National population size estimation of illicit drug users through the network scale-up method in 2013 in Iran

Research

Paper

National

population

size

estimation

of

illicit

drug

users

through

the

network

scale-up

method

in

2013

in

Iran

Ali

Nikfarjam

,

Mostafa

Shokoohi

,

Armita

Shahesmaeili

*

,

Ali

Akbar

Haghdoost

,

Mohammad

Reza

Baneshi

_,

_Saiedeh

_{Haji-Maghsoudi}

_,

_Azam

_Rastegari

_,

Abbas

Ali

Nasehi

,

Nadereh

Memaryan

,

Termeh

Tarjoman

a

SupervisorResearchandDevelopmentEmergencyMedicalCenter,MinistryofHealth,Tehran,Iran

b

RegionalKnowledgeHub,andWHOCollaboratingCentreforHIVSurveillance,InstituteforFuturesStudiesinHealth,KermanUniversityofMedicalSciences, Kerman,Iran

c

Epidemiology&Biostatistics,SchulichSchoolofMedicine&Dentistry,TheUniversityofWesternOntario,London,Canada

d

ResearchCenterforModelinginHealth,InstituteforFuturesStudiesinHealth,BiostatisticsandEpidemiologyDepartment,HealthSchool, KermanUniversityofMedicalSciences,Kerman,Iran

e_{IranHelalInstituteofApplied-Science&Technology,Tehran,Iran} f

SchoolofBehavioralSciencesandMentalHealth,IranUniversityofMedicalSciences,Tehran,Iran

g

CommunityMedicineSpecialist,MinistryofHealth,Tehran,Iran

Introduction

Iran’s geographical location, especially its long border with Afghanistan(whichisthemaingrowerofopiumintheworld)and its proximity with Pakistan, has turnedit into a major transit

ARTICLE INFO Articlehistory: Received20April2015

Receivedinrevisedform18January2016 Accepted21January2016 Keywords: Druguse Iran Networkscaleup Prevalence Sizeestimation ABSTRACT

Background: ForabetterunderstandingofthecurrentsituationofdruguseinIran,weutilizedthe networkscale-upapproachtoestimatetheprevalenceofillicitdruguseintheentirecountry. Methods:Weimplementedaself-administered,street-basedquestionnaireto7535passersbyfromthe generalpublicover18yearsofagebystreetbasedrandomwalkquotasampling(basedongender,age and socio-economic status) from 31 provinces in Iran. The sample size in each province was approximately400,rangingfrom200to1000.Ineachprovince75%ofsamplewasrecruitedfromthe capitaland theremaining25%was recruitedfromoneofthelargecitiesofthatprovincethrough stratifiedsampling. Thequestionnairecomprisedquestionsondemographicinformationaswellas questionstomeasurethetotalnetworksizeofparticipantsaswellasthenetworksizeineachofseven drugusegroups includingOpium,Shire (combinationof Opiumresidue andpureopium), Crystal Methamphetamine,heroin/crack(whichinIraniancontextisacocaine-freedrugthatmostlycontains heroin,codeine,morphineandcaffeinewithorwithoutotherdrugs),Hashish,Methamphetamine/LSD/ ecstasy,andinjectingdrugs.Theestimatedsizeforeachgroupwasadjustedfortransmissionandbarrier ratios.

Results:Themostcommontypeofillicitdrugusedwasopiumwiththeprevalenceof1500per100,000 populationfollowedbyshire(660),crystalmethamphetamine(590),hashish(470),heroin/crack(350), methamphetamine,LSDandecstasy(300)andinjectingdrugs(280).Alltypesofsubstancesweremore commonamongmenthanwomen.Theuseofopium,shireandinjectingdrugswasmorecommonin individualsover30whereastheuseofstimulantsandhashishwaslargestamongindividualsbetween 18and30yearsofage.

Conclusion:Itseemsthatyoungerindividualsandwomenaremoredesiredtousenewsyntheticdrugs suchascrystalmethamphetamine.Extendingthepreventiveprogramsespeciallyinyouthaslikeas scalingupharmreductionserviceswouldbethemainprioritiesinpreventionandcontrolofsubstance useinIran.Becauseofpoorservicecoverageandhighstigmainwomen,moretargetedprogramsinthis affectedpopulationareneeded.

ß2016ElsevierB.V.Allrightsreserved.

* Correspondingauthor.Tel.:+989133406291.

E-mailaddress:[email protected](A.Shahesmaeili).

ContentslistsavailableatScienceDirect

International

Journal

of

Drug

Policy

j ou rna l h om e pa ge : w w w. e l s e v i e r. co m/ l oc a t e / dru gpo

http://dx.doi.org/10.1016/j.drugpo.2016.01.013

country for illicit drugs (UNODC, 2011). After inflation and unemployment, substance smuggling and substance use is the thirdmajordilemmainIraniansocietyresultinginserioussocial, economical and public health consequences ( Moghanibashi-Mansourieh&Deilamizade,2014).

Mostofourknowledgeregardingtheepidemiologyandextent ofsubstanceuseinIranisbasedonofficialreportsoftheMinistry ofHealth(MOH)andIranianDrugControl(DCHQ).Discrepancies betweenthesereports,though,makethepictureofsubstanceuse moreambiguous.In2002,MOHandDCHQestimatedthenumber ofsubstanceusersinthepopulationabove15yearsBecauseof some biases and methodological errors (e.g. sampling from emergency clinics) the results were never announced or distributed(Tafreshi, 2012). A rapidsituation assessmentand analysisin2007estimatedthenumberofsubstanceuserstobe 1.2million(correspondingtoanadult prevalenceof2.4%).The mostcommontypeofdrugusedwasopium(34%)followedby crack(whichinIranisacocaine-freedrugthatmostlycontains heroin,codeine, morphine and caffeinewith or withoutother drugs)(26.6%),heroin(19.9%), shire(a combination of Opium residue and pure opium) (4.4%), neurjezik (4.1%), crystal methamphetamine (3.6%) and hashish (2%) (Narenjiha et al., 2007).Inthissurvey theopiateuserswererecruited fromthe street,prisonsandtreatmentclinicsin27citiesinIran,whichmay limitthegeneralizabilityoftheresults.Theresultsofthelatest nationalhouseholdsurveyin2011,conductedbyDCHQon15,000 households,suggestthatthereare1.325millionsubstanceusers (correspondingtoaprevalenceof2.65%)amongpeoplebetween 15and65yearsold.Basedontheirreport, themost prevalent substances were opium (52.02%), crystal methamphetamine (26.22%),crack(15.95%),heroin(9.77%),hashish(6.43%),ecstasy (3.08%)andshire(2.83%)(Sarramietal.,2013;Tafreshi,2012). Althoughtheseofficialreportsprovidepolicymakerssomewhat ofa snapshotregardingthesituationof substanceusein Iran, somepracticalandmethodologicalproblems(suchasobtaining data from prisoners, treatment seekers and streetdrug users, stigmatowardsubstanceuseinhouseholdsurveys,andextraction ofinformationdirectlyfromthestudypopulation)leadtocertain challenges and debates regarding the pragmatic extent of substanceuse in Iran. Furthermore, thediversity of the study populationandthemethodologythatwasusedinthesesurveys maylimitthemonitoringoftrendsovertime.Thelackofvalidand generalizableinformationonsubstanceuse,therefore,isthemain challengeofdrugusesurveillanceinIran.

Network scale-up (NSU) is an indirect method of size estimationinhiddenpopulationsthathasbeenusedtoestimate the size of hidden populations such as HIV-positive persons (Killworth etal.,1998)substanceusers(Kadushinetal.,2006;

Salganiketal.,2011a),menwhohavesexwithmen(MSM)(Ezoe

etal.,2012)and(other)groupsathigherriskofHIV(Moldova

UNAIDS Country Office, 2010;Paniotto et al., 2009; Shokoohi

etal.,2012)aroundtheworld. Generally, everypersonknows how many people he/she identifies as his/her active social network.Theaveragenumberofhiddensubgroupsreportedby everyrespondentconstitutesafractionoftheirnetwork.Weare thereforeabletoscaleupthisfractiontoafractionofthetotal populationandthusobtainanestimatethenumberof hard-to-reachgroupsofinterest(Bernardetal.,2010;Johnsenetal.,1995, 1989).Theindirectnatureofdatacollection,theabilityofthis method to estimate the size of different subpopulations concurrently, its low cost and simplicity makes it a feasible methodforsizeestimationinsituationswheretheusefulnessof other methods such as enumeration, capture-recapture and multiplier is doubtful due to limited access to reliable data sources(Guoetal.,2013;Rastegarietal.,2013,2014;Salganik etal.,2011a)

Duetothesignificanceofdeterminingareliableandupdated estimateofthesizeofthedruguserpopulationbyage,genderand typesofsubstanceforpolicymaking,planningand surveillance purposes,wedesignedanationalsurveytoestimatethesizeof illicitdruguseintheentirecountryusingNSUmethodology.The policymakersandotherhealthprofessionalsmaybenefitfromthe results of this study in planning prevention and treatment interventionsaswellasinresourceallocationandinmonitoring thetrendovertime.ItseemsthatnotonlyinIran,butalsointhe EasternMediterraneanRegion,thisisthefirsttimesuchastudy wasimplementedonanationalscale.

Methods

Samplinganddatacollection

Inthiscross-sectionalstudywerecruited7535individualsfrom allprovincesofIran.BasedonapilotstudyinKerman,thelowest prevalenceofaspecificdrugusewas1.2%,whichcorrespondedto injectingdrugs(Shokoohietal.,2012).Wesettheprecisiontobe 0.2 of prevalence and type one error at 0.05. The sample size nationally was estimated at 7000. In the end we recruited 7535participantsfrom31provinces.Theaveragesamplesizein each province was approximately 400, ranging from 200 to 1000.Becauseaboutthree-quartersoftheIranianpopulationlive incapitalcities,wetriedtokeepthisratioinoursample.So,75%of thesamplefromeachprovincewasrecruitedfromthecapitaland theremaining25%wasrecruitedfromoneofthelargecitiesofthat province.

Eligiblepersonswereindividualsover18yearsofagewhohad lived in Iran for at least five years prior to the survey. We implementedaself-administered,street-basedquestionnaireon passersbyfromthegeneralpublicwhomettheeligibilitycriteria. Therationalforchoosingtheserespondentswasthatinaprevious study,wecomparedtheoddsofinformationdisclosureinresponse tosensitivequestionsinthreeinterviewmethods(street-based, household,and telephone interviews). The resultsshowedthat street-based interviewing provides a higher rate of disclosing drug-relatedbehaviorsandsexualpracticesthantelephoneand householdinterviewing(Haghdoostetal.,2013).Basedonsocial andeconomicclasses,eachcitywasstratifiedintothreezones.In eachstratum,twotofourstreetswereselectedrandomly.Because we recruited participantsnon-randomly, we asked ourtrained interviewerstorecruitallage,sexandsocio-economicgroupsinto thesampleviaquotasampling.Quotasweresetforgender(50% male,50%female),age(50%between18and30,50%above30%) andsocio-economicstatus.Furthermore,wecomparedthesex-age distribution of sample with whole country census data, the difference wasnot significant.To maximizetheconfidentiality, weaskedtheinterviewersnottoselectrespondentsfromoffices, shops, or similar places. If any invited subject refused to participate,replacementwasdone.Thequestionnaireswerefilled out anonymously. The questionnaire consisted of questions on demographic information as well as questions to measure the networksizeofparticipantsandthenumberofparticipantsineach ofthesevendrugusesubgroups(Rastegarietal.,2013). Applyingnetworkscale-upmethodology

TheNSUmethodisbasedontheassumptionthatbycalculating the proportion of drug users in the social network of a representative sample of the general population (m/C), the prevalenceofdruguseinthewholepopulationcanbeestimated. Wecalculatedtheaveragenumberofdrugusersinthepersonal networkofparticipants(m)byaskingthemhowmanypeoplethey knewwhobelongtoeachofseventargetgroupsincluding:

-Opiumusers(locallynamedTeriakwhichconsistedoftwotypes ofrawopiumandSookhtehwhichisopiumdrossremainingafter theopiumissmokedortakenorally).

-Opiumsapusers(locallynamedShirewhichisarefinedproduct ofopiumthatisobtainedmainlyfromSookhteh,withorwithout adding opium, boiled in water and filtered several times to removetheinsolublematerials(Khademietal.,2012).

-New synthetic drug users including: amphetamine and/or ecstasyand/orLSD(A-E-L)users.

-Crystalmethamphetamineusers(locallycalledShisheh). -Heroinand/orcrack(H-C)users(Thecomponentsofcrackthatis

availableinIranarenotthesameasincrackcocaine.Insteadof cocaine,Iraniancrackcontainsmostlyheroin,codeine,morphine and caffeine with or without other drugs. So because of similaritiesbetweenheroinandIraniancrack,wegroupedthese twotypesofdrugstogether)(Farhoudianetal.,2014;Kazemifar etal.,2011).

-Hashishand/orcannabis(H-M)users. -Peoplewhoinjectdrugs(PWID).

Thedefinitionof‘‘know’’was‘‘peoplewhomyouknowandwho knowyou,inappearanceorbyname,withwhomyoucaninteract, ifneeded,andwithwhomyouhavecontactedoverthelasttwo yearspersonally,orbytelephoneore-mail’’(Midanik&Greenfield, 2003;Rastegarietal.,2013).‘‘Druguse’’wasdefinedasatleastone occasion of substance use during the one year preceding the survey. We estimated the average network size of the Iranian populationat 308(Rastegariet al.,2013).We askedmorethan 7000respondentstodescribethefrequencyof23referencegroups withknownsizes(forexample,thenumberofmennamedHamed andthenumberofpeoplewhoworkinanelementaryschool)in theirnetwork.Toestimatethesizeofsubstanceusersbasedonage andsexgroups,weaskedparticipanttocharacterizethesexand age group (<18 years, 18–30 years, >30 years) of nominated persons.Weassumedthattheprevalenceofthereferencegroups in the country is proportional to that of our respondents. To exclude unreliablereferencegroups, regressionand ratio-based approacheswerefollowed.Tovalidatetheresults,datasplitting wasapplied.Weconcludedthat308wasasolidfiguretodescribe the network size of the Iranian population. The details on estimating the network size of participants (C) and validation processareavailableelsewhere(Rastegarietal.,2013).

Dataanalysis

SincetheNSUmethodassumesthatrespondentsareawareof thebehavior of othermembers of their network, and that the members of the general population have an equal chance of knowing anyone in the target group (McCormick et al., 2010;

Salganik et al., 2011b; Shelley et al., 2006), we adjusted the

estimates based on two correction factors: transmission and barrierratios.Asanexample,avisibilityof0.50meansthatcrude NSUestimatesmustbedividedby0.5,ordoubled.Thesecorrection factorswereextractedfrompreviousstudiesonpeoplewhoinject drugsinIran(Maghsoudietal.,2014).Accordingly,transmission ratioand barrier ratio for all typesof drug were0.54and 0.7, respectively.Theonlytwoexceptionsweretransmissionratiofor Heroin/Crack (0.65) and barrier ratio for Amphetamine/LSD/ Ecstasy(1).WedividedthecrudeNSUestimatebyvisibilityand popularity factors to adjust for these two sources of bias. To providetheuncertaintylevelforestimates,weappliedtheMonte Carlotechnique.Weusedthebelowequationtoestimatethe95% confidenceintervalforestimates.

j ˆe¼ P imij P ici t1

v

Inthisequationt(thetotalnumberofthegeneralpopulation) andPicithenetworksizeofrespondents)arefixedbuttheother

three components Pimij (number of people in a particular subgroupjthat therespondentknows),vf(visibilityfactor)and bf(barrierfactor)arerandomcomponents.WeassumedthatPimij (measuredasfrequency)followPoissondistributionwhilevfand bf (which their values range from 0 to 1) follow uniform distribution.Toprovideuncertaintyranges(95%CI)forestimates, wegeneratedrandomly1000Poissondistributionwiththemean of Pimij and allowed vfs vary by 10%. Percentiles of 2.5 and 97.5wereconsideredaslowerand upperbounds ofconfidence interval, respectively. Allanalysis was done using SPSS.20 and Stata.11software.

Ethicalconsideration

Thestudyprotocolwasapprovedandreviewedbytheresearch ethicscommitteesofbothKermanUniversityofMedicalsciences (ethicno:163/90/KA)andtheIranianMinistryofHealth.Before gatheringsufficientdataand completingtherequired question-naires,aninformedconsentwasobtainedfromallrespondents. They were assured that all information and discussions would remain confidential and wereinformed that their participation wasvoluntary.Duringtheinterviewing,participantswerefreeto discontinuethestudyatanytime.

Results

The analysis of data related to 7535 participants from 31 provinces showedthat 48.2% of them weremenand 51.8% werewomen(Table1).Themeanage(standarddeviation)ofmen andwomenwas30.79(11.28)and30.80(10.11),respectively.Age rangedbetween18and87yearsold.Overhalfoftheparticipants weremarried(52.3%)andhadahighschooldiploma/lessthanhigh schooldiplomalevelofeducation(56.9%).

Ourresultsindicatethemostcommontypeofillicitdrugused wasopiumwiththeprevalenceof1500per100,000persons,that was followed by shire (660), crystal methamphetamine (590), hashish(470),heroin/crack(350/100,000),stimulants (metham-phetamine,LSDand ecstasy) (300/100,000)and injecting drugs (280/100,000).Alltypesofsubstanceusewereatleastfourtimes (forstimulants)morecommonamongmenthanwomen(180,000 vs.44,000,respectively).Thelargestdifferencebetweenthetwo genderswasrelatedtoinjectingdrugsinwhichtheestimatedsize formenwasmorethan12timeshigherthaninwomen(193,000 vs.16,000,respectively).Table2showstheestimatedsizeofeach substance use group in the countrybased on gender and age. Amongthreeagegroups,thelowestestimatedsizewasrelatedto thepopulationunder18.Thenumberofpeoplewhouseopiumand shire in the above 30 age group was 784,000 and 352,000, respectivelywhichishigherthancorrespondingestimatesbothin

Table1

Demographicinformationofparticipants.

Demographiccharacteristics Frequency(%) Gender Male 3584(48.2)

Female 3853(51.8)

Age 18–30 3996(53.03)

>30 3539(46.9) Maritalstatus Single 3155(42.3) Married 3899(52.3) Divorced/widowed 295(4) Education Highschooldiploma/underdiploma 4237(56.9)

University/Bachelor 2847(38.3) Masterandupper 358(4.8)

population under18 (5400 and 1600) and between 18 and 30 (340,000and140,000).Whereasregardingstimulantsandhashish, the largest size pertained to individuals between 18 and 30 (253,000forcrystalmethamphetamineand254,000forhashish) which is comparable to corresponding estimates in under 18 (15,000and900,respectively)andabove30(176,000and106,000 inthatorder)groups.Thelowestestimatesforheroin/crackand injecting drugs was related to population under 18 (2900 for heroin/crack and 900 for injecting drugs) but there were no substantialdifferencesbetweenagegroupsabove30and18–30 concerningheroin/crack(124,000vs.135,000)andinjectingdrugs (106,000vs.102,000).Table2showsthetotalestimatedsizeand prevalenceofeachtypeofdruguseper100,000persons. Discussion

Thisstudyrevealedthatopium,shireandcrystal methamphet-aminearethemostcommontypesofdrugsusedinIran,butthe patternvariessomewhatbasedonageandsex.Inbothgenders, opium,shire andcrystal methamphetamineare thethreemost commontypesofdrugsused.Theonlydifferencebetweenthetwo genders is that after opium, crystal methamphetamine is the secondmostcommonsubstanceforwomenbutthethirdformen. Generally, women are less affected by drugs. The estimated numberofmenwasbetweenfourtimes(foramphetamine,ecstasy andLSD)and12times(forinjectingdrugs)higherthanwomen. While older individualstend to use traditional types of drugs, youngerpeoplearemorelikelytousenewsyntheticones.

Althoughwedonothaveanyupdatedofficialestimationwith whichtocomparetheresultsandcross-validatethefindings,the latestrapidsituationassessmentofdruguseanddependencyin Iran(RSA),conductedin2007,estimatedtheprevalenceofdrug usetobe2.4%.Theyalsoreportedthetotalnumberofdrugusersto be1.2million,ofwhich200,000ofthemarePWID(Narenjihaetal., 2007).Analogoustoourresults,themostcommondrugsinthat studywereopium,crackandheroin,inthatorder.Additionally,the resultsofanationalhouseholdsurveyin2011indicatedthatnearly 3%oftheIranianpopulationusessometypeofdrug.Basedontheir reports,opium,followedbycrystalmethamphetamine,thencrack andheroinwerethemostprevalenttypesofdrugsused(Sarrami etal.,2013).

Analyzingtheevidence,itseemsthattheuseofnew,synthetic drugs,especiallycrystalmethamphetamine,hasbecomeahealth concerninrecentyears(Alam-mehrjerdietal.,2015;Noori,2011). TheresultsoftheRSAin2007indicated that6.3%ofsubstance usersusedmethamphetamine.Althoughtherearenearly6000 cen-ters supervised by medical sciences universities, state welfare

organizationsandprisonorganizationswhichprovidemethadone maintenance therapy (MMT) and buprenorphine maintenance therapy(BMT)fortheapproximatelyhalfmillionpeoplewhouse opiates(includingopium,crack/heroinandshireusers),thereis only a limited number of treatment clinics for other types of substances(Alam-mehrjerdietal.,2014;IranianMinistryofHealth

andMedical Education,2015). Moreover,manyindividualswho

undergoopiatemaintenancetreatment concurrentlyuse stimu-lantsformanyreasons(e.g.feelinggoodandgettinghighinthe absence of the drug) (Shariatirad et al., 2013). This situation potentiallyreducestheeffectivenessofMMTorBMT(Shariatirad etal.,2013).Inexperienceofcliniciansandpsychologistswhoare involved in MMT/BMT regarding the treatment of stimulants makesthetreatmentoftheaffectedpopulationmorecomplicated. This results in most synthetic drug users having no access to appropriatetreatment services.Thereforetheextensionofsuch treatments nationwide and revising the available treatment protocols into comprehensive guidelines which also cover modalitiesforparalleldruguseareneeded.

Wehavelearnedthatthepatternofdrugusevariesbytheageof theindividual.Opium,amphetamine/ecstasy/LSDandshirearethe mostcommontypesofdrugsamongthepopulationover30.Inthe groups 18–30and under18,themostcommon type ofdrugis opium then amphetamine/ecstasy/LSD. In both of these age groups, hashish and crystal methamphetamine are the second andthirdmostcommondrugused.Itseemsthatyoungerpeople tendtousenew,syntheticdrugsmorethantraditionalsubstances while individuals in older age groups are more likely to use traditionalsubstances suchasopiumandshire.Consideringthe psycho-social and physical complications and complexities of treatment,useofsyntheticdrugshasturnedintooneofthemain challengesinrelationshiptosubstanceuseamongyouthandtheir families(Russelletal.,2008).Lackofappropriateknowledge,the reasonableprice,easeofuseandthepositiveattitudetowardthese drugsarepotentialexplanationsfortheelevatedtendencyofyouth toward newsyntheticdrugs. Mostof substanceuseprevention programs among youth in Iran are implemented through the media.Educationinschoolonlycoversabout8%ofschools(Islamic

RepublicNewsAgency,2015).

GenderisanotherchallengeincontrolofdruguseinIran.Asthe resultsshow,inbothgenders,opium,shireandcrystal metham-phetaminearethemostcommontypesofdrugsused.

Itisreportedthatbetween6%and10%ofsubstanceusersinIran arefemale(Noori,2011;Khajedaluee&Moghadam,2013;Sarrami etal.,2013; Tavakolietal.,2014).Our estimatesshowthatthe differencebetweenmalesandfemalesmayvarybasedontypesof drugs.Althoughgenerallywomenarelessaffectedbydrugsthan

Table2

Theprevalenceandestimatedsizeofvariousgroupsofdrugusersaccordingtogenderandage. Prevalence/

100,000 population

Totalestimate (CI95%)

Men(CI95%) Women(CI95%) Lessthan18 (CI95%)

18-30(CI95%) Morethan30 (CI95%) Opium 1500 1,100,000 1,033,000 79,000 5,400 340,000 784,000 (973,000,1,273,000) (912,000,1,186,000) (68,000,91,000) (4000,6900) (299,000,390,000) (693,000,897,000) Shire 660 493,000 449,000 45,000 1600 140,000 352,000 (438,000,566,000) (398,000,515175,000) (39,000,53,000) (1000,2000) (123,000,160,000) (311,000,399,000) Amphetamine/ Ecstasy/LSD 300 224,000 180000 44,000 15,000 167,000 40,000 (206,000,247,000) (164,000,200,000) (39,000,49,000) (13,000,17,000) (153,000,185,000) (35,000,45,000) Crystal methamphetamine 590 440,000 394,000 46,000 8000 253,000 176,000 (387,000,502,000) (349,000,451,000) (40,000,54,000) (7000,10,000) (221,000,289,000) (156,000,202,000) Heroin/crack 350 262,000 241,000 22,000 2900 135,000 124,000 (235,000,296,000) (214,000,273,000) (19,000,25,000) (2000,3600) (121,000,154,000) (111,000,140,000) Hashish (cannabi0073) 470 353,000 326,000 28,000 11,000 254,000 87,000 (312,000,403,000) (286,000,373,000) (25,000,33,000) (9000,13,000) (224,000,292,000) (76,000,100,000) Injectingdrugs 280 208,000 193,000 16,000 900 102,000 106,000 (183,000,238,000) (170,000,220,000) (13,000,19,000) (500,1400) (89,000,117,000) (92,000,122,000)

men,currentevidencesuggests,despitetheincreasingprevalence ofsubstanceuseinbothgenders,theriseismoreacceleratedin women, especially regarding stimulants (Noori, 2011). Women constituteonly2–6%ofthepopulationwhoseektreatmentindrug treatmentclinicsinIran(Tavakolietal.,2014),whichcorresponds to9600–28,800individualsoutofall4800treatmentseekers.If weonlyconsiderthenumberofwomenwhouseopium(78,800) thepoor availabilityoftreatment servicesforwomenwould be evident. Indeed, Iranianwomen experience more stigma being drugusersandconsequentlyaremorereluctanttoseektreatment (Dolanetal.,2011).

Regardingtheregisteredcasesofopiateusersindrugtreatment clinics (0.5 million) (Iranian Ministry of Health and Medical Education,2015)ourestimatesindicatethatlessthanhalfofopiate users are under the coverage of drug treatment centers. Additionally basedon a MOH reportof thehalfmillion people who received drug maintenance treatment, only 26,000 were

PWID(IranianMinistryofHealthandMedicalEducation,2015).

ConsideringtheestimatedsizeofPWIDinourstudy(208,000),this corresponds to 12.5% of all PWID. A cross-sectional study on 572PWIDinTehranalsorevealedthatonlyasmallportionofPWID (9%)useddrugtreatmentservices,whilemajorityofthem(55%) hadaccesstoneedleexchangeprograms(NEP)(Rahnamaetal., 2014).Itisestimatedthat13.5%ofPWIDinIranareHIVpositive. Because this estimate has been extracted from national bio-behavioralsurveillanceamongPWIDwhoarereferredtofacilities, itcannotreflecttheHIVprevalenceofallPWIDinthecommunity. WethereforecannotestimatethenumberofHIV-positivecases basedon theestimatednumberof this subgroup.Nevertheless, increasingharmreductionservicescouldhavesubstantialeffect on reducing HIV incidence and transmission in this key group (Nasirianetal.,2012).

Estimations based on NSU, in comparison to both direct interviewandusingsecondarydatasources,mayprovidehigher numbersofprevalenceofdruguse(Salganiketal.,2011a).Thisis becauseindirectinterview,respondentsusuallytendto underre-port their stigmatized behaviour.The main limitation of some secondarydatasources,suchasmandatorydrugscreeningthatare implemented before marriage, getting governmental jobs or obtaining licences, is the foreknowledge of individuals about screenings.Inothersourcesofdata,suchasprisoninmatesand policearrestees,genderimbalancemaylimitthegeneralizability oftheresults(Razzaghietal.,1999).

Weacknowledgethelimitationsofthisstudy.Asstatedlater, theprevalenceofalltypesofdrugusewasverylowinagegroups under18.Someamountsofunderestimationarepossibleinthis age group.On the otherhand about 42.5% of our sample had universitydegreewhichdifferswithcorrespondingnumberof18% based on latestcensus in Iran. These may be because in each province,werecruitedrespondentsfromthecapitalcityandalso oneofthelargecitiesofthatprovince.Thusbecauseofselection biastheresultscouldnotbegeneralizabletoruralpopulations. Lastly,despitetheusefulnessofNSUinhighstigmatizedsetting, there is still some sourceof biases. Because of some practical constrains,wecouldnotconductseparatestudiesamongdifferent drugusergroupsineachprovincetoestimatethetransmissionand barrierratiousedtoadjusttheresults.Itispossiblethatcultural contextsaffectthesocialtransparencyofdrugusebehavior.This effectcandifferslightlyinvariousareasofthecountry.Selection biasandsocialdesirabilitybiascouldnotberuledoutdueto non-randomsamplingand interview-basednatureof datacollection respectively. Recall bias is another threat of validity in NSU.

Salganiketal.(2011a)showedthatIndividualsmayhavetendency

tooverestimatethesizesof smallerpopulationsand underesti-matethesizesoflargerpopulations.Ourgoalforfuturestudiesis toreducethesecircumscriptionsasmuchaspossible.Nonetheless,

theindirectnatureofdatagathering,lowcostandpossibilityof concurrent size estimation of various groups are the main strengthsofNSUforsizeestimationofhiddenpopulations.The nationwideessenceofstudyandconsequentlylargesamplesizeis another advantage of ourstudy. Overall, we believe,given the scopeofthestudy,ourfindingshaveimportantimplicationsfor bothresearchandpolicy.Tomonitorthetrendandevaluationof programs, we highly recommend this survey to be conducted repeatedly.

Funding

Thisworkwassupported byMinistryofHealthandMedical Education,Tehran,Iran.

Contributors

Allauthorshavereadandapprovedthefinalmanuscript. Acknowledgment

The authorswouldlike tothanktheMinistryof Health and MedicalEducation,Tehran,Iranforfinancialsupport.

Conflictofinterest:Allofauthorsareindependentresearchers withoutanyconflictofinterestexceptDr.Nikfarjamwhoatthe time of design and implementation of project was director in Deputy of MentalHealth &Social Health and substance Abuse Office,MinistryofHealthandMedicalEducation.

References

Alam-mehrjerdi,Z.,Mokri,A.,&Dolan,K.(2015).Methamphetamineuseand treat-mentinIran:AsystematicreviewfromthemostpopulatedPersianGulfcountry. AsianJournalofPsychiatry.

Alam-mehrjerdi,Z.,Noori,R.,&Dolan,K.(2014).Opioiduse,treatmentandharm reductionservices: The firstreport from the Persian Gulfregion. Journal of SubstanceUse,1–7.

Bernard,H.R.,Hallett,T.,Iovita,A.,Johnsen,E.C.,Lyerla,R.,McCarty,C.,etal.(2010).

Countinghard-to-countpopulations:Thenetworkscale-upmethodforpublic health.SexuallyTransmittedInfections,86(Suppl.2),ii11–ii15.

Dolan,K.,Salimi,S.,Nassirimanesh,B.,Mohsenifar,S.,Allsop,D.,&Mokri,A.(2011).

CharacteristicsofIranianwomenseekingdrugtreatment.JournalofWomen’s Health,20(11),1687–1691.

Ezoe,S.,Morooka,T.,Noda,T.,Sabin,M.L.,&Koike,S.(2012).Populationsizeestimation ofmenwhohavesexwithmenthroughthenetworkscale-upmethodinJapan. PLoSONE,7(1),e31184.

Farhoudian,A.,Sadeghi,M.,Vishteh,H.R.K.,Moazen,B.,Fekri,M.,&Movaghar,A.R. (2014).ComponentanalysisofIraniancrack;anewlyabusednarcoticsubstancein Iran.IranianJournalofPharmaceuticalResearch:IJPR,13(1),337–344.

Guo,W.,Bao,S.,Lin,W.,Wu,G.,Zhang,W.,Hladik,W.,etal.(2013).Estimatingthesize ofHIVkeyaffectedpopulationsinChongqing,China,usingthenetworkscale-up method.PLOSONE,8(8),e71796.

Haghdoost,A.A.,Baneshi,M.R.,Eybpoosh,S.,&Khajehkazemi,R.(2013).Comparison ofthreeinterviewmethodsonresponsepatterntosensitiveandnon-sensitive questions.IranianRedCrescentMedicalJournal,15(6),500–506.

IranianMinistryofHealthandMedicalEducation(2015).AIDSprogressreport(inFarsi).

http://www.unaids.org/sites/default/files/country/documents/ IRN_narrative_report_2015_persian.pdf

IslamicRepublicNewsAgency(2015).Educationonfightingaddictionin8%ofschools.

http://www.irna.ir/fa/News/81713411/

Johnsen,E.C.,Bernard,H.R.,Killworth,P.D.,Shelley,G.A.,&McCarty,C.(1995).Asocial networkapproachtocorroboratingthenumberofAIDS/HIV+victimsintheUS. SocialNetworks,17(3),167–187.

Johnsen,E.C.,Killworth,P.D.,&Robinson,S.(1989).Estimatingthesizeofanaverage personalnetworkandofaneventsubpopulation.InM.Kochen(Ed.),Thesmall world(pp.159–175).AblexPress.

Kadushin,C.,Killworth,P.D.,Bernard,H.R.,&Beveridge,A.A.(2006).Scale-upmethods asappliedtoestimatesofheroinuse.JournalofDrugIssues,36(2),417–440.

Kazemifar,A.M.,Solhi,H.,&Badakhshan,D.(2011).CrackinIran:Isitreallycocaine. JournalofAddictionResearh&Therapy,2,107. http://dx.doi.org/10.4172/2155-6105.1000107

Khademi,H.,Malekzadeh,R.,Pourshams,A.,Jafari,E.,Salahi,R., Semnani,S.,etal. (2012).OpiumuseandmortalityinGolestanCohortStudy:Prospectivecohort studyof50,000adultsinIran.BMJ:BritishMedicalJournal,344.

Khajedaluee,M.,&Moghadam,M.D.(2013).Methodsandpatternsofdrugabuse amongyoungaddictwomen.JournalofResearch&Health,3(4),527–535.

Killworth,P.D.,Johnsen,E.C.,McCarty,C.,Shelley,G.A.,&Bernard,H.R.(1998).Asocial networkapproachtoestimatingseroprevalenceintheUnitedStates.Social Net-works,20(1),23–50.

Maghsoudi,A.,Baneshi,M.R.,Neydavoodi,M.,&Haghdoost,A.(2014).Network scale-upcorrectionfactorsforpopulationsizeestimationofpeoplewhoinjectdrugsand femalesexworkersinIran.PLOSONE,9(11),e110917.http://dx.doi.org/10.1371/ journal.pone.0110917

McCormick,T.H.,Salganik,M.J.,&Zheng,T.(2010).Howmanypeopledoyouknow? Efficientlyestimatingpersonalnetworksize.JournaloftheAmericanStatistical Association,105(489),59–70.

Midanik,L.T.,&Greenfield,T.K.(2003).Telephoneversusin-personinterviewsfor alcoholuse: Resultsofthe 2000 NationalAlcohol Survey. DrugandAlcohol Dependence,72(3),209–214.

Moghanibashi-Mansourieh,A.,&Deilamizade,A.(2014).Thestateofdatacollectionon addictioninIran.Addiction,109(5),854.

MoldovaUNAIDSCountryOffice(2010).SizeestimationforhighriskgroupsMoldova.

http://aids.md/aids/files/166/

report-estimations-sizes-vulnerable-populations-mission-2010-en.pdf

Narenjiha,H.,Rafiei,H.,Baghestani,A.R.,Nouri,R.,Shirinbayan,P.,&Farhadi,M.H. (2007).RapidsituationassessmentofdrugabuseinIran(year2004)Researchreport. Tehran:DariushInstitute.

Nasirian,M., Doroudi,F.,Gooya,M.M.,Sedaghat, A.,&Haghdoost,A.A.(2012).

ModelingofhumanimmunodeficiencyvirusmodesoftransmissioninIran.Journal ofResearchinHealthSciences,12(2),81–87.

Noori,R.(2011).Situationanalysisofwomendruguse(inFarsi).(Strategicreport). Tehran:IranExpediencyCouncilSecretariat.

Paniotto,V.,Petrenko,T.,Kupriyanov,V.,&Pakhok,O.(2009).Estimatingthesizeof populationswithhighriskforHIVusingthenetworkscale-upmethod.http://nersp. osg.ufl.edu/~ufruss/scale-up/Ukraine_Final%20Report_scale-up_eng_July24.pdf

Rahnama,R.,Mohraz,M.,Mirzazadeh,A.,Rutherford,G.,McFarland,W., Akbari,G., etal.(2014).Accesstoharmreductionprogramsamongpersonswhoinjectdrugs: Findingsfromarespondent-drivensamplingsurveyinTehran,Iran.International JournalofDrugPolicy,25(4),717–723.

Rastegari,A.,Baneshi,M.R.,Haji-Maghsoudi,S.,Nakhaee,N.,Eslami,M., Malekafzali, H.,etal.(2014).EstimationoftheannualincidenceofabortionsinIranapplying NetworkScaleUpapproach.Acceptedinircmj.

Rastegari,A.,Haji-Maghsoudi,S.,Haghdoost,A.,Shatti,M.,Tarjoman,T.,&Baneshi,M. R.(2013).TheestimationofactivesocialnetworksizeoftheIranianpopulation. GlobalJournalofHealthScience,5(4),217–227.

Razzaghi, E.M., Rahimi,M. A.,Hosseni,M.,&Madani,S. (1999).Rapidsituation assessment(RSA)ofdrugabuseinIran(1998–1999).PreventionDepartment,State WelfareOrganization,MOHandUNODC.

Russell,K.,Dryden,D.M.,Liang,Y.,Friesen,C.,O’Gorman,K.,Durec,T.,etal.(2008).Risk factorsformethamphetamineuseinyouth:Asystematicreview.BMCPediatrics, 8(1),48.

Salganik,M.J.,Fazito,D.,Bertoni,N.,Abdo,A.H.,Mello,M.,&Bastos,F.I.(2011).

Assessing networkscale-upestimates for groupsmost atriskfor HIV/AIDS: EvidencefromamultiplemethodstudyofheavydrugusersinCuritiba,Brazil. AmericanJournalofEpidemiology,174(10),1190–1196.

Salganik,M.J.,Mello,M.B.,Abdo,A.H.,Bertoni,N.,Fazito,D.,&Bastos,F.I.(2011).The gameofcontacts:Estimatingthesocialvisibilityofgroups.SocialNetworks,33(1), 70–78.

Sarrami,H.,Ghorbani,M.,&Minooei,M.(2013).Surveyoffourdecadesofaddiction prevalenceresearchesinIran.ResearchonAddiction(inFarsi),7(26),29–52.

Shariatirad, S., Maarefvand, M., & Ekhtiari, H. (2013). Methamphetamine use andmethadonemaintenancetreatment:Anemergingprobleminthedrug addictiontreatmentnetworkinIran.InternationalJournalofDrugPolicy,24, e115–e116.

Shelley,G.A.,Killworth,P.D.,Bernard,H.R.,McCarty,C.,Johnsen,E.C.,&Rice,R.E. (2006).WhoknowsyourHIVstatusII?Informationpropagationwithinsocial networksofseropositivepeople.HumanOrganization,65(4),430–444.

Shokoohi,M.,Baneshi,M.R.,&Haghdoost,A.-a.(2012).Sizeestimationofgroupsat highriskofHIV/AIDSusingnetworkscaleupinKerman,Iran.InternationalJournal ofPreventiveMedicine,3(7),471.

Tafreshi,H.(2012).TheStatisticsonAddictioninIran(inFarsi).RaziJournal,23(2),49– 56.

Tavakoli,M.,Mohammadi,L.,Yarmohammadi,M.,Farhoudian,A.,Ja’fari,F.,&Farhadi, M.H.(2014).StatusandtrendofsubstanceabuseanddependenceamongIranian women.JournalofRehabilitation,14(5),30–37.

UNODC(2011).TechnicalcooperationprogrammeintheIslamicRepublicofIran2011– 2014.https://www.unodc.org/documents/islamicrepublicofiran/CP_Iran11-14.pdf