两步移动搜索法案例.ppt

Medical Geography:Health Care Accessibility and ImplicationsFahui WangFred B.Kniffen Professor of Geography Louisiana State UniversityJune 17,2010 CNU.ContentsoThe issue of accessoGIS-based accessibility measuresoOptimization towards equal accessoApplication in healthcare access nPHSA designation and late-stage canceroConcluding comments.Access Matters!oConvenience of access to activitiesnjob,school,healthcare,park,public service,shopping oAccess defines“location”oSpatialnWhere you areoNonspatialnWho you areoSocial justice and public policy.MeasurementoHow scarce are the services?nSupply vs.Demand(ratio)oHow far are demands from supplies?nDistancenRoad networknTransportation means.Primitive measuresoDistance(time)from supplyoSimply potential modeloSupply-demand ratio within an areaoFCA with a radius/boxoFCA with a travel time range.From S/D Ratio to FCAoSimply S/D RatiooFCA with a RadiusnImprovements oWithin-unit variationoCross-bordernLimitationsoStraight-lineoS-D beyond radius.2-step floating catchment area:2SFCAoScale availability at each supply location oSum up accessible supplies around each demand location.2SFCA Recap1a.For each supply location j,select all demand locations k within a travel time range from j(catchment Cj).1b.Compute the supply to demand ratio within Cj:2a.For each demand location i,search all supply locations j within the travel time range from i(catchment Zi).2b.Sum up the supply to demand ratios at these supply locations:.Generalizing“distance decay”oas a continuous function na power or exponential function na Gaussian functionoas a discrete variablendichotomous as in 2SFCA nmultiple as in E2SFCAoa hybrid approachna kernel density function na three-zone approach.Generalizing“distance decay”.Validation oModeling actual supply-demand interactionsnIndividual travel behaviornAggregated travel pattern.Optimization oTowards equal accessibilityoFormulated as a planning problem ominimizing accessibility gaps among demand locations by adjusting the amounts of supplies at given supply locations.A Quadratic Programming ApproachoSubject towhere.Value of optimizationoEmphasizing the equality issue oIdentifying over-and under-capacity of supplyoGuiding planning and public policy in closing the gap.Application in Healthcare Access:HPSA designationoU.S.ranks behind in health performanceodeep disparities in access to care and health outcomesoFederal programs:MUP&HPSAs oCalls for better measure of accessibility.Non-spatialNon-spatialSocioeconomicSocioeconomicDisadvantagesDisadvantages(Factor 1)(Factor 1)Socio-culturalSocio-culturalBarriersBarriers(Factor 2)(Factor 2)High Healthcare High Healthcare NeedsNeeds(Factor 3)(Factor 3)Female-headed households(%)Female-headed households(%)Population in poverty(%)Population in poverty(%)Nonwhite minorities(%)Nonwhite minorities(%)Households w/o vehicles(%)Households w/o vehicles(%)Home ownership(%)Home ownership(%)Households with linguistic isolation(%)Households with linguistic isolation(%)Households with 1 person per room(%)Households with 1 person per room(%)Population w/o high-school diploma(%)Population w/o high-school diploma(%)Median income($)Median income($)High needs Population(%)High needs Population(%)Consolidating nonspatial variables.Integrating Spatial and Nonspatial FactorsoGeographic AreanPrimary Indicator(spatial accessibility score)1/3500nPrimary Indicator(spatial accessibility score)1/3000 AND secondary indicator(factor 3)1 standard deviation above mean oPopulation GroupnPrimary Indicator(factor 1)1 standard deviation above meannPrimary Indicator(factor 1)standard deviation above mean and secondary indicator(factor 2)1 standard deviation above mean.Application in healthcare access(vs.late-stage cancer).Risk breakdowns.CultureKnowledge(Dis)advantageHealth insuranceAccess to primary careAvailability and accessibilityof screening servicesQuality of health careMultilevel approachSocio-economic disadvantageSocial networks and institutionsLocal resourcesPERSONPLACE.Rural-urban continuumChicagoChicago suburbsOther metro areasLarge town(10-50k)Rural.Multilevel Model CoefficientsBreastColorectalLungProstateChic_sub-.181*-.065*-.001-.092*-.087*-.029-.007-.012-.017-.189*-.105*-.062Other_metro-.278*-.146*-.110*-.131*-.122*-.056-.172*-.183*-.252*-.128*-.039-.012Large town-.336*-.162*-.167*-.272*-.250*-.198*-.331*-.338*-.431*-.191*-.065-.063Rural-.207*-.032-.063-.169*-.156*-.085-.204*-.212*-.323*-.164*-.035-.038Age50.547*.541*.216*.216*.195*.188*.129.128Age70-.279*-.279*-.082*-.083*-.234*-.233*-.174*-.177*Black.373*.370*.083*.077*-.068*-.085*.289*.279*Income-.134*-.047-.110*-.087Non_Engli.012*.002-.001.003Access-37.6*20.0-27.2*-9.9Time.0020.000.Findings oEffects of individual attributesnOlder age linked to reduced risknBlacks linked to higher riskoEffects of contextual variablesnIncome&primary care access:significant in breast and lung cancernOthers:marginally or not significantoEffects of urban-rural location.“urban disadvantage”?.Public policy implication.Concluding remarksoRise of Computational Social Sciences(CSS)and Spatially-integrated Social Sciences(SSS)oTrends in social sciencesoMichael Batty:“to do good social science that is policy relevant,quantitative methods are essential and such methods,and the theory behind their practice,must be spatial.”.