Government & Politics

Geospatial tools for the identification of a malaria corridor in Estado Sucre, a Venezuelan north-eastern state

Abstract. Landscape ecology research relies on frameworks based on geographical information systems (GIS), geostatistics and spatial-feature relationships. With regard to health, the approach consists of systems analysis using a set of powerful tools
of 8
All materials on our website are shared by users. If you have any questions about copyright issues, please report us to resolve them. We are always happy to assist you.
Related Documents
  Geospatial tools for the identification of a malaria corridor inEstado Sucre, a Venezuelan north-eastern state Laura Delgado-Petrocelli 1 , Alberto Camardiel 2 , Víctor Hugo Aguilar 3 , Néstor Martinez 4 ,Karenia Córdova 3 , Santiago Ramos 1 1 Ecological Information Systems and Environmental Modelling Laboratory (SIMEA), Institute of Tropical Zoology, Central University of Caracas, Venezuela; 2 Postgraduate Area in Statistics and Actuarial Sciences,Faculty of Economy and Social Sciences, Central University of Caracas, Venezuela; 3 Institute of Geography and Regional Development, Faculty of Humanities and Education, Central University of Caracas, Venezuela; 4 School of Geography, Faculty of Humanities and Education, Central University of Caracas, Venezuela Abstract. Landscape ecology research relies on frameworks based on geographical information systems (GIS), geostatisticsand spatial-feature relationships. With regard to health, the approach consists of systems analysis using a set of powerfultools aimed at the reduction of community vulnerability through improved public policies. The north-oriental malaria focus,one of five such foci in Venezuela, situated in the north-eastern part of the Estado Sucre state, unites several social and envi-ronmental features and functions as an epidemiological corridor, i.e. an endemic zone characterised by permanent interac-tion between the mosquito vector and the human host allowing a continuous persistence of the malaria lifecycle. A GIS wasdeveloped based on official cartography with thematic overlays depicting malaria distribution, socio-economic conditions,basic environmental information and specific features associated with the natural wetlands present in the area. Generally,malaria foci are continuously active but when the malaria situation was modelled in the north-oriental focus, a differential,spatio-temporal distribution pattern situation was found, i.e. a situation oscillating between very active and dormant trans-mission. This pattern was displayed by spatial and statistical analysis based on the model generated in this study and theresults were confirmed by municipal and county malaria records. Control of malaria, keeping the incidence at a permanentlylow level within the regional population, should be possible if these results are taken into account when designing and imple-menting epidemiological surveillance policies. Keywords: malaria, geographical information system, remote sensing, spatial analysis, landscape ecology, epidemiologicalcorridor, Venezuela. Introduction Climate pattern change, natural disasters, cumula-tive environmental impact and the emergence and re-emergence of infectious diseases exert an increasingimpact on the world of today. Examples of vulnerablecommunities with humans living under critical condi-tions due to poverty include localities ravished bymalaria and/or other tropical diseases, and where theenvironment is negatively influenced by anthropogenicactivity. In such places, a systemic approach is morethan ever needed if we are to efficiently manage pub-lic health issues. The implementation of an effectiveenvironmental strategy based on the use of geospatialtechnologies such as digital terrain models (DTM),remote sensing, digital image processing and globalpositioning systems (GPS) is called for. These applica-tions produce superior spatial analysis through inte-grated approaches based on geographical informationsystems (GIS) and open new possibilities to solve exist-ing health problems. Advances in informatics andcomputational power, storage and handling of largevolumes of data, have exploded during the latestdecades and contributed to the development of new,powerful GIS approaches, capable of initiating novelways and means to deal with diseases such as malaria,Chagas disease, dengue, leishmaniasis and other trop-ical diseases.Early on, Kitron (1998) recognised the need for sys-temic approaches to study the epidemiology of para-sitic diseases. In the People’s Republic of China, thegenetic discontinuity of the intermediate host snails of schistosomiasis was shown to be ultimately influ-enced by landscape ecology (Li et al., 2009), whileDongus et al. (2009) expose how agricultural and Corresponding author:Santiago RamosEcological Information Systems andEnvironmental Modeling Laboratory (SIMEA)Caracas, VenezuelaTel. +58 212 605 1311; Fax +58 212 605 1204E-mail: Geospatial Health 5(2), 2011, pp. 169-176  L. Delgado-Petrocelli et al. - Geospatial Health 5(2), 2011, pp. 169-176 170 geographical features influence the presence of  Anopheles larvae in urban areas in Dar es Salam,Tanzania. Some authors (Beck et al., 2000; Kitron,2000) recommend spatial analysis and utilisation of the potential of GIS for infectious disease manage-ment, while others (Romaña et al., 2003) have con-ceptualised the epidemiological landscape notion in away that underlines the importance of ecology. Theidea of epidemic foci as part of landscape ecology isof paramount interest for our approach, as it holdsthat persistence of endemicity in a geographical local-ity is supported, indeed ensured, by environmentaland edaphic (soil characteristic) conditions furtheringthe coexistence of parasite, vector and host (Romañaet al., 2003). Barrera et al.,(1998) studied a malariahotspot with strong transmission in Venezuela fromthat point of view, and this research group alsoapplied this approach to characterise an epidemic of hemorrhagic dengue fever in this country (Barrera etal., 2000). Subsequently, Delgado and Ramos (2007)developed a systemic framework derived from theconcept of landscape ecology by combining varioustechnologies. This perspective not only provides asuperior comprehension of the dynamics of ruralmalaria endemicity, but also highlights the relation-ship between spatial variables such as height andslope with malaria risk. The work presented here con-siders the malaria incidence in the Paria Peninsula, aregion of the Estado Sucre state in Venezuela, in thelight of geographical factors. Here, elementary land-scape components such as the north-south variationbetween highland and lowland and the topographicalcoast-to-inland gradients are treated as a matrix dis-playing patches and corridor patterns with respect tothis disease.The hypothesis put forward holds that sequentiallylocated malaria foci promote the establishment of con-ditions that support persistent disease, each focus gen-erating/contributing to an epidemiological, endemiccorridor by spatially relaying the disease along one ormore geographical features. This important compo-nent of the ecological landscape concept provides aperspective that we have named “panoramic epidemi-ology” (Delgado et al., 2007), an approach capable of pinpointing environmental factors which could bedecisive when designing effective policies for the con-trol and surveillance of parasitic diseases. Our objec-tive is to show, through the combination of spatial andstatistical analyses, the presence and dynamics of anepidemiological malaria corridor in the PariaPeninsula part of Estado Sucre in the north-easternpart of Venezuela. Material and methods Study area Paria Peninsula has a surface of 11,800 km 2 and islocated between latitudes 10°13’10” and 10°44’10”north and longitudes 61°50’44” and 64°30’00” west(Fig. 1). The Estado Sucre state is divided into 15municipalities, in turn subdivided into 55 smaller,administrative units called “parroquias”. Spatial and epidemiological data The study was conducted using malaria incidencedata gathered from records for the 28 parroquias thatmake up Paria Peninsula. The parroquia was taken asthe spatial unit for the statistical analyses of themonthly records of the number of malaria casesreported to the local health centres between 1986 and1999. These data were gathered from the malaria divi-sion of the former Ministry for Sanitation and SocialAssistance (SAS) and the current state office in chargeof regional public health (FUNDASALUD-SUCRE).The total number of cases included in the investigationwas 4,704. Fig. 1. Geographical location of the study area in Venezuela.  L. Delgado-Petrocelli et al. - Geospatial Health 5(2), 2011, pp. 169-176 171 The meteorological variables related to the El NiñoSouthern Oscillation (ENSO) phenomenon for eachparroquia included in the study, were considered.Criteria from the National Oceanographic andAtmospheric Administration (NOAA) of the UnitedStates of America were used and the data correspon-ding to climatic variability came from registers pub-lished in reports by the NOAA Earth System ResearchLaboratory (ESRL; http://www.esrl. Forthe GIS approach we relied on basic vectorial covers(overlays) which include roads, main hydrographicinformation, human population centres as well astopographic information such as ground level curvesplus height and slope. We included information onlocal climes including isohyets, i.e. lines drawnthrough geographical points recording equal amountsof precipitation during a specific period. GIS was alsoused to draw a response surface, whose line densitywas proportional to the number of cases per year andplace. The spatial scale was 1:100,000. Methodology The full study period was 1986-1999, but we alsogenerated some epidemiologic scenarios from 1997 to1999 using the “Malaria Incidence Spatial Model”(MISM), previously developed by Delgado et al.(2001, 2007). We used ANOVA ( for variance analysis to assess the overall sig-nificance of sets of main effects and interactions andside-by-side box-plots to display all the components of value-splitting by considering their mean squares. Theusefulness of this approach to the study of the malar-ia incidence in the Paria Peninsula region was enhacedusing a particularly robust and comprehensive versionof ANOVA (Hoaglin et al.,1991). This approachemphasises the exploratory point of view, while retain-ing the traditional methods of least-squares estima-tion. In addition to the ANOVA table, numerical andgraphical displays were added to gain insight into thedata structure. This technique facilitates the explana-tion of the effects of spatial and climatic variabilityproduced by ENSO events taking intensity and fre-quency into account.In order to confirm whether or not there is a statis-tically significant association between the number of malaria cases in the parroquias and the spatio-tempo-ral factors under consideration, we analysed values,consisting of the natural logarithms + 0.5, of the num-ber of malaria cases as a 4-way layout without repli-cation. The transformation of the response was donein order to induce a more symmetric distribution of the transformed variable than the srcinal one and toapproximate the basic assumptions required to per-form a valid ANOVA analysis (Figs. 2 and 3). Factorsused to investigate the response variable included“Month” with 12 levels, the “Time Period” with twolevels (1986-1989 and 1990-1999), the “ENSOPhenomenon” with seven levels (i.e. weak Niña, mod-erate Niña, strong Niña, neutral, weak Niño, moder-ate Niño, strong Niño) and the “Parroquia” unit with28 levels.Since the spatial factor presented a considerableeffect on the number of malaria cases, we furtheranalysed data using the multiple comparison testbased on Tukey’s honest significant difference (HSD)according to Hoaglin et al. (1991). All comparisonsbetween the parroquias were done pair-wise control-ling for the experimental error between the levels of the geographical factor. This procedure evaluates thedifference between pairs of averages of the responsevariable assuring extra protection to control the over- Fig. 2. Box-plots based on the number of malaria cases (left) and on the natural log of the case number + 0.5 (right).  DF, degrees of freedom; SS, sum of squares; MS, mean square. L. Delgado-Petrocelli et al. - Geospatial Health 5(2), 2011, pp. 169-176 172 all error rate under the null hypothesis of no differencein the set of effects. The error rate is the probabilitythat one or more comparisons lead to an incorrectconclusion of having found a significant difference. Results Table 1 gives the summary of the ANOVA resultsand Figure 3 displays side-by-side box-plots of the fit-ted main effects, double-factor interactions and resid-uals. The “Month” variable was found to have verysmall effect relative to the “Time Period”, “ENSOPhenomenon” and “Parroquia” ones, and was there-fore deemed to be unimportant. Among the double-factor interactions, “Period” + “ENSO” produced thelargest effects. The largest values, positive or negative,issuing from the ANOVA analysis and the side-by-sidebox-plots (Fig. 3.), came from the “ENSO” and“Parroquia” variables, which also showed the largestmain effects. Thus, there exists a strong statisticallysignificant association (P <0.001) between the levels of the climatic factor, those of the spatial factor and thoseof the time factor (though this may be explained by thefact that the duration of each period was different). Table 1. ANOVA results for the number of malaria cases (expressed as natural log + 0.5) of the effects of the variables and combi-nations thereof investigated. Source factorsDFSSMSF ratioP valueMonthPeriodParroquiaENSO PhenomenonMonth * PeriodMonth * ParroquiaMonth * ENSOPeriod * ParroquiaPeriod * ENSOParroquia * ENSOMonth * Period * ParroquiaMonth * Period * ENSOMonth * Parroquia * ENSOPeriod * Parroquia * ENSOErrorTotal1112761129766276162297661,7821,621,7824,70374.48317.253,364.581,765.3461.11128.351,102.86399.13181.371,161.24146.461,205.561,307.84242.731,274.6112,732.926.77317.25124.61294.225.560.4316.7114.7830.237.170.4918.270.731.500.729.47443.54174.22411.357.770.623.3620.6742.2610.020.6925.541.032.09<0.001<0.001<0.001<0.001<0.0011.00<0.001<0.001<0.001<0.0011.00<0.0010.29<0.001 Fig. 3. Comparative box-plots of the adjusted factorial effects.  L. Delgado-Petrocelli et al. - Geospatial Health 5(2), 2011, pp. 169-176 173 Furthermore, we found significant interactionsbetween the climatic phenomenon and the temporaldimension, and between the temporal and the spatialdimensions, but they were much weaker than whatwas found for the main effects. The goodness of fit of the model given by the square multiple correlationcoefficient (R 2 ) was 90% and the adjusted R 2 for thenumber of factors in the model was 74%.The HSD-Tukey multiple comparison test confirmedthe presence of a true, epidemiological corridor that isparticularly evident due to the assisted effect of theroad axis between the villages El Pilar and Guiria, sit-uated within their homonymous parroquias (Table 2).We were able to define 11 homogeneous groups of parroquias that did not differ in terms of the responsevariable. The existence of a series of sets with differentaverage numbers of malaria cases was verified. Thesets could be ranked according to the average numberof malaria cases starting with set no. 1 (G1), whichhad the highest average number of cases(Yaguaraparo, Irapa, El Paujil and El Pilar) down toset no. 10 (G10), which included the Guiria andAntonio José de Sucre parroquias. A zero set, consist-ing of 13 parroquias, had the lowest number of newmalaria cases (Table 2).With regard to slope, the gradient ranges between 0-2% and 2-5% showed the highest number of malariacases. Figure 4 shows the malaria intensity during1997, when foci appeared near the Campoma Lagoonin the Cariaco coastal zone as well as along the PariaPeninsula. A relationship was shown between locali-ties situated in the lowlands (i.e. elevations not sur-passing 100 m above the mean sea level). When themalaria dynamics were analysed and comparedsequentially over the years it was possible, asdescribed previously (Delgado et al., 2004; Delgado,2005; Delgado and Ramos, 2007), to identify locali-ties that correspond to particular malaria foci (Fig. 4).Interestingly, the same approach also worked whenattempting to correlate malaria to height and slopewith 200 m above the mean sea levelappearing as thethreshold variable in an ecological interpretation forthe particular distribution of malaria in the region.Figure 4 also shows that the northern coastal focihave diminished between 1997 and 1999, while thosein the Paria zone persisted. When GIS overlays wereproduced, using contour lines simultaneously with theMISM (Delgado et al., 2001), clearer areas correspon-ding to higher localities appeared, most of them at thesouthern part of the state, in contrast to the darkerareas that correspond to the lowlands particularly inthe area of the Paria Peninsula. The risk for malariatransmission there is reinforced by nearby wetlandsconstitute ideal vector habitats promoting reproduc-tion and maintenance of stable mosquito populations. Fig. 4. The “Malaria Incidence Spatial Model” (MISM) for 1997 and 1999.Table 2. The parroquias grouped according to the main effectsgradient obtained by the HSD-Tukey procedure based on thenumber of malaria cases (expressed as natural log + 0.5). NumberParroquiaGRUPO 0182125311612416102379Puerto SantoSan Antonio de IrapaTunapuicitoBideauFrancisco Antonio VásquezCristóbal ColónGuaraúnosCampo ClaroMarabalEl RincónSoroEl Morro de Puerto SantoEl Pilarn. s.n. s.n. s.n. s.n. s.n. s.n. s.n. s.n. s.n. s.n. s.n. s.n. s. n. s., not significant
Similar documents
View more...
Related Search
We Need Your Support
Thank you for visiting our website and your interest in our free products and services. We are nonprofit website to share and download documents. To the running of this website, we need your help to support us.

Thanks to everyone for your continued support.

No, Thanks

We need your sign to support Project to invent "SMART AND CONTROLLABLE REFLECTIVE BALLOONS" to cover the Sun and Save Our Earth.

More details...

Sign Now!

We are very appreciated for your Prompt Action!