Quantifying cross-border movements and migrations for guiding the strategic planning of malaria control and elimination

MISSING IMAGE

Material Information

Title:
Quantifying cross-border movements and migrations for guiding the strategic planning of malaria control and elimination
Physical Description:
Mixed Material
Language:
English
Creator:
Pindolia, Deepa K.
Garcia, Andres J.
Huang, Zhuojie
Fik, Timothy
Smith, David L.
Tatem, Andrew J.
Publisher:
Bio Med Central (Malaria Journal)
Publication Date:

Notes

Abstract:
Background: Identifying human and malaria parasite movements is important for control planning across all transmission intensities. Imported infections can reintroduce infections into areas previously free of infection, maintain ‘hotspots’ of transmission and import drug resistant strains, challenging national control programmes at a variety of temporal and spatial scales. Recent analyses based on mobile phone usage data have provided valuable insights into population and likely parasite movements within countries, but these data are restricted to sub-national analyses, leaving important cross-border movements neglected. Methods: National census data were used to analyse and model cross-border migration and movement, using East Africa as an example. ‘Hotspots’ of origin-specific immigrants from neighbouring countries were identified for Kenya, Tanzania and Uganda. Populations of origin-specific migrants were compared to distance from origin country borders and population size at destination, and regression models were developed to quantify and compare differences in migration patterns. Migration data were then combined with existing spatially-referenced malaria data to compare the relative propensity for cross-border malaria movement in the region. Results: The spatial patterns and processes for immigration were different between each origin and destination country pair. Hotspots of immigration, for example, were concentrated close to origin country borders for most immigrants to Tanzania, but for Kenya, a similar pattern was only seen for Tanzanian and Ugandan immigrants. Regression model fits also differed between specific migrant groups, with some migration patterns more dependent on population size at destination and distance travelled than others. With these differences between immigration patterns and processes, and heterogeneous transmission risk in East Africa and the surrounding region, propensities to import malaria infections also likely show substantial variations. Conclusion: This was a first attempt to quantify and model cross-border movements relevant to malaria transmission and control. With national census available worldwide, this approach can be translated to construct a cross-border human and malaria movement evidence base for other malaria endemic countries. The outcomes of this study will feed into wider efforts to quantify and model human and malaria movements in endemic regions to facilitate improved intervention planning, resource allocation and collaborative policy decisions.
General Note:
Pindolia et al. Malaria Journal 2014, 13:169 http://www.malariajournal.com/content/13/1/169; Pages 1-11
General Note:
doi:10.1186/1475-2875-13-169 Cite this article as: Pindolia et al.: Quantifying cross-border movements and migrations for guiding the strategic planning of malaria control and elimination. Malaria Journal 2014 13:169.

Record Information

Source Institution:
University of Florida
Holding Location:
University of Florida
Rights Management:
© 2014 Pindolia et al.; licensee BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly credited. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.
System ID:
AA00024683:00001

Full Text

PAGE 1

1 Comparing cross border and within country (internal) migration in Kenya, stratified by age 2 Comparing cross border and within country (internal) migration in Kenya, stratified by age and gender Methods used to generate networks of internal and cross border migrants were similar to methods developed and appli ed in Pindolia et al [1] W ith no data on cross border migrant origins, mean in degree and mean in graph strength were used instead of mean degree and mean graph strength, which incorporate HPM in both directions References : 1. Pindolia DK, Garcia AJ, Wesolowski A, Smith DL, Buckee CO, Noor AM, Snow RW, Tatem AJ: Human movement data for malaria control and elimination strategic planning. Malar J 2012, 11: 205.



PAGE 1

(a)Ethiopian in migrant hotspots (b)Somali in migrants (c)Sudanese in migrants (d)Ugandan in migrants (e)Tanzanian in migrants

PAGE 2

Hotspots indicating possible between country collaborations.



PAGE 1

Equations for model 2 split by destination country Kenya: = + + + + + + + + + + + + + + + Tanzania: = + + + + + + + + + + + + + + + + + + + + + + + + Uganda: = + + + + + + + + + + + + + + + Mi,j: total number of migrants from origin, i, to destination, j. Pj: total population size of destination location within each destination country. Di,j: Shortest distance between origin country border and destination location. K1: Somalia, K2: Ethiopia, K3: Sudan, K4: Uganda (do not need a dummy for the fifth country, Tanzania) T1: Kenya, T2: Uganda, T3: Rwanda, T4: Burundi, T5: DRC, T6: Zambia, T7: Malawi (do not a dummy for the fifth country, Mozambique) U1: Kenya, U2: Sudan, U3: DRC, U4: Rwanda (do not a dummy for the fifth country, Tanzania) i,j are exponents, specific to each origin destination pair or destination, estimated from the data.



PAGE 1

M odel 2 for Kenya, extending model to include age and gender : = + + + + + + + + + + + + + + + + + + + + + + + Aa: malariarelevant age categories; a=8; 1: < 5 years; 2: 5 9 years; 3: 10 19 years; 4: 20 29 years; 5: 3039 years; 6: 4049 years; 7: 5059 years; 8: 60 plus Gg: gender categories; g=2; 1: male, 2: female Regression output: Explanatory variables Regression Coef ficient S tandard Error t p value logPop 0.3318945 0.0092155 36.015 < 2e 16 logDist 0.3924936 0.0070194 55.916 < 2e 16 factor(AGE)2 0.0491894 0.0365404 1.346 0.178255 factor(AGE)3 0.3969941 0.0317654 12.498 < 2e 16 factor(AGE)4 0.5879676 0.0310477 18.938 < 2e 16 factor(AGE)5 0.4560387 0.0317142 14.380 < 2e 16 factor(AGE)6 0.1964364 0.0339049 5.794 6.92e 09 factor(AGE)7 0.0479760 0.0380357 1.261 0.207191 factor(AGE)8 0.0280316 0.0389207 0.720 0.471392 factor(GENDER)M 0.0009421 0.0158129 0.060 0.952494 KE_ET:logPop 0.0575988 0.0173310 3.323 0.000890 KE_SM:logPop 0.0762631 0.0153896 4.956 7.24e 07 KE_SU:logPop 0.0836244 0.0142484 5.869 4.41e 09 KE_UG:logPop 0.0801133 0.0122896 6.519 7.15e 11 KE_ET:logDist 0.2292590 0.0124716 18.382 < 2e 16 KE_SM:logDist 0.2433380 0.0133431 18.237 < 2e 16 KE_SU:logDist 0.3116560 0.0254647 12.239 < 2e 16 KE_UG:logDist 0.5303973 0.0113675 46.659 < 2e 16 KE_ET:factor(AGE)2 0.0038053 0.0699597 0.054 0.956623 KE_ET:factor(AGE)3 0.1433140 0.0625309 2.292 0.021916 KE_ET:factor(AGE)4 0.2848937 0.0616237 4.623 3.79e 06 KE_ET:factor(AGE)5 0.2495923 0.0646461 3.861 0.000113 KE_ET:factor(AGE)6 0.0773828 0.0701320 1.103 0.269864 KE_ET:factor (AGE)7 0.0008427 0.0767247 0.011 0.991237 KE_ET:factor(AGE)8 0.0499196 0.0751771 0.664 0.506677 KE_SM:factor(AGE)2 0.0093038 0.0636394 0.146 0.883767 KE_SM:factor(AGE)3 0.1397553 0.0570119 2.451 0.014236 KE_SM:factor(AGE)4 0.3212290 0.0567203 5.663 1.49e 08 KE_SM:factor(AGE)5 0.2887977 0.0581382 4.967 6.81e 07 KE_SM:factor(AGE)6 0.0198199 0.0622346 0.318 0.750129 KE_SM:factor(AGE)7 0.0967190 0.0656521 1.473 0.140702 KE_SM:factor(AGE)8 0.1598639 0.0656984 2.433 0.014965 KE_SU:factor(AGE)2 0.0057587 0.0546070 0.105 0.916013 KE_SU:factor(AGE)3 0.1539909 0.0489669 3.145 0.001663 KE_SU:factor (AGE)4 0.3438848 0.0491193 7.001 2.57e 12 KE_SU:factor(AGE)5 0.2580136 0.0510843 5.051 4.42e 07

PAGE 2

KE_SU:factor(AGE)6 0.1144147 0.0542600 2.109 0.034981 KE_SU:factor(AGE)7 0.0157266 0.0600515 0.262 0.793411 KE_SU:factor(AGE)8 0.0059988 0.0591863 0.101 0.919269 KE_UG:factor(AGE)2 0.1316750 0.0458392 2.873 0.004074 KE_UG:factor(AGE)3 0.1069047 0.0416495 2.567 0.010268 KE_UG:factor(AGE)4 0.5140590 0.0413092 12.444 < 2e 16 KE_UG:factor(AGE)5 0.6951047 0.0421692 16.484 < 2e 16 KE_UG:factor(AGE)6 0.8076488 0.0443712 18.202 < 2e 16 KE_UG:factor(AGE)7 0.9628390 0.0484411 19.876 < 2e 16 KE_UG:factor(AGE)8 0.8576411 0.0491423 17.452 < 2e 16 KE_ET:factor(GENDER)M 0.0442031 0.0321400 1.375 0.169036 KE_SM:factor(GENDER)M 0.0080601 0.0287825 0.280 0.779453 KE_SU:factor(GENDER)M 0.0455020 0.0258460 1.761 0.078329 KE_UG:factor(GENDER)M 0.0218555 0.0213185 1.025 0.305279

PAGE 3

Age stratified origin specific in migrant occurrence plotted against Euclidean distance from shared border between origin countries and destination country, Kenya:

PAGE 8

Gender stratified origin specific in migrant occurrence plotted against Euclidean distance from shared border between origin countries and destination country, Kenya:



PAGE 1

RESEARCHOpenAccessQuantifyingcross-bordermovementsand migrationsforguidingthestrategicplanning ofmalariacontrolandeliminationDeepaKPindolia1,2,3*,AndresJGarcia1,2,ZhuojieHuang4,5,TimothyFik2,DavidLSmith6,7andAndrewJTatem7,8AbstractBackground: Identifyinghumanandmalariaparasitemoveme ntsisimportantforcontrolplanningacrossall transmissionintensities.Importedinfectionscanreint roduceinfectionsintoareaspreviouslyfreeofinfection, maintain ‘ hotspots ’ oftransmissionandimportdrugresistantstra ins,challengingnationalcontrolprogrammes atavarietyoftemporalandspatialscales.Recentana lysesbasedonmobilephoneusagedatahaveprovided valuableinsightsintopopulationandlikelyparasitemov ementswithincountries,butthesedataarerestricted tosub-nationalanalyses,leavingimportantcross-bordermovementsneglected. Methods: Nationalcensusdatawereusedtoanalyseandmod elcross-bordermigrationandmovement,using EastAfricaasanexample. ‘ Hotspots ’ oforigin-specificimmigrantsfromneighbouringcountrieswereidentified forKenya,TanzaniaandUganda.Populationsoforigin-s pecificmigrantswerecomparedtodistancefromorigin countrybordersandpopulationsizeatdestination,an dregressionmodelsweredevelopedtoquantifyand comparedifferencesinmigrationpatterns.Migrationdatawerethencombinedwithexistingspatially-referenced malariadatatocomparetherelativepropensityforcross-bordermalariamovementintheregion. Results: Thespatialpatternsandprocessesforimmigrationweredifferentbetweeneachoriginanddestination countrypair.Hotspotsofimmigration,forexample,wer econcentratedclosetoorigincountrybordersformost immigrantstoTanzania,butforKenya,asimilarpatte rnwasonlyseenforTanzanianandUgandanimmigrants. Regressionmodelfitsalsodifferedbetweenspecific migrantgroups,withsomemigrationpatternsmore dependentonpopulationsizeatdestinationanddistanc etravelledthanothers.Withthesedifferencesbetween immigrationpatternsandprocesses,andheterogeneous transmissionriskinEastAfricaandthesurrounding region,propensitiestoimportmalariainfectionsalsolikelyshowsubstantialvariations. Conclusion: Thiswasafirstattempttoquantifyandmodel cross-bordermovementsrelevanttomalaria transmissionandcontrol.Withnationalcensusavailableworldwide,thisapproachcanbetranslatedtoconstruct across-borderhumanandmalariamovementevidenceba seforothermalariaendemiccountries.Theoutcomes ofthisstudywillfeedintowidereffortstoquantifyandmodelhumanandmalariamovementsinendemic regionstofacilitateimprovedinterventionplanning,resourceallocationandcollaborativepolicydecisions.BackgroundFundingformalariacontrolhassubstantiallyincreased inthepastdecade,reducingmalariaburdensacross transmissionzones[1-3].However,financialresources remainlowerthanthelevelsrequiredtomeetglobal eradicationgoals[4]and,therefore,improvementsinthe quantitativeevidencebaseareimportantforguidingthe strategicallocationofinterventions.Humanpopulation movement(HPM)thatleadstothemovementofinfections,overvaryingspatialandtemporalscales,playsan importantroleinmalariadynamicsacrossthefullrange oftransmissionintensitiesandepidemiologicalphases [5-9].Inmalaria-freerecept ivesettings(post-elimination),infectionimportatio nthreatensreintroduction andresurgence[10],whilstinareasofheterogeneous risk(pre-elimination),highertransmission ‘ hotspots ’ may *Correspondence: dpindolia@gmail.com1EmergingPathogensInstitute,UniversityofFlorida,Gainesville,Florida,USA2DepartmentofGeography,UniversityofFlorida,Gainesville,Florida,USA Fulllistofauthorinformationisavailableattheendofthearticle 2014Pindoliaetal.;licenseeBioMedCentralLtd.ThisisanOpenAccessarticledistributedunderthetermsoftheCreative CommonsAttributionLicense(http://creativecommons.org/licenses/by/2.0),whichpermitsunrestricteduse,distribution,and reproductioninanymedium,providedtheoriginalworkisproperlycredited.TheCreativeCommonsPublicDomain Dedicationwaiver(http://creativecommons.org/publicdomain/zero/1.0/)appliestothedatamadeavailableinthisarticle, unlessotherwisestated.Pindolia etal.MalariaJournal 2014, 13 :169 http://www.malariajournal.com/content/13/1/169

PAGE 2

serveainfectionsources(exportinginfections)[5].HPM mayalsoleadtotheemergenceofdrug-resistantstrains ofmalariathatchallengecontrolprogramsinbothhigh andlowtransmissionareas[11-13].Quantifyingboth withincountryandcross-bordermovementsis,therefore,importantforstrategicinterventionplanningand surveillanceatanationallevel,andencouragingand facilitatingcountrycollaborationsataregionallevel. Thefailureofpreviouseliminationprogrammeswas partlyattributedtoimportedinfectionsfromneighbouring highertransmissionriskcountries[14].Inareascloseto elimination,HPMfromhighertransmissionneighbouring regions,combinedwithlimitedandunsustainablefunding, continuetochallengetheachievementandsustainability ofmalaria-freestatus[15].BasedonWHOrecommendations,aneliminationfeasibilityassessmentconducted inZanzibarillustratedtheimportanceofquantifying HPMforstrategiceliminationplanning[16,17].Countries withhighermalariaprevalenceneighbours,suchasthe DominicanRepublic,SouthAfricaandChinaoftenexhibit higherprevalence ‘ hotspots ’ closetobordersasaresult ofcross-bordermovementscarryinginfections.HPM inandoutofthesehighertransmissionregionsmay leadtoinfectionflowsthatthreatenonwardtransmission andburdenhealthsystems[5,7].Drugresistancehasbeen amajorchallengeamongmigrantgroupsnearborder areasinAsiaandmorerecentlyinAfrica[11].Betweencountrycollaborations,suchastheLumomboMalaria ControlInitiativebetweenborderingSouthAfrica, SwazilandandMozambique[18],andthecollaborative malaria-freeinitiativelaunchedintheArabianPeninsula [19,20],weredevelopedtotacklemalariaataregional scale.SuchprogrammesbenefitfromquantitativeevidenceonHPMtobetterdevisenationalandregional interventionandsurveillancestrategies[21],andrefrain fromrepeatingtheinefficienciesofsingle-countrystrategiesofthepast[14]. Inrecentyears,therehasbeenagrowthintheavailabilityofdataformeasuringHPMacrossspatialand temporalscalesthatareimportantformalariacontrol [7].Theuseofmobilephonecalldatarecordstomodel parasitemovements,bycombiningHPMtrajectorieswith malariametricdataoffersoneofthemostpromising approaches,providingfinescaleestimatesinspace andtime,andcoveringlargepercentagesofnational populations[5,22-24].Analysesofmobilephonedata howeverareconstrainedtowithin-countrymovements duetophonenetworkcompanyrestrictionsanddonot containinformationonindividual-leveldemographics andothermalaria-levelcharacteristics,suchastheuse ofpreventivemeasures.Otherdatatypes,suchastravel historysurveys,whichmaycontainthistypeofdata, arerestrictedtosmallgeographicareasandspecific sub-populations[25].Cross-borderquestionnairesremain expensivetoundertakeandinmanymalariouscountries, bordersareporous,withHPMthroughremoteland bordercrossingpointsand ‘ unofficial ’ borderpoints[26]. Morewidelyusedbutlessspatiallyandtemporallyrefined arecensusandsurveydata,whichcontaindemographic andcross-bordermigrationdata.Migrationdatafrom censuseshaverecentlybeenshowntostronglycorrelate withmovementpatternsacrosstemporalscales[27], highlightingthatsuchdatamaybeusefulforquantifying malaria-relevantHPM.Quantitativecross-borderHPM evidencehasrarelybeenusedforunderstandinghuman andmalariamovementsandprovidingguidanceonextent andnatureofbetween-countrycooperationforcontrol andelimination. Here,toexploreandillustratethepotentialofcensusderivedmigrationdatainquantifyingcross-borderhuman andmalariaconnectivitiesandmovements,analysesof datafromEastAfricawereundertaken.Nationalcensus dataforKenya,TanzaniaandUgandawereanalysedto highlightpatternsincross-bordermigrationbymapping significantorigin-specificimmigrant ‘ hotspots ’ andsubnationalareasthatshouldconsidercollaboratingon controlandeliminationstr ategieswithneighbouring countries.Thedatawerefittedtoaregressionmodelto helpexplainandcompareobservedpatternsanddescribe processesofimmigration.Existingspatialmalariaprevalencedataandmathematicalmodelswerethencombined withHPMdatatoillustratedifferencesinmalariamovementpropensitiesintoKenya,TanzaniaandUgandafrom theirneighbouringregions.MethodsCensusdataCross-bordercensusmigrationdatawereobtainedfor Kenya,TanzaniaandUganda(Table1).TheKenya1999 censuswasobtainedfromtheKenyaNationalStatistics Bureau(KNBS).Individual-levelrecordsforallindividuals enumeratedwereavailableforselectedvariables,including currentsub-location(administrativelevel5boundary)of residence,birthandpreviousresidencelocation(district/ administrativelevel2boundaryforinternalmigrants andborderingcountrynameforcross-bordermigrants), anddemographicdataonageandgender.ForTanzania, aggregateddataonthenumberofresidentsineach sub-locationandtheirnationalitywereobtainedfrom the2002census.Demographicstratificationswerenot availableforTanzania.ForUganda,2002censusmicrodata,asystematicselectedsubsetofcountrywidenational housingandpopulationcensusdataobtainedfromIntegratedPublicUseMicrodataSe ries,International(IPUMS) [28]wereobtainedonline.Thesamplecontainedrecords forallcensusquestionsfora10%sampleofallindividuals enumerated.TomakemigrationdefinitioncomparablePindolia etal.MalariaJournal 2014, 13 :169 Page2of11 http://www.malariajournal.com/content/13/1/169

PAGE 3

betweencountries,migrantsweredefinedbasedonplace ofbirthandcurrentresidencelocation.Therespective countrycensuseswerealsousedtoextracttotalpopulationsizeperadministrativeboundary.MalariadataCountry-levelmalariatransmissionmapsforKenya, TanzaniaandUgandaandtheirrespectiveneighbouring countrieswereobtainedfromthe2010global Plasmodium falciparum endemicitymaps(with P.falciparum parasite rate,standardizedfor2-10yearolds( Pf PR2-10),for11km pixels)fromtheMalariaAtlasProject(MAP)[29,30]. Toobtainpopulation-weighted Pf PR2-10,anAfrica-wide populationdistributiongrid(withpopulationdensity foreach11kmpixel)wasobtainedfromtheWorldPop Project[31]andcountryspecificgridsextracted.The endemicitymapsandpopulationgridswerealignedby overlayingeachcountryendemicitymapoverthepopulationdistributiongrid.Foreachpixelonthemap, population-weighted Pf PR2-10wascalculated.SpatialanalysisOrigin-specificdataonnumbersofmigrants(basedon birthcountryandcurrentresidencelocationcomparisons) wereobtainedforeachadministrativeunitinthethree destinationcountries,Kenya,TanzaniaandUganda(at differentadministrativeresolutions,asdescribedabove). TheGetis-OrdGstatisticwasusedtoestimatelocal ‘ hotspots ’ oforigin-specificimmigrants(basedonspatial characteristicsastemporaldescriptionswerenotavailable forallcountries).Statisticallysignificanthotspotswere determinedbasedonaGiZScore>1.96(highZscores areameasureofstandarddeviationassociatedwithlow pvalues.AGiZScoreof1.96correspondstopvalue<0.05 anda95%confidenceintervalusingthestandardNormal distributionassumptionoftheoreticalspatialrandomness) [32].Significanthotspotsweremappedtoillustratesingleorigin,aswellasmultiple-origin,over-lappinghotspots. Administrativeunitswereclassifiedintosingle-origin hotspotsifalocationwasahotspotformigrantsfrom onlyoneorigin,andmultiple-originhotspotsifalocation wasahotspotformigrantsfrommorethanoneorigin.ModellingmigrationModellingmigration(basedonbirthcountryandcurrent residencelocationcomparisons)flowscanprovidemigrationinformationforlocationsandtimeperiodswheredata arenotavailable[27].Traditionally,humanmovement modelshavebeenbasedontheconceptofgravity,that assumesapositiverelationshipbetweenmigrantflowand theproductofpopulationsizesasoriginanddestination, andanegativerelationshipbetweenmigrantflowand distancetravelled[33,34].Toexploreapossiblegravitylikepatternincross-bordermigrationinEastAfrica, origin-specificimmigrantoccurrencewasplottedagainst Euclideandistancefromsharedbordersbetweenorigin anddestinationcountries.Withoriginsdefinedatabroad resolution(countrylevel,administrativeunit0),atraditionalgravitymodelcouldnotbefitted.Instead,asimple positiverelationshipbetweenmigrantflowbetweenorigin i ,anddestination j ,andtotalpopulationsizeatthedestinationwasassumed,withanegativerelationshipbetween migrantflowandEuclideandistancebetweenorigin countryborderanddestinationlocationforeachorigindestinationpair(Equation1). Migrantflowi ; je Totalpopulationsizej i ; jEuclideandistancei ; j i ; j 1 Asetof3linearregressionmodels(oneforeach destinationcountry)weredevelopedtoquantifythe variabilityinmigrantflowasdeterminedbydestination populationsizeanddistancetravelled,basedonEquation 2.(RefertoAdditionalfile1forexpandeddestinationspecificequations).Toachievealinearrelationship, allvariableswerelog-tran sformed.Toallowcomparisonsoftheeffectofdestinationpopulationsizeand distancetravelledonmigrantflowsbetweenorigindestinationpairs(differentgroupsofmigrants),dummy variablesandinteractiontermswereincorporatedinto eachmodel. Table1DescriptionofthetypeofcensusdataavailableforeachcountryKenyaTanzaniaUganda Datatype IndividuallevelAggregateddataIndividuallevel Migrationdata LifetimemigrationLifetimemigrationLifetimemigration Additionalmigrationdata Recentmigration-Recentmigration Spatialresolutionatdestination Administrativelevel5Administrativelevel5Administrativelevel2 Spatialresolutionatorigin Administrativelevel0Administrativelevel0Administrativelevel0 Yearofdatacollection 199920022002 Othervariables Age,Gender-Pindolia etal.MalariaJournal 2014, 13 :169 Page3of11 http://www.malariajournal.com/content/13/1/169

PAGE 4

i = 1 … n ;n=numberofneighbouringcountriesforeach destination j = 1 2 3 ;1=Kenya,2=Tanzania,3=Uganda Mi j:Totalnumberofmigrantsfromorigin,i,todestination,j. Pj:Totalpopulationsizeofdestinationlocationwithin eachdestinationcountry. Di j:Shortestdistancebetweenorigincountryborder anddestinationlocation. R1 … r:Dummyvariablesrepresentingneighbouring countriesperdestination(r=Numberofcountries-1). 0, … r; 0, … r; 0, … r:Exponentsestimatedfrom thedata. i j:Errorterm. TheoverallfitoftheregressionmodelswasquantifiedusinganadjustedR-squaredvalue.Theeffectsof populationsizeanddistanceonmigrantflowforeach origin-specificvariablewereestimatedbyaddingthe origin-specificcoefficientineachmodeltothereference variable(representinganarbitrarilychosenorigincountry) ineachmodel.Significanceofregressioncoefficientswas basedonp-valuesfromt-teststodeterminedifferencesin theeffectofpopulationsizeanddistanceonmigrantflow betweenthereferenceoriginandallotheroriginsforeach destination. ForKenya,thisanalysiswasextendedtoincludeage andgender,assuchinformationwereavailablehere. Ageandgenderstratifiedor igin-specificimmigrant occurrencewasplottedagainstEuclideandistancefrom sharedbordersbetweenorigincountriesandthedestinationcountry,Kenya.Theregressionmodelwasextended toincludeageandgenderasadditionalexplanatoryvariables,withcorrespondingdummyvariablescreatedfor agegroupandgendercategories(RefertoAdditionalfile2 forextendedequations).MalariaconnectivitiesMalariaimportationpropensityquantifieslikelyimported infectionroutes,asmigrantsarelikelytomaintainconnectionswiththeirhomelocationsandmayengagein shorttermtravelthat,duetolostimmunity,maylead toimportedinfectionsatdestination.Previousanalyses haveshownthestrongrelationshipsbetweenthestrengths oflonger-termspatialmigrationconnectivitiesand shortertermmovements[27].Importationpropensity estimateswerebasedontwoendemicitymetricsi)mean population-weighted Pf PRobtainedatadministrative0 level(fortheentireorigincountry)andii)mean Pf PR withina100kmbufferfromdestinationcountryborder foreachneighbouringorigincountry. Pf PRprovidesa usefulmeasureforendemicityatlarge-scales,andasthe migrationdatadoesnotincludespecificoriginlocations, aggregatedoriginendemicityestimatesatanationaland sub-nationalwereused.Thetwotypesofendemicityestimatesforeachneighbouringorigincountrywerethen multipliedbythenumberoforigin-specificmigrantsin eachdestinationcountrytoobtainmalariaimportation propensity,whichwasrelativelycomparedwithinand betweenthedestinationcountries.ResultsMigrationpatternsPatternsofsignificantorigin-specificimmigranthotspots differedbothbetweenandwithindestinationcountries (Figure1).InTanzania(Figure1A),thehighestnumber oforigin-specifichotspotswereseenclosetotheborders withtheorigincountriesoftherespectivesetsofmigrants, exceptformigrantsfromMalawi,forwhichthemajorityof hotspotswerenearTanzania ’ scapitalandlargesturbancity, DaresSalaam.Somelocationsweresignificanthotspots formigrantsfromtwodifferentcountries,forexample,in thenorthwestregion,variousTanzaniansub-locations werehotspotsforbothRwandanandBurundianimmigrants.NearDaresSalaam,hotspotsoverlappedfor KenyanandMalawiimmigrants.Immigrantpatternsin Kenya(Figure1B)differedfromthoseinTanzania.Most distinctively,hotspotoverlapwasmoreprevalent,with somelocationsbeingquantifiedashotpotsforallfive neighbouringcountries.Additionally,Ugandanimmigrant hotspotswerewidespreadacrossKenya,whilstEthiopian andSomaliimmigranthotspotsweremainlyinthecentral regionofthecountry.Tanzanianhotspotswereseenin themostpopulatedregionsacrosstheborderandnear largeurbancentressuchasNairobiandMombasa, whilstSudanesehotspotsweremostprominentnear thesharedborder,inthecentralregionsaroundNairobi andnearMombasa(Additionalfile3).Evenwiththelower resolutionimmigrantdataavailableforUganda(administrative2levelhotspots,comparedtoadministrative5 levelhotspotsinKenyaandTanzania),overlapbetween origin-specificmigranthotspotswaslessfrequentthan inKenya(Figure1C).Nevertheless,5districtsinsouthern Ugandawereoverlappinghotspotsformigrantsfromboth TanzaniaandRwanda,andonenorth-westerndistrictwas ahotspotformigrantsfromSudanandtheDemocratic RepublicofCongo. log Mi ; j 0 1R1 … rRr 0log Pj 1R1logPj … rRrlogPj 0logDi ; j 1R1logDi ; j … rRrlogDi ; j i ; j 2 Pindolia etal.MalariaJournal 2014, 13 :169 Page4of11 http://www.malariajournal.com/content/13/1/169

PAGE 5

MigrationprocessesEachorigin-destinationpairshowedadifferentrelationshipbetweenorigin-specificimmigrantabundanceat thedestinationandthedistancebetweentheorigin countryborderandthedestinationlocation(Figure2).In Tanzania(Figure2A),thelargestorigin-specificimmigrant populationswerefoundtobeclosetotheirrespective origincountryborders,illustratinganinverserelationship betweenimmigrantpopulationanddistancebetween originanddestination.Clustersofimmigrantswere alsoseeninareasaroundandinthecapitalcity,with theinverserelationshipbetweendistanceandmigrant sizebecominglessrelevant.InKenya,theinverserelationshipbetweendistanceandmigrantpopulationabundance wasmainlyseenforimmigrantsfromTanzania(Figure2B). ClustersclosetothecapitalseenforEthiopian,Somaliand Sudanesemigrantshoweverwerelessdistinctthanfor Tanzania,whilstUgandansshowedamoreevendispersionacrossKenya.Patternswerelessobviousfor Ugandaduetothelowresolutionofimmigrantdata (Figure2C). Overallfitsofdestination-specificregressionmodels differed,withadjustedR-squarevaluesforKenyabeing 27.65%,23.49%forTanzaniaand18.05%forUganda. Acrossalloriginsanddestinations,populationsizewas positivelyassociatedwithmigrationwhilstdistanceshowed aninverserelationship,exceptforUgandansinKenya (Table2).ForTanzaniaandU ganda,distancewasamore importantdeterminantformig rationcomparedtopopulationsizeatdestination,howeverinKenya,populationsize atthedestinationlocationwasasignificantdeterminant forallmigrantgroups.ForTanzania,significanteffectsof distancecorrelatedwithmostmigrantpopulationsbeing concentratedalongborders,asillustratedinFigures1and 2.Similarly,populationsizesatdestinationlocationsasa significantdeterminantofmigrationinKenyacorrelated withimmigrationpatternsillustratedinFigures1and 2.Withindestination-specificregressionmodels,the importanceofpopulationsizeanddistancedescribing thevariationinorigin-specificimmigrantsalsoshowed heterogeneitythroughdifferencesineffectssizes.For example,inKenya,populationsizehadthelargest effectforSomaliandSudanesemigrants,comparedto migrantsfromotherorigins.AsUgandanimmigrants showedamoredisperseddistribution(Figure1),the effectofdestinationpopulationsizewasthesmallest comparedtomigrantsfromot herorigins.Byincluding ageandgenderasadditionalexplanatoryvariables,the modelfitforKenyaimprovedfrom27.65%to33.14%, highlightingtheimporta nceofaccountingfordemographicdifferences[9].Significantdifferencesbetween agegroupswereidentifiedfororigincountries,however differencesingenderremainedinsignificantlydifferent throughout(Additionalfile4).MalariaconnectivitiesBasedonthevariationsseeninimmigrantpatternsand heterogeneityinmalariatransmissionriskacrossthe EastAfricanandneighbouringregions,propensitiesto importinfectionslikelydifferssubstantiallybetween Figure1 Origin specificimmigranthotspotsinthethreedestinationcountries:A)TanzaniaB)KenyaC)Uganda. Statisticalsignificance basedonGiZScore>1.96,usingtheGetis-OrdGstatistic.Hotspotscolouredbasedonorigincountryofmigrants.Countrycodes:TZ:Tanzania,KE: Kenya,UG:Uganda,RW:Rwanda,BU:Burundi,DRC:DemocraticRepublicofCongo,ZA:Zambia,MW:Malawi,MZ:Mozambique,SM:Somalia,ET: Ethiopia,SU:Sudan. Pindolia etal.MalariaJournal 2014, 13 :169 Page5of11 http://www.malariajournal.com/content/13/1/169

PAGE 6

destinationcountries.AsseeninFigure3,underthe assumptionsusedhere,forallthreeEastAfricancountries, theurbanareassuchasNairobi,Mombasa,DaresSalaam andKampalahadhigherestimatedpropensitiestoimport infections(likelysinksofinfection).Thedistributionof importationpotentialbyUgandanimmigrantswaswidespreadinKenya,however,itwasfocusedtonorthwestern borderregionsinTanzania.Overall,basedonthelargest immigrantpopulationsizesandhigherendemicityin Uganda,comparedtoKenya ’ sotherneighbouringcountries, propensitytoimportwassigni ficantlyhigherbyUgandans (Figure4).InTanzania,thelargestpropensitieswere estimatedtobefromimmigrantsfromBurundiandthe DemocraticRepublicofCongo,whilstinUganda,estimates werelargestforSudaneseandCongoleseimmigrants.DiscussionQuantifyingHPMcanprovideusefulinformationfor evidence-basedmalariacontrolandeliminationplanning. Asshownhere,thepatternsandprocessesofmovements candiffersignificantlyoverspaceaswellasbetween countriesanddemographicgroups[9],whichleadsto heterogeneitiesininfectionimportationpropensities, underliningtheimportanceof accountingforlocalcontext. Quantifyingthesedifferencescanaidtheidentification ofpopulationgroupsmostlikelytoimportinfections, neighbouringcountriesandregionsthataremostlikely toexportinfections( “ sources ” )andwithincountrylocationsthatareatelevatedriskofimportationononward transmission( “ sinks ” ).Identifyingkeypopulationgroups, sourcesandsinksallowsnationalcontrolandsurveillance resourcestobestrategicallytailoredandtargeted[35,36], andhighlightsareasandpopulationswherefurtherdata collectionstudies[37]anddetailedassessmentscanbe made.Asdrugresistancecontinuestocreatechallenges formalariacontrol,particularlyinborderregions,data oncross-bordermovementscaninformcontainment strategies[35].Moreover,quantifyingthesecross-border linkagesandconnectivitiescanprovideindicatorson whenandwhereneighbouringcountriesmightcollaborate Figure2 Distance migrantfunctions illustratingtherelationshipbetweenthenumberofmigrantsineachdestinationcountry:A) TanzaniaB)KenyaC)Uganda,comparedtotheEuclideandistancefromorigincountryborder. Y-axisisrepresentedunderalogarithmic scaletoillustratevariation.Countrycodes:TZ:Tanzania,KE:Kenya,UG:Uganda,RW:Rwanda,BU:Burundi,DRC:DemocraticRepublicofCongo, ZA:Zambia,MW:Malawi,MZ:Mozambique,SM:Somalia,ET:Ethiopia,SU:Sudan. Pindolia etal.MalariaJournal 2014, 13 :169 Page6of11 http://www.malariajournal.com/content/13/1/169

PAGE 7

toplaninterventionsandshareinformationatboth nationalandregionallevels(Figure1).Forexample,based ontheassumptionthatsomemigrantgroupsmayhave higherfluxesoftraveltohomecountries,prophylaxismay bemadeavailableinregionswherethesetypesofmigrant populationsaremostabundant.InEastAfrica,sucha strategymaybeadequateforDRCmigrantslivingin TanzaniaasDRCmigrantsareconcentratedinthe westernregionsclosetoborders,howeverwouldbe difficulttoadministerforUgandanslivinginKenya,as populationsaremorespreadacrossthecountry.In generalhowever,theoperationalchallengesforprophylaxisprovisionatthisscaleneedconsideration. Migrationpatternsareheterogeneous,bothwithinand betweendestinationcountries.Migrantflowstrengths havebeenshowntocorrelatewithshorttermmovement patterns(mayresultduetomigrantsmaintainingties andvisitingfamilyatoriginlocations)[27],whichareof importanceintermsofimportedinfections[x],depending onendemicitylevelsatorigins(Figures3and4).Differencesintherural-urbandistributionofmigrantpopulations maythereforeimplythatsomemigrantgroupsmaybe morelikelytoimportinfectionsintourbanareas comparedtoruralareas,aresultthathaspreviously beenshown[9].Duetodifferencesinreceptivity,the likelihoodofonwardtransmissiondiffersbetweenrural andurbansettings[38].Withheterogeneoustransmissionwithincountrybordersandlikelysignificantlylargeramountsofinternalmigration(Additionalfile4),it canbeimportanttocollectivelyassessbothinternal andcross-borderimportation.Nevertheless,theabundanceofcross-borderimmigrantpopulationsprovide usefulindicationsonwherecountriescancollaborateto developcontext-specificandtargetedinterventions.For example,basedonthemigrationhotspotsidentified here,forKenyanmalariacontrolstrategies,itmaybe beneficialtohighlightcollaborationwithneighbouring countriesasanationalpolicy,aspreviouslydonein SouthernAfrica[18],however,inTanzania,collaborativeworkmaybestfocusedinareasclosetoborders. Thedataandmethodologyusedintroducessomelimitationsintothisstudy.Issueswithcensusdatainclude thedifficultyofcapturingup-to-datemigranttrends, migrantswhoarefleeingfromconflictorpoliticalinstability[39]andotherhigh-riskgroupsforinfectionimportation,suchashighlymobilepopulationsandillegal Table2Regressionanalysisoutputsforthreedestination-specificmodels,whichmodelmigrationasthedependent variableanddestinationpopulationsizeanddistancetravelledastheindependentvariablesPopulationsizeDistance Model1.destination:KEOriginsEffectsetp-valueEffectsetp-value TZ^0.320.0133.23<0.05*-0.370.01-51.55<0.05* ET0.280.02-2.37<0.05*-0.160.0116.36<0.05* SM0.400.025.43<0.05*-0.150.0115.95<0.05* SU0.410.016.43<0.05*-0.090.0310.89<0.05* UG0.240.01-5.74<0.05*0.120.0142.07<0.05* Model2.Destination:TZ KE^0.530.0146.55<0.05*-0.440.01-55.07<0.05* MZ0.370.04-4.10<0.05*-0.320.025.23<0.05* MW0.370.06-2.480.01-0.150.039.59<0.05* ZA0.230.06-4.76<0.05*-0.340.032.94<0.05* DRC0.610.061.480.14-0.750.05-6.38<0.05* BU0.470.04-1.560.12-0.520.02-3.35<0.05* RW0.450.06-1.360.17-0.800.04-9.80<0.05* UG0.400.06-2.13<0.05*-0.590.04-3.78<0.05* Model3.Destination:UG TZ^0.400.0141.44<0.05*-0.350.01-58.89<0.05* KE0.990.331.790.07-1.150.22-3.62<0.05* SU0.430.540.050.96-1.690.31-4.38<0.05* DRC1.130.342.13<0.05*-1.190.20-4.18<0.05* RW0.770.351.040.30-1.030.25-2.70<0.05*^referencecategoryforeachdestination-specificregressionmodel. *significant,basedona5%significancelevel. Countrycodes:TZ:Tanzania,KE:Kenya,UG:Uganda,RW:Rwanda,BU:Burundi,DRC:DemocraticRepublicofCongo,ZA:Zambia,MW:Malawi,MZ:Mozambiqu e, SM:Somalia,ET:Ethiopia,SU:Sudan. Effectsofeachindependentvariable,stratifiedbyorigincountry,wereestimatedusingtheregressioncoefficientsoftheinteractionterms(int eractionbetween explanatoryvariablesanddummyvariablescreatedtorepresentrespectiveorigincountries).Pindolia etal.MalariaJournal 2014, 13 :169 Page7of11 http://www.malariajournal.com/content/13/1/169

PAGE 8

immigrants,demonstratingtheinadequacyofcensusdata capturingthefullrangeofHPM.Moreover,censusdata doesnotrecorddetailedmalaria-relevantcharacteristics, suchasbednetuseandaccesstohealthcare,whichwould allowmoredetailedstratificationofhigh-riskgroups,but theintegrationofsuchdatasetswithgeoreferencedhouseholdsurveyinformationofferspossibilitiestoovercome this[7,9].Migrationratescanbeusedasanindication forcomparingtherelativelikelihoodsofshorterterm travel[27],however,frequenciesoftravelto/fromhome locationsandelsewhereareunknownandthereforedifficultiesremaininestimatingabsolutenumbersofimported infections.Furthermore,usingmigrationasanindicator forfutureshortertermHPMmaybelessapplicablefor certaingroups,suchasthosefleeingfromconflict,as theyarelesslikelytoreturnhome.Censusesgenerally recordinternationalmigrantor iginsatacoarserresolution (countryname)comparedtowithincountrylocations (smalleradministrativeboundaries),makingitdifficult toestimaterelativeparasitecarriageratesthroughmalaria prevalencemaps.Limitationsalsoariseinthestructureof theregressionmodelspresentedhere,whichonlyinclude effectsofdistanceandpopulationsizeatdestination onmigrantflows.Otherpushandpullfactors,suchas demographics,occupationandsocioeconomicfactors [40],arelikelytobeimportanttoinclude,asdemonstrated hereforKenya(Additionalfile2).Finally,somelimitations existintheuseof Pf PRdataasamalariametricinthis context.Mean Pf PRendemicitymapsprovidehighresolutionspatially-referencedmetricsatlargescales,but Pf PR isapoormeasureforlowtransmissionareas(requiring largesurveysamplestodetectcases)[41].Additionally, thecontemporarymapdatausedheredonotprovide measuresofreceptivityandthereforearelimitedinterms ofassessingtheeffectsandimplicationsonlocaltransmissionfromimportedcasesinanarea[16]. Wehavepresentedhereaframeworkbuiltoncensusderivedmigrationdataforprovidingbroadassessments ofcross-borderhumanandmalariamovements.While theexampleanalyseswerefocusedonimportationto Figure3 Comparingspatialpatternsoforigin specificpropensitiesofmalariaimportationintoKenya TanzaniaandUgandafrom neighbouringcountries basedontwotypesofmalariaendemicityestimateassumptionsatorigins(i)population weightedmean Pf PRand(ii)mean Pf PRwithin100kmfromdestinationcountryborder. Propensityofimportation=numberoforigin-specificmigrants*origin Pf PRestimate.Scalerepresentsonestandarddeviationfromtheestimatedvalue,dividedintofourcategories.Countrycodes:TZ:Tanzania,KE: Kenya,UG:Uganda,RW:Rwanda,BU:Burundi,DRC:DemocraticRepublicofCongo,ZA:Zambia,MW:Malawi,MZ:Mozambique,SM:Somalia, ET:Ethiopia,SU:Sudan. Pindolia etal.MalariaJournal 2014, 13 :169 Page8of11 http://www.malariajournal.com/content/13/1/169

PAGE 9

Kenya,TanzaniaandUganda,withcensusdatawidely available[42]andexistingglobalmalariaendemicitydata [29],thesemethodscanbeexpandedtocontinental scales,throughtheassemblyofcensus,microdata,mobile phonecallrecordsandhouseholdsurveysthatrecord cross-bordermigrationandHPM[7].Ifmovementdata canbestratifiedbyagegroupsandifageatwhichmovementsoccurcanbeobtained,mathematicalmodelscan beusedtoestimateage-specific Pf PRestimatesandrefine estimatesofpropensitiesofgroupsandroutesformalaria importation[43-46].Theanalysespresentedhererepresentastartingpointformobilityassessments,andideally shouldbesupplementedwithcross-bordersurveys[47], andothersurveyswithquestionnairedesignsthatinclude adequatetravelhistoryquestions,targetingspecificmobile populationsandhigh-risklocations.Censusmigration datacanalsobeintegratedwithHPMestimatesfrom mobilephoneusagedataandmalariasurveillancedatato refineimportationestimates[5,22,23],thoughsuchphone dataareoftendifficulttoobtainandexpensivetoprocess, whichrepresentsaconstraintformanypoorly-resourced malariousregions.Throughtheadditionofmigrantcharacteristicdescriptions,fore xampleoccupationalgroups andimprovedspatialpopulationdescriptions,more complexspatialanalysesandinteractionmodelsmaybe utilized[48,49].Novelanalysisandmodellingmethods couldalsobedevelopedtocombinemigrationdatawith spatially-referenceddrugresistancedata[50]tounderstand migrationasadeterminantofdrugresistanceemergence [12].Finally,withhumanmovementsplayinganimportant roleinthetransmissionofo therdiseasesandarangeof healthconcerns,theframeworkputforwardheremayalso beofvalueinunderstandingepidemiologicaldynamicsand designinginterventionstrategiesbeyondmalaria.ConclusionWithnationalandinternation alfundingunderthreat,novel toolsandtechniquesthatimprovetheevidence-base fordesigningmoreefficientinterventionandsurveillance strategiesareimportant.Here,aframeworkforutilizing existingHPMdatafromcensuseshasbeendeveloped, andcombinedwithreadilyavailablemalariaendemicity mapstoillustratehowexistingretrospectivelygathered datacanbeusedforquantifyingcross-bordermovements relevantformalariainterventionandsurveillancestrategies. Significantvariationsbetween countries,withincountries andbetweenmigrantgroupswerefound,highlightingthe importanceoflocalcontextinmobilityassessmentsand thevalueofsuchdata.Identifyingkeyregionsandmigrants groupsenablessurveillancea ndinterventionstrategiesto bebuiltaroundavailableevidence,andprovidesuseful guidanceforcountriesembarkingoncollaborativeefforts.AdditionalfilesAdditionalfile1: Equationsformodel2splitbydestinationcountry. Additionalfile2: Model2forKenya,extendingmodeltoinclude ageandgender. Additionalfile3: Hotspotsindicatingpossiblebetweencountry collaborations. Additionalfile4: Methodsusedtogeneratenetworksofinternal andcross-bordermigrantsweresimilartomethodsdevelopedand Figure4 Relativemagnitudesoforigin-specificmalariaimportationpropensityintoeachdestinationcountry(Kenya,Tanzaniaand Uganda),basedontwotypesofendemicityestimateassumptions((i)population weightedmean Pf PRand(ii)mean Pf PRwithin100km fromdestinationcountryborder). Theyaxisshowsorigin-specificmalariaimportationpropensityasapercentageofthetotalmalariaimportation propensityintherespectivedestinationcountry.Countrycodes:TZ:Tanzania,KE:Kenya,UG:Uganda,RW:Rwanda,BU:Burundi,DRC:Democratic RepublicofCongo,ZA:Zambia,MW:Malawi,MZ:Mozambique,SM:Somalia,ET:Ethiopia,SU:Sudan. Pindolia etal.MalariaJournal 2014, 13 :169 Page9of11 http://www.malariajournal.com/content/13/1/169

PAGE 10

appliedinPindoliaetal[1]. Withnodataoncross-bordermigrant origins,meanin-degreeandmeanin-graphstrengthwereusedinstead ofmeandegreeandmeangraphstrength,whichincorporateHPMin bothdirections. Competinginterests Theauthorsdeclarethattheyhavenocompetinginterests. Authors ’ contributions DKPdidtheliteraturesearch,identifieddatasets,carriedouttheanalysisand wrotethefirstdraftofthemanuscript.AJG,ZH,TFcontributedtotheanalysis ofthemanuscript.DLScontributedtotheanalysisandreviewofthe manuscript.AJTcontributedtothewriting,analysisandreviewofthe manuscript.Allauthorsreadandapprovedthefinalversionofthemanuscript. Acknowledgements TheauthorsacknowledgetheSpatialEpidemiologyUnitattheDepartmentof PublicHealthResearch,KEMRI-WellcomeTrustinKenya,fordataacquisition supportandthankVictorAlegana,AbdisalanNoorandRobertSnowfortheir contributionsduringthedatacompilationandconceptualizationstagesofthis manuscript.AJT&DLSacknowledgefundingsupportfromtheEmerging PathogensInstitute,UniversityofFlorida,theRAPIDDprogramoftheScience andTechnologyDirectorate,DepartmentofHomelandSecurity,andthe FogartyInternationalCenter,NationalInstitutesofHealth,andarealso supportedbygrantsfromNIH/NIAID(U19AI089674)andtheBillandMelinda GatesFoundation(#49446and#1032350).DLSacknowledgesfundingsupport fromBloombergFamilyFoundation.Thefundershadnoroleinstudydesign, datacollectionandanalysis,decisiontopublish,orpreparationofthe manuscript.ThispaperformspartoftheoutputoftheWorldPoppopulation mappingproject(www.worldpop.org.uk),Flowminder(www.flowminder.org) andthehumanmobilitymappingproject(www.thummp.org). Authordetails1EmergingPathogensInstitute,UniversityofFlorida,Gainesville,Florida,USA.2DepartmentofGeography,UniversityofFlorida,Gainesville,Florida,USA.3ClintonHealthAccessInitiative,Boston,MA,USA.4CenterforInfectious DiseaseDynamics,PennsylvaniaStateUniversity,UniversityPark, Pennsylvania,USA.5DepartmentofBiology,PennsylvaniaStateUniversity, UniversityPark,Pennsylvania,USA.6DepartmentofEpidemiology,Johns HopkinsBloombergSchoolofPublicHealth,Baltimore,USA.7Fogarty InternationalCentre,NationalInstitutesofHealth,Bethesda,MD20892,USA.8DepartmentofGeographyandEnvironment,UniversityofSouthampton, Southampton,UK. Received:23October2013Accepted:28March2014 Published:3May2014 References1.KatzI,KomatsuR,Low-BeerD,AtunR: Scalinguptowardsinternational targetsforAIDS,tuberculosis,andmalaria:contributionofglobal fund-supportedprogramsin2011-2015. PLoSOne 2011, 6: e17166. 2.TannerM,deSavignyD: Malariaeradicationbackonthetable. BullWorld HealthOrgan 2008, 86: 82. 3.MurrayCJL,RosenfeldLC,LimSS,AndrewsKG,ForemanKJ,HaringD, FullmanN,NaghaviM,LozanoR,LopezAD: Globalmalariamortality between1980and2010:asystematicanalysis. Lancet 2012, 379: 413 – 431. 4.SnowRW,GuerraCA,MutheuJJ,HaySI: Internationalfundingformalaria controlinrelationtopopulationsatriskofstable Plasmodiumfalciparum transmission. PLoSMed 2008, 5: e142. 5.WesolowskiA,EagleN,TatemAJ,SmithDL,NoorAM,SnowRW,BuckeeCO: Quantifyingtheimpactofhumanmobilityonmalaria. Science 2012, 338: 267 – 270. 6.MartensP,HallL: Malariaonthemove:humanpopulationmovement andmalariatransmission. EmergInfectDis 2000, 6: 103 – 109. 7.PindoliaDK,GarciaAJ,WesolowskiA,SmithDL,BuckeeCO,NoorAM,SnowRW, TatemAJ: Humanmovementdataformalariacontrolandelimination strategicplanning. MalarJ 2012, 11: 205. 8.StoddardST,MorrisonAC,Vazquez-ProkopecGM,PazSoldanV,KochelTJ, KitronU,ElderJP,ScottTW: Theroleofhumanmovementinthe transmissionofvector-bornepathogens. PLoSNeglTropDis 2009, 3: e481. 9.PindoliaDKGA,HuangZ,SmithDL,AleganaVA,NoorAM,SnowRW,TatemAJ: Thedemographicsofhumanandmalariamovementandmigration patternsinEastAfrica. MalarJ 2013, 12: 397. 10.CohenJM,SmithDL,CotterC,WardA,YameyG,SabotOJ,MoonenB: Malariaresurgence:asystematicreviewandassessmentofitscauses. MalarJ 2012, 11: 122. 11.RoperC,PearceR,NairS,SharpB,NostenF,AndersonT: Intercontinental spreadofpyrimethamine-resistantmalaria. Science 2004, 305: 1124. 12.LynchC,RoperC: Thetransitphaseofmigration:circulationofmalaria anditsmultidrug-resistantformsinAfrica. PLoSMed 2011, 8: e1001040. 13.BhumiratanaA,IntarapukA,Sorosjinda-NunthawarasilpP,ManeekanP, KoyadunS: Bordermalariaassociatedwithmultidrugresistanceon Thailand-myanmarandThailand-cambodiaborders:transmissiondynamic, vulnerability,andsurveillance. BiomedResInt 2013, 2013: 363417.14.NajeraJA,Gonzalez-SilvaM,AlonsoPL: Somelessonsforthefuturefrom theGlobalMalariaEradicationProgramme(1955-1969). PLoSMed 2011, 8: e1000412. 15.SmithDL,CohenJM,MoonenB,TatemAJ,SabotOJ,AliA,MugheirySM: Infectiousdisease.SolvingtheSisypheanproblemofmalariainZanzibar. Science 2011, 332: 1384 – 1385. 16.WorldHealthOrganization: Malariaelimination:afieldmanualforlowand moderateendemiccountries. Geneva:WorldHealthOrganization;2007. 17.MalariaEliminationGroup: MalariaEliminationinZanzibar:AFeasibility Assessment ;2009. 18.SharpBL,KleinschmidtI,StreatE,MaharajR,BarnesKI,DurrheimDN,RidlFC, MorrisN,SeocharanI,KuneneS,LAGrangeJJ,MthembuJD,MaartensF, MartinCL,BarretoA: Sevenyearsofregionalmalariacontrolcollaboration – Mozambique,SouthAfrica,andSwaziland. AmJTropMedHyg 2007, 76: 42 – 47. 19.MeleigyM: ArabianPeninsulastateslaunchplantoeradicatemalaria. BMJ 2007, 334: 117. 20.SnowRW,AmratiaP,ZamaniG,MundiaCW,NoorAM,MemishZA, AlZahraniMH,AlJasariA,FikriM,AttaH: Themalariatransitiononthe ArabianPeninsula:progresstowardamalaria-freeregionbetween 1960-2010. AdvParasitol 2013, 82: 205 – 251. 21.KhamsiriwatcharaA,SudathipP,SawangS,VijakadgeS,PotithavorananT, SangvicheanA,SatimaiW,DelacolletteC,SinghasivanonP,LawpoolsriS, KaewkungwalJ: ArtemisininresistancecontainmentprojectinThailand. (I):Implementationofelectronic-basedmalariainformationsystemfor earlycasedetectionandindividualcasemanagementinprovincesalong theThai-Cambodianborder. MalarJ 2012, 11: 247. 22.TatemAJ,QiuY,SmithDL,SabotO,AliAS,MoonenB: Theuseofmobile phonedatafortheestimationofthetravelpatternsandimported Plasmodiumfalciparum ratesamongZanzibarresidents. MalarJ 2009, 8: 287. 23.TatemAJ,HuangZ,NaribC,KumarU,KandulaD,PindoliaDK,SmithDL, CohenJM,GraupeB,UusikuP,LourenoC: Integratingrapidriskmapping andmobilephonecallrecorddataforstrategicmalariaelimination planning. MalarJ 2014, 13: 52. 24.LeMenachATA,CohenJM,HaySI,RandellH,PatilA,SmithDL: Travelrisk, malariaimportation,andmalariatransmissioninZanzibar. Scientic Reports 2011, 1: 93. 25.ShanksGD,BiomndoK,GuyattHL,SnowRW: Travelasariskfactorfor uncomplicated Plasmodiumfalciparum malariainthehighlandsof westernKenya. TransRSocTropMedHyg 2005, 99: 71 – 74. 26.IOM: TheImmigrationandbordermanagementprogramme. Geneva,Switzerland:InternationalOrganizationforMigration;2011. 27.WesolowskiA,BuckeeCO,PindoliaDK,EagleN,SmithDL,GarciaAJ,TatemAJ: Theuseofcensusmigrationdatatoapproximatehumanmovement patternsacrosstemporalscales. PLoSOne 2013, 8: e52971. 28.MinnesotaPopulationCenter: IntegratedPublicUseMicrodataSeries, International:Version6.1[Machine-readabledatabase]. Minneapolis: UniversityofMinnesota;2011. 29.GethingPW,PatilAP,SmithDL,GuerraCA,ElyazarIR,JohnstonGL,TatemAJ, HaySI: Anewworldmalariamap: Plasmodiumfalciparum endemicityin 2010. MalarJ 2011, 10: 378. 30. TheMalariaAtlasProject .www.map.ox.ac.uk. 31. TheWorldPopProject. www.worldpop.org.uk. 32.AnselinL,GetisA: Spatialstatistical-analysisandgeographicinformationsystems. AnnRegSci 1992, 26: 19 – 33. 33.GreenwoodMJ: HumanMigration-Theory,Models,andEmpirical-Studies. JRegSci 1985, 25: 521 – 544. 34.RoyJR,ThillJC: Spatialinteractionmodelling. PapRegSci 2004, 83: 339 – 361.Pindolia etal.MalariaJournal 2014, 13 :169 Page10of11 http://www.malariajournal.com/content/13/1/169

PAGE 11

35.SturrockHJ,HsiangMS,CohenJM,SmithDL,GreenhouseB,BousemaT, GoslingRD: Targetingasymptomaticmalariainfections:active surveillanceincontrolandelimination. PLoSMed 2013, 10: e1001467. 36.CohenJM,DlaminiS,NovotnyJM,KandulaD,KuneneS,TatemAJ: Rapid case-basedmappingofseasonalmalariatransmissionriskforstrategic eliminationplanninginSwaziland. MalarJ 2013, 12: 61. 37.KoitaK,NovotnyJ,KuneneS,ZuluZ,NtshalintshaliN,GandhiM,GoslingR: Targetingimportedmalariathroughsocialnetworks:apotentialstrategy formalariaeliminationinSwaziland. MalarJ 2013, 12: 219. 38.HaySI,GuerraCA,TatemAJ,AtkinsonPM,SnowRW: Urbanization,malaria transmissionanddiseaseburdeninAfrica. NatRevMicrobiol 2005, 3: 81 – 90. 39.IOM: Inpursuitofthesoutherndream:victimsofnecessityassessmentofthe irregularmovementofmenfromEastAfricaandtheHorntoSouthAfrica. Geneva:InternationalOrganizationforMigration;2009. 40.AdepojuA: MigrationandSocioeconomicChangeinAfrica. IntSocSciJ 1979, 31: 207 – 225. 41.HaySI,SmithDL,SnowRW: Measuringmalariaendemicityfromintense tointerruptedtransmission. LancetInfectDis 2008, 8: 369 – 378. 42.KingMRS,AlexanderJT,FloodS,GenadekK,SchroederMB,TrampeB, VickR: IntegratedPublicUseMicrodataSeries,CurrentPopulationSurvey: Version3.0.[Machine-readabledatabase]. Minneapolis,MN:Minnesota PopulationCenter;2010. 43.SmithDL,McKenzieFE,SnowRW,HaySI: Revisitingthebasicreproductive numberformalariaanditsimplicationsformalariacontrol. PLoSBiol 2007, 5: e42. 44.SmithDL,McKenzieFE: Staticsanddynamicsofmalariainfectionin Anophelesmosquitoes. MalarJ 2004, 3: 13. 45.SmithDL,GuerraCA,SnowRW,HaySI: Standardizingestimatesofthe Plasmodiumfalciparum parasiterate. MalarJ 2007, 6: 131. 46.SmithDL,DushoffJ,SnowRW,HaySI: Theentomologicalinoculationrate and Plasmodiumfalciparum infectioninAfricanchildren. Nature 2005, 438: 492 – 495. 47.PongvongsaT,HaH,ThanhL,Marchan dRP,NonakaD,TojoB,PhongmanyP, MojiK,KobayashiJ: Jointmalariasurveysleadtowardsimprovedcross-border cooperationbetweenSavannakhetprovince,LaosandQuangTriprovince. Vietnam.MalarJ 2012, 11: 262. 48.FikTJ: Hierarchicalinteraction-themodelingofacompetingcentralplacesystem. AnnRegSci 1988, 22: 48 – 69. 49.FikTJ,AmeyRG,MulliganGF: Labormigrationamongsthierarchically competingandinterveningoriginsanddestinations. EnvironPlannA 1992, 24: 1271 – 1290. 50.FleggJA,PatilAP,VenkatesanM,RoperC,NaidooI,HaySI,SibleyCH, GuerinPJ: Spatiotemporalmathematicalmodellingofmutationsofthe dhpsgeneinAfrican Plasmodiumfalciparum MalarJ 2013, 12: 249.doi:10.1186/1475-2875-13-169 Citethisarticleas: Pindolia etal. : Quantifyingcross-bordermovements andmigrationsforguidingthestrategicplanningofmalariacontroland elimination. MalariaJournal 2014 13 :169. Submit your next manuscript to BioMed Central and take full advantage of: € Convenient online submission € Thorough peer review € No space constraints or color “gure charges € Immediate publication on acceptance € Inclusion in PubMed, CAS, Scopus and Google Scholar € Research which is freely available for redistribution Submit your manuscript at www.biomedcentral.com/submit Pindolia etal.MalariaJournal 2014, 13 :169 Page11of11 http://www.malariajournal.com/content/13/1/169