BMC Health Services Research This Provisional PDF corresponds to the article as it appeared upon acceptance. The fully-formatted PDF version will become available shortly after the date of publication, from the URL listed below. Cost and efficiency of public sector sexually transmitted infection clinics in Andhra Pradesh, India BMC Health Services Research 2005, 5:69 doi:10.1186/1472-6963-5-69 Lalit Dandona (dandona@asci.org.in) Pratap Sisodia (pratapsisodia@asci.org.in) T. L. N. Prasad (drellen2002@yahoo.com) Elliot Marseille (emarseille@comcast.net) M. Chalapathi Rao (mcrao@asci.org.in) A. Anod Kumar (anod@asci.org.in) S. G. Prem Kumar (prem@asci.org.in) Y. K. Ramesh (ramesh@asci.org.in) Mead Over (meadover@worldbank.org) M. Someshwar (someshmfa@yahoo.com) James G. Kahn (jgkahn@ucsf.edu) ISSN 1472-6963 Article type Research article Submission date 14 Apr 2005 Acceptance date 5 Nov 2005 Publication date 5 Nov 2005 Article URL http://www.biomedcentral.com/content/5/1/69 Like all articles in BMC journals, this peer-reviewed article was published immediately upon acceptance. It can be downloaded, printed and distributed freely for any purposes (see copyright notice below). Articles in BMC journals are listed in PubMed and archived at PubMed Central. For information about publishing your research in BMC journals or any BioMed Central journal, go to http://www.biomedcentral.com/info/authors/ © 2005 Dandona et al., licensee BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. Costandefficiencyofpublicsectorsexuallytransmittedinfectionclinics inAndhraPradesh,India LalitDandona1,PratapSisodia1,T.L.N.Prasad2,ElliotMarseille3,M.ChalapathiRao1, A.AnodKumar1,S.G.PremKumar1,Y.K.Ramesh1,MeadOver4,M.Someshwar1, JamesG.Kahn3 1HealthStudiesArea,CentreforHumanDevelopment,AdministrativeStaffCollegeof India,Hyderabad,India 2AndhraPradeshStateAIDSControlSociety,Hyderabad,India 3InstituteforHealthPolicyStudiesandAIDSResearchInstitute,Universityof California,SanFrancisco,USA 4DevelopmentResearchGroup,WorldBank,WashingtonDC,USA Email: LD:dandona@asci.org.in,lalit.dandona@rediffmail.com;PS:pratapsisodia@asci.org.in; TLNP:drellen2002@yahoo.com;EM:emarseille@comcast.net;MCR: mcrao@asci.org.in;AAK:anod@asci.org.in;SGPK:prem@asci.org.in;YKR: ramesh@asci.org.in;MO:meadover@worldbank.org;MS:someshmfa@yahoo.com; JGK:jgkahn@ucsf.edu Correspondenceto: LalitDandona,HealthStudiesArea,CentreforHumanDevelopment,Administrative StaffCollegeofIndia,RajBhavanRoad,Hyderabad–500082,India E-mail:dandona@asci.org.in,lalit.dandona@rediffmail.com;Tel:(+9140)23376958; Fax:(+9140)23312954 Cost-efficiencyofSTIclinicsinIndia 2 Abstract Background:Controlofsexuallytransmittedinfections(STIs)isanimportantpartof theefforttoreducetheriskofHIV/AIDS.STIclinicsinthegovernmenthospitalsin Indiaprovideservicespredominantlytothepoor.Dataonthecostandefficiencyof providingSTIservicesinIndiaarenotavailabletohelpguideefficientuseofpublic resourcesfortheseservices. Methods:Standardisedmethodswereusedtoobtaindetailedcostandoutputdataforthe 2003-2004fiscalyearfromwrittenrecordsandinterviewsin14governmentSTIclinics intheIndianstateofAndhraPradesh.TheeconomiccostperpatientreceivingSTI treatmentwascalculated,andthevariationsoftotalandunitcostsacrosstheSTIclinics analysed.Multivariateregressiontechniquewasusedtoestimateincrementalunitcosts. TheoptimalnumberofSTIsthatcouldbehandledbytheclinicswasestimated. Results:18807STIswerediagnosedandtreatedatthe14STIclinicsinfiscalyear 2003-2004(range323-2784,median1199).TheeconomiccostoftreatingeachSTI varied5-foldfromIndianRupees(INR)225.5(US$4.91)toINR1201.5(US$26.15) between13clinics,withoneotherclinichavingaveryhighcostofINR2478.5(US$ 53.94).TheaveragecostperSTItreatedforall14clinicscombinedwasINR729.5(US$ 15.88).Personnelsalariesmadeup76.2%ofthetotalcost.ThenumberofSTIstreated perdoctorfull-timeequivalentandcost-efficiencyforeachSTItreatedhadasignificant directnon-linearrelation(p<0.001,R2=0.81;powerfunction).Withamultiple regressionmodel,apartfromthefixedcosts,theincrementalcostforeachSTIdetected andcostoftreatmentwasINR55.57(US$1.21)andforeachfollow-upvisitwasINR 3.75(US$0.08).BasedonestimatesofoptimalSTIcasesthatcouldbehandledwithout compromisingqualitybyeachdoctorfull-timeequivalentavailable,itwasprojectedthat at8ofthe14clinicssubstantiallymoreSTIcasescouldbehandled,whichcouldincrease thetotalSTIcasestreatedatthe14clinicscombinedby38%atanadditionalcostofonly 3.5%forserviceprovision. Conclusion:Thereisun-utilisedcapacityinthepublicsectorSTIclinicsinthisIndian state.Effortstofacilitateutilisationofthiscapacitywouldbeuseful,asthiswould enablemorepoorpatientswithSTIstobeservedatminimaladditionalcost,andwould alsoreducethecostperSTItreatedleadingtomoreefficientuseofpublicresources. Cost-efficiencyofSTIclinicsinIndia 3 Background IndiahasoneofthehighestnumberofpersonslivingwithHIVintheworld[1,2].The presenceofsexuallytransmittedinfectionsotherthanHIV(referredtoasSTIsinthis paper)increasessubstantiallytheriskofacquiring/transmittingHIV,andtherefore,the controlofSTIscanbeanimportantpartoftheefforttocontrolemergingHIVepidemics [3].ThestateofAndhraPradeshwith80millionpopulationhasoneofthehighest estimatedburdenofHIVamongtheIndianstatesbasedonantenatalsentinel surveillance,andalsooneofthehighestHIVprevalence(16-30%)amongpublicsector STIclinicattendeesincludedinthesentinelsurveillanceduring1998-2004[4]. STIclinicsinthepublicsectorhospitalsrunbythegovernmentprovidecare predominantlytothepoorsegmentofsocietyinIndia,andtherefore,theirrolein controllingSTIsandHIVisparticularlyimportant.Recognisingtheimportantrolethat thepublicsectorSTIclinicscouldplayinHIVcontrol,thenumberofformally designatedSTIclinicsinthegovernmenthospitalsinAndhraPradeshwasincreasedfrom 28to85intheyear2002(oralcommunication,AndhraPradeshStateAIDSControl Society,Hyderabad,India,14December2004). WithHIV/AIDScontrolhavingbecomeamajorpublichealthissueinIndia[5], thefundingavailableforthishasbeenincreasing[6,7].However,assessmentofthecost andefficiencyofvariousstrategiestocontrolHIVinIndiaisnotreadilyavailable[8]. ThisinformationisneededforefficientutilisationofresourcesavailableforHIVcontrol. AspartofastudytoassessthecostandefficiencyofvariousHIVpreventionstrategiesin AndhraPradesh,weassessedthecostandefficiencyofSTIclinicsinthepublicsector. Cost-efficiencyofSTIclinicsinIndia 4 Methods Thisstudywaspartofalargermulti-countryefforttostudycostandefficiencyofHIV preventioninIndia,Mexico,Russia,SouthAfricaandUgandabythePreventAIDS NetworkforCost-EffectivenessAnalysis(PANCEA)[9].Detailsofthemethodsforthe overallmulti-countrystudyaredescribedelsewhere[10].Descriptionofthemethods relevantforthispaperfollows.ThisstudywasapprovedbytheEthicsCommitteeofthe AdministrativeStaffCollegeofIndia,Hyderabad,IndiaandtheCommitteeonHuman ResearchoftheUniversityofCalifornia,SanFrancisco,USA. SelectionofSTIclinics WeincludedSTIclinicsinthegovernment-runpublicsectorhospitalsforthisstudy. Theseclinicsprovideservicespredominantlytothepooratnoorminimalfee,andmany personsreceivingservicesherehaverelativelyadvancedSTIs.Atthetimeofstarting datacollectionforthisstudyinmid-2004,85publicsectorclinicswerefunctioning.Of these,28STIclinicshadbeenfunctioningsincethe1960’sinthestatecapitalandthe districtheadquartersaspartofthemedicalcollegehospitalsandothermajorpublic hospitalsofAndhraPradesh.Inaddition,57STIclinicswereformallydesignatedin mid-2002inpublichospitals,indistrictheadquarterhospitalsandinareahospitals locatedinsmallerjurisdictions.Generally,theSTIcareinmedicalcollegehospitalsand districtheadquarterhospitalsisprovidedindistinctseparateclinics,whereasinarea hospitalsitisprovidedaspartofthegeneraloutpatient. AndhraPradeshhasthreegeographicregions:thenorthernTelanganaregionhas thestatecapitalandnineotherdistricts,theeasternCoastalregionhasninedistricts,and thesouthernRayalseemaregionhasfourdistrictswithapopulationofnearlyhalfthatin theothertworegions.InordertoobtainabroadsampleofSTIclinicsinpublic hospitals,weusedthethreegeographicregionsofAndhraPradeshandtheold-new clinicsasthetwostrataforsampling.FifteenSTIclinicswererandomlysampledto obtainsixclinicseachintheTelanganaandCoastalregionsandthreeclinicsinthe Rayalseemaregion,whichincludedthreeeachofoldandnewclinicsinTelanganaand CoastalregionsandtwooldandonenewclinicsinRayalseemaregion. Cost-efficiencyofSTIclinicsinIndia 5 Datacollectionprocedures TheinitialversionsofthedatacollectioninstrumentsfromtheglobalPANCEAstudy werereviewedandadaptedtosuitthecontextofAndhraPradesh.Thedatacollection team,consistingofsixresearcherswithbackgroundineconomicsorfinance,was involvedwiththeadaptationoftheinstrumentsandreceivedextensivetrainingtoensure astandardisedapproachtodatacollection.Apilotstudywasdonetomakefinal refinementsinthedatacollectionformatandapproach. DatawerecollectedfortheApril2003–March2004fiscalyearatthe15sampled STIclinicsduringJune–August2004.Datacollectionincludedahistoryofthe evolutionoftheSTIclinic,detailedcostandoutputdatabymonth,andpatientexit interviews.Formalinformedconsenttocollectdatawasobtainedfromthesenior-most personresponsibleforeachSTIclinic,generallythesuperintendentofthehospitalin whichtheclinicwaslocated.TheSTIclinicmedicalofficer(s),medico-socialworker, counsellor,andlaboratorytechnicianwereinterviewedandavailablewrittenrecords reviewedtoobtaindata.Eachvisitstartedwithaninterviewcontainingstructuredopen- endedquestionsonthehistoryoftheSTIclinic,andoperationalorcommunityfactors thatmayhaveaffectedthedemandfororsupplyofservices.Datacollectionthen proceededtocostandoutputdata,inparallelwithexitinterviewsofpatients.Data collectionatanSTIclinicbythreeinvestigatorslastedoneweek.Datawererecordedin thefieldonlaptopcomputersinMSExcelandMSWordfiles,whichafterreviewwere enteredintoanMSAccessdatabase. Costdata ThecostoftheSTIclinicwasdividedintofivecategories:salaries,recurrentgoods, capitalgoods,recurrentservices,andrentals.Thesecostdatawerecollectedforeach month,asfaraspossible.Economiccostwascomputed,i.e.thetrueresourcecost incurredratherthanjustthefinancialcost.Theinpatientcostwasalsoincludedforthe inpatientfacilitiesthatwerebeingutilisedforsomeoftheSTIclinicpatients.Similar costingmethodswereusedacrosstheclinicsforthefivecostcategories. Salarycostwascomputedforallpersonnelcontributingtotheworkofthe outpatientandinpatientservicesoftheSTIclinic,whichincludedthemedicalofficer(s), Cost-efficiencyofSTIclinicsinIndia 6 medico-socialworker/counsellor,nurse(s),laboratorytechnician,andinsomecases attender.IfaparticularpersonwascontributingparttimetotheSTIclinicwork,only thatproportionofsalarywasincludedintheSTIclinicsalarycost.Forexample,ifthe medicalofficerwascontributingonlyhalftimetotheSTIclinicandtheotherhalfto teachingthenonlyhalfhis/hersalarywasincludedinthepersonnelcost.Thesalaryof thepersonnelwasnotedfromtheofficialrecordsofhospitalofwhichtheSTIclinicwas apart.Iffringebenefitswerepaidinadditiontotheregularsalary,thesewereincluded. Themajorcomponentofrecurrentgoodswasmedicines.Themedicines prescribedtopatientsatthesepublicclinicsareprovidedfree.Inordertocalculatethe costofmedicines,firstthelowestmarketretailratesofthemedicinesusuallyprescribed atthesepublicclinicsforeachSTIdiagnosisduringthe2003-2004wereobtained.This wasdiscountedby30%,whichistheestimateddiscountforbulkpurchaseofmedicines bythegovernmentassuggestedbytheprocurementagency.Theserateswerethen appliedtothenumberofeachtypeofSTIreportedbyaSTIclinic.Theexperienceofthe STIinchargeattheAndhraPradeshStateAIDSControlSocietysuggestedthat complicatedcasesneededfurthertreatmentandthattherewerevariationsintreatment regimens,whichwouldenhancedtheoverallmedicationcostby25%.Accordingly,this enhancementwasappliedtoobtainthetotalmedicinecostforeachSTIclinic.Theother recurrentgoodsutilisedattheSTIclinicsincludedinpatientfoodexpenses,testkitsfor STIs,malecondoms,behaviouralchangecommunicationmaterials,needlesandsyringes, gloves,spirit,sodiumhypochloritesolution,cotton,soap,dettol,phenyl,distilledwater, testtubes,antigen,blottingpaper,alcohol,liquidparaffin,buffersolution,stationeryand someotherlaboratorymaterialsandmiscellaneousitems.Theinpatientfoodexpenses werecalculatedbasedoncostestimatesforitemsonthemenuandthenumberof inpatients.ThetestkitsforSTIswerecostedbyapplying30%discounttothelowest marketrateforbulkpurchasebygovernment,asforthemedicines.Themarketpriceof thecondomssuppliedtotheSTIclinicissubsidizedby70%bythegovernment.We consideredastheeconomiccostofthecondomwhatitwouldhavebeenwithoutthe subsidy.InformationwasobtainedfromtheAndhraPradeshStateAIDSControlSociety aboutthecostofthebehaviouralchangecommunicationmaterialssuppliedtotheSTI clinic,whichwasusedforouranalysis.Attemptsweremadetogetthecostoftheother Cost-efficiencyofSTIclinicsinIndia 7 goodsfromtheSTIclinicrecords,whichifavailablewereusedforanalysis.Ifnot available,threequotationsforthesegoodswereobtainedfromthemarketforthe2003- 2004fiscalyearandtheaverageofthesetakenasthecost. CapitalgoodsusedfortheworkofSTIclinicsincludedoutpatientandinpatient furniture,electricalfixtures,refrigerator,centrifuge,microscope,needleandsyringe destroyer,waterbath,sliderotator,waterfilter,sterilizer,andweighingmachine.Asfor recurrentgoods,ifinformationaboutthecostofcapitalgoodswasnotavailablefromthe STIclinicoritsparenthospital,themarketpricewasdeterminedfromretailsellersof thesegoods.Thelifeofthecapitalgoodswasassumedtobefiveyears,andtherefore, one-fifthofthecostwasallocatedtothe2003-2004fiscalyearifthegoodwasusedfor thefullyear.Ifagoodwaspurchasedinthemiddleofthisfiscalyearandusedonlyfor halftheyear,thecostallocatedtothisgoodwashalfoftheyearlycost.Ifacapitalgood, forexamplerefrigeratororcentrifuge,wasalsobeingusedforworkotherthanthatofthe STIclinic,adeterminationwasmadefromtheSTIclinicstaffabouttheproportionofuse forSTIworkandthatproportionalcostwasallocatedtotheSTIclinic. Recurrentservicesincludedcleaningandbuildingmaintenance,electricity,water, telephone,gas/oil,wastedisposal,theoccasionaltrainingofstaffduringthatfiscalyear, andsomemiscellaneousitems.Thecostofbuildingmaintenancewascalculatedbased onthespaceoccupiedbytheoutpatientandinpatientcomponentsoftheSTIclinic. Electricityandwatercostswerebasedonapplyingthemarketratestotheestimated usage.Theestimatedusageofelectricitywascalculatedfromtheelectricalfixturesand thenumberofhoursofuseperdayandforwaterfromthedailyestimatedusebyclinic staffandpatients.Telephoneandotherrecurrentservicescostswerecalculatedbasedon actualexpenditure. AllSTIclinicswerelocatedinaparentpublichospital,andnorentwasbeing paid.Inordertocalculateeconomicrentalcost,theactualfloorareaoccupiedbythe outpatientandinpatientservicesoftheSTIclinicwasdetermined,rentratesobtained fromthreesourcesinthatareaforhealthfacilitiesforthe2003-2004period,andthe averageoftheseratesappliedtothisarea. Forareahospitals,whereSTIcareisgenerallyprovidedaspartofthegeneral outpatient,thecostsforSTIcarewerecalculatedbasedonestimatesoftheproportionof Cost-efficiencyofSTIclinicsinIndia 8 personneltime,goods,andservicesusedforSTIcare,andtherentalcostwas apportionedbasedontheratioofSTIcasestoalloutpatientcases.Thisestimationdid notposesubstantialdifficulties. TheaverageexchangerateofIndianRupees(INR)45.95toaUS$forthe2003- 2004fiscalyearwasusedtoconverttheINRcosttoUS$[11]. Outputdata DetaileddatawereobtainedfromthewrittenmonthlysummaryrecordsoftheSTIclinics regardingtheservicesprovidedeverymonth.Theseincludedthenumberandtypeof STIsdetectedandtreated,totalpatientvisits,typeoftestsdone,treatmentgivenforSTIs, andnumberofpatientsreceivinginpatientservicesforSTIsandinpatientdays.Forarea hospitals,whereSTIcareisgenerallyprovidedaspartofthegeneraloutpatient,dataon visitsrelatedtoSTIsweretakenfromtheoutpatientrecordsthatspecifythediagnosesfor thevisits.Characteristicsofprogrammeoperationwereascertained,includingclient characteristicsandwhethertherewereanyhurdlestotheprovisionofservices. AteachSTIclinic,20patientsusingtheSTIserviceswereinterviewedregarding theirperceptionsabouttheseservices.Thefirstpatientsavailabletotheinvestigatorsat theSTIclinicduringtheweekofdatacollectionwereselectedforinterview,after obtainingverbalinformedconsentthatexplainedthepurposeoftheinterviewand assuredanonymityoftherespondent.Theinterviewswereconductedinaquietcorner outofthehearingrangeofothersinordertoencouragehonestresponses. Qualitycontrol Qualitycontrolmeasuresincludedathoroughpilotstudybeforecommencingformaldata collection,comprehensivetrainingofaqualifieddatacollectionteamincludingtheir conceptualunderstandingofalldataissues,fullback-upandjustificationforanydata recorded,supervisionofdatacollectionateachSTIclinicbytheprojectcoordinator, thoroughreviewbythestudyteamofthedataobtainedateachSTIclinic,andcontacting theSTIclinicsagaintoobtaininformationaboutdataissuesthatneededclarification afterthereview. Cost-efficiencyofSTIclinicsinIndia 9 Dataanalysis AnalysisofthedatawasdoneusingSPSSstatisticalsoftware.Theaverageeconomic costperpatientdiagnosedandtreatedforSTIwascalculatedasthemeasureofcost- efficiencyofeachSTIclinic.Therelationofthismeasureofefficiencywiththecost componentswasassessedthroughregressionanalyses.Incrementalcostsforinitialand follow-upvisitswereassessedusingamultipleregressionmodel.Thepersonnel,capital goodsandrentalcostswererelativelyfixedforeachSTIclinicirrespectiveofthevolume ofservicesprovided.Wethereforetreatedtheseasrelativelyfixedcosts,andtreatedthe recurrentgoodsandrecurrentservicesasvariablecostsforeachclinic.Inadditionto usingthenumberofinitialandfollow-upvisitsasindependentvariablesforeachSTI clinic,wealsousedthefixedcostsasavariableintherightsideoftheequationasthe constantalonemaynotestimatetheentirefixedcosts[12].Weusedthisregression modeltoassessthetotalcostfunction.AhighR2valuewouldbeexpectedforthis model,sincetheleftsideoftheequation(totalcost)wasdeterminedinourcosting largelybasedoninformationpresentintherightsideoftheequation.Thatis,fixedcosts areclinic-specificandamajorportionofvariablecosts(medicines)wascalculated assumingstandardinputsandcostspervisitacrossclinics.Thisisconsistentwiththe objectiveofourmodel,whichwastodeterminetheincrementalcostsforinitialand follow-upvisits. TheoptimalnumberofpatientsthatcouldhaveSTIdetectedandtreatedinayear wascalculatedusingthefollowingassumptions,whichwerebasedoninputofSTIclinic staffandpersonsfamiliarwiththeworkingoftheseSTIclinics: 1. ThenumberofSTIsdetectedandtreateddependedmainlyontheavailabilityoffull- timeequivalentsofthedoctorintheoutpatientoftheSTIclinic.Sincethemajority ofthepatientsattendingtheseclinicshaveSTIs,theproportionofinitialvisitsnot relatedtoSTIsissmall.Therefore,thelatterwouldnothaveasubstantialimpacton thenumberofSTIsthatcouldbedetectedandtreatedperfull-timeequivalentof doctor. 2. Anaverageoftwofollow-upvisitsforeachSTI,i.e.atotalofthreevisitsforeach STIdiagnosis,wouldbeusefulforqualitycareforpatientsattendingtheseclinics. Cost-efficiencyofSTIclinicsinIndia 10 3. Duringtheusual5workinghoursoftheSTIclinicoutpatientinaday,adoctorcould satisfactorilysee5patientswithnewSTIdiagnosisand10follow-ups,without compromisingquality. 4. Considering250workingdaysinayear,foreachdoctorfull-timeequivalent availableintheoutpatientclinic,1250newSTIscouldbediagnosedandtreated withoutcompromisingquality. TheeconomiccostforprovidingservicestotheoptimalnumberofSTIpatientswasthen estimated. ReliabledataonthenumberofSTIsdetectedandtreatedatoneofthesampled STIclinicslocatedinanareahospitalcouldnotbeobtained.Therefore,thisclinicwas excludedfromtheanalysisanddataarepresentedfor14ofthe15sampledSTIclinics. Cost-efficiencyofSTIclinicsinIndia 11 Results Atotalof18807diagnosesandassociatedtreatmentforSTIswerereportedbythe14STI clinicsinthe2003-2004fiscalyear,withamedianvalueof1199andmeanof1343 (Table1).ThisnumberwasgenerallyhighestatSTIclinicslocatedinthetertiary hospitalsofmedicalcolleges,followedbytheSTIclinicsindistrictheadquarter hospitals,andleastintheSTIclinicsinareahospitals.Thistrendwaslargelyduetothe differencesintherelativesizesofthecatchmentpopulationsforthesethreecategories. OfthetotalSTIsdetectedandtreated,59.2%wereinmales.Themaletofemaleratio washighestat3:2formedicalcollegehospitalSTIclinicsbutwasreverseinarea hospitalSTIclinics(Table1).Theratiooftotalout-patientvisitstothecasesofSTIs detectedandtreatedwasreportedtobehighestbythemedicalcollegehospitalSTI clinics(3.33)ascomparedwiththedistrictheadquarterhospitalSTIclinics(1.40)andfor theareahospitalSTIclinics(1.76),suggestingthatthenumberoffollow-upvisitsis relativelyhigherinthemedicalcollegeSTIclinics.Dataregardingtheexactreasonfor eachvisitwerenotavailable.However,itisestimatedthatthemajorityoftheinitial visitsintheseclinicsresultinSTIdiagnoses,andthemajorityofothervisitsarefollow- upvisitsrelatedtoSTItreatmentandcounselling. Allsampledmedicalcollegehospitalsandmostdistrictheadquartershospitalshad specialist(s)trainedinSTIsavailableforoutpatientsintheirSTIclinics,whereasmost areahospitalsdidnothaveaspecialistfortheirSTIclinics.ThediagnosisofSTIswas mostfrequentlymadebasedonclinicalassessmentwithoutcompletelaboratory investigations.InvestigationsweredonemostofteninSTIclinicsinmedicalcollege hospitals,infrequentlyindistrictheadquarterhospitals,andalmostneverinarea hospitals.ThediagnosesofSTIsreportedbytheseSTIclinicsareshowninTable2.Of theclassicSTIs,thediagnosisofherpeswasreportedtobethehighest.Consideringall diagnosestogether,thehighestreporteddiagnosiswasthemiscellaneouscategorythat includedmostlyscabiesandsomepediculosisandothers. Thetotaleconomiccostofservicesprovidedbythe14STIclinicsduringthe 2003-2004fiscalyearwasINR13,719,992(US$298,585),ofwhich9.9%wasfor inpatientservicesprovidedfor1106(5.9%)ofthetotal18807STIstreated.Ofthetotal cost,personnelmadeup76.2%,recurrentgoods10.2%(58.5%ofthiswasforSTI Cost-efficiencyofSTIclinicsinIndia 12 medicines),rentals9.9%,recurrentservices2.1%,andcapitalgoods1.7%.Therewere modestvariationsintheseproportionalcostsamongsomeoftheSTIclinics(Table3).If thefinancialcostswereconsidered,byexcludingthecostsforrentalsandcondom subsidy,thesewouldbe10.4%lessthantheeconomiccostsforthe14STIclinics combined. TheeconomiccostforeachpatientdetectedtohaveSTIandtreatedforitvaried 5-foldbetween13ofthe14STIclinicsfromINR225.5(US$4.91)toINR1201.5(US$ 26.15),withamedianofINR609.9(US$13.27)(Table3).TheremainingSTIclinichad anunusuallyhighcostperSTIdetectedandtreated(INR2478.5,US$53.94).The averagecostofdetectingandtreatingeachSTIforpatientsatallthe14STIclinics combinedwasINR729.5(US$15.88).Theaveragecostofdetectingandtreatingeach STIwas77%higherattheSTIclinicsinthemedicalcollegescomparedwiththoseat districtheadquarterandareahospitals(Table3). Sincepersonnelaccountedforthemajorportionofcost,anddoctorshadthe highestrelativesalaryamongpersonnel,therewasasignificantdirectnon-linearrelation betweenthenumberofSTIcasestreatedperdoctorfull-timeequivalentinayearandthe cost-efficiencyforeachSTIcasetreatedatthe14STIclinics,thebestfitforwhichwas obtainedwithapowerfunction(p<0.001,R2=0.81)asshowninFigure1.Ifdatafor theoneclinicwithanexceptionallyhighcostperSTItreatedwereexcluded(DHQ4), consideringitasanoutlier,thebestfitforthisrelationwasobtainedwithanexponential function(p<0.001,R2=0.81)asshowninFigure2. Consideringthe2003-2004fiscalyeardatafromeachSTIclinicasadatapoint, themultipleregressionmodelexplainedinthemethodssectionwasusedtoassesstotal economiccostasafunctionoffixedcosts,costofinitialvisitsinwhichSTIcaseswere detected(includingtreatmentcost),andcostofothervisits(mostlyfollow-upvisits), whichrevealedthefollowingrelation: Ccircumflex=3036.35+1.04X+55.57Y+3.75Z whereCcircumflexistotaleconomiccostinINR,Xisfixedcosts,Yisthenumberofinitialvisitsin whichSTIcasesweredetectedandtreatment/medicinesgiven,andZisthenumberof othervisits.Asexpected,thismodelhadahighR2valueofalmost1.00.Thefitofthe modelwassignificantatp<0.001(F=15906;degreesoffreedom:3forregression,10 Cost-efficiencyofSTIclinicsinIndia 13 forresidual,13total).Inthismodel,twovariables(fixedcostsandinitialvisits)were statisticallysignificantwhereasonevariable(follow-upvisits)wasnot(Table4).This modelsuggeststhatapartfromtheconstant/fixedcosts,theadditionalcostforeachinitial visitinwhichSTIwasdetectedandtreatmentgiven(includingcostoftreatment)was INR55.57(US$1.21)andtheadditionalcostforeachfollow-upvisitwasverysmallat INR3.75(US$0.08). Personnelat13ofthe14(92.9%)STIclinicsrespondedthatmorepatientscould beservedbytheirSTIclinicswiththeavailablepersonnelandinfrastructureiftherewere moredemand.Basedontheassumptionsmentionedinthemethodssectionaboutthe numberofSTIcasesthatcouldbedetectedandtreatedwithoutcompromisingqualityby eachdoctorinayeariftherewereoptimaldemand,4ofthe6medicalcollegeSTI clinics,2ofthe4districtheadquarterSTIclinics,and2ofthe4areahospitalSTIclinics couldincreasethenumberofSTIsdetectedandtreatedperyearsubstantiallyby54-287% withtheavailablepersonnelandinfrastructure(Table5).Ifthiswereachieved,the overallnumberofSTIcasesdetectedandtreatedbythe14STIclinicscouldincreaseby 38%from18807to25916.Thetotalcostforthiswouldincreaseonlyby3.5%fromthe 2003-2004costofINR13,719,992(US$298,585)toINR14,203,222(US$309,102), usingtheequationCcircumflex=3036.35+1.04X+55.57Y+3.75Zfromthemultipleregression modelmentionedpreviouslyforincrementalcosts. Inadditiontoinadequatedemandasthemajorhurdletoprovisionofservices,two STIclinicsmentionedinadequatesuppliesandtwoclinicsmentionedinadequatestaffing ashurdles.Thetwoclinicsthatmentionedinadequatesuppliesashurdle(AH1andAH3) hadthehighestcostperSTItreatedamongtheareahospitals(Table3)butthiswasclose tothemediancostperSTItreatedforallclinicsconsideredtogether.Oneoftheclinics thatmentionedinadequatestaffingashurdle(DHQ4)hadthehighestcostofallperSTI treatedandtheother(AH2)hadoneofthelowestcostsofallperSTItreated(Table3). Thesmallproportionofclinicsreportinghurdlesotherthaninadequatedemandprevent generalisationsabouttherelationofthesehurdlestoefficiency. Theproportionoffull-timeequivalentsforfemaledoctorsavailableintheSTI clinicsofmedicalcollegeswasmuchlowerthanintheSTIclinicsofdistrictheadquarters Cost-efficiencyofSTIclinicsinIndia 14 andareahospitals(Table5),whichwouldpartlyberesponsiblefortherelativelylower proportionoffemalepatientstreatedinthemedicalcollegeSTIclinics(Table1). ThepatientinterviewsattheSTIclinicsrevealed70.4%wereverysatisfiedwith theservicesprovided,andtheremainingweresomewhatsatisfied,withnoone mentioningthathe/shewasnotsatisfied(Table6). AllmedicalcollegehospitalSTIclinicsandtwodistrictheadquarterSTIclinics (DHQ2andDHQ4)wereoldSTIclinics(establishedsince1960s),andtheremaining werenewlydesignatedSTIclinicsin2002.Becausemajordifferencesintheservices andcost-efficiencywereobservedbetweenSTIclinicslocatedinthethreecategoriesof hospitals,andtheonlycategorythathadbotholdandnewclinicswasthedistrict headquarterhospitals,separateanalysisforoldandnewclinicswasnotdone. Cost-efficiencyofSTIclinicsinIndia 15 Discussion IntheAndhraPradeshstateofIndia,analysisof14publicsectorSTIclinicsrevealedthat theaveragecostofdiagnosingandtreatingeachSTIwasINR729.5(US$15.88),and thiscostrangedmany-foldbetweenthe14clinics.FixedcostsoftheSTIclinicsmadeup thepredominantproportion,withpersonnelcostsexceedingthree-fourthsofthetotal cost.Althoughitisunusualtothinkofpersonnelasfixedcosts,thisconceptapplieshere asthepersonnelwereemployedattheSTIclinicsregardlessoftheamountofservices delivered.Basedonestimatesofoptimalworkloadthatcouldbehandledbytheavailable personnel,weestimatedthatalargeproportionoftheseSTIclinicscoulddealwithmore STIcaseswithoutcompromisingquality,whichwouldincreasethenumberofSTIs treatedatthe14clinicsbyoverone-thirdthenumbertreatedinthe2003-2004fiscalyear ataminimaladditionalcostofabout3.5%forprovisionofservices.Thisdoesnot includetheadditionalcoststhatmaybeinvolvedinincreasingthedemandforthese services.Thesearedifficulttoassessandneedtobeassessedoveraperiodoftime,and mayormaynotbelarge.SincethepublicsectorSTIclinicspredominantlyservepoor patientsinIndiaatalmostnodirectcosttothepatients,itisimportantthatmechanisms bedevelopedtoutilisetheunusedcapacityintheseclinics.Thiswouldresultinmore poorpatientstobeservedandmoreefficientuseofthepublicresourcesavailableforSTI andHIVcontrol. Substantialdifferenceswerefoundintheservicesandcost-efficiencybetweenthe STIclinicsatthethreedifferenttypeoflocations,i.e.medicalcollege,district headquarterandareahospitals.ThenumberofSTIstreatedperfull-timedoctor equivalent,aswellasthecost-efficiencyperSTItreated,wasleastinthemedicalcollege hospitalSTIclinics.However,itisimportanttonotethatmedicalcollegeclinicsalso hadthehighestnumberoffollow-upvisitsperSTI,availabilityofSTIspecialists,and levelofinvestigationsfordiagnosis,suggestingthatthethoroughnessofSTIcarewould behigherattheseclinics.Inaddition,thetertiarymedicalcollegeclinicsareestimatedto getmorecomplicatedpatientsandalsoserveavitalroleintheteachingofresidentsand medicalstudents.Therefore,therelativelyhighercostperSTItreatedinmedicalcollege clinicsisbyitselfnotanegativething,ifitisassociatedwithprovisionofhigherquality careandcaretomorecomplicatedpatients.Attemptsatimprovingcost-efficiency Cost-efficiencyofSTIclinicsinIndia 16 shouldnotbeattheexpenseofqualityofcare.Wecalculatedtheoptimalworkloadthat couldbehandledbythepersonnelattheSTIclinicstakingintoaccounttheneedto maintainquality.Thehighestrelativeunusedcapacitywasinthemedicalcollege hospitalSTIclinics.Ontheotherhand,thelownumberoffollow-upvisitsatthedistrict headquarterandareahospitalSTIclinicssuggeststhatemphasisonencouragingfollow- upsandqualityofcareisneeded.Inaddition,theexceptionallyhighnumberofSTI patientstreatedattwodistrictheadquarterhospitalSTIclinics(DHQ1andDHQ2)and thehighnumberatoneareahospitalSTIclinic(AH2)perfull-timedoctorequivalent availablesuggeststhattheneedforensuringqualityofcaremustbeemphasised, assumingthatqualityisimpairedbyanexcessivecaseload. Therelativelylowproportionoffemalepatientstreatedatmedicalcollege hospitalSTIclinicscanperhapsbeattributedtothelowerproportionoffemaledoctors availableattheseclinicsascomparedwithdistrictheadquarterandareahospitalSTI clinics,andalsotothefactthatgynaecologyclinicsareavailableatmedicalcollege hospitalsthatareusuallystaffedbyfemaledoctorsandmanyfemalesgototheseclinics forSTI-relatedsymptoms.Anotherfactorfortherelativelyhigherproportionoffemale STIpatientsseenattheareahospitalsisthefactthattheSTIclinicsinthesehospitalsrun aspartofthegeneraloutpatientandnotasadistinctlyseparateclinic,whichencourages morefemalestoutilisetheseservices. InpatientservicesareusedbytheSTIclinicstoadmitpatientsofveryadvanced STIswhoarequitepoorandhavealowchanceofcomingbackforfollow-up.Although theseinpatientservicesmakeup10%ofthetotalcostfor6%ofthetotalSTIstreated, thisisgenerallyutilisedfortheSTIpatientswhoaremostinneedandwhoareathigh riskofnotcompletingthetreatmentandfollow-upintheabsenceoftheseservices,and therefore,maybejustified. ThereporteddistributionofSTIdiagnosesshouldbeinterpretedcautiouslyinthe backgroundthatinvestigationsaredonealmostneveratareahospitalclinicsand infrequentlyatdistrictheadquarterhospitalclinics,andthatSTIspecialistsaremostlynot availableatareahospitalclinics.Theveryhighproportionofscabiesreportedbythese publicsectorSTIclinicsisgenerallysuggestiveoftheverylowsocioeconomicstatusof thepatientsutilisingtheservicesoftheseclinics. Cost-efficiencyofSTIclinicsinIndia 17 Anencouragingfindinginthisstudywasthatofthepatientsinterviewedatthe STIclinicsovertwo-thirdwereverysatisfiedwiththeclinicservicesandtheremaining somewhatsatisfied,withnorespondentreportingdissatisfaction.Thisimpliesthatonce thepatients,whoaremostlypoor,reachtheSTIclinicanduseitsservices,theyare generallysatisfiedwiththeservices.Animportantfactorinthisperceptionislikelytobe thefactthatfreemedicinesaregivenfortreatment. Considerationoftheabovementionedissueswouldhavetobekeptinmindwhile planningmethodstoenhanceutilisationoftheunusedcapacityinthepublicsectorSTI clinicsinthisIndianstatesuchthatthequalityofservicesisalsoadequate.Onemajor issuethatcomesinthewayofmoredemandforservicesoftheseclinicsisthestigmathat isattachedwithattendinganSTIclinic,whichiscommonlyperceivedbythesocietyto beassociatedwithpoorpersonalcharacterorothernegativeattributes.Anoptionthat couldbeexploredtoreducethisstigmawouldbetogiveamoreinnocuousnametothese clinics,suchas,familyhealthclinics,andprovideSTIservicesaspartofa comprehensivesexualhealthpackage.Suchanapproachwouldrequiresome restructuring,butmaybeworthconsideringasalong-termsolutionforbetterutilisation ofSTIservicesinthepublicsectorhospitals.Anotherissuethatneedstobeaddressedto increasetheutilisationofservicesatpublicsectorSTIclinicsisthedevelopmentof systematiclinkageswithnon-governmentalorganisationsworkingwithhigh-riskgroups suchassexworkers,suchthattheycaneasilyavailtheservicesoftheseSTIclinics. ThemajorityofoutpatienthealthcareinIndia,includingSTItreatment,is providedbytheprivatesector[13].Itisalsowidelybelievedthatthepatientswhouse thepublicsectorSTIclinicsarepredominantlypoorandwithrelativelyadvancedSTIs thathaveoftenbeenunsuccessfullytreatedelsewhere.Inthisbackground,thecostand efficiencyanalysispresentedinthispaperpertainstothismostvulnerablegroup.The highrateof16-30%HIVinthepublicsectorSTIclinicattendeeswhoparticipatedinthe sentinelsurveillanceinAndhraPradeshoverthepastfewyears[4]supportsthe interpretationthatthesepatientsarehighlyvulnerable.Itispossiblethatthecost- efficiencytrendsfortheprivatesectorSTIcareinIndiamaybedifferentfromthosein thepublicsector.However,itmaynotbeunreasonabletoassumethatthecost-efficiency estimatesofpublicsectorSTIcareapplytothosewhoareatmostriskofacquiringor Cost-efficiencyofSTIclinicsinIndia 18 transmittingHIV,andtherefore,areimportantforassessingandcomparingthecost- efficiencyandcost-effectivenessofthevariousHIVpreventionstrategies. EstimatesofcostandefficiencyofSTIservicesinIndia,andtheireffectivenessin preventingHIV,havenotbeenpreviouslyreadilyavailablefromIndia[14].Infact,the costestimatesofSTItreatmentfromtheMwanzastudyinTanzania[15]wereused previouslybytheWorldBanktoestimatetheSTIcareresourcesneededinIndiawhile preparingtheloanagreementforsupportingIndia’sAIDScontrolprogramme[16].The cost-efficiencyestimatespresentedinthispapercouldbeusedtoestimatetheresources neededforSTIcarebythepublichealthcaresysteminAndhraPradesh,andthesedata alsopointtosomeissueswhich,ifaddressed,couldleadtoenhancementintheprovision oftheseservicesandalsotheircost-efficiency.Thesecostandefficiencyestimatesof STItreatmentcouldalsobeusedtoestimatecost-effectivenessofSTItreatmentforHIV preventioninIndiaeitherbyusingpublishedestimatesofeffectivenessfromotherparts oftheworldorbyconductingstudiesoneffectivenessofSTItreatmentinpreventing HIVinIndia.Inaddition,thesecost-efficiencyestimatesofSTItreatmentwouldbe comparedwithsimilarestimatesfromothercountriesinthePANCEAstudythatare usingsimilarmethodology[10],andalsowithestimatesforotherHIVprevention strategiesinAndhraPradesh,suchasHIVvoluntarycounsellingandtesting[17]and HIVpreventionprogrammesforfemalesexworkers[18]usingsimilarmethodology. Suchstandardisedcross-countryandcountry-specificcost-efficiencyestimateswouldbe veryusefulinestimatingtheresourcesneededforHIVcontrolinspecificcountriesas wellasgloballywithmoreconfidencethanhasbeenpossiblesofarwiththeavailable data.Thisisimportantinthebackgroundoftheextensivedebateinrecenttimesabout theresourcesneededtocontrolHIV. Cost-efficiencyofSTIclinicsinIndia 19 Conclusions IntheIndianstateofAndhraPradesh,thereisun-utilisedcapacityinthepublicsector STIclinicsthatprovideservicespredominantlytothepoor.Effortstofacilitate utilisationofthiscapacitywouldbeuseful,asthiswouldenablemorepoorpatientswith STIstobeservedatminimaladditionalcost,andwouldalsoreducethecostperSTI treatedleadingtomoreefficientuseofpublicresources.Comprehensiveanddynamic analysisoftheefficiencyofHIVpreventionservicesusingstandardisedmethodsis necessarytomakeoptimaluseoftheincreasingresourcesthatarebecomingavailablefor thispurposeinIndiaandotherpartsofthedevelopingworld. Cost-efficiencyofSTIclinicsinIndia 20 Competinginterests Theauthorsdeclarethattheyhavenocompetinginterests. Authors’contributions LDledthePANCEAstudyinIndia,guidedthedesign,datacollectionandanalysis,and wrotetheinitialdraftofthispaper.PScontributedtothedesign,datacollectionand analysis.TLNPandEMcontributedtothedesignandadvisedondatacollection, analysisandpresentation.MCR,AAK,SGPK,YKRandSMcontributedtodata collectionandanalysis.MOcontributedtothedesignandadvisedonanalysis.JGK oversawthePANCEAdesignandcontributedtotheanalyticdesignandpresentation. Allauthorsreadandapprovedthefinalmanuscript. Acknowledgments WethanktheAndhraPradeshStateAIDSControlSocietyforfacilitatingthisstudy,and thestaffandclientsoftheSTIclinics/hospitalsforparticipatinginthisstudy.Thisstudy wassupportedbytheWorldBankcontract7130248andbytheU.S.NationalInstitutesof HealththroughTaskOrder#7contract282-98-0026andgrantR01DA15612.Theviews expressedinthispaperarethoseoftheauthorsanddonotnecessarilyreflecttheviewsof thefundingagencies,organisationsthatfacilitatedthisstudy,ortheinstitutions employingtheauthors. Presentationatmeeting ThispaperwaspresentedattheAnnualMeetingoftheGlobalForumforHealth Research–Forum9,Mumbai,India,12-16September2005. Cost-efficiencyofSTIclinicsinIndia 21 References [1] JointUnitedNationsProgrammeonHIV/AIDS(UNAIDS):2004Reportonthe GlobalAIDSEpidemic.Geneva;2004. [2] NationalAIDSControlOrganisation(NACO):HIVEstimates(2003).New Delhi:NACO,MinistryofHealth&FamilyWelfare,GovernmentofIndia. [http://www.nacoonline.org/facts_hivestimates.htm] [3] SanganiP,RutherfordG,WilkinsonD:Population-basedinterventionsfor reducingsexuallytransmittedinfections,includingHIVinfection.Cochrane DatabaseSystRev2004;(2):CD001220. [4] NationalAIDSControlOrganisation(NACO):ObservedHIVprevalencelevels statewise:1998-2004.NewDelhi:NACO,MinistryofHealth&FamilyWelfare, GovernmentofIndia.[http://www.nacoonline.org/facts_statewise.htm] [5] RaoJVRP,GangulyNK,MehendaleSM,BollingerRC:India’sresponsetothe HIVepidemic.Lancet2004;364:1296-1297. [6] NationalAIDSControlOrganisation(NACO):NationalAIDScontrol programmephaseII(1999-2006).NewDelhi:NACO,MinistryofHealth& FamilyWelfare,GovernmentofIndia. [http://www.nacoonline.org/abt_phase2.htm] [7] Bill&MelindaGatesFoundation:Avahan:IndiaAIDSInitiative. [http://www.gatesfoundation.org/GlobalHealth/Pri_Diseases/HIVAIDS/HIVProgr amsPartnerships/Avahan.htm] [8] DandonaL:EnhancingtheevidencebaseforHIV/AIDScontrolinIndia.Natl MedJIndia2004;17:160-166. [9] PreventAIDSNetworkforCost-EffectivenessAnalysis:Homepage. [http://hivinsite.ucsf.edu/InSite?page=pancea] [10] MarseilleE,DandonaL,SabaJ,McConnelC,RollinsB,GaistP,LundbergM, OverM,BertozziS,KahnJG:AssessingtheefficiencyofHIVprevention aroundtheworld:methodsofthePANCEAproject.HealthServRes2004; 39:1993-2012. [11] ReserveBankofIndia:ExchangerateoftheIndianRupee:Table145. [http://www.rbi.org.in/scripts/PublicationsView.aspx?Id=6709] Cost-efficiencyofSTIclinicsinIndia 22 [12] KleinLR:AnIntroductiontoEconometrics.EnglewoodCliffs,USA:Prentice- Hall;1977. [13] MahalA,SinghJ,AfridiF,LambaV,GumberA,SelvarajuV:Whobenefitsfrom publichealthspendinginIndia:resultsofabenefitincidenceanalysisforIndia. NewDelhi:NationalCouncilofAppliedEconomicResearch;2002. [14] WalkerD:Costandcost-effectivenessofHIV/AIDSpreventionstrategiesin developingcountries:isthereanevidencebase?HealthPolicyPlan2003; 18:4-17. [15] GilsonL,MkanjeR,GrosskurthH,MoshaF,PicardJ,GavyoleA,ToddJ, MayaudP,SwaiR,FransenL,MabeyD,MillsA,HayesR:Cost-effectivenessof improvedtreatmentservicesforsexuallytransmitteddiseasesinpreventing HIV-1infectioninMwanzaRegion,Tanzania.Lancet1997;350:1805-1809. [16] WorldBank:Projectappraisaldocumentonaproposedcreditintheamountof SDR140.82milliontoIndiaforsecondnationalHIV/AIDSproject.Washington, DC;1999. [17] DandonaL,SisodiaP,RameshYK,KumarSGP,KumarAA,RaoMC, SomeshwarM,HanslB,MarshallN,MarseilleE,KahnJG:Costandefficiency ofHIVvoluntarycounsellingandtestingcentresinAndhraPradesh,India. NatlMedJIndia2005;18:26-31. [18] DandonaL,SisodiaP,KumarSGP,RameshYK,KumarAA,RaoMC,Marseille E,SomeshwarM,MarshallN,KahnJG:HIVpreventionprogrammesfor femalesexworkersinAndhraPradesh,India:outputs,costandefficiency. BMCPublicHealth2005;5:98.[http://www.biomedcentral.com/1471- 2458/5/98/] Cost-efficiencyofSTIclinicsinIndia 23 Legendsforfigures. Figure1.RelationshipbetweenSTIstreatedperdoctorfull-timeequivalentandthecost perSTItreated(p<0.001,powerfunction).FTEisfull-timeequivalent,INRisIndian Rupee. Figure2.RelationshipbetweenSTIstreatedperdoctorfull-timeequivalentandthecost perSTItreatedafterexcludingoneextremevaluethatcouldbeconsideredanoutlier(p< 0.001,exponentialfunction).FTEisfull-timeequivalent,INRisIndianRupee. Cost-efficiencyofSTIclinicsinIndia 24 Table1.NumberofSTIstreatedatSTIclinicsinthefiscalyear2003-2004. STIsdiagnosedandtreated Male FemaleSTIclinic Total number Number Percent Number Percent Total visits Ratioof totalvisits toSTIs treated Medicalcollegehospitals MC1 2784 1827 65.6 957 34.4 9093 3.27 MC2 1078 792 73.5 286 26.5 7260 6.73 MC3 1921 1189 61.9 732 38.1 7204 3.75 MC4 2186 1514 69.3 672 30.7 7064 3.23 MC5 1743 1333 76.5 410 23.5 4638 2.66 MC6 2223 1255 56.5 968 43.5 4543 2.04 AllMC 11935 7910 66.3 4025 33.7 39802 3.33 Districtheadquarterhospitals DHQ1 1759 851 48.4 908 51.6 2058 1.17 DHQ2 1320 788 59.7 532 40.3 1365 1.03 DHQ3 1012 447 44.2 565 55.8 1530 1.51 DHQ4 323 257 79.6 66 20.4 1220 3.78 AllDHQ 4414 2343 53.1 2071 46.9 6173 1.40 Areahospitals AH1 551 76 13.8 475 86.2 650 1.18 AH2 582 253 43.5 329 56.5 1215 2.09 AH3 558 220 39.4 338 60.6 631 1.13 AH4 767 331 43.2 436 56.8 1842 2.40 AllAH 2458 880 35.8 1578 64.2 4338 1.76 Total 18807 11133 59.2 7674 40.8 50313 2.68 Cost-efficiencyofSTIclinicsinIndia 25 Table2.DistributionofreportedSTIdiagnoses. No. TypeofSTI Number Percent 1 Herpes 2507 13.3 2 Chlamydialinfection 2212 11.8 3 Gonorrhea 1389 7.4 4 Nonspecificgenitalulcer 1249 6.6 5 Nonspecificvaginaldischarge 1139 6.1 6 Candidiasis 1123 6.0 7 Syphilis 1007 5.4 8 Genitalwarts 808 4.3 9 Trichomonas 643 3.4 10 Chancroid 586 3.1 11 Lymphogranulomavenereum 228 1.2 12 Bacterialvaginosis 225 1.2 13 Donovanosis 160 0.9 14 Miscellaneousincludingscabiesand pediculosis 5532 29.4 Total 18807 100 Cost-efficiencyofSTIclinicsinIndia 26 Table3.EconomiccostofSTIclinicsinthe2003-2004fiscalyear. Totaleconomic cost Percentofeconomiccost CostperSTI treatedSTI clinic* INR US$ Personnel Recurrentgoods Rentals Recurrentservices Capitalgoods Percent costdue to inpatient services† Number ofSTIs treated INR US$ Medicalcollegehospitals MC6 1283694 27937 76.5 12.2 6.2 2.7 2.4 8.3 2223 577.5 12.6 MC5 1345786 29288 82.7 7.4 6.7 1.7 1.6 2.6 1743 772.1 16.8 MC4 1746668 38012 75.0 10.0 11.3 1.8 1.9 9.8 2186 799.0 17.4 MC3 1569840 34164 80.6 9.7 6.0 2.0 1.7 5.3 1921 817.2 17.8 MC2 1061341 23098 66.8 8.8 19.2 3.0 2.1 10.2 1078 984.5 21.4 MC1 3345020 72797 79.8 8.0 9.9 1.2 1.1 17.2 2784 1201.5 26.1 AllMC 10352349 225296 77.7 9.1 9.6 1.9 1.6 10.4 11935 867.4 18.9 Districtheadquarterhospitals DHQ1 396627 8632 63.9 25.3 6.4 2.9 1.4 11.0 1759 225.5 4.9 DHQ2 349816 7613 62.5 22.7 8.2 3.4 3.3 15.0 1320 265.0 5.8 DHQ3 612111 13321 73.2 9.8 14.1 1.8 1.1 9.1 1012 604.9 13.2 DHQ4 800543 17422 75.3 4.9 16.7 1.8 1.3 5.2 323 2478.5 53.9 AllDHQ 2159097 46988 70.5 12.9 12.7 2.2 1.6 9.0 4414 489.1 10.6 Areahospitals AH2 198736 4325 60.2 18.4 11.9 5.9 3.6 14.2 582 341.5 7.4 AH4 329889 7179 71.5 14.4 9.0 2.8 2.2 9.7 767 430.1 9.4 AH3 340348 7407 78.7 11.2 5.4 3.0 1.7 6.2 558 609.9 13.3 AH1 339572 7390 76.0 14.2 5.0 3.0 1.7 0.6 551 616.3 13.4 AllAH 1208545 26301 72.9 14.1 7.4 3.4 2.2 6.9 2458 491.7 10.7 Total 13719992 298585 76.2 10.2 9.9 2.1 1.7 9.9 18807 729.5 15.9 INRisIndianRupee 1US$=INR45.95(averageexchangerateinthe2003-2004fiscalyear)[11] *STIclinicsarrangedwithineachhospitaltypecategoryindecreasingorderofcost-efficiency, i.e.increasingorderofcostperSTItreated. †ThisinpatientcostdoesnotincludecostofmedicinesfortreatmentofSTIs,whichwould havebeenincurredevenifthesepatientsweretreatedasoutpatient. Cost-efficiencyofSTIclinicsinIndia 27 Table4.Coefficientsinthemultipleregressionmodelandtheirsignificance. Variable Coefficient(betai) Standarderror t Significance Constant 3036.35 8049.64 0.377 0.714 Fixedcosts 1.04 0.01 95.761 0.000 VisitsinwhichSTIwasdetected 55.57 8.28 6.712 0.000 Othervisits(mostlyfollow-up) 3.75 2.88 1.300 0.223 Dependentvariable:Totalcost Cost-efficiencyofSTIclinicsinIndia 28 Table5.OptimalnumberofSTIsthatcouldbetreatedattheSTIclinics. Doctorfull-timeequivalentsavailable STI clinic* Total Male(%) Female(%) Number ofSTIs treatedin fiscalyear 2003-2004 Optimal numberof STIsthat couldbe treated† Percent increaseif optimal number treated Medicalcollegehospitals MC6 1.20 1.20(100) 0(0) 2223 1500 MC5 1.50 1.25(83.3) 0.25(16.7) 1743 1875 8 MC4 3.00 1.50(50.0) 1.50(50.0) 2186 3750 72 MC3 2.46 1.96(79.7) 0.50(20.3) 1921 3075 60 MC2 1.50 1.50(100) 0(0) 1078 1875 74 MC1 5.50 5.00(90.9) 0.50(9.1) 2784 6875 147 AllMC 15.16 12.41(81.9) 2.75(18.1) 11935 18950 59 Districtheadquarterhospitals DHQ1 0.40 0.40(100) 0(0) 1759 500 DHQ2 0.33 0(0) 0.33(100) 1320 413 DHQ3 1.25 0.25(20.0) 1.00(80.0) 1012 1563 54 DHQ4 1.00 1.00(100) 0(0) 323 1250 287 AllDHQ 2.98 1.65(55.4) 1.33(44.6) 4414 3725 Areahospitals AH2 0.20 0.20(100) 0(0) 582 250 AH4 0.58 0.33(56.9) 0.25(43.1) 767 725 AH3 1.00 0.70(70.0) 0.30(30.0) 558 1250 124 AH1 0.81 0.31(38.5) 0.50(61.5) 551 1016 84 AllAH 2.59 1.54(59.5) 1.05(40.5) 2458 3241 32 Total 20.73 15.60(75.3) 5.13(24.7) 18807 25916 38 *STIclinicsarrangedwithineachhospitaltypecategoryindecreasingorderofcost- efficiency,i.e.increasingorderofcostperSTItreated. †Basedontheassumptionthatonefull-timeequivalentdoctorcouldsatisfactorilytreat1250 STIsinayear,asdescribedinthemethodssection. Cost-efficiencyofSTIclinicsinIndia 29 Table6.SatisfactionofpatientswithservicesprovidedbySTIclinics. Very satisfied Somewhat satisfied NotsatisfiedSTIclinic Total numberof patients interviewed No. % No. % No. % Medicalcollege hospitals 120 79 65.8 41 34.2 0 0 Districtheadquarter hospitals 80 57 71.2 23 28.8 0 0 Areahospitals 80 61 76.2 19 23.8 0 0 Total 280 197 70.4 83 29.6 0 0 y=85264x-0.702 R2=0.81 0 500 1000 1500 2000 2500 3000 0 1000 2000 3000 4000 5000 STIcasestreatedperdoctorFTEinayear CostperSTItreatedinINR y=966.58e-0.0003x R2=0.81 0 250 500 750 1000 1250 1500 0 500 1000 1500 2000 2500 3000 3500 4000 4500 5000 STIcasestreatedperdoctorFTEinayear CostperSTItreatedinINR Figure2