Balancing Treatment Allocation over Continuous Covariates A New Imbalance Measure for Minimization英文文献资料.docVIP

HindawiPublishingCorporation JournalofProbabilityandStatistics Volume2012,ArticleID842369,13pages doi:10.1155/2012/842369 ResearchArticle BalancingTreatmentAllocationover Continuous Covariates:ANewImbalance MeasureforMinimization YanqingHu andFeifangHu 1 2 1Department ofStatistics,WestVirginiaUniversity,HodgesHall,Morgantown, WV26506,USA 2Department ofStatistics,University ofVirginia,HalseyHall,Charlottesville, VA22904,USA Correspondence shouldbeaddressedtoFeifangHu,fh6e@ Received12January2012;Revised10April2012;Accepted 7May2012 Academic Editor:PetervanderHeijden Copyright q2012Y.HuandF.Hu. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use,distribution, andreproduction in anymedium,providedtheoriginalworkisproperlycited. Inmanyclinical trials,itisimportant tobalance treatment allocation overcovariates. Although a greatmanypapershavebeenpublished onbalancing overdiscretecovariates, theprocedures for continuous covariates have been less well studied. Traditionally,a continuous covariate usually needstobetransformedtoadiscreteonebysplittingitsrangeintoseveralcategories.Suchpractice mayleadtolossofinformation andissusceptibletomisspeci?cation ofcovariatedistribution. The morerecentpapers seektode?ne animbalance measure thatpreserves thenature ofcontinuous covariates and set the allocation rule in order to minimize that measure. We propose a new design, which de?nes the imbalance measure by the maximum assignment di?erence when all possible divisions ofthecovariate range areconsidered. This measure depends only onranks of thecovariate values and istherefore freeofcovariate distribution. Inaddition, wedeveloped an e?cientalgorithm toimplement thenewprocedure.Bysimulation studiesweshowthatthenew procedureisabletokeepgoodbalancepropertiesincomparison withotherpopulardesigns. 1.Introduction Balanced allocation among treatment groups is often desirable in