DiscordDiscordanceandConcurrencyComparingIndividualandPartnershipL - PDF document

Discord,Discordance,andConcurrency:ComparingIndividualandPartnership-LevelAnalysesofNewPartnershipsofYoungAdultsatRiskofSexuallyTransmittedInfectionsPAMINAM.GORBACH,MHS,DPH,*LYDIAN.DRUMRIGHT,MPH,†ANDKINGK.HOLMES,MD,PBackgroundandObjectives:Partnershipsthatarediscordantbydemographicandbehavioralcharacteristicsorareconcurrentmayfacilitatetransmissionofsexuallytransmittedinfections(STIs)bybridgingsexualnetworks.ThegoalofthisstudywastoexamineifSTIriskwithinpartnershipsisassociatedwithdiscordanceandconcurrencyusingthepartnershipastheunitofanalysis. TheauthorsthanktheresearchinterviewersJulieYasuda,NanceeEvan-galista,KristinBorchardt,andJennieFosterandDr.RobertGunn,Dr.RobertGilchick,Dr.ChrisPeter,andothermembersoftheSanDiegoDepartmentofPublicHealth,SexuallyTransmittedDiseaseProgramfortheircontributionstothisstudy.ThisstudywassupportedbyNIAID1K22AI01737–01.Correspondence:PaminaM.Gorbach,MHS,DrPH,Departmentof SexuallyTransmittedDiseases,January2005,Vol.32,No.1,p.7–12DOI:10.1097/01.olq.0000148302.81575.fcCopyright©2004,AmericanSexuallyTransmittedDiseasesAssociationAllrightsreserved. networks,thatpersonbecomesaparticularlyefcientbridge,potentiallycarryingSTIsbetweengroups.Therearefewdataontheproportionofconcurrentpartnershipsthatinvolvebridgingofnetworksorontheproportionsofpotentialbridgerswhohaveconcurrentpartnerships(asopposedtoserialpartnerships)withmembersofdifferentsexualnetworks.Partnershipsthatinvolvediscordance,aswellasconcurrentpartners,havethepotentialtorapidlyacceleratethespreadofSTIs/HIVinapopulation.PreviousstudiesshowingassociationsbetweendiscordanceandSTIwerebasedonpartnerdatafromindexreports,usedself-reportedSTIhistory,orusedhistoricaldata.ToconsiderwhetherdiscordanceandconcurrencywithinpartnershipsrepresentrisksforhavingSTI,wecollectedbehav-ioralandbiologicdatafrombothmembersofnewpartnershipstoexamineassociationswithSTI.Weconsidereddiscordancebyethnicity,numberoflifetimepartners,educationlevel,andcon-templationofpartnershiptermination.Concurrencyatthestartofthepartnershipwasalsoexamined.BetweenAugust2000andSeptember2001,96younghetero-sexualadults(aged18–25years)seekingcareatSTDandfamilyplanning(FP)clinicsinSanDiegoCountyandtheirnewsexualpartners(aged18–30years)wererecruitedtoparticipateina12-monthlongitudinalstudyofsexualpartnerships.Anewsexualpartnerwasdenedassomeonewithwhomtheindexpartnerhadhadvaginalintercourseforthersttimewithintheprevious3months.For43partnerships,theindexpartnerwasmaleandfor53,theindexpartnerwasfemale;51partnershipswererecruitedfromFPclinicsand45fromSTDclinics.Participationratesofthosewhometeligibilitycriteria(newsexualpartnership,aged18–25andheterosexual)were80%to90%;allofthosewhodeclinedparticipationreportedthattheirpartnerwasnotwillingtoparticipatewiththem.Eligibleparticipantsattendedtheirinitialvisittogetherasacoupleandwereinterviewedusinga240-itemaudio-computer-assistedself-interview(ACASI)of60minutes’durationsimulta-neouslyinseparate,privatelocations.UrinespecimenscollectedfromallparticipantsattheinitialvisitweretestedforNeisseriagonorrhoeaebyligasechainreaction(LCR;AbbottDiagnostics,AbbottPark,IL).Currentinfectionwasdeterminedusingculture(InPouch;BiomedDiagnostics,SanJose,CA)ofurinespecimensfrommenandself-collectedvaginalswabsfromwomen.For90patients(including45participantsrecruitedfromtheSTDclinic,11oftheirpartners,and34recruitedfromFPclinicswhovisitedtheSTDclinic),medicalrecordswerereviewedforprevious3-monthhistoryofgonorrhea,chlamydialinfection,trichomoniasis,non-gonococcalurethritis(NGU),mucopurulent,cervicitisandpelvicinammatorydisease.OfthoseparticipantswithSTI-positivere-portsintheirmedicalrecords,most(89%;n25)hadbeendiagnosedwithin1weekbeforeorafterthebaselineinterview.Ofthe3remainingSTI-positiveindividuals,allwerediagnosed15to70daysbeforethebaselineinterviewandafterinitiationofthepartnershipwiththeinterviewpartner.Allparticipantsweredeemedmentallytatthetimeofenroll-ment,receivedastudycodetoprotectcondentiality,andreceivednancialcompensationforparticipation.TheInstitutionalReviewBoardsoftheUniversityofCalifornia,LosAngeles,andSanDiegoStateUniversityreviewedandapprovedtheprotocol.Allparticipantsgaveinformedconsent.Theseanalysesusedatafr

omtheinitialbaselinevisit.Allpartnershipcharacteristicanalyses(ie,discordance)werebasedonpartnershiplevelvariables.Discordancevariableswerecreatedbycomparingpartners’responsestothefollowing:lifetimenumberofpartners(groupedby9orfewerlifetimepartnersvs.10orgreater,aspreviouslyexaminedbyLaumannandYoum),contemplationofpartnershiptermination(evervs.nevercontemplatedtermina-tionofpartnership),educationlevel(lessthanhighschool,com-pletedhighschool,ormorethanhighschool),andethnicity(self-reportofanyof7differentethnicgroups).Partnershipsinwhichresponsestothesequestionswerethesamewereconsideredconcordant;partnershipsinwhichpartnersreporteddifferentre-sponsestothesamequestionwereconsidereddiscordant.Addi-tionally,partnershipswereclassiedas1)concurrentif1orbothpartnersreportedadditionalpartnersatthestartoftheirpartner-ship;2)usingcondomsconsistentlyifbothpartnersreportedconsistentcondomusewithinthepastmonth;and3)havinganSTIwithinthepartnershipif1orbothmembershadanSTIatbaselineorwithintheprior3months.InanalysesexaminingSTIriskinthe192individualsbyindi-vidualcharacteristics,chi-squaredanalysis,tests,andlogisticregressionwereusedtodeterminetheassociationsbetweenSTIwithinthepast3monthsandage,education,ethnicity,consider-ationofpartnershiptermination,numberoflifetimepartners,con-domuse,andconcurrency.Forallindividual-levelanalyses,sur-veyestimationtechniquesinSTATA7.0SE,whichweredesignedtocorrectfornonindependenceofclustersintheregressionmod-wereusedtocontrolforthepresenceofbothmembersofthepartnership.Atthepartnershiplevel,usingthe96partnershipsastheunitofanalysis,chi-squaredanalysisandmultivariatelogisticregressionwereusedtodetermineassociationsbetweenpartner-shipSTIandpartnershipdiscordancebyethnicity,educationallevel,numberoflifetimepartners,orconsiderationofpartnershiptermination;partnershipconcurrency;andconsistentcondomuse.DatawereenteredbyparticipantresponseintoCi3(SawtoothTechnologies,Northbrook,IL)andanalyzedusingSTATAver-sion7.0SE(STATACorp.,CollegeStation,TX).Ofthe192participants,themeanagewas21.8years(Table1).Themajority(60.9%)reportededucationbeyondhighschool.Participantsrepresentedawiderangeofethnicgroups,withlessthanhalfidentifyingthemselvesaswhite;and75%reportedatleastpart-timeemployment.Fifty-threepercentofparticipantsreportedhaving10ormorelifetimesexualpartnersand62%reportedevercontemplatingpartnershipterminationwiththeirenrollmentpartner.Only26%ofparticipantsreportedconsistentcondomuseinthepastmonthwiththeirenrollmentpartner;14%reportedhavingaconcurrentpartneratthestartofthispartnership.STIs(chlamydialinfection,gonorrhea,ortrichomoniasis)weredetectedin11.5%ofindividualsatbaseline,andanother3.1%hadarecenthistoryofSTIs(identiedbyreviewofthepast3monthsoftheirmedicalrecordsintheSTDclinic),producingatotal14.6%ofindividualswithrecentorcurrentSTI.Meanageofsexualdebut,andmeannumberofsexualpartnersduringthepastmonthandpast12monthsdidnotdiffersignicantlyforindivid-ualswithandwithoutSTI.ComparedwiththosewithoutanSTI,signicantlymoreparticipantswithSTIsreportedconcurrencyatrstsexualcontactwiththeirenrollmentpartner(0.02)andreportedHispanicethnicity(0.03;Table1).Bylogisticregression(datanotshowninthetable),theoddsofhavingacurrentorrecentSTIwas3.1forthosewhoreportedaconcurrentpartneratthestartoftheirpartnershipascomparedwiththosewhodidnot(0.02)andwas3.2forthosewhoreportedHispanicethnicitycomparedwithallothers(0.04).Age,education,SexuallyTransmittedDiseasesJanuary2005GORBACHETAL numberoflifetimepartners,contemplationofpartnershiptermi-nation,andcondomusewerenotassociatedwithSTIsattheindividuallevelbylogisticregression.Ofthe96partnerships,22%hadaleast1memberwhohadhadacurrentorrecentSTI(Fig.1);inonethirdofthosepartnerships,membershadacurrentorrecentSTI.In82%ofpartnerships,atleast1memberreportedinconsistentcondomuseinthepastmonth.In26%ofpartnerships,1orbothpartnersreportedhavingconcurrentsexpartnersatthestartofthepartnershipwiththeenrollmentpartner;bothpartnersreportedhavingaconcurrentpartnerin13%ofpartnerships.Therewerehighlevelsofdiscor-dancewithinpartnerships:43%werediscordantbynumberoflifetimepartners,41%bycontemplationofpartnershiptermina-tion,50%byeducationallevel,and49%byethnicity(Fig.1).B

ychi-squaredanalysis(Table2),currentorrecentSTIinthepartnershipwassignicantlyassociatedwithdiscordancebycon-templationofpartnershiptermination(33%vs.14%concordant,0.02)andlifetimenumberofpartners(35%vs.13%ofconcordantpartnerships,0.01).Whendiscordancebylifetimenumberofpartnerswasbrokendowninto1)discordant,2)con-cordantfor9orfewerpartners,and3)concordantby10orgreaterpartners,thosemostlikelytohaveanSTIeitherindividuallyorwithintheirpartnershipwereindiscordantpartnerships(8%ofpartnershipsconcordantfor9orfewerpartners,9%concordantfor10orgreaterpartners,and22%ofthosediscordanthadSTI;0.03).Additionally,agreaterpercentageofpartnershipsthatwerediscordantbyethnicityhadanSTIthanthosethatwereconcordant(30%vs.14%),whichwasnearlystatisticallysignicant(TABLE1.IndividualBackgroundCharacteristicsbyCurrentorRecent*SexuallyTransmittedInfections(n No.(%)NoSTINo.(%)No.(%)Age,mean(median)20.8(20.5)21.5(22.0)21.4(22.0)Education,meanyears(median)12.1(12.5)13.2(13.0)13.1(13.0)Education(categorical)Lessthanhighschool6(21.4)27(16.5)33(17.2)Completedhighschool8(28.6)34(20.7)42(21.9)Morethanhighschool14(50.0)103(62.8)117(60.9)AfricanAmerican/Black2(7.1)14(8.5)25(13.0)Asian/PacicIslander6(21.4)19(11.6)16(8.3)Caucasian/white-non-Hispanic9(32.1)25(15.2)89(46.4)Hispanic/Latino8(28.6)81(49.4)34(17.7)Otherethnicity1(3.6)8(4.9)9(4.7)Mixedethnicity2(7.1)17(10.4)19(9.9)HispanicEthnicityvs.allothers†9(32.1)25(15.2)34(17.7)CurrentemploymentFull14(50.0)66(40.2)80(41.7)Part6(21.4)60(36.6)66(34.4)None8(28.6)38(23.2)46(23.9)Contemplatingpartnershiptermination19(67.9)100(61.0)119(62.0)Meanagesexualdebut(median)15.8(16)15.9(16)16(15.9)Meanno.ofpartnerspastmonth(median)1.3(1.0)1.5(1.0)1.4(1.0)Meanno.ofpartnerspastyear(median)3.7(3.0)4.5(3.0)4.3(3.0)Meanno.oflifetimesexpartners(median)14.8(9.5)17.5(10.0)17.1(10.0)Reports10ormorelifetimepartners14(50.0)87(53.1)101(52.6)Alwayscondomusepastmonth–newpartner8(28.6)40(24.4)48(25.5)Nevercondomusepastmonth–newpartner8(28.6)48(29.3)56(29.8)Reportconcurrentpartneratstartofpartnership8(28.6)19(11.6)27(14.1)Female12(42.6)84(51.2)96(50.0)*RecentSTIobtainedfrom3-moreviewofSTDclinicmedicalcharts.sexuallytransmittedinfection. Fig.1.Percentageofpartnershipsdiscordant,concurrentatthestartofthepartnership,usingcondomsconsistently,andwithsex-uallytransmittedinfectioninthepartnerships(n96).()Discor-dantbynumberoflifetimepartners9orfewerpartnersversus10orgreater.()Discordantbypartnershipterminationeverversusneverconsideredseparationorterminationofpartnership.()Dis-cordantbyeducationlessthanhighschoolversushighschoolversusmorethanhighschool.()Discordantbyethnicityreportofanyofthe7differentethnicgroups.Vol.32No.1DISCORDANCEANDCONCURRENCYASSTIRISKSINPARTNERSHIPS 0.07).SignicantlymorepartnershipswithconcurrencyatthestartoftheirpartnershiphadarecentorcurrentSTI(36%withcon-currencyvs.17%withoutconcurrency,0.05).DiscordancebyeducationlevelandreportedcondomusewithinthepartnershipwerenotassociatedwithcurrentSTI(Table2).Multivariatelogisticregressionmodelsgaveresultssimilartotheunivariateanalyses(Table3).CurrentorrecentSTIwithinthepartnershipwassignicantlyassociatedwithethnicdiscordance(oddsratio[OR],3.4;0.04),discordancebycontemplationofpartnershiptermination(OR,4.2;0.02),discordancebylifetimenumberofpartners(OR,4.9;0.01),andpartnershipconcurrency(1orbothmemberswithconcurrentpartnersatstartofthepartnership;OR,3.8;0.03).EducationaldiscordanceandpartnershipcondomusewerenotsignicantlyassociatedwithanSTIinthepartnershipateithertheunivariateormultivariatelevels.Itisimportanttonote,however,thatwheneachoftheidentiedpartnershipriskfactorswasstratiedbycondomuse,thesizeoftheeffectoftheriskwasgreateramongpartnershipswithinconsistentcondomuse.Foryoungheterosexualswithnewsexualpartners,characteris-ticsoftheirpartnershipsmayproviderepresentationsofSTIriskbeyondindividualcharacteristicsbyidentifyingwhethertheirpart-nersinhabitdifferentsexualnetworks.AnalysesofindividualSTIbyindividualcharacteristicsinthisstudyonlyrevealedthatre-portingaconcurrentpartneratthestartofthepartnershiporHispanicethnicitywereassociatedwithSTI.However,whenthepartnershipservedastheunitofanalysis,thosemorelikelytohavearecentorcurrentSTIweredisco

rdantbyethnicity,numberoflifetimepartners,andcontemplationofpartnershiptermination.Moreover,ifoneorbothmembersofthepartnershipreportedaconcurrentpartneratthetimeofrstsexualcontactwiththeirenrollmentpartner,thepartnershipwas3.8timesmorelikelytohavearecentorcurrentSTI.Thesendingssuggestthatyoungadultswithbridgingbehaviors,whethertemporal(ie,concurrency)orsocial(bydiscordance),aremorelikelytoacquireandtransmitSTIs.Therefore,thereisapotentialtoenhancetheSTIriskassessmentsconductedinSTDandFPclinicsbycapturingtheriskembeddedinpatients’partnerships.OurresultsareconsistentwithpreviousstudiesofpartnershipdiscordanceandSTIthatsuggestthatdiscordancemaybeassoci-atedwithSTIacquisitionbecauseitisamarkerforbridgingbetweenhighandlowSTIprevalencegroups.DiscordancebynumberoflifetimepartnersrepresentsaconnectionbetweenhighandlowSTIprevalencepopulations.PartnershipsdiscordantbynumberoflifetimepartnersweremorelikelytohaveanSTIthanpartnershipsthatwereconcordant,regardlessofsexualactivityclass(8%hadSTIsamongthoseconcordantforfewerthan10lifetimepartners,9%hadSTIsamongthoseconcordantfor10ormorepartnersvs.22%withSTIsamongthosediscordantbylifetimepartners).Fortheseyoungadults,beingwithapartnerfromanetworkwithadifferentlevelofpartnerchangerepresentsagreaterriskforSTIsthanhavingapartnerfromanetworkwiththesamelevelofpartnerchangeasoneself,reiteratingthatapatient’sSTIriskmaynotonlybearesultoftheirbehavior,buttheirpartner’sbehavioraswell.DiscordancebyethnicgroupisanotherproxyforbridgingbetweenhigherandlowerSTIprevalencepopulationsandwasassociatedwithcurrentSTIinthisstudyandothers.OurdataonethnicdiscordancemayalsosupporttheconceptofSTIper-sistencewithingroupsinwhichmostpartnersareethnicallycon-andmaydemonstratetheimportanceofSTIbridgingbetweengroupsdiscordantbyethnicityinspreadinginfectionacrossapopulation.Wefoundthatpartnershipsthatwerediscor-dantbyethnicityandhadarecentorcurrentSTIweremostoftencomposedofwhiteandblackorwhiteandHispanicindividuals,whereasthosewhichwereconcordantbyethnicityandhadanSTIweremostoftenHispanic.InColoradoSprings,discordantpart-nershipsbetweenwhiteandblackindividualswasamarkerofbridgingfromthecoregrouptogeographicallyadjacentandperi-pheralpopulations.Inourdata,wecouldbeobservingasimilartrendinwhichourHispanicparticipantswereasubsetforahighSTIprevalencegroupandthewhiteandblackparticipantswithdiscordantpartnerswerethebridgers.FortheseyoungadultsinSanDiego,itappearsthatassortativemixingbyHispanicpartic-ipantswasariskforSTI,whereasdisassortativemixingincreasedTABLE3.LogisticRegressionAnalysis:AssociationsBetweenPartnershipsWithSTIandPartnershipDiscordance,Concurrency,andCondomUse(n Ratio95%CIDiscordantbyethnicity3.421.1–11.20.043*Discordantbyeducation0.850.3–2.70.776Discordantbypartnershiptermination4.231.3–13.60.015*Discordantbylifetimepartners4.891.5–16.50.010*Concurrencyinpartnership3.771.2–12.20.026*Condomuse100%timepastmonth2.470.6–9.70.185sexuallytransmittedinfection;CIcondenceinterval.TABLE2.UnivariateAnalysis:AssociationsBetweenPartnershipCurrentorRecentSTIandDiscordance,Concurrency,andCondomUse(n No(%STI-Positive)(%STI-Positive)(%STI-Positive)OddsRatio(95%CI)Ethnicity14.330.00.0662.55(0.9–7.0)Education20.822.90.8051.13(0.4–2.0)Contemplatingpartnershiptermination14.033.30.022*3.06(1.1–8.3)No.oflifetimepartners13.235.00.013*3.17(1.2–8.8)No(%STI-Positive)Yes(%STI-Positive)Concurrency16.936.00.047*2.8(1.0–7.7)Condomuse100%timepastmonth20.026.90.4661.64(0.5–5.3)sexuallytransmittedinfection;CIcondenceinterval.SexuallyTransmittedDiseasesJanuary2005GORBACHETAL theriskofSTIforwhiteandblackparticipants.Additionally,ethnicdiscordancehaspreviouslybeenshowntobeassociatedwithdecreasedcondomuseamongadolescentpartnerships,whichmaycontributetoincreasedriskofSTI.AlthoughcondomusewasnotassociatedwithSTIorethnicdiscordanceinourstudy,whenstratiedbycondomuse,theassociationbetweenSTIandethnicdiscordancewassignicantonlyamongthosepartnershipswithinconsistentcondomuse(29.4%hadSTIsthatwerediscor-dantbyethnicitycomparedwith11.1%whowereconcordant;Discordancebycontemplationofpartnershipterminationmayhavemoretodowithpartnershipdynamicsorpartnership“discord”thanbridgingbetweenhighandlowSTIprevalencepopulations.P

artnershipsconcordantbyacommitmenttore-maintogetherorinwhichbotharecontemplatingseparationare2differenttypesofpartnerships,therstrepresentingcommit-tedpartnershipsandthesecondmorecasualpartnerships.Part-nershipsdiscordantoncontemplatingrelationshipterminationarethoseinwhicheachpartnerseesthepartnershipdifferently,oneregardingitascasualorendingwhiletheotherbelievesitiscommitted.CondomusewasnotsignicantlyassociatedwithSTIorwithdiscordancebypartnershiptermination,anditwasnotloweramongpartnersconcordantbycommitment;however,veryfewpartnershipsreportedconsistentcondomusesotheremayhavebeentoolittlepowertodetectaneffect.Additionally,concurrencywasnotassociatedwithdiscordancebycontem-platingpartnershiptermination;however,concurrencymaynotnecessarilyreectalackofrelationshipcommitment.Ourpre-viousworkhasshownthatindividualsmaypracticeconcur-rencytomeetneeds1partnermaynotbeabletofulll,althoughtheyremaincommittedtoandinvolvedsexuallywiththispartner.Therefore,theassociationbetweenpartnershiptermination,ordiscordinthesepartnershipsandSTIriskmaybetheresultofriskfactorsthatliebeneaththisdiscord,whichmaynothavebeendirectlymeasuredinthestudy.Partnership-levelconcurrency,aswellasindividual-levelcon-currency,wasindependentlyassociatedwithSTIinthisstudy.Thisisconsistentwithpreviousstudiesofconcurrencyand12,19–21Moreover,withinoursample,concurrencyanddiscor-dancecoexistedandwerenotmutuallyexclusive.Thiscombina-tionofconcurrencyandbridgingmayenhancethespreadofSTIatthenetworklevel.BridgingalonecanbringSTIfromhigherSTIprevalencepopulationstolowerSTIprevalencepopulations;andconcurrencyalonelinksindividualsintoclustersofsexualconnec-tionsatonepointintime,increasingtheprobabilityofSTIspreadtomanymembersofthenetworkwithinashortperiodoftime.BothscenariosincreasethespreadofSTIindifferentways.Thus,if2populations(ornetworks)werebridgedbyindividualswhohaveconcurrentpartners,theincreaseinSTIwouldbeatleastadditive,ifnotmultiplicative,innature.However,STItransmis-sionwillonlybeobservedifconcurrencyorbridgingoccurswithinoracrossnetworksinwhichSTIalreadyexists.Regardless,measuringconcurrencyamongbothmembersofapartnershipmayallowforamorecompleteunderstandingoftheinteractionbe-tweenconcurrencyandbridging,aswellasanindividual’sriskasaresultoftheirpartner’sbehavior.ThisstudyislimitedinitsabilitytofullyexploreassociationsbetweenSTIandageoreducationaldiscordanceasaresultofthestudydesign.First,thoseinthemostage-discordantpart-nerships(withmorethan10yearsapart)wereexcluded.Sec-ond,theabilitytoexamineeducationaldifferenceswaslimitedbecausemanyofourparticipantsintheirlateteensandearly20smaynotyethavecompletedtheireducationalcareers,againlimitingtheabilitytofullyexaminetheeffectofthistypeofdiscordance.Theremayalsobelimitationsindetectingsmall,signicantdifferencesasaresultofsamplesize.Forexample,condomusewasnotassociateddirectlywithSTI;however,itwasapowerfulstratifyingvariablethatenhancedassociationsbetweenSTIanddiscordance.Thesendings,aswellasdatafromotherstudiesexaminingbothmembersofsexualpartnerships,mayprovidevaluableinsightformathematicalmodelsthatexaminetheroleofsexualmixingpatternsandconcurrencyintheinvasion,persistence,andelimi-nationofSTIs.Althoughmodelsthathaveexaminedeitherbridg-orconcurrencyhavegreatlyadvancedtheunderstandingofSTItransmission,theinformationtheyprovidemaynotbecomplete.Thisstudyindicatesthatconcurrency(tem-poralmixing)andbridging(socialmixing)arenotindependentofoneanother.However,concurrencyanddiscordancebynumberoflifetimepartners,ethnicity,andconsiderationofpartnershipter-minationwereeachindependentlyassociatedwithanincreasedriskofSTIinoursample.Therefore,modelsthatcanaccountfortheoverlapintemporal(concurrency)andsocial(bridging)mixingpatternsarelikelytoprovideamorerealisticimpressionoftrans-missiondynamics.Moredataonhowindividualschoosepartnersanddecideonconcurrentpartnersmaybeneededbeforemodelscanbeconstructedthatbetteraccountforthecoexistenceofdiscordanceandconcurrency.Individualswithdiscordantandconcurrentpartners(bridges)representextremelyefcientroutesforSTI/HIVepidemicstospreadacrossapopulationbycarryingdiseaseacrossdistinctsexualnetworks;therefore,itisimportantthatcliniciansandcounselorsinSTDan

dFPclinicsareabletoincorporatepartnershiplevelassessmentstohelpidentifythemandfocusspecialattentiononfuturepreventionofSTIsonthese1.AralSO.Patternsofsexualmixing:MechanismsfororlimitstothespreadofSTIs?SexTransmInfect2000;76:415–416.2.AndersonRM.Transmissiondynamicsofsexuallytransmittedinfec-tions.In:HolmesKK,SparlingPF,MardhP-A,LemonSM,StammWE,PiotP,WasserheitJN,eds.SexuallyTransmittedDiseases.NewYork:McGraw-HillCo,Inc,1999:25–38.3.RothenbergRB,PotteratJJ.Temporalandsocialaspectsofgonorrheatransmission:Theforceofinfectivity.SexTransmDis1987;15:88–4.PotteratJJ,MuthSQ,etal.Sexualnetworkstructureasanindicatorofepidemicphase.SexTransmInfect2002;78:i152–i158.5.LaumannEO,GagnonJH,MichaelRT,MichaelsS.TheSocialOrganizationofSexuality.Chicago:UniversityofChicagoPress,6.AralSO,HughesJP,StonerB,etal.Sexualmixingpatternsinthespreadofgonococcalandchlamydialinfections.AmJPublicHealth1999;89:825–833.7.LaumannEO,YoumY.Racial/ethnicgroupdifferenceinthepreva-lenceofsexuallytransmitteddiseasesintheUnitedStates:Anet-workexplanation.SexTransmDis1999;26:250–261.8.FordK,LepkowskiJM.CharacteristicsofsexualpartnersandSTDinfectionamongAmericanadolescents.IntJSTDAIDS2004;15:260–265.9.LeeJK,JenningsJM,EllenJM.Discordantsexualpartnering:Astudyofhigh-riskadolescentsinSanFrancisco.SexTransmDis2003;30:234–340.10.ServiceSM,BlowerS.HIVtransmissioninsexualnetworks:Anempiricalanalysis.ProcRSocLondBBiolSci1995;260:237–244.11.GregsonS,NyamukapaCA,GarnettGP,etal.Sexualmixingpatternsandsex-differentialsinteenageexposuretoHIVinfectioninruralZimbabwe.Lancet2002;359:1896–1903.12.KoumansEH,FarleyTA,GibsonJJ,etal.CharacteristicsofpersonswithsyphilisinareasofpersistingsyphilisintheUnitedStates:Vol.32No.1DISCORDANCEANDCONCURRENCYASSTIRISKSINPARTNERSHIPS Sustainedtransmissionassociatedwithconcurrentpartnerships.SexTransmDis2001;28:497–503.13.PotteratJJ,Zimmerman-RogersH,etal.Chlamydiatransmission:Concurrency,reproductionnumber,andtheepidemictrajectory.AmJEpidemiol1999;150:1331–1339.14.STATAUser’sGuide,release7.CollegeStation,TX:STATAPressPublication,2003..15.FordK,SohnW,LepkowskiJM.Characteristicsofadolescent’ssexualpartnersandtheirassociationwithuseofcondomsandothercontraceptivemethods.FamPlannPerspect2001;33:100–16.StonerBP,WhittingtonWLH,AralSO,etal.Avoidingriskysexpartners:Perceptionofpartners’risksvpartners’selfreportedrisks.SexTransmInfect2003;79:197–201.17.DrumrightLN,GorbachPM,HolmesKK.Dopeoplereallyknowtheirsexualpartners?Concurrency,knowledgeofpartnerbehaviorandsexuallytransmittedinfectionswithinpartnerships.SexTransmDis2004;31:437–442.18.GorbachPM,StonerBP,AralS,CoronadoN,ConnorS,HolmesK.Ittakesavillage:UnderstandingconcurrentpartnershipsinSeattle,WA.SexTransmDis2002;29.19.RosenbergMDGJ,GurveyJE,AdlerN,etal.Concurrentsexpartnersandriskforsexuallytransmitteddiseasesamongadolescents.SexTransmDis1999;26:208–212.20.CastorD,JollyP,FurlongeC,etal.DeterminantsofgonorrheainfectionamongSTIclinicattendersinTrinidad–II:Sexualbehav-iouralfactors.IntJSTDAIDS2002;13:46–51.21.ManhartL,AralSO,HolmesKK,etal.Sexpartnerconcurrency:Measurement,prevalenceandcorrelatesamongurban18–39yearolds.SexTransmDis2002;29:133–143.22.MorrisM.Concurrentpartnershipsandsyphilispersistence:Newthoughtsonanoldpuzzle.SexTransmDis2001;28:504–507.23.GuptaS,AndersonRM,MayRM.Networksofsexualcontacts:ImplicationsofthepatternofspreadofHIV.AIDS1989;3:807–817.24.BlowerSM,McLeanAR.Mixingecologyandepidemiology.ProcRSocLond1991;245:187–192.25.MorrisM,KretzschmarM.ConcurrentpartnershipsandthespreadofHIV.AIDS1997;11:641–648.26.WattsC,MayR.TheinuenceofconcurrentpartnershipsonthedynamicsofHIV/AIDS.MathBiosci1992;108:89–104.27.ChickS,AdamsA,KoopmanJ.Analysisandsimulationofastochas-tic,discrete-individualmodelofSTItransmissionwithpartnershipconcurrency.MathBiosci2000;166:45–68.28.KretzschmarM,MorrisM.Measuresofconcurrencyinnetworksandthespreadofinfectiousdisease.MathBiosci1996;133:165–195.29.GhaniA,GarnettG.Risksofacquiringandtransmittingsexuallytransmitteddiseasesinsexualpartnernetworks.SexTransmDis2000;27:579–587.30.BauchCT.AversatileODEapproximationtoanetworkmodelforthespreadofsexuallytransmitteddiseases.JMathBiol2002;45:375–395.SexuallyTransmittedDiseasesJanuary2005GORBACH