabase of malaria parasite prevalence for the Malaria AtlasProject Malar J 2007617
15 Brun R Blum J Chappuis F Burri C Human African trypanosomiasisLancet 2010375148ndash59
16 Simarro PP Cecchi G Paone M et al The Atlas of human Africantrypanosomiasis a contribution to global mapping of neglectedtropical diseases Int J Health Geogr 2010957
17 Bern C Maguire JH Alvar J Complexities of assessing the diseaseburden attributable to leishmaniasis PLoS Negl Trop Dis 20082e313
18 WHO Report of the Scientific Working Group Meeting onLeishmaniasis (Geneva 2ndash4 February 2004) Geneva World HealthOrganization 2004
19 Faye B Bucheton B Banuls AL et al Seroprevalence of Leishmaniainfantum in a rural area of Senegal analysis of risk factors involvedin transmission to humans Trans R Soc Trop Med Hyg2011105333ndash40
20 Slater H Michael E Predicting the current and future potentialdistributions of lymphatic filariasis in Africa using maximumentropy ecological niche modelling PLoS One 20127e32202
21 Basanez MG Pion SD Churcher TS et al River blindness a successstory under threat PLoS Med 20063e371
22 Noma M Nwoke BE Nutall I et al Rapid epidemiological mapping ofonchocerciasis (REMO) its application by the African Programme forOnchocerciasis Control (APOC) Ann Trop Med Parasitol 200296(Suppl 1)S29ndash39
23 Zoure HG Wanji S Noma M et al The geographic distribution of Loaloa in Africa results of large-scale implementation of the RapidAssessment Procedure for Loiasis (RAPLOA) PLoS Negl Trop Dis20115e1210
24 Favier C Chalvet-Monfray K Sabatier P et al Rift Valley fever in WestAfrica the role of space in endemicity Trop Med Int Health2006111878ndash88
25 Jouan A Coulibaly I Adam F et al Analytical study of a Rift Valleyfever epidemic Res Virol 1989140175ndash86
26 Diallo M Ba Y Sall AA et al Amplification of the sylvatic cycle ofdengue virus type 2 Senegal 1999ndash2000 entomologicfindings and epidemiologic considerations Emerg Infect Dis20039362ndash7
27 Jentes ES Poumerol G Gershman MD et al The revised global yellowfever risk map and recommendations for vaccination 2010consensus of the Informal WHO Working Group on Geographic Riskfor Yellow Fever Lancet Infect Dis 201111622ndash32
28 Rogers DJ Wilson AJ Hay SI Graham AJ The global distribution ofyellow fever and dengue Adv Parasitol 200662181ndash220
29 Grard G Drexler JF Fair J et al Re-emergence of CrimeanndashCongohemorrhagic fever virus in Central Africa PLoS Negl Trop Dis20115e1350
30 Nabeth P Thior M Faye O Simon F Human CrimeanndashCongohemorrhagic fever Senegal Emerg Infect Dis 2004101881ndash2
31 Vial L Diatta G Tall A et al Incidence of tick-borne relapsing fever inWest Africa longitudinal study Lancet 200636837ndash43
32 Jensenius M Davis X von Sonnenburg F et al Multicenter GeoSentinelanalysis of rickettsial diseases in international travelers 1996ndash2008Emerg Infect Dis 2009151791ndash8
33 Socolovschi C Mediannikov O Sokhna C et al Rickettsiafelis-associated uneruptive fever Senegal Emerg Infect Dis2010161140ndash2
34 Cazorla C Socolovschi C Jensenius M Parola P Tick-borne diseasestick-borne spotted fever rickettsioses in Africa Infect Dis Clin NorthAm 200822p531ndash44ixndashx
35 Merritt RW Walker ED Small PL et al Ecology and transmission ofBuruli ulcer disease a systematic review PLoS Negl Trop Dis20104e911
36 Tissot-Dupont H Brouqui P Faugere B Raoult D Prevalence ofantibodies to Coxiella burnetti Rickettsia conorii and Rickettsia typhiin seven African countries Clin Infect Dis 1995211126ndash33
37 Mediannikov O Fenollar F Socolovschi C et al Coxiella burnetii inhumans and ticks in rural Senegal PLoS Negl Trop Dis 20104e654
38 Kitron U Risk maps transmission and burden of vector-bornediseases Parasitol Today 200016324ndash5
Transactions of the Royal Society of Tropical Medicine and Hygiene
11 of 12
at University of L
iverpool on May 9 2013
httptrstmhoxfordjournalsorg
Dow
nloaded from
39 Thomson MC Connor SJ Milligan P Flasse SP Mapping malaria risk inAfrica what can satellite data contribute Parasitol Today199713313ndash8
40 Matthys B Koudou BG NrsquoGoran EK et al Spatial dispersion andcharacterisation of mosquito breeding habitats in urbanvegetable-production areas of Abidjan Cote drsquoIvoire Ann Trop MedParasitol 2010104649ndash66
41 Machault V Vignolles C Pages F et al Spatial heterogeneity andtemporal evolution of malaria transmission risk in Dakar Senegalaccording to remotely sensed environmental data Malar J 20109252
42 Dambach P Sie A Lacaux JP et al Using high spatial resolutionremote sensing for risk mapping of malaria occurrence in theNouna District Burkina Faso Glob Health Action 20092 doi103402ghav2io2094
43 Bogh C Lindsay SW Clarke SE et al High spatial resolution mappingof malaria transmission risk in The Gambia West Africa usingLANDSAT TM satellite imagery Am J Trop Med Hyg 200776875ndash81
44 Bicout DJ Sabatier P Mapping Rift Valley fever vectors and prevalenceusing rainfall variations Vector Borne Zoonotic Dis 2004433ndash42
45 Clements AC Pfeiffer DU Martin V et al Spatial risk assessment of RiftValley fever in Senegal Vector Borne Zoonotic Dis 20077203ndash16
46 Rogers DJ Hay SI Packer MJ Predicting the distribution of tsetse fliesin West Africa using temporal Fourier processed meteorologicalsatellite data Ann Trop Med Parasitol 199690225ndash41
47 Cecchi G Mattioli RC Slingenbergh J de la Rocque S Land cover andtsetse fly distributions in sub-Saharan Africa Med Vet Entomol200822364ndash73
48 Thomson MC Connor SJ Environmental information systems for thecontrol of arthropod vectors of disease Med Vet Entomol200014227ndash44
49 Hay SI Tucker CJ Rogers DJ Packer MJ Remotely sensed surrogatesof meteorological data for the study of the distribution andabundance of arthropod vectors of disease Ann Trop Med Parasitol1996901ndash19
50 Randolph SE Rogers DJ A generic population model for theAfrican tick Rhipicephalus appendiculatus Parasitology 1997115265ndash79
51 Cumming GS Comparing climate and vegetation as limiting factorsfor species ranges of African ticks Ecology 200283255ndash68
52 Ostfeld RS Glass GE Keesing F Spatial epidemiology an emerging (orre-emerging) discipline Trends Ecol Evol 200520328ndash36
53 Hay SI Tatem AJ Graham AJ et al Global environmental data formapping infectious disease distribution Adv Parasitol 20066237ndash77
54 Riedel N Vounatsou P Miller JM et al Geographical patterns andpredictors of malaria risk in Zambia Bayesian geostatisticalmodelling of the 2006 Zambia national malaria indicator survey(ZMIS) Malar J 2010937
55 Gosoniu L Veta AM Vounatsou P Bayesian geostatisticalmodeling of Malaria Indicator Survey data in Angola PLoS One20105e9322
56 Kalluri S Gilruth P Rogers D Szczur M Surveillance of arthropodvector-borne infectious diseases using remote sensing techniquesa review PLoS Pathog 200731361ndash71
57 Clements AC Pfeiffer DU Emerging viral zoonoses frameworks forspatial and spatiotemporal risk assessment and resource planningVet J 200918221ndash30
58 Phillips SJ Anderson RP Schapire RE Maximum entropy modelingof species geographic distributions Ecological Modelling 2006190231ndash59
59 Fenollar F Neglected and emerging diseases in sub-Saharan AfricaClin Microbiol Infect 201117957ndash8
60 Eisen RJ Eisen L Lane RS Predicting density of Ixodes pacificusnymphs in dense woodlands in Mendocino County Californiabased on geographic information systems and remotesensing versus field-derived data Am J Trop Med Hyg 200674632ndash40
P Ratmanov et al
12 of 12
at University of L
iverpool on May 9 2013
httptrstmhoxfordjournalsorg
Dow
nloaded from
Gambia potential Anopheles breeding sites were mapped at thevillage level using satellite imagery4243
A critically important review regarding the potential for GISRStechnologies and methodologies to be used for the preventionsurveillance and control of the mosquito Ae aegypti thedengue virus vector can be found in an article by Eisen et al7
In Senegal a clear association between the amount of rainfallthe abundance of vectors and the prevalence of RVF has beendemonstrated44 In another study in Senegal the average totalmonthly rainfall from December to February was the most im-portant spatial predictor of the risk of RVF45
GIS and RS techniques are promising and powerful tools for de-scribing tsetse distribution because thermal data appeared to bethe most useful predictor variable followed by indices of vegeta-tion and rainfall46 Recently a study demonstrated the potentialof high-resolution images for mapping the habitat of Glossina ona local scale as well as in larger areas47
Applications of GISRS have been summarised with examplesof studies on various vectors such as the malaria vector Anoph-eles the arbovirus vector culicine mosquitoes (Aedes spp andCulex spp) the leishmaniasis vector Phlebotomus sandflies thetrypanosomiasis vector tsetse and the loiasis vector Chrysopsin a review by Thomson and Connor48
Previous studies have used multivariate analyses to estimatethe spatial prevalence of particular tick species and to makeinferences about different environmental variables that deter-mine their distribution in Africa4950 Another approach to predict-ing tick distribution was established using a set of methods in
which the author concluded that on average climatic variablesare better predictors of tick distribution than vegetation-relatedvariables and the key to describing tick distribution is the covari-ance of temperature and rainfall51
It should be taken into account that risk mapping based onvectors has serious limitations Primarily disease risk or incidenceis most closely correlated with the abundance of pathogen-infected vectors rather than simply the presence of vectors orthe total abundance of vectors52
Perspectives of application of geographicinformation systemsremote sensingtechnologies for studying vectorbornediseases remaining challenges and conclusionGIS facilitate comparisons between disease patterns and envir-onmental data while RS technologies can use high-resolutionsatellite data to provide estimates of variables such as tempera-ture vegetation and humidity53 During the past two decadessatellite RS technology has shown promising results in assessingthe risk of various VBDs at different spatial scales Satellite-basedimagery is characterised by its spatial spectral and temporalresolution Despite some limited attempts to apply RS to epi-demiology the methods have not yet demonstrated theirexpected potential GISRS technologies are undoubtedly a valu-able source of information for epidemiologists GISRS technolo-gies should not be overestimated and it is necessary to take all of
Figure 3 Geographic distribution of cutaneous leishmaniasis in West African countries in 2010 (data from the WHO retrieved from httpwwwwhoint)
P Ratmanov et al
6 of 12
at University of L
iverpool on May 9 2013
httptrstmhoxfordjournalsorg
Dow
nloaded from
Figure 4 Geographic distribution of lymphatic filariasis Loa loa filariasis (loiasis) and onchocerciasis in West African countries (data from the WHOretrieved from httpwwwwhoint)
Figure 5 Geographic distribution of Rift Valley fever (RVF) in West African countries (data from the WHO retrieved from httpwwwwhoint)
Transactions of the Royal Society of Tropical Medicine and Hygiene
7 of 12
at University of L
iverpool on May 9 2013
httptrstmhoxfordjournalsorg
Dow
nloaded from
Figure 6 Geographic distribution of dengue in West African countries in 2011 (data from the WHO retrieved from httpwwwwhoint)
Figure 7 Geographic distribution of yellow fever in West African countries in 2011 (data from the WHO retrieved from httpwwwwhoint)
P Ratmanov et al
8 of 12
at University of L
iverpool on May 9 2013
httptrstmhoxfordjournalsorg
Dow
nloaded from
Figure 8 Geographic distribution of CrimeanndashCongo haemorrhagic fever in West African countries (data from the WHO retrieved from httpwwwwhoint)
Figure 9 Geographic distribution of Buruli ulcer in 2010 in West African countries (data from the WHO retrieved from httpwwwwhoint)
Transactions of the Royal Society of Tropical Medicine and Hygiene
9 of 12
at University of L
iverpool on May 9 2013
httptrstmhoxfordjournalsorg
Dow
nloaded from
their advantages and limitations into account Moreover VBDinterventions have to be identified and recognised as they canconfound the relationship between the environment anddisease In the recent malaria indicator surveys in Zambia andAngola for example no relationship was observed between re-motely sensed data and malaria risk5455
The typical modelling approach investigates associationsbetween multivariate environmental data and patterns ofvector presence or absence for mapping vectors and VBDs Atthe first step simple statistical models could be a good startingpoint for linking the limited number of environmental variablesthat can be derived from satellite data Simple statisticalmodels are restricted because they often fit linear functionsbetween environmental variables and presenceabsence datawhen it is most likely that such associations are highlycomplex and non-linear20 At the second step sophisticatedprocess-based models that rely on vector biology as predictorsof diseases and their risk should be developed56
Regression models have been widely applied in landscape epi-demiology The type of regression model (logistic Poisson linearetc) is determined by the type of outcome variable to be pre-dicted (eg binary count continuous) and environmental vari-ables measured at sampled locations are entered ascovariates The resultant model is then either used to predictthe outcome variable at non-sampled locations based onobserved values of the covariates at the prediction locations orto explain observed patterns of disease on the basis of themodel covariates57
Besides regression modelling there are many otherapproaches in spatial epidemiology spatial clustering discrimin-ant analysis generalised linear models generalised additivemodels Bayesian estimation methods and others Howeversuch traditional models require both disease presence anddisease absence data At the same time there are differentlsquopresence-onlyrsquo models that could be used when no lsquoabsencedatarsquo are available One of these presence-only machine learning(rule-based) algorithms is an ecological niche modelling algo-rithm based on maximum entropy (Maxent) This method canbe used successfully with very small sample sizes and does notrequire independence of covariates58 Moreover these advan-tages could be crucial in Africa where complete surveillancedata from all regions are not available
Basic spatial modelling approaches include interpolationbased on spatial dependence in vector or VBD data and extrapo-lation based on associations between vector or VBD data and en-vironmental or socioeconomic predictor variables The latterapproach can be a powerful tool to gain insights into levels ofrisk within areas where surveillance data are lacking or unreli-able However it should be noted that model extrapolation isrestricted to areas with ecological and climatic characteristicssimilar to those of the model development area1
Global strategies for controlling infections in sub-SaharanAfrica have focused on the lsquobig threersquo diseases namely AIDSTB and malaria Other causes of fever and infection fall intothe category of NTDs In the twenty-first century it is importantto compile a comprehensive list of the infections in sub-SaharanAfrica and to study their epidemiology59
Few investigations have addressed the unknown causes offever in West Africa The majority of fevers have long been con-sidered to be related to malaria Recent studies have shown that
NTDs are the most frequent causes of fever in both indigenes andtravellers In particular TBRF due to B crocidurae has been redis-covered931 and the rickettsiae including R felis have beenreported33 Overall new pathogens among the rickettsiae wereidentified in the environment8
The geographic distribution of ticksmdashvectors of diseases inWest Africamdashis a neglected area of study However applyingGISRS technologies to the study of Ixodes ticks in North Americahas yielded significant results For example the use of geospatialmodelling has revealed that high concentrations of I pacificusand I scapularis which are the key tick vectors of Lyme diseasein North America can be predicted by GISRS-based environmen-tal factors related to elevation slope of the landscape vegetationtype soil type temperature and moisture60 The same approachescould be used in epidemiological studies of R felis infection anemerging disease in West Africa
Today epidemiologists often use new GISRS techniques tostudy a variety of VBDs The associations between satellite-derived environmental variables (such as temperature humidityelevation vegetation rainfall surface water land use land covertype and soil moisture) and vector density are used to identifyand characterise vector habitats A wide variety of RS datapowerful GIS and statistical software packages are available fordesktop computing environments making it affordable and feas-ible for epidemiologists to experiment with new spatial analysistechniques56
Maps that show the seasonal risk of VBDs will be necessary tomonitor the impacts of changes on vector ecology Using GISRSadvanced analytical tools and a landscape ecology approachthese risk maps could play a major role in defining research ques-tions and surveillance needs and in guiding control efforts andfield studies38
From the other side this approach is based on ecological ana-lysis and requires assembling all epidemiologic available dataand moreover it should be supported by fieldwork The most suc-cessful GISRS applications to study VBDs in West Africa areeither local (country)-level studies using community-based sur-veillance data or continental (sub-continental)-level studiesusing published scientific literature data Cooperation of the epi-demiological community could allow enhancing studies of VBDsin West Africa The most important goal of the application of GISRS technologies to the study of VBDs is to reduce the burden ofdisease by generating information that empowers the public totake protective action and helps public health agencies to effect-ively allocate their limited prevention surveillance and controlresources There is great potential to use GISRS technologiesto improve the surveillance prevention and control of vector-borne infectious diseases in West Africa
Supplementary dataSupplementary data are available at Transactions Online(httptrstmhoxfordjournalsorg)
Authorsrsquo contributions All authors contributed equally to themanuscript DR conceived the review PR and OM researched the dataand literature PR OM and DR analysed and interpreted the data for
P Ratmanov et al
10 of 12
at University of L
iverpool on May 9 2013
httptrstmhoxfordjournalsorg
Dow
nloaded from
the maps and discussed the content of the review PR and DR drafted themanuscript All authors critically revised the manuscript for intellectualcontent and read and approved the final version DR is guarantor ofthe paper
Funding None
Competing interests None declared
Ethical approval Not required
References1 Eisen L Eisen RJ Using geographic information systems and decision
support systems for the prediction prevention and control ofvector-borne diseases Annu Rev Entomol 20115641ndash61
2 WHO Neglected Tropical Diseases Geneva World HealthOrganization 2012 httpwwwwhointneglected_diseasesdiseasesen [accessed 8 January 2013]
3 Hay SI An overview of remote sensing and geodesy for epidemiologyand public health application Adv Parasitol 2000471ndash35
4 Pavlovsky EN Natural Nidality of Transmissible Diseases with SpecialReference to the Landscape Epidemiology of ZooanthroponosesUrbana IL University of Illinois Press 1966
5 Kitron U Landscape ecology and epidemiology of vector-bornediseases tools for spatial analysis J Med Entomol 199835435ndash45
6 Rogers DJ Randolph SE Studying the global distribution of infectiousdiseases using GIS and RS Nat Rev Microbiol 20031231ndash7
7 Eisen L Lozano-Fuentes S Use of mapping and spatial andspace-time modeling approaches in operational control of Aedesaegypti and dengue PLoS Negl Trop Dis 20093e411
8 Mediannikov O Diatta G Fenollar F et al Tick-borne rickettsiosesneglected emerging diseases in rural Senegal PLoS Negl Trop Dis20104e821
9 Parola P Diatta G Socolovschi C et al Tick-borne relapsing feverborreliosis rural Senegal Emerg Infect Dis 201117883ndash5
10 WHO World Malaria Report 2010 Geneva World HealthOrganization 2010
11 Sullivan D Uncertainty in mapping malaria epidemiologyimplications for control Epidemiol Rev 201032175ndash87
12 Machault V Vignolles C Borchi F et al The use of remotely sensedenvironmental data in the study of malaria Geospat Health20115151ndash68
13 Snow RW Marsh K le Sueur D The need for maps of transmissionintensity to guide malaria control in Africa Parasitol Today199612455ndash7
14 Guerra CA Hay SI Lucioparedes LS et al Assembling a globaldatabase of malaria parasite prevalence for the Malaria AtlasProject Malar J 2007617
15 Brun R Blum J Chappuis F Burri C Human African trypanosomiasisLancet 2010375148ndash59
16 Simarro PP Cecchi G Paone M et al The Atlas of human Africantrypanosomiasis a contribution to global mapping of neglectedtropical diseases Int J Health Geogr 2010957
17 Bern C Maguire JH Alvar J Complexities of assessing the diseaseburden attributable to leishmaniasis PLoS Negl Trop Dis 20082e313
18 WHO Report of the Scientific Working Group Meeting onLeishmaniasis (Geneva 2ndash4 February 2004) Geneva World HealthOrganization 2004
19 Faye B Bucheton B Banuls AL et al Seroprevalence of Leishmaniainfantum in a rural area of Senegal analysis of risk factors involvedin transmission to humans Trans R Soc Trop Med Hyg2011105333ndash40
20 Slater H Michael E Predicting the current and future potentialdistributions of lymphatic filariasis in Africa using maximumentropy ecological niche modelling PLoS One 20127e32202
21 Basanez MG Pion SD Churcher TS et al River blindness a successstory under threat PLoS Med 20063e371
22 Noma M Nwoke BE Nutall I et al Rapid epidemiological mapping ofonchocerciasis (REMO) its application by the African Programme forOnchocerciasis Control (APOC) Ann Trop Med Parasitol 200296(Suppl 1)S29ndash39
23 Zoure HG Wanji S Noma M et al The geographic distribution of Loaloa in Africa results of large-scale implementation of the RapidAssessment Procedure for Loiasis (RAPLOA) PLoS Negl Trop Dis20115e1210
24 Favier C Chalvet-Monfray K Sabatier P et al Rift Valley fever in WestAfrica the role of space in endemicity Trop Med Int Health2006111878ndash88
25 Jouan A Coulibaly I Adam F et al Analytical study of a Rift Valleyfever epidemic Res Virol 1989140175ndash86
26 Diallo M Ba Y Sall AA et al Amplification of the sylvatic cycle ofdengue virus type 2 Senegal 1999ndash2000 entomologicfindings and epidemiologic considerations Emerg Infect Dis20039362ndash7
27 Jentes ES Poumerol G Gershman MD et al The revised global yellowfever risk map and recommendations for vaccination 2010consensus of the Informal WHO Working Group on Geographic Riskfor Yellow Fever Lancet Infect Dis 201111622ndash32
28 Rogers DJ Wilson AJ Hay SI Graham AJ The global distribution ofyellow fever and dengue Adv Parasitol 200662181ndash220
29 Grard G Drexler JF Fair J et al Re-emergence of CrimeanndashCongohemorrhagic fever virus in Central Africa PLoS Negl Trop Dis20115e1350
30 Nabeth P Thior M Faye O Simon F Human CrimeanndashCongohemorrhagic fever Senegal Emerg Infect Dis 2004101881ndash2
31 Vial L Diatta G Tall A et al Incidence of tick-borne relapsing fever inWest Africa longitudinal study Lancet 200636837ndash43
32 Jensenius M Davis X von Sonnenburg F et al Multicenter GeoSentinelanalysis of rickettsial diseases in international travelers 1996ndash2008Emerg Infect Dis 2009151791ndash8
33 Socolovschi C Mediannikov O Sokhna C et al Rickettsiafelis-associated uneruptive fever Senegal Emerg Infect Dis2010161140ndash2
34 Cazorla C Socolovschi C Jensenius M Parola P Tick-borne diseasestick-borne spotted fever rickettsioses in Africa Infect Dis Clin NorthAm 200822p531ndash44ixndashx
35 Merritt RW Walker ED Small PL et al Ecology and transmission ofBuruli ulcer disease a systematic review PLoS Negl Trop Dis20104e911
36 Tissot-Dupont H Brouqui P Faugere B Raoult D Prevalence ofantibodies to Coxiella burnetti Rickettsia conorii and Rickettsia typhiin seven African countries Clin Infect Dis 1995211126ndash33
37 Mediannikov O Fenollar F Socolovschi C et al Coxiella burnetii inhumans and ticks in rural Senegal PLoS Negl Trop Dis 20104e654
38 Kitron U Risk maps transmission and burden of vector-bornediseases Parasitol Today 200016324ndash5
Transactions of the Royal Society of Tropical Medicine and Hygiene
11 of 12
at University of L
iverpool on May 9 2013
httptrstmhoxfordjournalsorg
Dow
nloaded from
39 Thomson MC Connor SJ Milligan P Flasse SP Mapping malaria risk inAfrica what can satellite data contribute Parasitol Today199713313ndash8
40 Matthys B Koudou BG NrsquoGoran EK et al Spatial dispersion andcharacterisation of mosquito breeding habitats in urbanvegetable-production areas of Abidjan Cote drsquoIvoire Ann Trop MedParasitol 2010104649ndash66
41 Machault V Vignolles C Pages F et al Spatial heterogeneity andtemporal evolution of malaria transmission risk in Dakar Senegalaccording to remotely sensed environmental data Malar J 20109252
42 Dambach P Sie A Lacaux JP et al Using high spatial resolutionremote sensing for risk mapping of malaria occurrence in theNouna District Burkina Faso Glob Health Action 20092 doi103402ghav2io2094
43 Bogh C Lindsay SW Clarke SE et al High spatial resolution mappingof malaria transmission risk in The Gambia West Africa usingLANDSAT TM satellite imagery Am J Trop Med Hyg 200776875ndash81
44 Bicout DJ Sabatier P Mapping Rift Valley fever vectors and prevalenceusing rainfall variations Vector Borne Zoonotic Dis 2004433ndash42
45 Clements AC Pfeiffer DU Martin V et al Spatial risk assessment of RiftValley fever in Senegal Vector Borne Zoonotic Dis 20077203ndash16
46 Rogers DJ Hay SI Packer MJ Predicting the distribution of tsetse fliesin West Africa using temporal Fourier processed meteorologicalsatellite data Ann Trop Med Parasitol 199690225ndash41
47 Cecchi G Mattioli RC Slingenbergh J de la Rocque S Land cover andtsetse fly distributions in sub-Saharan Africa Med Vet Entomol200822364ndash73
48 Thomson MC Connor SJ Environmental information systems for thecontrol of arthropod vectors of disease Med Vet Entomol200014227ndash44
49 Hay SI Tucker CJ Rogers DJ Packer MJ Remotely sensed surrogatesof meteorological data for the study of the distribution andabundance of arthropod vectors of disease Ann Trop Med Parasitol1996901ndash19
50 Randolph SE Rogers DJ A generic population model for theAfrican tick Rhipicephalus appendiculatus Parasitology 1997115265ndash79
51 Cumming GS Comparing climate and vegetation as limiting factorsfor species ranges of African ticks Ecology 200283255ndash68
52 Ostfeld RS Glass GE Keesing F Spatial epidemiology an emerging (orre-emerging) discipline Trends Ecol Evol 200520328ndash36
53 Hay SI Tatem AJ Graham AJ et al Global environmental data formapping infectious disease distribution Adv Parasitol 20066237ndash77
54 Riedel N Vounatsou P Miller JM et al Geographical patterns andpredictors of malaria risk in Zambia Bayesian geostatisticalmodelling of the 2006 Zambia national malaria indicator survey(ZMIS) Malar J 2010937
55 Gosoniu L Veta AM Vounatsou P Bayesian geostatisticalmodeling of Malaria Indicator Survey data in Angola PLoS One20105e9322
56 Kalluri S Gilruth P Rogers D Szczur M Surveillance of arthropodvector-borne infectious diseases using remote sensing techniquesa review PLoS Pathog 200731361ndash71
57 Clements AC Pfeiffer DU Emerging viral zoonoses frameworks forspatial and spatiotemporal risk assessment and resource planningVet J 200918221ndash30
58 Phillips SJ Anderson RP Schapire RE Maximum entropy modelingof species geographic distributions Ecological Modelling 2006190231ndash59
59 Fenollar F Neglected and emerging diseases in sub-Saharan AfricaClin Microbiol Infect 201117957ndash8
60 Eisen RJ Eisen L Lane RS Predicting density of Ixodes pacificusnymphs in dense woodlands in Mendocino County Californiabased on geographic information systems and remotesensing versus field-derived data Am J Trop Med Hyg 200674632ndash40
P Ratmanov et al
12 of 12
at University of L
iverpool on May 9 2013
httptrstmhoxfordjournalsorg
Dow
nloaded from
Figure 4 Geographic distribution of lymphatic filariasis Loa loa filariasis (loiasis) and onchocerciasis in West African countries (data from the WHOretrieved from httpwwwwhoint)
Figure 5 Geographic distribution of Rift Valley fever (RVF) in West African countries (data from the WHO retrieved from httpwwwwhoint)
Transactions of the Royal Society of Tropical Medicine and Hygiene
7 of 12
at University of L
iverpool on May 9 2013
httptrstmhoxfordjournalsorg
Dow
nloaded from
Figure 6 Geographic distribution of dengue in West African countries in 2011 (data from the WHO retrieved from httpwwwwhoint)
Figure 7 Geographic distribution of yellow fever in West African countries in 2011 (data from the WHO retrieved from httpwwwwhoint)
P Ratmanov et al
8 of 12
at University of L
iverpool on May 9 2013
httptrstmhoxfordjournalsorg
Dow
nloaded from
Figure 8 Geographic distribution of CrimeanndashCongo haemorrhagic fever in West African countries (data from the WHO retrieved from httpwwwwhoint)
Figure 9 Geographic distribution of Buruli ulcer in 2010 in West African countries (data from the WHO retrieved from httpwwwwhoint)
Transactions of the Royal Society of Tropical Medicine and Hygiene
9 of 12
at University of L
iverpool on May 9 2013
httptrstmhoxfordjournalsorg
Dow
nloaded from
their advantages and limitations into account Moreover VBDinterventions have to be identified and recognised as they canconfound the relationship between the environment anddisease In the recent malaria indicator surveys in Zambia andAngola for example no relationship was observed between re-motely sensed data and malaria risk5455
The typical modelling approach investigates associationsbetween multivariate environmental data and patterns ofvector presence or absence for mapping vectors and VBDs Atthe first step simple statistical models could be a good startingpoint for linking the limited number of environmental variablesthat can be derived from satellite data Simple statisticalmodels are restricted because they often fit linear functionsbetween environmental variables and presenceabsence datawhen it is most likely that such associations are highlycomplex and non-linear20 At the second step sophisticatedprocess-based models that rely on vector biology as predictorsof diseases and their risk should be developed56
Regression models have been widely applied in landscape epi-demiology The type of regression model (logistic Poisson linearetc) is determined by the type of outcome variable to be pre-dicted (eg binary count continuous) and environmental vari-ables measured at sampled locations are entered ascovariates The resultant model is then either used to predictthe outcome variable at non-sampled locations based onobserved values of the covariates at the prediction locations orto explain observed patterns of disease on the basis of themodel covariates57
Besides regression modelling there are many otherapproaches in spatial epidemiology spatial clustering discrimin-ant analysis generalised linear models generalised additivemodels Bayesian estimation methods and others Howeversuch traditional models require both disease presence anddisease absence data At the same time there are differentlsquopresence-onlyrsquo models that could be used when no lsquoabsencedatarsquo are available One of these presence-only machine learning(rule-based) algorithms is an ecological niche modelling algo-rithm based on maximum entropy (Maxent) This method canbe used successfully with very small sample sizes and does notrequire independence of covariates58 Moreover these advan-tages could be crucial in Africa where complete surveillancedata from all regions are not available
Basic spatial modelling approaches include interpolationbased on spatial dependence in vector or VBD data and extrapo-lation based on associations between vector or VBD data and en-vironmental or socioeconomic predictor variables The latterapproach can be a powerful tool to gain insights into levels ofrisk within areas where surveillance data are lacking or unreli-able However it should be noted that model extrapolation isrestricted to areas with ecological and climatic characteristicssimilar to those of the model development area1
Global strategies for controlling infections in sub-SaharanAfrica have focused on the lsquobig threersquo diseases namely AIDSTB and malaria Other causes of fever and infection fall intothe category of NTDs In the twenty-first century it is importantto compile a comprehensive list of the infections in sub-SaharanAfrica and to study their epidemiology59
Few investigations have addressed the unknown causes offever in West Africa The majority of fevers have long been con-sidered to be related to malaria Recent studies have shown that
NTDs are the most frequent causes of fever in both indigenes andtravellers In particular TBRF due to B crocidurae has been redis-covered931 and the rickettsiae including R felis have beenreported33 Overall new pathogens among the rickettsiae wereidentified in the environment8
The geographic distribution of ticksmdashvectors of diseases inWest Africamdashis a neglected area of study However applyingGISRS technologies to the study of Ixodes ticks in North Americahas yielded significant results For example the use of geospatialmodelling has revealed that high concentrations of I pacificusand I scapularis which are the key tick vectors of Lyme diseasein North America can be predicted by GISRS-based environmen-tal factors related to elevation slope of the landscape vegetationtype soil type temperature and moisture60 The same approachescould be used in epidemiological studies of R felis infection anemerging disease in West Africa
Today epidemiologists often use new GISRS techniques tostudy a variety of VBDs The associations between satellite-derived environmental variables (such as temperature humidityelevation vegetation rainfall surface water land use land covertype and soil moisture) and vector density are used to identifyand characterise vector habitats A wide variety of RS datapowerful GIS and statistical software packages are available fordesktop computing environments making it affordable and feas-ible for epidemiologists to experiment with new spatial analysistechniques56
Maps that show the seasonal risk of VBDs will be necessary tomonitor the impacts of changes on vector ecology Using GISRSadvanced analytical tools and a landscape ecology approachthese risk maps could play a major role in defining research ques-tions and surveillance needs and in guiding control efforts andfield studies38
From the other side this approach is based on ecological ana-lysis and requires assembling all epidemiologic available dataand moreover it should be supported by fieldwork The most suc-cessful GISRS applications to study VBDs in West Africa areeither local (country)-level studies using community-based sur-veillance data or continental (sub-continental)-level studiesusing published scientific literature data Cooperation of the epi-demiological community could allow enhancing studies of VBDsin West Africa The most important goal of the application of GISRS technologies to the study of VBDs is to reduce the burden ofdisease by generating information that empowers the public totake protective action and helps public health agencies to effect-ively allocate their limited prevention surveillance and controlresources There is great potential to use GISRS technologiesto improve the surveillance prevention and control of vector-borne infectious diseases in West Africa
Supplementary dataSupplementary data are available at Transactions Online(httptrstmhoxfordjournalsorg)
Authorsrsquo contributions All authors contributed equally to themanuscript DR conceived the review PR and OM researched the dataand literature PR OM and DR analysed and interpreted the data for
P Ratmanov et al
10 of 12
at University of L
iverpool on May 9 2013
httptrstmhoxfordjournalsorg
Dow
nloaded from
the maps and discussed the content of the review PR and DR drafted themanuscript All authors critically revised the manuscript for intellectualcontent and read and approved the final version DR is guarantor ofthe paper
Funding None
Competing interests None declared
Ethical approval Not required
References1 Eisen L Eisen RJ Using geographic information systems and decision
support systems for the prediction prevention and control ofvector-borne diseases Annu Rev Entomol 20115641ndash61
2 WHO Neglected Tropical Diseases Geneva World HealthOrganization 2012 httpwwwwhointneglected_diseasesdiseasesen [accessed 8 January 2013]
3 Hay SI An overview of remote sensing and geodesy for epidemiologyand public health application Adv Parasitol 2000471ndash35
4 Pavlovsky EN Natural Nidality of Transmissible Diseases with SpecialReference to the Landscape Epidemiology of ZooanthroponosesUrbana IL University of Illinois Press 1966
5 Kitron U Landscape ecology and epidemiology of vector-bornediseases tools for spatial analysis J Med Entomol 199835435ndash45
6 Rogers DJ Randolph SE Studying the global distribution of infectiousdiseases using GIS and RS Nat Rev Microbiol 20031231ndash7
7 Eisen L Lozano-Fuentes S Use of mapping and spatial andspace-time modeling approaches in operational control of Aedesaegypti and dengue PLoS Negl Trop Dis 20093e411
8 Mediannikov O Diatta G Fenollar F et al Tick-borne rickettsiosesneglected emerging diseases in rural Senegal PLoS Negl Trop Dis20104e821
9 Parola P Diatta G Socolovschi C et al Tick-borne relapsing feverborreliosis rural Senegal Emerg Infect Dis 201117883ndash5
10 WHO World Malaria Report 2010 Geneva World HealthOrganization 2010
11 Sullivan D Uncertainty in mapping malaria epidemiologyimplications for control Epidemiol Rev 201032175ndash87
12 Machault V Vignolles C Borchi F et al The use of remotely sensedenvironmental data in the study of malaria Geospat Health20115151ndash68
13 Snow RW Marsh K le Sueur D The need for maps of transmissionintensity to guide malaria control in Africa Parasitol Today199612455ndash7
14 Guerra CA Hay SI Lucioparedes LS et al Assembling a globaldatabase of malaria parasite prevalence for the Malaria AtlasProject Malar J 2007617
15 Brun R Blum J Chappuis F Burri C Human African trypanosomiasisLancet 2010375148ndash59
16 Simarro PP Cecchi G Paone M et al The Atlas of human Africantrypanosomiasis a contribution to global mapping of neglectedtropical diseases Int J Health Geogr 2010957
17 Bern C Maguire JH Alvar J Complexities of assessing the diseaseburden attributable to leishmaniasis PLoS Negl Trop Dis 20082e313
18 WHO Report of the Scientific Working Group Meeting onLeishmaniasis (Geneva 2ndash4 February 2004) Geneva World HealthOrganization 2004
19 Faye B Bucheton B Banuls AL et al Seroprevalence of Leishmaniainfantum in a rural area of Senegal analysis of risk factors involvedin transmission to humans Trans R Soc Trop Med Hyg2011105333ndash40
20 Slater H Michael E Predicting the current and future potentialdistributions of lymphatic filariasis in Africa using maximumentropy ecological niche modelling PLoS One 20127e32202
21 Basanez MG Pion SD Churcher TS et al River blindness a successstory under threat PLoS Med 20063e371
22 Noma M Nwoke BE Nutall I et al Rapid epidemiological mapping ofonchocerciasis (REMO) its application by the African Programme forOnchocerciasis Control (APOC) Ann Trop Med Parasitol 200296(Suppl 1)S29ndash39
23 Zoure HG Wanji S Noma M et al The geographic distribution of Loaloa in Africa results of large-scale implementation of the RapidAssessment Procedure for Loiasis (RAPLOA) PLoS Negl Trop Dis20115e1210
24 Favier C Chalvet-Monfray K Sabatier P et al Rift Valley fever in WestAfrica the role of space in endemicity Trop Med Int Health2006111878ndash88
25 Jouan A Coulibaly I Adam F et al Analytical study of a Rift Valleyfever epidemic Res Virol 1989140175ndash86
26 Diallo M Ba Y Sall AA et al Amplification of the sylvatic cycle ofdengue virus type 2 Senegal 1999ndash2000 entomologicfindings and epidemiologic considerations Emerg Infect Dis20039362ndash7
27 Jentes ES Poumerol G Gershman MD et al The revised global yellowfever risk map and recommendations for vaccination 2010consensus of the Informal WHO Working Group on Geographic Riskfor Yellow Fever Lancet Infect Dis 201111622ndash32
28 Rogers DJ Wilson AJ Hay SI Graham AJ The global distribution ofyellow fever and dengue Adv Parasitol 200662181ndash220
29 Grard G Drexler JF Fair J et al Re-emergence of CrimeanndashCongohemorrhagic fever virus in Central Africa PLoS Negl Trop Dis20115e1350
30 Nabeth P Thior M Faye O Simon F Human CrimeanndashCongohemorrhagic fever Senegal Emerg Infect Dis 2004101881ndash2
31 Vial L Diatta G Tall A et al Incidence of tick-borne relapsing fever inWest Africa longitudinal study Lancet 200636837ndash43
32 Jensenius M Davis X von Sonnenburg F et al Multicenter GeoSentinelanalysis of rickettsial diseases in international travelers 1996ndash2008Emerg Infect Dis 2009151791ndash8
33 Socolovschi C Mediannikov O Sokhna C et al Rickettsiafelis-associated uneruptive fever Senegal Emerg Infect Dis2010161140ndash2
34 Cazorla C Socolovschi C Jensenius M Parola P Tick-borne diseasestick-borne spotted fever rickettsioses in Africa Infect Dis Clin NorthAm 200822p531ndash44ixndashx
35 Merritt RW Walker ED Small PL et al Ecology and transmission ofBuruli ulcer disease a systematic review PLoS Negl Trop Dis20104e911
36 Tissot-Dupont H Brouqui P Faugere B Raoult D Prevalence ofantibodies to Coxiella burnetti Rickettsia conorii and Rickettsia typhiin seven African countries Clin Infect Dis 1995211126ndash33
37 Mediannikov O Fenollar F Socolovschi C et al Coxiella burnetii inhumans and ticks in rural Senegal PLoS Negl Trop Dis 20104e654
38 Kitron U Risk maps transmission and burden of vector-bornediseases Parasitol Today 200016324ndash5
Transactions of the Royal Society of Tropical Medicine and Hygiene
11 of 12
at University of L
iverpool on May 9 2013
httptrstmhoxfordjournalsorg
Dow
nloaded from
39 Thomson MC Connor SJ Milligan P Flasse SP Mapping malaria risk inAfrica what can satellite data contribute Parasitol Today199713313ndash8
40 Matthys B Koudou BG NrsquoGoran EK et al Spatial dispersion andcharacterisation of mosquito breeding habitats in urbanvegetable-production areas of Abidjan Cote drsquoIvoire Ann Trop MedParasitol 2010104649ndash66
41 Machault V Vignolles C Pages F et al Spatial heterogeneity andtemporal evolution of malaria transmission risk in Dakar Senegalaccording to remotely sensed environmental data Malar J 20109252
42 Dambach P Sie A Lacaux JP et al Using high spatial resolutionremote sensing for risk mapping of malaria occurrence in theNouna District Burkina Faso Glob Health Action 20092 doi103402ghav2io2094
43 Bogh C Lindsay SW Clarke SE et al High spatial resolution mappingof malaria transmission risk in The Gambia West Africa usingLANDSAT TM satellite imagery Am J Trop Med Hyg 200776875ndash81
44 Bicout DJ Sabatier P Mapping Rift Valley fever vectors and prevalenceusing rainfall variations Vector Borne Zoonotic Dis 2004433ndash42
45 Clements AC Pfeiffer DU Martin V et al Spatial risk assessment of RiftValley fever in Senegal Vector Borne Zoonotic Dis 20077203ndash16
46 Rogers DJ Hay SI Packer MJ Predicting the distribution of tsetse fliesin West Africa using temporal Fourier processed meteorologicalsatellite data Ann Trop Med Parasitol 199690225ndash41
47 Cecchi G Mattioli RC Slingenbergh J de la Rocque S Land cover andtsetse fly distributions in sub-Saharan Africa Med Vet Entomol200822364ndash73
48 Thomson MC Connor SJ Environmental information systems for thecontrol of arthropod vectors of disease Med Vet Entomol200014227ndash44
49 Hay SI Tucker CJ Rogers DJ Packer MJ Remotely sensed surrogatesof meteorological data for the study of the distribution andabundance of arthropod vectors of disease Ann Trop Med Parasitol1996901ndash19
50 Randolph SE Rogers DJ A generic population model for theAfrican tick Rhipicephalus appendiculatus Parasitology 1997115265ndash79
51 Cumming GS Comparing climate and vegetation as limiting factorsfor species ranges of African ticks Ecology 200283255ndash68
52 Ostfeld RS Glass GE Keesing F Spatial epidemiology an emerging (orre-emerging) discipline Trends Ecol Evol 200520328ndash36
53 Hay SI Tatem AJ Graham AJ et al Global environmental data formapping infectious disease distribution Adv Parasitol 20066237ndash77
54 Riedel N Vounatsou P Miller JM et al Geographical patterns andpredictors of malaria risk in Zambia Bayesian geostatisticalmodelling of the 2006 Zambia national malaria indicator survey(ZMIS) Malar J 2010937
55 Gosoniu L Veta AM Vounatsou P Bayesian geostatisticalmodeling of Malaria Indicator Survey data in Angola PLoS One20105e9322
56 Kalluri S Gilruth P Rogers D Szczur M Surveillance of arthropodvector-borne infectious diseases using remote sensing techniquesa review PLoS Pathog 200731361ndash71
57 Clements AC Pfeiffer DU Emerging viral zoonoses frameworks forspatial and spatiotemporal risk assessment and resource planningVet J 200918221ndash30
58 Phillips SJ Anderson RP Schapire RE Maximum entropy modelingof species geographic distributions Ecological Modelling 2006190231ndash59
59 Fenollar F Neglected and emerging diseases in sub-Saharan AfricaClin Microbiol Infect 201117957ndash8
60 Eisen RJ Eisen L Lane RS Predicting density of Ixodes pacificusnymphs in dense woodlands in Mendocino County Californiabased on geographic information systems and remotesensing versus field-derived data Am J Trop Med Hyg 200674632ndash40
P Ratmanov et al
12 of 12
at University of L
iverpool on May 9 2013
httptrstmhoxfordjournalsorg
Dow
nloaded from
Figure 6 Geographic distribution of dengue in West African countries in 2011 (data from the WHO retrieved from httpwwwwhoint)
Figure 7 Geographic distribution of yellow fever in West African countries in 2011 (data from the WHO retrieved from httpwwwwhoint)
P Ratmanov et al
8 of 12
at University of L
iverpool on May 9 2013
httptrstmhoxfordjournalsorg
Dow
nloaded from
Figure 8 Geographic distribution of CrimeanndashCongo haemorrhagic fever in West African countries (data from the WHO retrieved from httpwwwwhoint)
Figure 9 Geographic distribution of Buruli ulcer in 2010 in West African countries (data from the WHO retrieved from httpwwwwhoint)
Transactions of the Royal Society of Tropical Medicine and Hygiene
9 of 12
at University of L
iverpool on May 9 2013
httptrstmhoxfordjournalsorg
Dow
nloaded from
their advantages and limitations into account Moreover VBDinterventions have to be identified and recognised as they canconfound the relationship between the environment anddisease In the recent malaria indicator surveys in Zambia andAngola for example no relationship was observed between re-motely sensed data and malaria risk5455
The typical modelling approach investigates associationsbetween multivariate environmental data and patterns ofvector presence or absence for mapping vectors and VBDs Atthe first step simple statistical models could be a good startingpoint for linking the limited number of environmental variablesthat can be derived from satellite data Simple statisticalmodels are restricted because they often fit linear functionsbetween environmental variables and presenceabsence datawhen it is most likely that such associations are highlycomplex and non-linear20 At the second step sophisticatedprocess-based models that rely on vector biology as predictorsof diseases and their risk should be developed56
Regression models have been widely applied in landscape epi-demiology The type of regression model (logistic Poisson linearetc) is determined by the type of outcome variable to be pre-dicted (eg binary count continuous) and environmental vari-ables measured at sampled locations are entered ascovariates The resultant model is then either used to predictthe outcome variable at non-sampled locations based onobserved values of the covariates at the prediction locations orto explain observed patterns of disease on the basis of themodel covariates57
Besides regression modelling there are many otherapproaches in spatial epidemiology spatial clustering discrimin-ant analysis generalised linear models generalised additivemodels Bayesian estimation methods and others Howeversuch traditional models require both disease presence anddisease absence data At the same time there are differentlsquopresence-onlyrsquo models that could be used when no lsquoabsencedatarsquo are available One of these presence-only machine learning(rule-based) algorithms is an ecological niche modelling algo-rithm based on maximum entropy (Maxent) This method canbe used successfully with very small sample sizes and does notrequire independence of covariates58 Moreover these advan-tages could be crucial in Africa where complete surveillancedata from all regions are not available
Basic spatial modelling approaches include interpolationbased on spatial dependence in vector or VBD data and extrapo-lation based on associations between vector or VBD data and en-vironmental or socioeconomic predictor variables The latterapproach can be a powerful tool to gain insights into levels ofrisk within areas where surveillance data are lacking or unreli-able However it should be noted that model extrapolation isrestricted to areas with ecological and climatic characteristicssimilar to those of the model development area1
Global strategies for controlling infections in sub-SaharanAfrica have focused on the lsquobig threersquo diseases namely AIDSTB and malaria Other causes of fever and infection fall intothe category of NTDs In the twenty-first century it is importantto compile a comprehensive list of the infections in sub-SaharanAfrica and to study their epidemiology59
Few investigations have addressed the unknown causes offever in West Africa The majority of fevers have long been con-sidered to be related to malaria Recent studies have shown that
NTDs are the most frequent causes of fever in both indigenes andtravellers In particular TBRF due to B crocidurae has been redis-covered931 and the rickettsiae including R felis have beenreported33 Overall new pathogens among the rickettsiae wereidentified in the environment8
The geographic distribution of ticksmdashvectors of diseases inWest Africamdashis a neglected area of study However applyingGISRS technologies to the study of Ixodes ticks in North Americahas yielded significant results For example the use of geospatialmodelling has revealed that high concentrations of I pacificusand I scapularis which are the key tick vectors of Lyme diseasein North America can be predicted by GISRS-based environmen-tal factors related to elevation slope of the landscape vegetationtype soil type temperature and moisture60 The same approachescould be used in epidemiological studies of R felis infection anemerging disease in West Africa
Today epidemiologists often use new GISRS techniques tostudy a variety of VBDs The associations between satellite-derived environmental variables (such as temperature humidityelevation vegetation rainfall surface water land use land covertype and soil moisture) and vector density are used to identifyand characterise vector habitats A wide variety of RS datapowerful GIS and statistical software packages are available fordesktop computing environments making it affordable and feas-ible for epidemiologists to experiment with new spatial analysistechniques56
Maps that show the seasonal risk of VBDs will be necessary tomonitor the impacts of changes on vector ecology Using GISRSadvanced analytical tools and a landscape ecology approachthese risk maps could play a major role in defining research ques-tions and surveillance needs and in guiding control efforts andfield studies38
From the other side this approach is based on ecological ana-lysis and requires assembling all epidemiologic available dataand moreover it should be supported by fieldwork The most suc-cessful GISRS applications to study VBDs in West Africa areeither local (country)-level studies using community-based sur-veillance data or continental (sub-continental)-level studiesusing published scientific literature data Cooperation of the epi-demiological community could allow enhancing studies of VBDsin West Africa The most important goal of the application of GISRS technologies to the study of VBDs is to reduce the burden ofdisease by generating information that empowers the public totake protective action and helps public health agencies to effect-ively allocate their limited prevention surveillance and controlresources There is great potential to use GISRS technologiesto improve the surveillance prevention and control of vector-borne infectious diseases in West Africa
Supplementary dataSupplementary data are available at Transactions Online(httptrstmhoxfordjournalsorg)
Authorsrsquo contributions All authors contributed equally to themanuscript DR conceived the review PR and OM researched the dataand literature PR OM and DR analysed and interpreted the data for
P Ratmanov et al
10 of 12
at University of L
iverpool on May 9 2013
httptrstmhoxfordjournalsorg
Dow
nloaded from
the maps and discussed the content of the review PR and DR drafted themanuscript All authors critically revised the manuscript for intellectualcontent and read and approved the final version DR is guarantor ofthe paper
Funding None
Competing interests None declared
Ethical approval Not required
References1 Eisen L Eisen RJ Using geographic information systems and decision
support systems for the prediction prevention and control ofvector-borne diseases Annu Rev Entomol 20115641ndash61
2 WHO Neglected Tropical Diseases Geneva World HealthOrganization 2012 httpwwwwhointneglected_diseasesdiseasesen [accessed 8 January 2013]
3 Hay SI An overview of remote sensing and geodesy for epidemiologyand public health application Adv Parasitol 2000471ndash35
4 Pavlovsky EN Natural Nidality of Transmissible Diseases with SpecialReference to the Landscape Epidemiology of ZooanthroponosesUrbana IL University of Illinois Press 1966
5 Kitron U Landscape ecology and epidemiology of vector-bornediseases tools for spatial analysis J Med Entomol 199835435ndash45
6 Rogers DJ Randolph SE Studying the global distribution of infectiousdiseases using GIS and RS Nat Rev Microbiol 20031231ndash7
7 Eisen L Lozano-Fuentes S Use of mapping and spatial andspace-time modeling approaches in operational control of Aedesaegypti and dengue PLoS Negl Trop Dis 20093e411
8 Mediannikov O Diatta G Fenollar F et al Tick-borne rickettsiosesneglected emerging diseases in rural Senegal PLoS Negl Trop Dis20104e821
9 Parola P Diatta G Socolovschi C et al Tick-borne relapsing feverborreliosis rural Senegal Emerg Infect Dis 201117883ndash5
10 WHO World Malaria Report 2010 Geneva World HealthOrganization 2010
11 Sullivan D Uncertainty in mapping malaria epidemiologyimplications for control Epidemiol Rev 201032175ndash87
12 Machault V Vignolles C Borchi F et al The use of remotely sensedenvironmental data in the study of malaria Geospat Health20115151ndash68
13 Snow RW Marsh K le Sueur D The need for maps of transmissionintensity to guide malaria control in Africa Parasitol Today199612455ndash7
14 Guerra CA Hay SI Lucioparedes LS et al Assembling a globaldatabase of malaria parasite prevalence for the Malaria AtlasProject Malar J 2007617
15 Brun R Blum J Chappuis F Burri C Human African trypanosomiasisLancet 2010375148ndash59
16 Simarro PP Cecchi G Paone M et al The Atlas of human Africantrypanosomiasis a contribution to global mapping of neglectedtropical diseases Int J Health Geogr 2010957
17 Bern C Maguire JH Alvar J Complexities of assessing the diseaseburden attributable to leishmaniasis PLoS Negl Trop Dis 20082e313
18 WHO Report of the Scientific Working Group Meeting onLeishmaniasis (Geneva 2ndash4 February 2004) Geneva World HealthOrganization 2004
19 Faye B Bucheton B Banuls AL et al Seroprevalence of Leishmaniainfantum in a rural area of Senegal analysis of risk factors involvedin transmission to humans Trans R Soc Trop Med Hyg2011105333ndash40
20 Slater H Michael E Predicting the current and future potentialdistributions of lymphatic filariasis in Africa using maximumentropy ecological niche modelling PLoS One 20127e32202
21 Basanez MG Pion SD Churcher TS et al River blindness a successstory under threat PLoS Med 20063e371
22 Noma M Nwoke BE Nutall I et al Rapid epidemiological mapping ofonchocerciasis (REMO) its application by the African Programme forOnchocerciasis Control (APOC) Ann Trop Med Parasitol 200296(Suppl 1)S29ndash39
23 Zoure HG Wanji S Noma M et al The geographic distribution of Loaloa in Africa results of large-scale implementation of the RapidAssessment Procedure for Loiasis (RAPLOA) PLoS Negl Trop Dis20115e1210
24 Favier C Chalvet-Monfray K Sabatier P et al Rift Valley fever in WestAfrica the role of space in endemicity Trop Med Int Health2006111878ndash88
25 Jouan A Coulibaly I Adam F et al Analytical study of a Rift Valleyfever epidemic Res Virol 1989140175ndash86
26 Diallo M Ba Y Sall AA et al Amplification of the sylvatic cycle ofdengue virus type 2 Senegal 1999ndash2000 entomologicfindings and epidemiologic considerations Emerg Infect Dis20039362ndash7
27 Jentes ES Poumerol G Gershman MD et al The revised global yellowfever risk map and recommendations for vaccination 2010consensus of the Informal WHO Working Group on Geographic Riskfor Yellow Fever Lancet Infect Dis 201111622ndash32
28 Rogers DJ Wilson AJ Hay SI Graham AJ The global distribution ofyellow fever and dengue Adv Parasitol 200662181ndash220
29 Grard G Drexler JF Fair J et al Re-emergence of CrimeanndashCongohemorrhagic fever virus in Central Africa PLoS Negl Trop Dis20115e1350
30 Nabeth P Thior M Faye O Simon F Human CrimeanndashCongohemorrhagic fever Senegal Emerg Infect Dis 2004101881ndash2
31 Vial L Diatta G Tall A et al Incidence of tick-borne relapsing fever inWest Africa longitudinal study Lancet 200636837ndash43
32 Jensenius M Davis X von Sonnenburg F et al Multicenter GeoSentinelanalysis of rickettsial diseases in international travelers 1996ndash2008Emerg Infect Dis 2009151791ndash8
33 Socolovschi C Mediannikov O Sokhna C et al Rickettsiafelis-associated uneruptive fever Senegal Emerg Infect Dis2010161140ndash2
34 Cazorla C Socolovschi C Jensenius M Parola P Tick-borne diseasestick-borne spotted fever rickettsioses in Africa Infect Dis Clin NorthAm 200822p531ndash44ixndashx
35 Merritt RW Walker ED Small PL et al Ecology and transmission ofBuruli ulcer disease a systematic review PLoS Negl Trop Dis20104e911
36 Tissot-Dupont H Brouqui P Faugere B Raoult D Prevalence ofantibodies to Coxiella burnetti Rickettsia conorii and Rickettsia typhiin seven African countries Clin Infect Dis 1995211126ndash33
37 Mediannikov O Fenollar F Socolovschi C et al Coxiella burnetii inhumans and ticks in rural Senegal PLoS Negl Trop Dis 20104e654
38 Kitron U Risk maps transmission and burden of vector-bornediseases Parasitol Today 200016324ndash5
Transactions of the Royal Society of Tropical Medicine and Hygiene
11 of 12
at University of L
iverpool on May 9 2013
httptrstmhoxfordjournalsorg
Dow
nloaded from
39 Thomson MC Connor SJ Milligan P Flasse SP Mapping malaria risk inAfrica what can satellite data contribute Parasitol Today199713313ndash8
40 Matthys B Koudou BG NrsquoGoran EK et al Spatial dispersion andcharacterisation of mosquito breeding habitats in urbanvegetable-production areas of Abidjan Cote drsquoIvoire Ann Trop MedParasitol 2010104649ndash66
41 Machault V Vignolles C Pages F et al Spatial heterogeneity andtemporal evolution of malaria transmission risk in Dakar Senegalaccording to remotely sensed environmental data Malar J 20109252
42 Dambach P Sie A Lacaux JP et al Using high spatial resolutionremote sensing for risk mapping of malaria occurrence in theNouna District Burkina Faso Glob Health Action 20092 doi103402ghav2io2094
43 Bogh C Lindsay SW Clarke SE et al High spatial resolution mappingof malaria transmission risk in The Gambia West Africa usingLANDSAT TM satellite imagery Am J Trop Med Hyg 200776875ndash81
44 Bicout DJ Sabatier P Mapping Rift Valley fever vectors and prevalenceusing rainfall variations Vector Borne Zoonotic Dis 2004433ndash42
45 Clements AC Pfeiffer DU Martin V et al Spatial risk assessment of RiftValley fever in Senegal Vector Borne Zoonotic Dis 20077203ndash16
46 Rogers DJ Hay SI Packer MJ Predicting the distribution of tsetse fliesin West Africa using temporal Fourier processed meteorologicalsatellite data Ann Trop Med Parasitol 199690225ndash41
47 Cecchi G Mattioli RC Slingenbergh J de la Rocque S Land cover andtsetse fly distributions in sub-Saharan Africa Med Vet Entomol200822364ndash73
48 Thomson MC Connor SJ Environmental information systems for thecontrol of arthropod vectors of disease Med Vet Entomol200014227ndash44
49 Hay SI Tucker CJ Rogers DJ Packer MJ Remotely sensed surrogatesof meteorological data for the study of the distribution andabundance of arthropod vectors of disease Ann Trop Med Parasitol1996901ndash19
50 Randolph SE Rogers DJ A generic population model for theAfrican tick Rhipicephalus appendiculatus Parasitology 1997115265ndash79
51 Cumming GS Comparing climate and vegetation as limiting factorsfor species ranges of African ticks Ecology 200283255ndash68
52 Ostfeld RS Glass GE Keesing F Spatial epidemiology an emerging (orre-emerging) discipline Trends Ecol Evol 200520328ndash36
53 Hay SI Tatem AJ Graham AJ et al Global environmental data formapping infectious disease distribution Adv Parasitol 20066237ndash77
54 Riedel N Vounatsou P Miller JM et al Geographical patterns andpredictors of malaria risk in Zambia Bayesian geostatisticalmodelling of the 2006 Zambia national malaria indicator survey(ZMIS) Malar J 2010937
55 Gosoniu L Veta AM Vounatsou P Bayesian geostatisticalmodeling of Malaria Indicator Survey data in Angola PLoS One20105e9322
56 Kalluri S Gilruth P Rogers D Szczur M Surveillance of arthropodvector-borne infectious diseases using remote sensing techniquesa review PLoS Pathog 200731361ndash71
57 Clements AC Pfeiffer DU Emerging viral zoonoses frameworks forspatial and spatiotemporal risk assessment and resource planningVet J 200918221ndash30
58 Phillips SJ Anderson RP Schapire RE Maximum entropy modelingof species geographic distributions Ecological Modelling 2006190231ndash59
59 Fenollar F Neglected and emerging diseases in sub-Saharan AfricaClin Microbiol Infect 201117957ndash8
60 Eisen RJ Eisen L Lane RS Predicting density of Ixodes pacificusnymphs in dense woodlands in Mendocino County Californiabased on geographic information systems and remotesensing versus field-derived data Am J Trop Med Hyg 200674632ndash40
P Ratmanov et al
12 of 12
at University of L
iverpool on May 9 2013
httptrstmhoxfordjournalsorg
Dow
nloaded from
Figure 8 Geographic distribution of CrimeanndashCongo haemorrhagic fever in West African countries (data from the WHO retrieved from httpwwwwhoint)
Figure 9 Geographic distribution of Buruli ulcer in 2010 in West African countries (data from the WHO retrieved from httpwwwwhoint)
Transactions of the Royal Society of Tropical Medicine and Hygiene
9 of 12
at University of L
iverpool on May 9 2013
httptrstmhoxfordjournalsorg
Dow
nloaded from
their advantages and limitations into account Moreover VBDinterventions have to be identified and recognised as they canconfound the relationship between the environment anddisease In the recent malaria indicator surveys in Zambia andAngola for example no relationship was observed between re-motely sensed data and malaria risk5455
The typical modelling approach investigates associationsbetween multivariate environmental data and patterns ofvector presence or absence for mapping vectors and VBDs Atthe first step simple statistical models could be a good startingpoint for linking the limited number of environmental variablesthat can be derived from satellite data Simple statisticalmodels are restricted because they often fit linear functionsbetween environmental variables and presenceabsence datawhen it is most likely that such associations are highlycomplex and non-linear20 At the second step sophisticatedprocess-based models that rely on vector biology as predictorsof diseases and their risk should be developed56
Regression models have been widely applied in landscape epi-demiology The type of regression model (logistic Poisson linearetc) is determined by the type of outcome variable to be pre-dicted (eg binary count continuous) and environmental vari-ables measured at sampled locations are entered ascovariates The resultant model is then either used to predictthe outcome variable at non-sampled locations based onobserved values of the covariates at the prediction locations orto explain observed patterns of disease on the basis of themodel covariates57
Besides regression modelling there are many otherapproaches in spatial epidemiology spatial clustering discrimin-ant analysis generalised linear models generalised additivemodels Bayesian estimation methods and others Howeversuch traditional models require both disease presence anddisease absence data At the same time there are differentlsquopresence-onlyrsquo models that could be used when no lsquoabsencedatarsquo are available One of these presence-only machine learning(rule-based) algorithms is an ecological niche modelling algo-rithm based on maximum entropy (Maxent) This method canbe used successfully with very small sample sizes and does notrequire independence of covariates58 Moreover these advan-tages could be crucial in Africa where complete surveillancedata from all regions are not available
Basic spatial modelling approaches include interpolationbased on spatial dependence in vector or VBD data and extrapo-lation based on associations between vector or VBD data and en-vironmental or socioeconomic predictor variables The latterapproach can be a powerful tool to gain insights into levels ofrisk within areas where surveillance data are lacking or unreli-able However it should be noted that model extrapolation isrestricted to areas with ecological and climatic characteristicssimilar to those of the model development area1
Global strategies for controlling infections in sub-SaharanAfrica have focused on the lsquobig threersquo diseases namely AIDSTB and malaria Other causes of fever and infection fall intothe category of NTDs In the twenty-first century it is importantto compile a comprehensive list of the infections in sub-SaharanAfrica and to study their epidemiology59
Few investigations have addressed the unknown causes offever in West Africa The majority of fevers have long been con-sidered to be related to malaria Recent studies have shown that
NTDs are the most frequent causes of fever in both indigenes andtravellers In particular TBRF due to B crocidurae has been redis-covered931 and the rickettsiae including R felis have beenreported33 Overall new pathogens among the rickettsiae wereidentified in the environment8
The geographic distribution of ticksmdashvectors of diseases inWest Africamdashis a neglected area of study However applyingGISRS technologies to the study of Ixodes ticks in North Americahas yielded significant results For example the use of geospatialmodelling has revealed that high concentrations of I pacificusand I scapularis which are the key tick vectors of Lyme diseasein North America can be predicted by GISRS-based environmen-tal factors related to elevation slope of the landscape vegetationtype soil type temperature and moisture60 The same approachescould be used in epidemiological studies of R felis infection anemerging disease in West Africa
Today epidemiologists often use new GISRS techniques tostudy a variety of VBDs The associations between satellite-derived environmental variables (such as temperature humidityelevation vegetation rainfall surface water land use land covertype and soil moisture) and vector density are used to identifyand characterise vector habitats A wide variety of RS datapowerful GIS and statistical software packages are available fordesktop computing environments making it affordable and feas-ible for epidemiologists to experiment with new spatial analysistechniques56
Maps that show the seasonal risk of VBDs will be necessary tomonitor the impacts of changes on vector ecology Using GISRSadvanced analytical tools and a landscape ecology approachthese risk maps could play a major role in defining research ques-tions and surveillance needs and in guiding control efforts andfield studies38
From the other side this approach is based on ecological ana-lysis and requires assembling all epidemiologic available dataand moreover it should be supported by fieldwork The most suc-cessful GISRS applications to study VBDs in West Africa areeither local (country)-level studies using community-based sur-veillance data or continental (sub-continental)-level studiesusing published scientific literature data Cooperation of the epi-demiological community could allow enhancing studies of VBDsin West Africa The most important goal of the application of GISRS technologies to the study of VBDs is to reduce the burden ofdisease by generating information that empowers the public totake protective action and helps public health agencies to effect-ively allocate their limited prevention surveillance and controlresources There is great potential to use GISRS technologiesto improve the surveillance prevention and control of vector-borne infectious diseases in West Africa
Supplementary dataSupplementary data are available at Transactions Online(httptrstmhoxfordjournalsorg)
Authorsrsquo contributions All authors contributed equally to themanuscript DR conceived the review PR and OM researched the dataand literature PR OM and DR analysed and interpreted the data for
P Ratmanov et al
10 of 12
at University of L
iverpool on May 9 2013
httptrstmhoxfordjournalsorg
Dow
nloaded from
the maps and discussed the content of the review PR and DR drafted themanuscript All authors critically revised the manuscript for intellectualcontent and read and approved the final version DR is guarantor ofthe paper
Funding None
Competing interests None declared
Ethical approval Not required
References1 Eisen L Eisen RJ Using geographic information systems and decision
support systems for the prediction prevention and control ofvector-borne diseases Annu Rev Entomol 20115641ndash61
2 WHO Neglected Tropical Diseases Geneva World HealthOrganization 2012 httpwwwwhointneglected_diseasesdiseasesen [accessed 8 January 2013]
3 Hay SI An overview of remote sensing and geodesy for epidemiologyand public health application Adv Parasitol 2000471ndash35
4 Pavlovsky EN Natural Nidality of Transmissible Diseases with SpecialReference to the Landscape Epidemiology of ZooanthroponosesUrbana IL University of Illinois Press 1966
5 Kitron U Landscape ecology and epidemiology of vector-bornediseases tools for spatial analysis J Med Entomol 199835435ndash45
6 Rogers DJ Randolph SE Studying the global distribution of infectiousdiseases using GIS and RS Nat Rev Microbiol 20031231ndash7
7 Eisen L Lozano-Fuentes S Use of mapping and spatial andspace-time modeling approaches in operational control of Aedesaegypti and dengue PLoS Negl Trop Dis 20093e411
8 Mediannikov O Diatta G Fenollar F et al Tick-borne rickettsiosesneglected emerging diseases in rural Senegal PLoS Negl Trop Dis20104e821
9 Parola P Diatta G Socolovschi C et al Tick-borne relapsing feverborreliosis rural Senegal Emerg Infect Dis 201117883ndash5
10 WHO World Malaria Report 2010 Geneva World HealthOrganization 2010
11 Sullivan D Uncertainty in mapping malaria epidemiologyimplications for control Epidemiol Rev 201032175ndash87
12 Machault V Vignolles C Borchi F et al The use of remotely sensedenvironmental data in the study of malaria Geospat Health20115151ndash68
13 Snow RW Marsh K le Sueur D The need for maps of transmissionintensity to guide malaria control in Africa Parasitol Today199612455ndash7
14 Guerra CA Hay SI Lucioparedes LS et al Assembling a globaldatabase of malaria parasite prevalence for the Malaria AtlasProject Malar J 2007617
15 Brun R Blum J Chappuis F Burri C Human African trypanosomiasisLancet 2010375148ndash59
16 Simarro PP Cecchi G Paone M et al The Atlas of human Africantrypanosomiasis a contribution to global mapping of neglectedtropical diseases Int J Health Geogr 2010957
17 Bern C Maguire JH Alvar J Complexities of assessing the diseaseburden attributable to leishmaniasis PLoS Negl Trop Dis 20082e313
18 WHO Report of the Scientific Working Group Meeting onLeishmaniasis (Geneva 2ndash4 February 2004) Geneva World HealthOrganization 2004
19 Faye B Bucheton B Banuls AL et al Seroprevalence of Leishmaniainfantum in a rural area of Senegal analysis of risk factors involvedin transmission to humans Trans R Soc Trop Med Hyg2011105333ndash40
20 Slater H Michael E Predicting the current and future potentialdistributions of lymphatic filariasis in Africa using maximumentropy ecological niche modelling PLoS One 20127e32202
21 Basanez MG Pion SD Churcher TS et al River blindness a successstory under threat PLoS Med 20063e371
22 Noma M Nwoke BE Nutall I et al Rapid epidemiological mapping ofonchocerciasis (REMO) its application by the African Programme forOnchocerciasis Control (APOC) Ann Trop Med Parasitol 200296(Suppl 1)S29ndash39
23 Zoure HG Wanji S Noma M et al The geographic distribution of Loaloa in Africa results of large-scale implementation of the RapidAssessment Procedure for Loiasis (RAPLOA) PLoS Negl Trop Dis20115e1210
24 Favier C Chalvet-Monfray K Sabatier P et al Rift Valley fever in WestAfrica the role of space in endemicity Trop Med Int Health2006111878ndash88
25 Jouan A Coulibaly I Adam F et al Analytical study of a Rift Valleyfever epidemic Res Virol 1989140175ndash86
26 Diallo M Ba Y Sall AA et al Amplification of the sylvatic cycle ofdengue virus type 2 Senegal 1999ndash2000 entomologicfindings and epidemiologic considerations Emerg Infect Dis20039362ndash7
27 Jentes ES Poumerol G Gershman MD et al The revised global yellowfever risk map and recommendations for vaccination 2010consensus of the Informal WHO Working Group on Geographic Riskfor Yellow Fever Lancet Infect Dis 201111622ndash32
28 Rogers DJ Wilson AJ Hay SI Graham AJ The global distribution ofyellow fever and dengue Adv Parasitol 200662181ndash220
29 Grard G Drexler JF Fair J et al Re-emergence of CrimeanndashCongohemorrhagic fever virus in Central Africa PLoS Negl Trop Dis20115e1350
30 Nabeth P Thior M Faye O Simon F Human CrimeanndashCongohemorrhagic fever Senegal Emerg Infect Dis 2004101881ndash2
31 Vial L Diatta G Tall A et al Incidence of tick-borne relapsing fever inWest Africa longitudinal study Lancet 200636837ndash43
32 Jensenius M Davis X von Sonnenburg F et al Multicenter GeoSentinelanalysis of rickettsial diseases in international travelers 1996ndash2008Emerg Infect Dis 2009151791ndash8
33 Socolovschi C Mediannikov O Sokhna C et al Rickettsiafelis-associated uneruptive fever Senegal Emerg Infect Dis2010161140ndash2
34 Cazorla C Socolovschi C Jensenius M Parola P Tick-borne diseasestick-borne spotted fever rickettsioses in Africa Infect Dis Clin NorthAm 200822p531ndash44ixndashx
35 Merritt RW Walker ED Small PL et al Ecology and transmission ofBuruli ulcer disease a systematic review PLoS Negl Trop Dis20104e911
36 Tissot-Dupont H Brouqui P Faugere B Raoult D Prevalence ofantibodies to Coxiella burnetti Rickettsia conorii and Rickettsia typhiin seven African countries Clin Infect Dis 1995211126ndash33
37 Mediannikov O Fenollar F Socolovschi C et al Coxiella burnetii inhumans and ticks in rural Senegal PLoS Negl Trop Dis 20104e654
38 Kitron U Risk maps transmission and burden of vector-bornediseases Parasitol Today 200016324ndash5
Transactions of the Royal Society of Tropical Medicine and Hygiene
11 of 12
at University of L
iverpool on May 9 2013
httptrstmhoxfordjournalsorg
Dow
nloaded from
39 Thomson MC Connor SJ Milligan P Flasse SP Mapping malaria risk inAfrica what can satellite data contribute Parasitol Today199713313ndash8
40 Matthys B Koudou BG NrsquoGoran EK et al Spatial dispersion andcharacterisation of mosquito breeding habitats in urbanvegetable-production areas of Abidjan Cote drsquoIvoire Ann Trop MedParasitol 2010104649ndash66
41 Machault V Vignolles C Pages F et al Spatial heterogeneity andtemporal evolution of malaria transmission risk in Dakar Senegalaccording to remotely sensed environmental data Malar J 20109252
42 Dambach P Sie A Lacaux JP et al Using high spatial resolutionremote sensing for risk mapping of malaria occurrence in theNouna District Burkina Faso Glob Health Action 20092 doi103402ghav2io2094
43 Bogh C Lindsay SW Clarke SE et al High spatial resolution mappingof malaria transmission risk in The Gambia West Africa usingLANDSAT TM satellite imagery Am J Trop Med Hyg 200776875ndash81
44 Bicout DJ Sabatier P Mapping Rift Valley fever vectors and prevalenceusing rainfall variations Vector Borne Zoonotic Dis 2004433ndash42
45 Clements AC Pfeiffer DU Martin V et al Spatial risk assessment of RiftValley fever in Senegal Vector Borne Zoonotic Dis 20077203ndash16
46 Rogers DJ Hay SI Packer MJ Predicting the distribution of tsetse fliesin West Africa using temporal Fourier processed meteorologicalsatellite data Ann Trop Med Parasitol 199690225ndash41
47 Cecchi G Mattioli RC Slingenbergh J de la Rocque S Land cover andtsetse fly distributions in sub-Saharan Africa Med Vet Entomol200822364ndash73
48 Thomson MC Connor SJ Environmental information systems for thecontrol of arthropod vectors of disease Med Vet Entomol200014227ndash44
49 Hay SI Tucker CJ Rogers DJ Packer MJ Remotely sensed surrogatesof meteorological data for the study of the distribution andabundance of arthropod vectors of disease Ann Trop Med Parasitol1996901ndash19
50 Randolph SE Rogers DJ A generic population model for theAfrican tick Rhipicephalus appendiculatus Parasitology 1997115265ndash79
51 Cumming GS Comparing climate and vegetation as limiting factorsfor species ranges of African ticks Ecology 200283255ndash68
52 Ostfeld RS Glass GE Keesing F Spatial epidemiology an emerging (orre-emerging) discipline Trends Ecol Evol 200520328ndash36
53 Hay SI Tatem AJ Graham AJ et al Global environmental data formapping infectious disease distribution Adv Parasitol 20066237ndash77
54 Riedel N Vounatsou P Miller JM et al Geographical patterns andpredictors of malaria risk in Zambia Bayesian geostatisticalmodelling of the 2006 Zambia national malaria indicator survey(ZMIS) Malar J 2010937
55 Gosoniu L Veta AM Vounatsou P Bayesian geostatisticalmodeling of Malaria Indicator Survey data in Angola PLoS One20105e9322
56 Kalluri S Gilruth P Rogers D Szczur M Surveillance of arthropodvector-borne infectious diseases using remote sensing techniquesa review PLoS Pathog 200731361ndash71
57 Clements AC Pfeiffer DU Emerging viral zoonoses frameworks forspatial and spatiotemporal risk assessment and resource planningVet J 200918221ndash30
58 Phillips SJ Anderson RP Schapire RE Maximum entropy modelingof species geographic distributions Ecological Modelling 2006190231ndash59
59 Fenollar F Neglected and emerging diseases in sub-Saharan AfricaClin Microbiol Infect 201117957ndash8
60 Eisen RJ Eisen L Lane RS Predicting density of Ixodes pacificusnymphs in dense woodlands in Mendocino County Californiabased on geographic information systems and remotesensing versus field-derived data Am J Trop Med Hyg 200674632ndash40
P Ratmanov et al
12 of 12
at University of L
iverpool on May 9 2013
httptrstmhoxfordjournalsorg
Dow
nloaded from
their advantages and limitations into account Moreover VBDinterventions have to be identified and recognised as they canconfound the relationship between the environment anddisease In the recent malaria indicator surveys in Zambia andAngola for example no relationship was observed between re-motely sensed data and malaria risk5455
The typical modelling approach investigates associationsbetween multivariate environmental data and patterns ofvector presence or absence for mapping vectors and VBDs Atthe first step simple statistical models could be a good startingpoint for linking the limited number of environmental variablesthat can be derived from satellite data Simple statisticalmodels are restricted because they often fit linear functionsbetween environmental variables and presenceabsence datawhen it is most likely that such associations are highlycomplex and non-linear20 At the second step sophisticatedprocess-based models that rely on vector biology as predictorsof diseases and their risk should be developed56
Regression models have been widely applied in landscape epi-demiology The type of regression model (logistic Poisson linearetc) is determined by the type of outcome variable to be pre-dicted (eg binary count continuous) and environmental vari-ables measured at sampled locations are entered ascovariates The resultant model is then either used to predictthe outcome variable at non-sampled locations based onobserved values of the covariates at the prediction locations orto explain observed patterns of disease on the basis of themodel covariates57
Besides regression modelling there are many otherapproaches in spatial epidemiology spatial clustering discrimin-ant analysis generalised linear models generalised additivemodels Bayesian estimation methods and others Howeversuch traditional models require both disease presence anddisease absence data At the same time there are differentlsquopresence-onlyrsquo models that could be used when no lsquoabsencedatarsquo are available One of these presence-only machine learning(rule-based) algorithms is an ecological niche modelling algo-rithm based on maximum entropy (Maxent) This method canbe used successfully with very small sample sizes and does notrequire independence of covariates58 Moreover these advan-tages could be crucial in Africa where complete surveillancedata from all regions are not available
Basic spatial modelling approaches include interpolationbased on spatial dependence in vector or VBD data and extrapo-lation based on associations between vector or VBD data and en-vironmental or socioeconomic predictor variables The latterapproach can be a powerful tool to gain insights into levels ofrisk within areas where surveillance data are lacking or unreli-able However it should be noted that model extrapolation isrestricted to areas with ecological and climatic characteristicssimilar to those of the model development area1
Global strategies for controlling infections in sub-SaharanAfrica have focused on the lsquobig threersquo diseases namely AIDSTB and malaria Other causes of fever and infection fall intothe category of NTDs In the twenty-first century it is importantto compile a comprehensive list of the infections in sub-SaharanAfrica and to study their epidemiology59
Few investigations have addressed the unknown causes offever in West Africa The majority of fevers have long been con-sidered to be related to malaria Recent studies have shown that
NTDs are the most frequent causes of fever in both indigenes andtravellers In particular TBRF due to B crocidurae has been redis-covered931 and the rickettsiae including R felis have beenreported33 Overall new pathogens among the rickettsiae wereidentified in the environment8
The geographic distribution of ticksmdashvectors of diseases inWest Africamdashis a neglected area of study However applyingGISRS technologies to the study of Ixodes ticks in North Americahas yielded significant results For example the use of geospatialmodelling has revealed that high concentrations of I pacificusand I scapularis which are the key tick vectors of Lyme diseasein North America can be predicted by GISRS-based environmen-tal factors related to elevation slope of the landscape vegetationtype soil type temperature and moisture60 The same approachescould be used in epidemiological studies of R felis infection anemerging disease in West Africa
Today epidemiologists often use new GISRS techniques tostudy a variety of VBDs The associations between satellite-derived environmental variables (such as temperature humidityelevation vegetation rainfall surface water land use land covertype and soil moisture) and vector density are used to identifyand characterise vector habitats A wide variety of RS datapowerful GIS and statistical software packages are available fordesktop computing environments making it affordable and feas-ible for epidemiologists to experiment with new spatial analysistechniques56
Maps that show the seasonal risk of VBDs will be necessary tomonitor the impacts of changes on vector ecology Using GISRSadvanced analytical tools and a landscape ecology approachthese risk maps could play a major role in defining research ques-tions and surveillance needs and in guiding control efforts andfield studies38
From the other side this approach is based on ecological ana-lysis and requires assembling all epidemiologic available dataand moreover it should be supported by fieldwork The most suc-cessful GISRS applications to study VBDs in West Africa areeither local (country)-level studies using community-based sur-veillance data or continental (sub-continental)-level studiesusing published scientific literature data Cooperation of the epi-demiological community could allow enhancing studies of VBDsin West Africa The most important goal of the application of GISRS technologies to the study of VBDs is to reduce the burden ofdisease by generating information that empowers the public totake protective action and helps public health agencies to effect-ively allocate their limited prevention surveillance and controlresources There is great potential to use GISRS technologiesto improve the surveillance prevention and control of vector-borne infectious diseases in West Africa
Supplementary dataSupplementary data are available at Transactions Online(httptrstmhoxfordjournalsorg)
Authorsrsquo contributions All authors contributed equally to themanuscript DR conceived the review PR and OM researched the dataand literature PR OM and DR analysed and interpreted the data for
P Ratmanov et al
10 of 12
at University of L
iverpool on May 9 2013
httptrstmhoxfordjournalsorg
Dow
nloaded from
the maps and discussed the content of the review PR and DR drafted themanuscript All authors critically revised the manuscript for intellectualcontent and read and approved the final version DR is guarantor ofthe paper
Funding None
Competing interests None declared
Ethical approval Not required
References1 Eisen L Eisen RJ Using geographic information systems and decision
support systems for the prediction prevention and control ofvector-borne diseases Annu Rev Entomol 20115641ndash61
2 WHO Neglected Tropical Diseases Geneva World HealthOrganization 2012 httpwwwwhointneglected_diseasesdiseasesen [accessed 8 January 2013]
3 Hay SI An overview of remote sensing and geodesy for epidemiologyand public health application Adv Parasitol 2000471ndash35
4 Pavlovsky EN Natural Nidality of Transmissible Diseases with SpecialReference to the Landscape Epidemiology of ZooanthroponosesUrbana IL University of Illinois Press 1966
5 Kitron U Landscape ecology and epidemiology of vector-bornediseases tools for spatial analysis J Med Entomol 199835435ndash45
6 Rogers DJ Randolph SE Studying the global distribution of infectiousdiseases using GIS and RS Nat Rev Microbiol 20031231ndash7
7 Eisen L Lozano-Fuentes S Use of mapping and spatial andspace-time modeling approaches in operational control of Aedesaegypti and dengue PLoS Negl Trop Dis 20093e411
8 Mediannikov O Diatta G Fenollar F et al Tick-borne rickettsiosesneglected emerging diseases in rural Senegal PLoS Negl Trop Dis20104e821
9 Parola P Diatta G Socolovschi C et al Tick-borne relapsing feverborreliosis rural Senegal Emerg Infect Dis 201117883ndash5
10 WHO World Malaria Report 2010 Geneva World HealthOrganization 2010
11 Sullivan D Uncertainty in mapping malaria epidemiologyimplications for control Epidemiol Rev 201032175ndash87
12 Machault V Vignolles C Borchi F et al The use of remotely sensedenvironmental data in the study of malaria Geospat Health20115151ndash68
13 Snow RW Marsh K le Sueur D The need for maps of transmissionintensity to guide malaria control in Africa Parasitol Today199612455ndash7
14 Guerra CA Hay SI Lucioparedes LS et al Assembling a globaldatabase of malaria parasite prevalence for the Malaria AtlasProject Malar J 2007617
15 Brun R Blum J Chappuis F Burri C Human African trypanosomiasisLancet 2010375148ndash59
16 Simarro PP Cecchi G Paone M et al The Atlas of human Africantrypanosomiasis a contribution to global mapping of neglectedtropical diseases Int J Health Geogr 2010957
17 Bern C Maguire JH Alvar J Complexities of assessing the diseaseburden attributable to leishmaniasis PLoS Negl Trop Dis 20082e313
18 WHO Report of the Scientific Working Group Meeting onLeishmaniasis (Geneva 2ndash4 February 2004) Geneva World HealthOrganization 2004
19 Faye B Bucheton B Banuls AL et al Seroprevalence of Leishmaniainfantum in a rural area of Senegal analysis of risk factors involvedin transmission to humans Trans R Soc Trop Med Hyg2011105333ndash40
20 Slater H Michael E Predicting the current and future potentialdistributions of lymphatic filariasis in Africa using maximumentropy ecological niche modelling PLoS One 20127e32202
21 Basanez MG Pion SD Churcher TS et al River blindness a successstory under threat PLoS Med 20063e371
22 Noma M Nwoke BE Nutall I et al Rapid epidemiological mapping ofonchocerciasis (REMO) its application by the African Programme forOnchocerciasis Control (APOC) Ann Trop Med Parasitol 200296(Suppl 1)S29ndash39
23 Zoure HG Wanji S Noma M et al The geographic distribution of Loaloa in Africa results of large-scale implementation of the RapidAssessment Procedure for Loiasis (RAPLOA) PLoS Negl Trop Dis20115e1210
24 Favier C Chalvet-Monfray K Sabatier P et al Rift Valley fever in WestAfrica the role of space in endemicity Trop Med Int Health2006111878ndash88
25 Jouan A Coulibaly I Adam F et al Analytical study of a Rift Valleyfever epidemic Res Virol 1989140175ndash86
26 Diallo M Ba Y Sall AA et al Amplification of the sylvatic cycle ofdengue virus type 2 Senegal 1999ndash2000 entomologicfindings and epidemiologic considerations Emerg Infect Dis20039362ndash7
27 Jentes ES Poumerol G Gershman MD et al The revised global yellowfever risk map and recommendations for vaccination 2010consensus of the Informal WHO Working Group on Geographic Riskfor Yellow Fever Lancet Infect Dis 201111622ndash32
28 Rogers DJ Wilson AJ Hay SI Graham AJ The global distribution ofyellow fever and dengue Adv Parasitol 200662181ndash220
29 Grard G Drexler JF Fair J et al Re-emergence of CrimeanndashCongohemorrhagic fever virus in Central Africa PLoS Negl Trop Dis20115e1350
30 Nabeth P Thior M Faye O Simon F Human CrimeanndashCongohemorrhagic fever Senegal Emerg Infect Dis 2004101881ndash2
31 Vial L Diatta G Tall A et al Incidence of tick-borne relapsing fever inWest Africa longitudinal study Lancet 200636837ndash43
32 Jensenius M Davis X von Sonnenburg F et al Multicenter GeoSentinelanalysis of rickettsial diseases in international travelers 1996ndash2008Emerg Infect Dis 2009151791ndash8
33 Socolovschi C Mediannikov O Sokhna C et al Rickettsiafelis-associated uneruptive fever Senegal Emerg Infect Dis2010161140ndash2
34 Cazorla C Socolovschi C Jensenius M Parola P Tick-borne diseasestick-borne spotted fever rickettsioses in Africa Infect Dis Clin NorthAm 200822p531ndash44ixndashx
35 Merritt RW Walker ED Small PL et al Ecology and transmission ofBuruli ulcer disease a systematic review PLoS Negl Trop Dis20104e911
36 Tissot-Dupont H Brouqui P Faugere B Raoult D Prevalence ofantibodies to Coxiella burnetti Rickettsia conorii and Rickettsia typhiin seven African countries Clin Infect Dis 1995211126ndash33
37 Mediannikov O Fenollar F Socolovschi C et al Coxiella burnetii inhumans and ticks in rural Senegal PLoS Negl Trop Dis 20104e654
38 Kitron U Risk maps transmission and burden of vector-bornediseases Parasitol Today 200016324ndash5
Transactions of the Royal Society of Tropical Medicine and Hygiene
11 of 12
at University of L
iverpool on May 9 2013
httptrstmhoxfordjournalsorg
Dow
nloaded from
39 Thomson MC Connor SJ Milligan P Flasse SP Mapping malaria risk inAfrica what can satellite data contribute Parasitol Today199713313ndash8
40 Matthys B Koudou BG NrsquoGoran EK et al Spatial dispersion andcharacterisation of mosquito breeding habitats in urbanvegetable-production areas of Abidjan Cote drsquoIvoire Ann Trop MedParasitol 2010104649ndash66
41 Machault V Vignolles C Pages F et al Spatial heterogeneity andtemporal evolution of malaria transmission risk in Dakar Senegalaccording to remotely sensed environmental data Malar J 20109252
42 Dambach P Sie A Lacaux JP et al Using high spatial resolutionremote sensing for risk mapping of malaria occurrence in theNouna District Burkina Faso Glob Health Action 20092 doi103402ghav2io2094
43 Bogh C Lindsay SW Clarke SE et al High spatial resolution mappingof malaria transmission risk in The Gambia West Africa usingLANDSAT TM satellite imagery Am J Trop Med Hyg 200776875ndash81
44 Bicout DJ Sabatier P Mapping Rift Valley fever vectors and prevalenceusing rainfall variations Vector Borne Zoonotic Dis 2004433ndash42
45 Clements AC Pfeiffer DU Martin V et al Spatial risk assessment of RiftValley fever in Senegal Vector Borne Zoonotic Dis 20077203ndash16
46 Rogers DJ Hay SI Packer MJ Predicting the distribution of tsetse fliesin West Africa using temporal Fourier processed meteorologicalsatellite data Ann Trop Med Parasitol 199690225ndash41
47 Cecchi G Mattioli RC Slingenbergh J de la Rocque S Land cover andtsetse fly distributions in sub-Saharan Africa Med Vet Entomol200822364ndash73
48 Thomson MC Connor SJ Environmental information systems for thecontrol of arthropod vectors of disease Med Vet Entomol200014227ndash44
49 Hay SI Tucker CJ Rogers DJ Packer MJ Remotely sensed surrogatesof meteorological data for the study of the distribution andabundance of arthropod vectors of disease Ann Trop Med Parasitol1996901ndash19
50 Randolph SE Rogers DJ A generic population model for theAfrican tick Rhipicephalus appendiculatus Parasitology 1997115265ndash79
51 Cumming GS Comparing climate and vegetation as limiting factorsfor species ranges of African ticks Ecology 200283255ndash68
52 Ostfeld RS Glass GE Keesing F Spatial epidemiology an emerging (orre-emerging) discipline Trends Ecol Evol 200520328ndash36
53 Hay SI Tatem AJ Graham AJ et al Global environmental data formapping infectious disease distribution Adv Parasitol 20066237ndash77
54 Riedel N Vounatsou P Miller JM et al Geographical patterns andpredictors of malaria risk in Zambia Bayesian geostatisticalmodelling of the 2006 Zambia national malaria indicator survey(ZMIS) Malar J 2010937
55 Gosoniu L Veta AM Vounatsou P Bayesian geostatisticalmodeling of Malaria Indicator Survey data in Angola PLoS One20105e9322
56 Kalluri S Gilruth P Rogers D Szczur M Surveillance of arthropodvector-borne infectious diseases using remote sensing techniquesa review PLoS Pathog 200731361ndash71
57 Clements AC Pfeiffer DU Emerging viral zoonoses frameworks forspatial and spatiotemporal risk assessment and resource planningVet J 200918221ndash30
58 Phillips SJ Anderson RP Schapire RE Maximum entropy modelingof species geographic distributions Ecological Modelling 2006190231ndash59
59 Fenollar F Neglected and emerging diseases in sub-Saharan AfricaClin Microbiol Infect 201117957ndash8
60 Eisen RJ Eisen L Lane RS Predicting density of Ixodes pacificusnymphs in dense woodlands in Mendocino County Californiabased on geographic information systems and remotesensing versus field-derived data Am J Trop Med Hyg 200674632ndash40
P Ratmanov et al
12 of 12
at University of L
iverpool on May 9 2013
httptrstmhoxfordjournalsorg
Dow
nloaded from
the maps and discussed the content of the review PR and DR drafted themanuscript All authors critically revised the manuscript for intellectualcontent and read and approved the final version DR is guarantor ofthe paper
Funding None
Competing interests None declared
Ethical approval Not required
References1 Eisen L Eisen RJ Using geographic information systems and decision
support systems for the prediction prevention and control ofvector-borne diseases Annu Rev Entomol 20115641ndash61
2 WHO Neglected Tropical Diseases Geneva World HealthOrganization 2012 httpwwwwhointneglected_diseasesdiseasesen [accessed 8 January 2013]
3 Hay SI An overview of remote sensing and geodesy for epidemiologyand public health application Adv Parasitol 2000471ndash35
4 Pavlovsky EN Natural Nidality of Transmissible Diseases with SpecialReference to the Landscape Epidemiology of ZooanthroponosesUrbana IL University of Illinois Press 1966
5 Kitron U Landscape ecology and epidemiology of vector-bornediseases tools for spatial analysis J Med Entomol 199835435ndash45
6 Rogers DJ Randolph SE Studying the global distribution of infectiousdiseases using GIS and RS Nat Rev Microbiol 20031231ndash7
7 Eisen L Lozano-Fuentes S Use of mapping and spatial andspace-time modeling approaches in operational control of Aedesaegypti and dengue PLoS Negl Trop Dis 20093e411
8 Mediannikov O Diatta G Fenollar F et al Tick-borne rickettsiosesneglected emerging diseases in rural Senegal PLoS Negl Trop Dis20104e821
9 Parola P Diatta G Socolovschi C et al Tick-borne relapsing feverborreliosis rural Senegal Emerg Infect Dis 201117883ndash5
10 WHO World Malaria Report 2010 Geneva World HealthOrganization 2010
11 Sullivan D Uncertainty in mapping malaria epidemiologyimplications for control Epidemiol Rev 201032175ndash87
12 Machault V Vignolles C Borchi F et al The use of remotely sensedenvironmental data in the study of malaria Geospat Health20115151ndash68
13 Snow RW Marsh K le Sueur D The need for maps of transmissionintensity to guide malaria control in Africa Parasitol Today199612455ndash7
14 Guerra CA Hay SI Lucioparedes LS et al Assembling a globaldatabase of malaria parasite prevalence for the Malaria AtlasProject Malar J 2007617
15 Brun R Blum J Chappuis F Burri C Human African trypanosomiasisLancet 2010375148ndash59
16 Simarro PP Cecchi G Paone M et al The Atlas of human Africantrypanosomiasis a contribution to global mapping of neglectedtropical diseases Int J Health Geogr 2010957
17 Bern C Maguire JH Alvar J Complexities of assessing the diseaseburden attributable to leishmaniasis PLoS Negl Trop Dis 20082e313
18 WHO Report of the Scientific Working Group Meeting onLeishmaniasis (Geneva 2ndash4 February 2004) Geneva World HealthOrganization 2004
19 Faye B Bucheton B Banuls AL et al Seroprevalence of Leishmaniainfantum in a rural area of Senegal analysis of risk factors involvedin transmission to humans Trans R Soc Trop Med Hyg2011105333ndash40
20 Slater H Michael E Predicting the current and future potentialdistributions of lymphatic filariasis in Africa using maximumentropy ecological niche modelling PLoS One 20127e32202
21 Basanez MG Pion SD Churcher TS et al River blindness a successstory under threat PLoS Med 20063e371
22 Noma M Nwoke BE Nutall I et al Rapid epidemiological mapping ofonchocerciasis (REMO) its application by the African Programme forOnchocerciasis Control (APOC) Ann Trop Med Parasitol 200296(Suppl 1)S29ndash39
23 Zoure HG Wanji S Noma M et al The geographic distribution of Loaloa in Africa results of large-scale implementation of the RapidAssessment Procedure for Loiasis (RAPLOA) PLoS Negl Trop Dis20115e1210
24 Favier C Chalvet-Monfray K Sabatier P et al Rift Valley fever in WestAfrica the role of space in endemicity Trop Med Int Health2006111878ndash88
25 Jouan A Coulibaly I Adam F et al Analytical study of a Rift Valleyfever epidemic Res Virol 1989140175ndash86
26 Diallo M Ba Y Sall AA et al Amplification of the sylvatic cycle ofdengue virus type 2 Senegal 1999ndash2000 entomologicfindings and epidemiologic considerations Emerg Infect Dis20039362ndash7
27 Jentes ES Poumerol G Gershman MD et al The revised global yellowfever risk map and recommendations for vaccination 2010consensus of the Informal WHO Working Group on Geographic Riskfor Yellow Fever Lancet Infect Dis 201111622ndash32
28 Rogers DJ Wilson AJ Hay SI Graham AJ The global distribution ofyellow fever and dengue Adv Parasitol 200662181ndash220
29 Grard G Drexler JF Fair J et al Re-emergence of CrimeanndashCongohemorrhagic fever virus in Central Africa PLoS Negl Trop Dis20115e1350
30 Nabeth P Thior M Faye O Simon F Human CrimeanndashCongohemorrhagic fever Senegal Emerg Infect Dis 2004101881ndash2
31 Vial L Diatta G Tall A et al Incidence of tick-borne relapsing fever inWest Africa longitudinal study Lancet 200636837ndash43
32 Jensenius M Davis X von Sonnenburg F et al Multicenter GeoSentinelanalysis of rickettsial diseases in international travelers 1996ndash2008Emerg Infect Dis 2009151791ndash8
33 Socolovschi C Mediannikov O Sokhna C et al Rickettsiafelis-associated uneruptive fever Senegal Emerg Infect Dis2010161140ndash2
34 Cazorla C Socolovschi C Jensenius M Parola P Tick-borne diseasestick-borne spotted fever rickettsioses in Africa Infect Dis Clin NorthAm 200822p531ndash44ixndashx
35 Merritt RW Walker ED Small PL et al Ecology and transmission ofBuruli ulcer disease a systematic review PLoS Negl Trop Dis20104e911
36 Tissot-Dupont H Brouqui P Faugere B Raoult D Prevalence ofantibodies to Coxiella burnetti Rickettsia conorii and Rickettsia typhiin seven African countries Clin Infect Dis 1995211126ndash33
37 Mediannikov O Fenollar F Socolovschi C et al Coxiella burnetii inhumans and ticks in rural Senegal PLoS Negl Trop Dis 20104e654
38 Kitron U Risk maps transmission and burden of vector-bornediseases Parasitol Today 200016324ndash5
Transactions of the Royal Society of Tropical Medicine and Hygiene
11 of 12
at University of L
iverpool on May 9 2013
httptrstmhoxfordjournalsorg
Dow
nloaded from
39 Thomson MC Connor SJ Milligan P Flasse SP Mapping malaria risk inAfrica what can satellite data contribute Parasitol Today199713313ndash8
40 Matthys B Koudou BG NrsquoGoran EK et al Spatial dispersion andcharacterisation of mosquito breeding habitats in urbanvegetable-production areas of Abidjan Cote drsquoIvoire Ann Trop MedParasitol 2010104649ndash66
41 Machault V Vignolles C Pages F et al Spatial heterogeneity andtemporal evolution of malaria transmission risk in Dakar Senegalaccording to remotely sensed environmental data Malar J 20109252
42 Dambach P Sie A Lacaux JP et al Using high spatial resolutionremote sensing for risk mapping of malaria occurrence in theNouna District Burkina Faso Glob Health Action 20092 doi103402ghav2io2094
43 Bogh C Lindsay SW Clarke SE et al High spatial resolution mappingof malaria transmission risk in The Gambia West Africa usingLANDSAT TM satellite imagery Am J Trop Med Hyg 200776875ndash81
44 Bicout DJ Sabatier P Mapping Rift Valley fever vectors and prevalenceusing rainfall variations Vector Borne Zoonotic Dis 2004433ndash42
45 Clements AC Pfeiffer DU Martin V et al Spatial risk assessment of RiftValley fever in Senegal Vector Borne Zoonotic Dis 20077203ndash16
46 Rogers DJ Hay SI Packer MJ Predicting the distribution of tsetse fliesin West Africa using temporal Fourier processed meteorologicalsatellite data Ann Trop Med Parasitol 199690225ndash41
47 Cecchi G Mattioli RC Slingenbergh J de la Rocque S Land cover andtsetse fly distributions in sub-Saharan Africa Med Vet Entomol200822364ndash73
48 Thomson MC Connor SJ Environmental information systems for thecontrol of arthropod vectors of disease Med Vet Entomol200014227ndash44
49 Hay SI Tucker CJ Rogers DJ Packer MJ Remotely sensed surrogatesof meteorological data for the study of the distribution andabundance of arthropod vectors of disease Ann Trop Med Parasitol1996901ndash19
50 Randolph SE Rogers DJ A generic population model for theAfrican tick Rhipicephalus appendiculatus Parasitology 1997115265ndash79
51 Cumming GS Comparing climate and vegetation as limiting factorsfor species ranges of African ticks Ecology 200283255ndash68
52 Ostfeld RS Glass GE Keesing F Spatial epidemiology an emerging (orre-emerging) discipline Trends Ecol Evol 200520328ndash36
53 Hay SI Tatem AJ Graham AJ et al Global environmental data formapping infectious disease distribution Adv Parasitol 20066237ndash77
54 Riedel N Vounatsou P Miller JM et al Geographical patterns andpredictors of malaria risk in Zambia Bayesian geostatisticalmodelling of the 2006 Zambia national malaria indicator survey(ZMIS) Malar J 2010937
55 Gosoniu L Veta AM Vounatsou P Bayesian geostatisticalmodeling of Malaria Indicator Survey data in Angola PLoS One20105e9322
56 Kalluri S Gilruth P Rogers D Szczur M Surveillance of arthropodvector-borne infectious diseases using remote sensing techniquesa review PLoS Pathog 200731361ndash71
57 Clements AC Pfeiffer DU Emerging viral zoonoses frameworks forspatial and spatiotemporal risk assessment and resource planningVet J 200918221ndash30
58 Phillips SJ Anderson RP Schapire RE Maximum entropy modelingof species geographic distributions Ecological Modelling 2006190231ndash59
59 Fenollar F Neglected and emerging diseases in sub-Saharan AfricaClin Microbiol Infect 201117957ndash8
60 Eisen RJ Eisen L Lane RS Predicting density of Ixodes pacificusnymphs in dense woodlands in Mendocino County Californiabased on geographic information systems and remotesensing versus field-derived data Am J Trop Med Hyg 200674632ndash40
P Ratmanov et al
12 of 12
at University of L
iverpool on May 9 2013
httptrstmhoxfordjournalsorg
Dow
nloaded from
39 Thomson MC Connor SJ Milligan P Flasse SP Mapping malaria risk inAfrica what can satellite data contribute Parasitol Today199713313ndash8
40 Matthys B Koudou BG NrsquoGoran EK et al Spatial dispersion andcharacterisation of mosquito breeding habitats in urbanvegetable-production areas of Abidjan Cote drsquoIvoire Ann Trop MedParasitol 2010104649ndash66
41 Machault V Vignolles C Pages F et al Spatial heterogeneity andtemporal evolution of malaria transmission risk in Dakar Senegalaccording to remotely sensed environmental data Malar J 20109252
42 Dambach P Sie A Lacaux JP et al Using high spatial resolutionremote sensing for risk mapping of malaria occurrence in theNouna District Burkina Faso Glob Health Action 20092 doi103402ghav2io2094
43 Bogh C Lindsay SW Clarke SE et al High spatial resolution mappingof malaria transmission risk in The Gambia West Africa usingLANDSAT TM satellite imagery Am J Trop Med Hyg 200776875ndash81
44 Bicout DJ Sabatier P Mapping Rift Valley fever vectors and prevalenceusing rainfall variations Vector Borne Zoonotic Dis 2004433ndash42
45 Clements AC Pfeiffer DU Martin V et al Spatial risk assessment of RiftValley fever in Senegal Vector Borne Zoonotic Dis 20077203ndash16
46 Rogers DJ Hay SI Packer MJ Predicting the distribution of tsetse fliesin West Africa using temporal Fourier processed meteorologicalsatellite data Ann Trop Med Parasitol 199690225ndash41
47 Cecchi G Mattioli RC Slingenbergh J de la Rocque S Land cover andtsetse fly distributions in sub-Saharan Africa Med Vet Entomol200822364ndash73
48 Thomson MC Connor SJ Environmental information systems for thecontrol of arthropod vectors of disease Med Vet Entomol200014227ndash44
49 Hay SI Tucker CJ Rogers DJ Packer MJ Remotely sensed surrogatesof meteorological data for the study of the distribution andabundance of arthropod vectors of disease Ann Trop Med Parasitol1996901ndash19
50 Randolph SE Rogers DJ A generic population model for theAfrican tick Rhipicephalus appendiculatus Parasitology 1997115265ndash79
51 Cumming GS Comparing climate and vegetation as limiting factorsfor species ranges of African ticks Ecology 200283255ndash68
52 Ostfeld RS Glass GE Keesing F Spatial epidemiology an emerging (orre-emerging) discipline Trends Ecol Evol 200520328ndash36
53 Hay SI Tatem AJ Graham AJ et al Global environmental data formapping infectious disease distribution Adv Parasitol 20066237ndash77
54 Riedel N Vounatsou P Miller JM et al Geographical patterns andpredictors of malaria risk in Zambia Bayesian geostatisticalmodelling of the 2006 Zambia national malaria indicator survey(ZMIS) Malar J 2010937
55 Gosoniu L Veta AM Vounatsou P Bayesian geostatisticalmodeling of Malaria Indicator Survey data in Angola PLoS One20105e9322
56 Kalluri S Gilruth P Rogers D Szczur M Surveillance of arthropodvector-borne infectious diseases using remote sensing techniquesa review PLoS Pathog 200731361ndash71
57 Clements AC Pfeiffer DU Emerging viral zoonoses frameworks forspatial and spatiotemporal risk assessment and resource planningVet J 200918221ndash30
58 Phillips SJ Anderson RP Schapire RE Maximum entropy modelingof species geographic distributions Ecological Modelling 2006190231ndash59
59 Fenollar F Neglected and emerging diseases in sub-Saharan AfricaClin Microbiol Infect 201117957ndash8
60 Eisen RJ Eisen L Lane RS Predicting density of Ixodes pacificusnymphs in dense woodlands in Mendocino County Californiabased on geographic information systems and remotesensing versus field-derived data Am J Trop Med Hyg 200674632ndash40
P Ratmanov et al
12 of 12
at University of L
iverpool on May 9 2013
httptrstmhoxfordjournalsorg
Dow
nloaded from