randomlyaccept

ORIGINALPAPERAge-RelatedDifferencesinSocio-demographicandBehavioralDeterminantsofHIVTestingandCounselinginHPTN043/NIMHProjectAcceptN.
Salazar-Austin1M.
Kulich2A.
Chingono3S.
Chariyalertsak4K.
Srithanaviboonchai4G.
Gray5,6L.
Richter6H.
vanRooyen7S.
Morin8M.
Sweat9J.
Mbwambo10G.
Szekeres11T.
Coates12D.
Celentano13TheNIMHProjectAccept(HPTN043)StudyTeamPublishedonline:6June2017TheAuthor(s)2017.
ThisarticleisanopenaccesspublicationAbstractYouthrepresentalargeproportionofnewHIVinfectionsworldwide,yettheirutilizationofHIVtestingandcounseling(HTC)remainslow.
Usingthepost-inter-vention,cross-sectional,population-basedhouseholdsur-veydonein2011aspartofHPTN043/NIMHProjectAccept,acluster-randomizedtrialofcommunitymobi-lizationandmobileHTCinSouthAfrica(SowetoandKwaZuluNatal),Zimbabwe,TanzaniaandThailand,weevaluatedage-relateddifferencesamongsocio-demo-graphicandbehavioraldeterminantsofHTCinstudyparticipantsbystudyarm,site,andgender.
Amultivariatelogisticregressionmodelwasdevelopedusingcompleteindividualdatafrom13,755participantswithrecentHIVtesting(prior12months)astheoutcome.
Youth(18–24years)wasnotpredictiveofrecentHTC,exceptforhigh-riskyouthwithmultipleconcurrentpartners,whowerelesslikely(aOR0.
75;95%CI0.
61–0.
92)tohaverecentlybeentestedthanyouthreportingasinglepartner.
Importantly,theinterventionwassuccessfulinreachingmenwithsitespecicsuccessrangingfromaOR1.
27(95%CI1.
05–1.
53)inSouthAfricatoaOR2.
30inThailand(95%CI1.
85–2.
84).
Finally,acrossadiverserangeofsettings,highereducation(aOR1.
67;95%CI1.
42,1.
96),highersocio-economicstatus(aOR1.
21;95%CI1.
08–1.
36),andmarriage(aOR1.
55;95%CI1.
37–1.
75)wereallpredictiveofrecentHTC,whichdidnotsigni-cantlyvaryacrossstudyarm,site,genderoragecategory(18–24vs.
25–32years).
ResumenLosjovenesrepresentanunagranproporciondenuevasinfeccionesporelVIHentodoelmundo,sinembargo,suutilizaciondelaspruebasdedetecciondelVIHyelasesoramiento(HTC)siguesiendobaja.
Utili-zandolaencuestadepoblaciontransversaldelapoblacionrealizadaen2011comopartedelproyectoHPTN043/ElectronicsupplementarymaterialTheonlineversionofthisarticle(doi:10.
1007/s10461-017-1807-5)containssupplementarymaterial,whichisavailabletoauthorizedusers.
&N.
Salazar-Austinnsalaza1@jhmi.
edu1DepartmentofPediatrics,JohnsHopkinsSchoolofMedicine,200N.
WolfeStRoom3147,Baltimore,MD21287,USA2FacultyofMathematicsandPhysics,DepartmentofProbabilityandStatistics,CharlesUniversity,Prague,CzechRepublic3UniversityofZimbabweCollegeofHealthSciences,Harare,Zimbabwe4DepartmentofCommunityMedicine,ChiangMaiUniversity,ChiangMai,Thailand5SouthAfricanMedicalResearchCouncil,CapeTown,SouthAfrica6UniversityoftheWitwatersrand,Johannesburg,SouthAfrica7HumanSciencesResearchCouncil,Pretoria,SouthAfrica8DepartmentofMedicine,UniversityofCalifornia,SanFrancisco,USA9DivisionofGlobalandCommunityHealth,MedicalUniversityofSouthCarolina,Charleston,SC,USA10MuhimbiliUniversityofHealthandAlliedSciences,DaresSalaam,Tanzania11UCLACenterforWorldHealth,UniversityofCalifornia,LosAngeles,USA12DepartmentofMedicine,UniversityofCalifornia,LosAngeles,CA,USA13DepartmentofEpidemiology,JohnsHopkinsBloombergSchoolofPublicHealth,Baltimore,USA123AIDSBehav(2018)22:569–579https://doi.
org/10.
1007/s10461-017-1807-5NIMHProjectAccept,unensayoaleatorizadoporgruposdemovilizacioncomunitariayHTCmovilenSudafrica(SowetoyKwaZuluNatal),Zimbabwe,TanzaniayTai-landia,evaluamoslasdiferenciasrelacionadasconlaedadentrelosdeterminantessociodemogracosyconductualesdeHTCenlosparticipantesdelestudioporbrazodeestudio,sitioysexo.
Sedesarrollounmodelomultivariadoderegresionlogsticautilizandodatosindividualescom-pletosde13,755participantesconpruebasdeVIHrecientes(antesde12meses)comoresultado.
Losjovenes(18-24anos)noeranpredictivosdeHTCrecientes,exceptolosjovenesdealtoriesgoconmultiplesparejasconcu-rrentes,queeranmenosprobables(aOR0,75;ICdel95%:0,61-0,92)companero.
Esimportantedestacarquelaintervencionfueexitosaenhombresconexitoespeccoenelsitio,desdeaOR1,27(IC95%1,05-1,53)enSudafricahastaaOR2,30enTailandia(IC95%:1,85-2,84).
Porultimo,enunaampliagamadecontextos,laeducacionsuperior(aO1,67;ICdel95%:1,42;1,96);mayorestatussocioeconomico(aOR1,21;ICdel95%:1,08-1,36)ymatrimonio(aOR1,55;ICdel95%:1,37-1.
75)erantodospredictivosdeHTCrecientes,quenovariaronsignicati-vamenteentrebrazodeestudio,sitio,sexoocategoradeedad(18-24vs25-32anos).
KeywordsMobileHIVtestingandcounselingYouthHigh-risksexualbehaviorProjectacceptDeterminantsHTCIntroductionYouthage15–24yearsaccountfor42%ofnewHIVdiagnosesworldwide[1].
UptakeofHIVtestingandcounseling(HTC)bytheseyouthremainsinadequate[2,3].
Itisestimatedthatonly15%ofyoungwomenand10%ofyoungmeninsub-SaharanAfricaknowtheirHIVstatus[4].
HIVincidenceamong15–24yearoldsishigh,particularlyamongyoungwomen,where7500youngwomenareestimatedtoacquireHIVeachweek[5].
From2010through2015,UNAIDSestimatesonlya6%declineinHIVincidenceamongfemalesage15–24years[5].
FocusedHIVpreventiononthisagegroupremainsapri-oritygiventheboldtargetofreducingtheannualrateofnewHIVinfectionsamongadolescentandyoungwomentounder100,000by2020[5].
HTCistherststepintoboththeHIVpreventionandtreatmentcascade.
HTCnotonlyidentiesone'sHIVstatus,butcanalsopositivelyinuenceone'ssexualriskbehaviorandreducethelikelihoodoffutureHIVacquisi-tionorfurthertransmission[6,7].
Reportedsocio-demo-graphicfactorsassociatedwithHTCincludeolderage,marriage,highereducationalstatus,urbanresidence,highersocio-economicstatus(SES),andreportingasinglepartner[8,9].
Foryouthage15–24years,studieshaveshownpregnancyoreverhavingmadesomeonepregnant,urbanresidence,highereducation(formen),andahigherfre-quencyofclinicvisitsareallpredictorsofHTCamongSouthAfricanyouth.
Additionally,beingHIV(amongmen)orknowingsomeonewhohasdiedofAIDS(formen),possessingknowledgeofHIV,havinghadaparentaldiscussionregardingHIV,andparticipatinginHIVpre-ventionprogrammingarealsopredictorsofHIVtestinginyouth[10,11].
YouthalsoreportsignicantpsychologicalbarrierstoHIVtestingincludinglackofcommunitysup-portandperceivednegativeattitudesofhealthcareworkers[12].
FurtherunderstandingthefactorsthatleadyouthtoundergoHIVtestingiscriticaltocreatingfocusedstrate-giestoincreaseHTCuptakeamongstthisat-riskpopulation.
NationalAIDSprogramshavetriedtomotivatehigh-riskyouthtoregularlyundergoHTC,buthavenotalwaysbeensuccessful,especiallyamongadolescentsandyoungadults[5].
Programshavereliedontraditionalfacility-basedHTCwhichhasevolvedtoincludebothprovider-initiatedtestingandroutine,oropt-out,testing.
Toreachthosecommunitymemberswhodonotregularlyaccesshealthcare,community-basedapproachessuchasmobileHTC,home-basedHTC,andself-testinghavebeendeveloped[13–19].
Foryouth,home-basedandself-testingstrategiesmayprovideincreasedcondentiality.
Nomatterwhatmethodisused,trulysupportiveservicesthatprovidenon-judgmental,empatheticcounselingservicesiscriticalforyouthbuy-inandtheircontinuedaccessofbothHIVpreventionandtreatmentservices[20].
Whileindividuallytheseapproachesholdpromise[16,21],acombinationofapproacheswilllikelybenecessarytoattainuniversalHTCcoveragetoreach90%ofthoselivingwithHIV,thegoalsetbyUNAIDSfor2020[22].
Eachofthesemethodsarebeingusedtotargetyouthage15–24years,buthavenotbeenevaluatedtodeterminetheiroptimaluseamongthisvulnerableandimportantagegroup[20].
FurtherstudiesareneededtodemonstrateacceptabilityandimprovementinHTCuptakeforyouth[20,21].
Traditionally,youngmenhavepoorlysoughtouthealthservices.
Voluntarymedicalmalecircumcisionprogramshavesuccessfullyreachedover11millionadolescentboysinAfricasince2008[5].
UNAIDSaimstousethisplatformtoprovideover90%ofmenage10–29yearswithcus-tomized,age-appropriatehealthservicesby2021[5].
ProjectAcceptwasacommunity-levelclusterrandom-izedtrialofamultilevelstructuralHIVpreventioninter-ventionwithmobileHTCconductedfrom2007to2010acrossavarietyofcommunitiesinThailand,Zimbabwe,Tanzania,andbothurban(Soweto)andrural(Vulindlela)SouthAfrica[23,24].
Theprimaryoutcome,community-570AIDSBehav(2018)22:569–579123levelHIVincidence,wascomparedbetweencommunitiesrandomizedtocommunity-basedvoluntarycounselingandtesting(CBVCT)withmobiletestingandstudy-supportedstigma-reducinginterventionsversustraditionalfacility-basedorstandardvoluntarycounselingandtesting(SVCT).
HIVprevalencevariedamongstudysitesfrom\1%inThailandto31%inKwaZulu-Natal,SouthAfrica.
AninterimevaluationofHTCin2009,usingHTCserviceutilizationdataavailablefromTanzania,ZimbabweandThailand,showed28%ofHTCclientsinCBVCTcom-munitieswerereceivingrepeatHTCandthatacrossthreecommunitypairs,HTCuptakewas40%higheramongclinicclientsinCBVCTcommunities[25].
ThisstudyutilizedlimiteddatacollectedonthesubjectswhousedHTCservicesandcouldthereforenotfullyevaluatesocio-demographicandbehavioraldeterminantsofHTC.
Usingthemoredetailedpost-interventioncross-sectionalsurveydata,theprimaryanalysisshoweda25%increaseinrecentHTC(overtheprior12months)inCBVCTversusSVCTcommunities.
Thisincreasewasmoreprofoundformen(45%increase)thanwomen(15%increase)[13].
ThediversityofProjectAcceptsitesprovidesauniqueoppor-tunitytoevaluatesocio-demographicandbehavioraldeterminantsofHTCacrossabroadspectrumofHIVepidemicsspanningtwocontinents,fourcountries,andruralandurbancommunities,toevaluatewhichsubgroupsmobileVCTmaybemosteffectiveattargeting.
ThisanalysisaimstodetermineagerelateddifferencesinHTCuptakeamongHPTN043ProjectAcceptpost-interventionsurveyparticipants,comparingyouthage18–24yearstothose25–32years,bysite,studyarmandgender,toinformfuturestrategiestoimproveHTCuptakeamongyouth.
MethodsProjectAcceptwasacommunity-levelcluster-randomizedtrialofcommunitymobilization,mobileHTC,andpost-testsupportservicesaimedatreducingcommunity-wideHIVincidenceandHIV-relatedstigmaconductedduring2007–2010in48communitiesinThailand,Zimbabwe,Tanzania,andtwositesinSouthAfrica(Vulindlela,KwaZuluNatalandSoweto,Gauteng).
Studyoutcomeswereassessedusingacross-sectional,population-based,post-interventionhouseholdsurveythatwasconductedfrom2009to2011.
Methodologyisdescribedindetailelsewhere[23,24].
Briey,menandwomenages18–32wererecruitedacrossall48communities,regardlessofparticipationinProjectAcceptactivitiesorparticipationinthebaselinesurvey.
Usingacompletelistingofcommunityhouseholds,householdswererandomlyselectedandvisitedbyinterviewteamsuntiltheyattainedthepre-speciedsamplesizetoassesstheprimaryoutcomeofcommunity-levelHIVincidence[24].
Afterpermissionwasobtainedfromtheheadofthehousehold,eligiblehouseholdmem-berswerethenlisted,andonewasrandomlyselectedforparticipationinadetailedsocio-demographicandbehav-ioralassessmentusingtheKishgridmethodandconsentedforparticipation.
AllsurveysandparticipantconsentswereapprovedbyallinvolvedUSinstitutionsandlocalethicscommittees.
Thissub-studywasconsideredIRB-exemptbytheJohnsHopkinsSchoolofPublicHealthInstitutionalReviewBoard.
MeasuresSurveyquestionsweredesignedcollaborativelywithallsites.
HIVtestingwasevaluatedrstbyhavingeverbeentested,secondbythetimingofthattesting([3years,1–3yearsor\1year)andnallybythefrequencyoftesting(onceversusrepeated).
RecentHIVtesting,denedasHTCovertheprior12months,wasusedastheoutcometoassessforboththeintervention'seffectandtheneedforrecurrentHTCamonghigh-risksub-groups.
Socio-demographicfactorsincludedage,gender,edu-cation,maritalstatus,SES,andemployment.
Maritalstatuswasclassiedascurrentlymarriedorunmarried.
SESwasassessedusingsite-specic,localdenitionscompositingincomeandhouseholdassetsandclassiedaslow,mediumorhigh.
Behavioralfactorsincludedsexualactivity,num-berofpartners,frequencyofsexualactivity,andfrequencyofcondomuse.
Recentsexualriskbehaviorwasassessedduringthe6monthspriortosurveyparticipation.
StatisticalAnalysesOutoftheoriginal14,291post-interventionsurveypar-ticipantswhocompletedthedetailedsocio-demographicandbehavioralinterview,536subjectswithincompletedatawereexcluded.
Allanalyseswereperformedontheremaining13,755participantswithcompletedata.
Multi-variatelogisticregressionmodelswereusedwithrecenttesting(denedasHIVtestingintheprior12months)astheoutcome.
Signicanceofpredictorswasassessedbylikelihoodratiotestsatthe0.
05level.
First,abasemodelincludingsite,gender,intervention,agecategory(18–24vs.
25–32years),andtheirsignicantinteractionswasbuilt.
Next,individualsocio-demographicandbehavioralfactorswereaddedonebyonetothebasemodel,includingitsinteractionswithgender,age,intervention,andsite.
Thenalmodelwasobtainedbysimultaneouslyaddingallsignicantfactorsandinteractionsfromthesesmallermodelstothebasemodelandremovingallinsignicantterms.
Twoversionsofthenalmodelweretted:unad-justedforcommunityeffects(59parameters)andadjustedforcommunityeffects(97parameters).
AdjustmentforAIDSBehav(2018)22:569–579571123communityeffectswasdonebycontraststhatsummedtozerowithineachsite-by-interventioncombinationsothatoverallsiteandinterventioneffectswouldnotbeaffectedbyadjustmentforcommunity.
CondenceintervalsarebasedonWaldtests.
TheanalysiswasperformedintheRsoftwareenvironment.
ResultsStudyPopulationTherewere13,755participantsage18–32yearswhocompletedProjectAccept'sdetailedsocio-demographicandbehavioralpost-interventionsurveyandhadcompletedatainallvariablesconsideredforanalysis(Table1).
Amongallsurveyparticipants,34.
6%reportedatleastonerecentHIVtestinCBVCTcommunitiesand29.
3%inSVCTcommunities(Table2).
Amongyouthparticipantsage18–24years,31.
8%reportedrecentHIVtestinginCBVCTcommunitiesand26.
9%inSVCTcommunities(Table3).
Only22.
0and16.
1%ofyoungmen(18–24years)reportedrecenttestinginCBVCTandSVCTcommunities,respectively.
Testingratesinyoungwomen(18–24years)wereatleasttwiceaslarge(Table3).
Socio-DemographicandBehavioralCharacteristicsofRecentTestingAhigherproportionofyouth(18–24years)inCBVCTcommunitiesunderwentrecentHIVtestingindependentofsite,gender,andallmeasuredsocio-demographicandbehavioralcovariates,(Table4),includinglow-riskgroupssuchasthosereportingneverhavinghadsex.
HigherTable1Studypopulation:ParticipantcharacteristicsofHPTN043projectacceptpost-interventioncross-sectionalcommunitysurveySiteThailandZimbabweTanzaniaKwaZulu-NatalSowetoAllsitesInterventionCBVCTSVCTCBVCTSVCTCBVCTSVCTCBVCTSVCTCBVCTSVCTCBVCTSVCTn1505(%)1569(%)1248(%)1243(%)1386(%)1379(%)1293(%)1234(%)1439(%)1459(%)6871(%)6884(%)GenderMale49.
151.
143.
746.
241.
943.
142.
541.
245.
244.
744.
645.
5Female50.
948.
956.
353.
858.
156.
957.
558.
854.
855.
355.
454.
5Agegroup18–2445.
648.
148.
149.
141.
939.
058.
563.
452.
554.
649.
250.
625–3254.
451.
951.
950.
958.
161.
041.
536.
647.
545.
450.
849.
4Education(years)0–533.
519.
53.
53.
523.
424.
62.
62.
60.
51.
013.
310.
76–937.
741.
440.
139.
362.
165.
89.
117.
36.
55.
331.
233.
910–1220.
726.
351.
450.
113.
68.
685.
076.
270.
068.
147.
344.
913ormore8.
112.
84.
97.
10.
91.
03.
34.
023.
025.
78.
310.
6SESgroupLow37.
927.
923.
828.
631.
235.
817.
522.
15.
84.
723.
423.
6Medium30.
231.
929.
325.
046.
645.
565.
061.
962.
165.
946.
646.
0High32.
040.
246.
946.
422.
218.
817.
516.
032.
129.
330.
030.
4EmploymentYes88.
785.
757.
063.
463.
064.
839.
638.
858.
258.
462.
163.
3No11.
314.
343.
036.
637.
035.
260.
461.
241.
841.
637.
936.
7MaritalstatusMarried55.
749.
253.
154.
751.
253.
42.
62.
08.
87.
334.
533.
7Unmarried44.
350.
846.
945.
348.
846.
697.
498.
091.
292.
765.
566.
3EverhadsexYes87.
484.
085.
787.
590.
592.
887.
685.
291.
491.
688.
688.
2No12.
616.
014.
312.
59.
57.
212.
414.
88.
68.
411.
411.
8nnumberofsubjectsparticipatinginthepost-interventionsurvey,CBVCTcommunity-basedvoluntarycounselingandtesting,SVCTstandardvoluntarycounselingandtesting572AIDSBehav(2018)22:569–579123percentagesofrecentHCTwereobservedamongpartici-pantswhoweresexuallyactive,hadatleast10yearsofeducation,orlivedinhouseholdswithmediumorhighSES.
Socio-demographicPredictorsofRecentTestingTheintervention'seffectonrecentHIVtestingwashighlysignicantandvariedbybothsiteandgender,butnotage(Table5).
ThestrongesteffectwasseeninThailandformen(aOR2.
30;95%CI1.
85–2.
84)and,toalessereffect,innon-pregnantwomen(aOR1.
92;95%CI1.
56–2.
36).
Theinter-ventionhadaminimaleffectonHIVtestingamongnon-pregnantwomeninTanzania(aOR1.
09;95%CI0.
91–1.
30),KwaZuluNatal,SouthAfrica(aOR1.
12;95%CI0.
91–1.
34)andSoweto,SouthAfrica(aOR1.
06;95%CI0.
89–1.
26).
Overall,interventioneffectswereevenweakeramongwomenreportingcurrentpregnancythanamongnon-preg-nantwomen(aOR0.
56,95%CI0.
39–0.
80).
Agewasnotdirectlyassociatedwithrecenttestingwhensocio-demographicandbehavioralfactorsweretakenintoaccount.
Participantswhowerenotmarried,includingTable2HIVtestinghistory:reportedHIVtestinghistoryinHPTN043projectacceptpost-interventionsurveySiteThailandZimbabweTanzaniaKwaZulu-NatalSowetoAllsitesInterventionCBVCTSVCTCBVCTSVCTCBVCTSVCTCBVCTSVCTCBVCTSVCTCBVCTSVCTn1505(%)1569(%)1248(%)1243(%)1386(%)1379(%)1293(%)1234(%)1439(%)1459(%)6871(%)6884(%)Notest36.
555.
150.
857.
233.
138.
440.
446.
838.
239.
339.
547.
3Testmorethan3yearsago19.
216.
89.
19.
720.
619.
612.
813.
013.
114.
315.
214.
8Test1–3yearsago20.
113.
67.
86.
810.
08.
36.
34.
68.
18.
210.
78.
6Onceinpastyear20.
912.
019.
919.
119.
521.
520.
318.
620.
721.
920.
318.
5Repeatedinpastyear3.
32.
512.
57.
316.
812.
320.
016.
919.
916.
314.
310.
8nnumberofparticipantswithknowntestingstatus,CBVCTcommunity-basedvoluntarycounselingandtesting,SVCTstandardvoluntarycounselingandtestingTable3Percenttestinginthelast12monthsbyageandgenderSiteThailandZimbabweTanzaniaKwaZulu-NatalSowetoAllsitesInterventionCBVCTSVCTCBVCTSVCTCBVCTSVCTCBVCTSVCTCBVCTSVCTCBVCTSVCTn1505(%)1569(%)1248(%)1243(%)1386(%)1379(%)1293(%)1234(%)1439(%)1459(%)6871(%)6884(%)GenderMale19.
610.
124.
614.
825.
319.
031.
822.
024.
025.
624.
717.
8Female28.
619.
038.
436.
344.
245.
046.
745.
054.
248.
542.
638.
9Agegroup18–2423.
312.
328.
023.
330.
530.
138.
333.
937.
034.
331.
826.
925–3224.
916.
436.
429.
440.
536.
143.
338.
544.
543.
137.
331.
8GenderandageMen18–24years19.
09.
219.
412.
122.
919.
330.
020.
218.
821.
122.
016.
1Men25–32years20.
211.
131.
618.
227.
518.
834.
826.
130.
331.
827.
819.
9Women18–24years28.
016.
137.
435.
637.
341.
345.
145.
053.
246.
041.
137.
5Women25–32years29.
121.
439.
136.
848.
546.
848.
845.
155.
351.
243.
840.
1nnumberofparticipantswithknownage,gender,andtestingstatus,CBVCTcommunity-basedvoluntarycounselingandtesting,SVCTstandardvoluntarycounselingandtestingAIDSBehav(2018)22:569–579573123Table4Percenttestinginthelast12monthsbyexplanatoryfactorsGender,ageMen18–24yearsMen25–32yearsWomen18–24yearsWomen25–32yearsInterventionCBVCTSVCTCBVCTSVCTCBVCTSVCTCBVCTSVCTn1643(%)1734(%)1422(%)1396(%)1738(%)1747(%)2068(%)2007(%)Yearsofeducation0–517.
711.
820.
213.
029.
228.
934.
929.
16–918.
211.
127.
715.
535.
735.
041.
739.
310–1224.
318.
430.
626.
544.
839.
749.
744.
913ormore24.
023.
230.
422.
846.
638.
246.
243.
4SESgroupLow18.
111.
523.
014.
439.
233.
138.
436.
5Medium23.
217.
828.
022.
042.
041.
145.
439.
3High22.
616.
831.
620.
940.
934.
546.
344.
8Employment(hasincomefromwork)Yes23.
516.
227.
819.
141.
536.
843.
640.
1No20.
016.
028.
024.
640.
738.
144.
340.
2MaritalstatusMarried33.
526.
227.
817.
640.
138.
341.
538.
4Unmarried20.
715.
027.
921.
541.
537.
246.
642.
2EverhadsexYes24.
218.
128.
420.
246.
542.
744.
240.
6No14.
59.
514.
813.
817.
113.
722.
921.
3Numberofpartnersinthelast6months016.
911.
423.
012.
332.
530.
848.
344.
0129.
522.
328.
122.
046.
341.
342.
439.
0219.
717.
930.
422.
946.
941.
253.
348.
73ormore23.
517.
936.
325.
960.
063.
655.
633.
3nnumberofparticipantsofgivengenderandagerangewhoreportedtestingstatus,CBVCTcommunity-basedvoluntarycounselingandtesting,SVCTstandardvoluntarycounselingandtestingTable5InterventioneffectonrecentHIVtesting(past12months)bygenderandsite:resultsofmultivariatelogisticregressionmodelSiteThailandZimbabweTanzaniaSouthAfricaKwaZulu-NatalSouthAfricaSoweton=13,755n=3074n=2491n=2765n=2527n=2898GenderaOR[95%CI]paOR[95%CI]paOR[95%CI]paOR[95%CI]paOR[95%CI]pMales2.
30[1.
85–2.
84]p\0.
0011.
60[1.
30–1.
97]p\0.
0011.
31[1.
07–1.
59]p=0.
0071.
34[1.
10–1.
64]p=0.
0041.
27[1.
05–1.
53]p=0.
012Females(non-pregnant)1.
92[1.
56–2.
36]p\0.
0011.
33[1.
10–1.
62]p=0.
0041.
09[0.
91–1.
30]p=0.
3371.
12[0.
93–1.
34]p=0.
2181.
06[0.
89–1.
26]p=0.
506Adjustedoddsratio(aOR)forrecentHIVtestinginCBVCTcommunitiesrelativetoSVCTcommunitiesbysiteandgenderwith95%condenceintervals[95%CI],andp-values(p)fornointerventioneffectOverallinterventioneffectsonrecentHIVtestingwerepreviouslyreported[13]aORswereadjustedformaritalstatus,education,employment,SES,sexualactivityandcondomuse574AIDSBehav(2018)22:569–579123thosewhoreportedbeingsingle,divorced,separatedorwidowed,werelesslikely(aOR0.
64,95%CI0.
57–0.
73)tohavebeenrecentlytestedthanparticipantswhoreportedbeingcurrentlymarried.
Moreeducatedparticipants,par-ticularlythosewhohadattendedatleast10yearsofschooling(aOR1.
67,95%CI1.
42–1.
96relativetoamaximumof5yearsofschooling),thosewhoearnedmoneyfromemployment(aOR1.
16,95%CI1.
06–1.
27),andthosewiththehighestsite-specicSES(aOR1.
21,95%CI1.
08–1.
36comparedtolowestSES)wereallmorelikelytohavetestedintheprior12months(Table6).
Importantly,theeffectsofsocio-demographicfactorsonrecentHIVtestingdidnotvarysignicantlybetweenthevesites,betweeninterventionandcontrolcommunities,betweengenders,orbetweenagecategories(18–24vs.
25–32years).
BehavioralPredictorsofRecentHIVTestingReportingmultiplepartnersinthepast6monthswastheonlyfactorthathadaneffectontestingratesthatvariedaccordingtoage.
Youth,age18–24years,withmultiplepartnerswerelesslikelytotest(aOR0.
75,95%CI0.
61–0.
92)thanyouthreportingasinglepartner.
However,olderparticipants,age25–32years,withmultiplepartnerswerenotlesslikelytotest(aOR1.
12,95%CI0.
91–1.
36)comparedtoolderparticipantswithasinglepartner(Table7).
Menwhowerenotsexuallyactiveinthe6monthspriortotheinterviewwerelesslikelytoreportrecenttesting(aOR0.
56,95%CI0.
45–0.
69).
Insexuallyactivewomen,testingrateswerenotassociatedwithpatternsofsexualactivity(aOR1.
03,95%CI0.
86–1.
25)(Table7).
Insexuallyactiveparticipants,higherratesofcondomuseweregenerallyassociatedwithhigherratesofrecenttesting,buttheeffectwassomewhatdifferentinmenthaninwomen.
Maleparticipantswhousecondomsabouthalfofthetimewerelesslikely(aOR0.
62;96%CI0.
47–0.
82)tohavebeenrecentlyHIVtestedcomparedtoregularmalecondomusers.
Femaleparticipantswhousedcondomsabouthalfthetimewerenotlesslikely(aOR1.
07;95%CI0.
84–1.
36)tohaverecentlybeenHIVtestedcomparedtoregularfemalecondomusers.
Bothmale(aOR0.
81;95%CI0.
66–0.
99)andfemale(aOR0.
71;95%CI0.
59–0.
84)participantswhoneverusedcondomswerelesslikelytohavetestedforHIVcomparedtoregularcondomusers.
Thereportedresultswereadjustedforsite,butnotcom-munity-leveleffects.
Whenxedcommunityeffectswereaddedtothemodel,theywerehighlysignicant,buttheresultsforallotherpredictorswereverysimilar(supple-mentaltable).
DiscussionThissub-analysisofProjectAcceptdatademonstratescommunitymobilization,stigmareductionandmobileHTCaresuccessfulatreachingmanyat-riskdemographicsincludingyouthage18–24yearsandmen.
NearlyoneinTable6Socio-demographicpredictorsofrecentHIVtesting(past12months):resultsofmultivariatelogisticregressionmodelSocio-demographicfactorAllparticipants(n=13,755)aOR95%CIpYearsofeducationAcceptincludesdatafrom48pairedcommunities,across5sites,4countriesand2continents,reectingthediversityoftheHIVepidemicandthedrivingforcesbehindtheirlocalepidemics.
ThishelpstoexplainthevariableinterventioneffectacrossdifferentsitesandTable7BehavioralpredictorsofrecentHIVtesting(past12months):resultsofmultivariatelogisticregressionmodelBehavioralfactorAge18–24(n=6862)Age25–32(n=6893)aOR95%CIpaOR95%CIpNumberofpartnersinthelastsixmonths0.
0070.
291partner(baselinelevel)1.
00––1.
00––multiplepartners0.
750.
61–0.
920.
0071.
120.
91–1.
360.
29Men(n=6195)Women(n=7560)Everhadsexacceptingatmosphereiscritical.
Condentialitycanbeaddressedthroughself-testing.
Acomfortabletestingenvironmentoutsideofthefacilityincludingatsocialcenters,orprovidinghome-basedandmobileHTCatyouth-targetedcommunityeventsareimportant[20].
Givenmultiplelayersofclusteringwithinsite,com-munitypairs,villagesandhouseholds,wewereunabletomodelcorrelationsbetweenindividualsusinglogisticregressionwithrandomeffects.
Themostimportantlevelofclustering(communities)wasaddedtothemodelasxedeffects.
Thislimitsgeneralizabilitytoothercommu-nitiesorevenotherindividualsinthesecommunities.
Thoughestimatedparametersshouldremainsimilar,thecalculated95%condenceintervalsmaybetoonarrow.
However,thestrongesthighlysignicanteffectsinourmodelshouldnotbeaffectedbythislimitation.
Inourmodel,lackofHIVtestingamongyouth18–24yearswasattributabletoothersocio-demographicandbehavioralcharacteristicsknowntobeassociatedwithlowerHIVtestingratesinallagegroups.
Importantly,community-basedmobiletestingdidimproveHTCuptakeinyouthincludingmen,whotraditionallydemonstratepoorutilizationoffacility-basedHTC.
YouthwithhighersexualriskwerelesslikelytohaveobtainedrecentHIVtesting.
Incorporatingthisknowledgeintoyouth-friendlyHIVpreventionmessaging,counselingandservicesisimportant.
ConclusionsMobileHTCwassuccessfulinreachingyouth,age18–24years,animportantat-riskpopulation.
Thiswasparticularlytrueforyoungmen.
Youth18–24yearswithhigh-risksexualbehavior,includingmultipleconcurrentpartners,accessedHTClesscommonlyinallcommunitiesincludingthosewithmobileandfacility-basedHTC.
AsHTCremainsanecessarygatewaytobothtreatmentandpreventionservices;improvingthequalityofyouth-basedHTCservicesmaynotonlyimproveaccessforthemajorityofyouth,butalsoyouthwithhigh-riskbehavior.
AcknowledgementsThissub-studywasfundedbytheHPTNscho-larprogram;salarysupportforprimaryauthorwassupportedbyT32AI052071.
PrimarysupportforProjectAcceptwasthroughNIMH(U01MH066687).
TheNIMHProjectAccept(HPTN043)StudyTeamHumanSciencesResearchCouncil,SouthAfrica(SalimAbdoolKarim,JanetFrohlich,PhilipJoseph,ThulaniNgubani,LindaRichter,HeidivanRooyen);UniversityofNorthCarolinaatChapelHill,NC,USA(LaurieAbler,SuzanneMaman,AudreyPettifor);MuhimbiliUniversityofHealthandAlliedSciences,Tanzania(ChristopherBamanyisa,LillianneChovenye,G.
P.
Kilonzo,NoraMargaretHogan,FlorenceP.
Lema,JessieK.
K.
Mbwambo,KhalifaM.
Mrumbi);JohnsHopkinsBloombergSchoolofPublicHealth,MD,USA(ChrisBeyrer,DavidD.
Celentano,BeckyGenberg,SurindaKawichai,BenjaminLink,CarlaE.
Zelaya);UniversityofCalifornia,SanFrancisco,CA,USA(AdamW.
Carrico,SebastianKevany,GertrudeKhumalo-Sakutukwa,TimLane,JoanneMickalian,SimonMort,StephenF.
Morin,WayneSteward);ChiangMaiUniversity,ResearchInstituteforHealthSciences,Thailand(ChonlisaChariyalertsak,SuwatChariyalertsak,SurindaKawichai,KriengkraiSrithanaviboonchai,SurasingVisrutaratna);UniversityofZimbabwe,Zimbabwe(AlfredChingono,TendayiJubenkanda,MemorySendah,TserayiMachinda,OliverMurima,AndrewTimbe,GodfreyWoelk);UniversityofCalifornia,LosAngeles,CA,USA(ThomasJ.
Coates,Agne`sFiamma,GregSzekeres);MedicalUniversityofSouthCarolina,SC,USA(KathrynCurran,AndrewM.
Sadowski,MichaelSweat,BasantSingh,MartaI.
Mulawa);FredHutchinsonCancerResearchCenter,StatisticalCenterforHIV/AIDSResearch&Prevention,WA,USA(DeborahDonnell);JohnsHopkinsUniversitySchoolofMedicine,MD,USA(SusanH.
Eshleman,LeTanyaJohnson-Lewis,OliverLaeyendecker,EstellePiwowar-Manning);InternationalCenterforResearchonWomen,USA(KatherineFritz,AmyGregowski);UniversityoftheWitwatersrand,ChrisHaniBaragwanathHospital,SouthAfrica(GlendaGray,SakhileMhlongo,PreciousModiba,GavinRobertson);TB/HIVCareAssociation,SouthAfrica(HarryHausler);CharlesUniversity,FacultyofMath-ematicsandPhysics,CzechRepublic(ZdenekHlavka,DanielHlubinka,MichalKulich);NationalInstituteofAllergyandInfec-tiousDiseases,USA(OliverLaeyendecker);UniversityofSouthampton,UK(NualaMcGrath);AnovaHealthInstitute,SouthAfrica(JamesMcIntyre).
CompliancewithEthicalStandardsConictofinterestAuthorsdeclarethattheyhavenoconictofinterest.
EthicalApprovalAllsurveysandparticipantconsentswereapprovedbyallinvolvedUSinstitutionsandbylocalethicscom-mitteesincludingtheJohnsHopkinsUniversityCommitteeonHumanResearch(Thailand),theChiangMaiUniversityResearchInstituteforHealthSciences(Thailand),theMinistryofPublicHealth(Thai-land),theMedicalUniversityofSouthCarolinaInstitutionalReviewBoardforHumanResearch(Tanzania),theinstitutionalreviewboardofMuhimbiliUniversityofHealthandAlliedSciences(Tanzania),theinstitutionalreviewboardoftheNationalInstituteofMedicalResearch(Tanzania),theUniversityofCaliforniaSanFranciscoCommitteeonHumanResearch(Zimbabwe),theMedicalResearchCouncilofZimbabwe(Zimbabwe),theUniversityoftheWitwater-srand,JohannesburgHumanResearchEthicsCommittee(Soweto,AIDSBehav(2018)22:569–579577123SouthAfrica)andtheHumanScienceResearchCouncilResearchEthicsCommittee(Vulindlela,SouthAfrica).
Thissub-studywasconsideredexemptfromIRBreviewbytheJohnsHopkinsSchoolofPublicHealthInstitutionalReviewBoard.
OpenAccessThisarticleisdistributedunderthetermsoftheCreativeCommonsAttribution4.
0InternationalLicense(http://creativecommons.
org/licenses/by/4.
0/),whichpermitsunrestricteduse,distribution,andreproductioninanymedium,providedyougiveappropriatecredittotheoriginalauthor(s)andthesource,providealinktotheCreativeCommonslicense,andindicateifchangesweremade.
References1.
UNAIDS.
FactSheet:Adolescents,YoungPeopleandHIV.
2012.
Availablefrom:http://les.
unaids.
org/en/media/unaids/contentassets/documents/factsheet/2012/20120417_FS_adolescentsyoungpeoplehiv_en.
pdf.
2.
PettiforA,ReesH,SteffensonA,Hlongwa-MadikizelaL,Mac-PhailC,VermaakK,etal.
HIVandsexualbehaviouramongyoungSouthAfricans:anationalsurveyof15-24yearolds.
Johannesburg,SouthAfrica:ReproductiveHealthResearchUnit,UniversityoftheWitwatersrand;2004.
3.
ShisanaO,RehleT,SimbayiL,ParkerW,ZumaK,BhanaA,etal.
SouthAfricanNationalHIVprevalence,HIVincidence,behaviourandcommunicationsurvey,2005.
CapeTown,SouthAfrica:HSRCPress;2005.
4.
WHO.
HIVandadolescents:guidanceforHIVtestingandcounsellingandcareofadolescentslivingwithHIV.
2013.
5.
UNAIDS.
GetontheFast-Track-Thelife-cycleapproachtoHIV.
2016.
Availablefrom:http://www.
unaids.
org/sites/default/les/media_asset/Get-on-the-Fast-Track_en.
pdf.
6.
CoatesTJ,KamengaMC,SweatMD,DeZoysaI.
EfcacyofvoluntaryHIV-1counsellingandtestinginindividualsandcou-plesinKenya,Tanzania,andTrinidad:arandomisedtrial.
ThevoluntaryHIV-1counselingandtestingefcacystudygroup.
Lancet.
2000;356(9224):103–12.
7.
FonnerVA,DenisonJ,KennedyCE,O'ReillyK,SweatM.
Voluntarycounselingandtesting(VCT)forchangingHIV-re-latedriskbehaviorindevelopingcountries.
CochraneDatabaseSystRev.
2012;9(9):CD001224.
8.
KaufmanMR,MasseyM,TsangSW,KamalaB,SerlemitsosE,LylesE,etal.
AnassessmentofHIVtestinginTanzaniatoinformfuturestrategiesandinterventions.
AIDSCare.
2015;27(2):213–7.
9.
CreminI,CauchemezS,GarnettGP,GregsonS.
PatternsofuptakeofHIVtestinginsub-SaharanAfricainthepre-treatmentera.
TropMedIntHealth.
2012;17(8):e26–37.
10.
MacPhailC,PettiforA,MoyoW,ReesH.
FactorsassociatedwithHIVtestingamongsexuallyactiveSouthAfricanyouthaged15-24years.
AIDSCare-PsycholSocio-MedAspAIDS/HIV.
2009;21(4):456–67.
11.
PeltzerK,MatsekeG.
DeterminantsofHIVtestingamongyoungpeopleaged18-24yearsinSouthAfrica.
AfrHealthSci.
2013;13(4):1012–20.
12.
MacPhailCL,PettiforA,CoatesT,ReesH.
''Youmustdothetesttoknowyourstatus'':attitudestoHIVvoluntarycounselingandtestingforadolescentsamongSouthAfricanyouthandpar-ents.
HealthEducBehav.
2008;35(1):87–104.
13.
CoatesTJ,KulichM,CelentanoDD,ZelayaCE,ChariyalertsakS,ChingonoA,etal.
Effectofcommunity-basedvoluntarycounsellingandtestingonHIVincidenceandsocialandbehaviouraloutcomes(NIMHProjectAccept;HPTN043):acluster-randomisedtrial.
LancetGlobalHealth.
2014;2(5):E267–77.
14.
GrabbeKL,MenziesN,TaegtmeyerM,EmukuleG,AngalaP,MwegaI,etal.
IncreasingaccesstoHIVcounselingandtestingthroughmobileservicesinKenya:strategies,utilization,andcost-effectiveness.
JAcquirImmuneDecSyndr.
2010;54(3):317–23.
15.
CorbettEL,DauyaE,MatamboR,CheungYB,MakamureB,BassettMT,etal.
UptakeofworkplaceHIVcounsellingandtesting:acluster-randomisedtrialinZimbabwe.
PLoSMed.
2006;3(7):e238.
16.
SabapathyK,VandenBerghR,FidlerS,HayesR,FordN.
Uptakeofhome-basedvoluntaryHIVtestinginsub-SaharanAfrica:asystematicreviewandmeta-analysis.
PLoSMed.
2012;9(12):e1001351.
17.
WachiraJ,KimaiyoS,NdegeS,MamlinJ,BraitsteinP.
Whatistheimpactofhome-basedHIVcounselingandtestingontheclinicalstatusofnewlyenrolledadultsinaalargeHIVcareprograminWesternKenyaClinInfectDis.
2012;54(2):275–81.
18.
WHO.
Planning,implementationandmonitoringhome-basedHIVtestingandcounselling.
ApracticalhandbookforSub-sa-haranAfrica.
2012.
19.
PantPaiN,SharmaJ,ShivkumarS,PillayS,VadnaisC,JosephL,etal.
Supervisedandunsupervisedself-testingforHIVinhigh-andlow-riskpopulaions:asystematicreview.
PLoSMed.
2013;10(4):e1001414.
20.
KurthAE,LallyMA,ChokoAT,InwaniIW,FortenberryJD.
HIVtestingandlinkagetoservicesforyouth.
JIntAIDSSoc.
2015;18(2Suppl1):19433.
21.
SutharAB,FordN,BachanasPJ,WongVJ,RajanJS,SaltzmanAK,etal.
TowardsuniversalvoluntaryHIVtestingandcoun-selling:asystematicreviewandmeta-analysisofcommunity-basedapproaches.
PLoSMed.
2013;10(8):e1001496.
22.
UNAIDS.
UNAIDS90-90-90,anambitioustreatmenttargettohelpendtheAIDSepidemic.
2014.
23.
Khumalo-SakutukwaG,MorinS,FritzK,CharleboisE,vanRooyenH,ChingonoA,etal.
Projectaccept(HPTN043):acommunity-basedinterventiontoreduceHIVincidenceinpop-ulationsatriskofHIVinSub-SaharanAfricaandThailand.
JAcquirImmuneDecSyndr.
2008;49(4):422–31.
24.
ProjectAcceptStudyGroup.
ProjectAccept(HPTN043),Acluster-randomizedtrialofcommunitymobilization,mobileHIVtesting,post-testsupportservices,andreal-timeperformancefeedback.
http://www.
cbvct.
med.
ucla.
edu/.
25.
SweatM,MorinS,CelentanoD,MulawaM,SinghB,MbwanboJ,etal.
Community-basedinterventiontoincreaseHIVtestingandcasedetectioninpeopleaged16–32yearsinTanzania,Zimbabwe,andThailand(NIMHProjectAccept,HPTN043):arandomizedstudy.
LancetInfectDis.
2011;11:525–32.
26.
SteinbergL.
Asocialneuroscienceperspectiveonadolescentrisk-taking.
DevRev.
2008;28(1):78–106.
27.
JohnsonS,BlumR,BieddJ.
Adolescentmaturityandthebrain:thepromiseandpitfallsofneuroscienceresearchinadolescenthealthpolicy.
JAdolescHealth.
2009;45(3):216–21.
28.
McGarrigleCA,MercerCH,FentonKA,CopasAJ,WellingsK,ErensB,etal.
InvestigatingtherelationshipbetweenHIVtestingandriskbehaviourinBritain:nationalsurveyofsexualattitudesandlifestyles2000.
Aids.
2005;19(1):77–84.
29.
SametJH,WinterMR,GrantL,HingsonR.
FactorsassociatedwithHIVtestingamongsexuallyactiveadolescents:aMas-sachusettssurvey.
Pediatrics.
1997;100(3):371–7.
30.
GoodmanE,TiptonAC,HechtL,ChesneyMA.
Perseverancepaysoff:healthcareproviders'impactonHIVtestingdecisionsbyadolescentfemales.
Pediatrics.
1994;94(6Pt1):878–82.
578AIDSBehav(2018)22:569–57912331.
ParkerL,MamanS,PettiforA,ChalachalaJL,EdmondsA,GolinCE,etal.
Barrierstoprovider-deliveredsexualbehaviorcounselingforyouthlivingwithHIV/AIDSinthedemocraticRepublicoftheCongo.
JHIVAIDSSocServ.
2013;12(3–4):3–4.
32.
TyleeA,HallerE,GrahamT,ChurchillR,SanciL.
Youth-friendlyprimary-careservices:howarewedoingandwhatmoreneedstobedoneLancet.
2007;369:1565–73.
33.
HelleringerS,KohlerHP,FrimpongJA,MkandawireJ.
IncreasinguptakeofHIVtestingandcounselingamongthepoorestinsub-saharancountriesthroughhome-basedserviceprovision.
JAcquirImmuneDecSyndr.
2009;51(2):185–93.
34.
MutaleW,MicheloC,JurgensenM,FylkesnesK.
Home-basedvoluntaryHIVcounsellingandtestingfoundhighlyaccept-ableandtoreduceinequalities.
BMCPublicHealth.
2010;10:347.
AIDSBehav(2018)22:569–579579123