long-run effects of temporary incentives on medical care productivity€¦ · ·...
TRANSCRIPT
Policy Research Working Paper 7348
Long-Run Effects of Temporary Incentives on Medical Care Productivity
Pablo CelhayPaul Gertler
Paula GiovagnoliChristel Vermeersch
Health Nutrition and Population Global Practice GroupJune 2015
WPS7348P
ublic
Dis
clos
ure
Aut
horiz
edP
ublic
Dis
clos
ure
Aut
horiz
edP
ublic
Dis
clos
ure
Aut
horiz
edP
ublic
Dis
clos
ure
Aut
horiz
ed
Produced by the Research Support Team
Abstract
The Policy Research Working Paper Series disseminates the findings of work in progress to encourage the exchange of ideas about development issues. An objective of the series is to get the findings out quickly, even if the presentations are less than fully polished. The papers carry the names of the authors and should be cited accordingly. The findings, interpretations, and conclusions expressed in this paper are entirely those of the authors. They do not necessarily represent the views of the International Bank for Reconstruction and Development/World Bank and its affiliated organizations, or those of the Executive Directors of the World Bank or the governments they represent.
Policy Research Working Paper 7348
This paper is a product of the Health Nutrition and Population Global Practice Group. It is part of a larger effort by the World Bank to provide open access to its research and make a contribution to development policy discussions around the world. Policy Research Working Papers are also posted on the Web at http://econ.worldbank.org. The authors may be contacted at [email protected].
The adoption of new clinical practice patterns by medical care providers is often challenging, even when the patterns are believed to be efficacious and profitable. This paper uses a randomized field experiment to examine the effects of temporary financial incentives paid to medical care clinics for the initiation of prenatal care in the first trimester of pregnancy. The rate of early initiation of prenatal care was 34 percent higher in the treatment group than in the control
group while the incentives were being paid, and this effect persisted at least 15 months and likely 24 months or more after the incentives ended. These results are consistent with a model where the incentives enable providers to address the fixed costs of overcoming organizational inertia in innova-tion, and suggest that temporary incentives may be effective at motivating improvements in long-run provider perfor-mance at a substantially lower cost than permanent incentives.
Long‐RunEffectsofTemporaryIncentivesonMedicalCareProductivity
PabloCelhayPaulGertler
PaulaGiovagnoliChristelVermeersch
JELClassification:I12,I13,I15,I18Keywords:Keywords:Pay‐for‐performance, results‐based financing,providerperformance,birthoutcomes,impactevaluation,maternalandchildhealth,organizationalinertia,temporaryincentivesAuthorAffiliation:PabloCelhay([email protected])isaPh.D.candidateattheUniversityofChicago.PaulGertler([email protected])istheLiKaShingProfessorattheUniversityofCalifornia,Berkeley.PaulaGiovagnoli([email protected])isanEconomistwiththeWorldBank.ChristelVermeersch([email protected])isaSeniorEconomistattheWorldBank.
2
Acknowledgements:Theexperimentdescribedinthispaperwasdevelopedundertheleadershipof Martin Sabignoso, National Coordinator of Plan Nacer and Humberto Silva, National Head ofStrategicPlanningofPlanNacer,MinistryofHealth,Argentina.Togetherwiththenationalteam,LuisLopez Torres and Bettina Petrella from the Misiones Office of Plan Nacer oversaw theimplementation of the pilot, facilitated access to provincial data, supported the authors ininterpretingdatasetsandtheprovinciallegalframeworkandincarryingoutthein‐depthinterviews.Fernando Bazán Torres, Ramiro Flores Cruz, Santiago Garriga, Alfredo Palacios, Rafael Ramirez,SilvestreRiosCenteno,GabrielaMoreno,andAdamRossprovidedexcellentassistanceandprojectmanagementsupport.AlvaroS.Ocariz,JavierMinskyandthestaffoftheInformationTechnologyunitat Central Implementation Unit (UEC) at the Ministry of Health provided valuable support inidentifyingsourcesofdata.TheauthorsacknowledgethecontributionsofSebastianMartinez,LuisPerezCampoy,VaninaCamporealeandDanielaRomerointheinitialdesignofthepilot.Theauthorsalso thankNedAugenblick,DanBlack,NickBloom,MeganBusse, StefanoDellaVigna,DamiendeWalque,EmanuelaGalasso, JeffGrogger,PetraVergeer, aswell asparticipants in seminars atUCBerkeley, Northwestern University and Chicago University for helpful comments. The authorsgratefullyacknowledgefinancialsupportfromtheHealthResultsInnovationTrustFund(HRITF)andtheStrategicImpactEvaluationFund(SIEF)oftheWorldBank.Theauthorsdeclarethattheyhavenofinancialormaterialinterestsintheresultsofthispaper.
1
1 INTRODUCTION
Successfulorganizationsareabletoefficientlyandreliablyproducehighqualityproductsthrough
theuseofreproducibleandstableroutines.1Routinesshapetheproductionprocessbydefiningeach
person’sroleandtheirpatternsofaction,andbycoordinatingthetasksperformedbythedifferent
teammembers.2 They can be thought of as organizational habits that reduce the complexity of
decision‐making,facilitatecoordinationacrossteammembers,andspeedproduction.However,once
established, routines are costly to change. The cost of adjustment includes the time andmoney
neededtoretoolroutines,anadjustmentperiodinwhichproductionislessreliablewhilethenew
routinesarebeinglearned,andpossiblypsychologicalresistancetochange.Asaresult,organizations
tendtoberesistanttoadoptingstructuralchangesthatarethoughttobeproductiveandprofitable
(HannanandFreeman1984;CarrollandHannan2000).Whileorganizationalroutinesarenecessary
forefficientandreliableproduction,theycanresultinorganizationalinertiatoinnovation.
Nowhereareorganizationalroutinesmoreimportantthanintheproductionofmedicalcare
services(Hoff2014).Medicalcareentailscoordinatingalarge,complexsetoftaskssuchasdeciding
what information to collect from the patient, assessing social and medical risks, deciding what
diagnostic tests to prescribe, interpreting symptoms and test results, and prescribing and
implementing treatments.3 Typically, a teamcoordinatedbyaphysician implements these tasks.
Nursesoftentakemedicalandsocialhistories,conductpreliminaryphysicalexams,andadminister
injections.Laboratorytechniciansanalyzebloodandurine.Pharmacistsdispensedrugsandmonitor
negative drug interactions. Physical and occupational therapists provide rehabilitation services.
Community health‐workers provide outreach, promotion and preventive services, and follow‐up
care to patients. Clinics establish practice routines that are consistent with their training and
experiencetostandardizeandcoordinatecare.
There is substantialevidenceoforganizational inertia inmedical careas indicatedby the
remarkablylowlevelofcompliancewithClinicalPracticeGuidelines(CPGs)worldwide(Figure1).
1OrganizationalroutinehasbeenstudiedextensivelysincepopularizedbyNelsonandWinter(1982).Ina
review of the literature Becker (2004) defines routines as “recurrent interaction patterns” within anorganization,oras“establishedrules,orstandardoperatingprocedures”.
2 Often relationships between team members and management are enforced by informal relationalcontracts(GibbonsandHenderson2012and2013).
3 Complex production technologies with sophisticated routines such as medical care require strongmanagementtobeefficientandproductive. Bloometal. (2014)provideevidencethatbettermanagementincreasespublichospitalproductivity.
2
CPGsdefinemedicalcareproductionpossibilityfrontiersinthattheyprescribetheclinicalcontent
ofcarethatmaximizesthelikelihoodofsuccessfulhealthoutcomesbasedonmedicalscience,clinical
trials,andpractitionerconsensus.LocalCPGsareregularlyupdatedandserveasthebasisoftraining
inmedicalschoolsandpractitionerrefreshercourses.WhilethelackofcompliancewithCPGsmay
inpartreflectalackofknowledge,evidenceshowsthatpractitionersoftenprovideastandardofcare
wellbelowtheirlevelofknowledgeofCPGs.4Inasystematicreviewoftheliteratureonreasonsfor
non‐complianceofCPGs,Cabanaetal. (1999)report thatresistancetochangingexistingpractice
patternsisoneofthemostimportantbarrierstoCPGadherence.Forexample,GrolandGrimshaw
(2003)surveyednursesanddoctorsintheUKabouttheadoptionofnewhandhygieneguidelines.
Forty‐ninepercentrespondedthatresistancetochangingoldroutineswasanobstacletocomplying
withnewguidelines.5
Changingdeep‐rootedhabitsishardandevensmallcostsofadjustmentmayinhibitchanges
in favor of maintaining the status quo, (DellaVigna 2009; Thaler and Sunstein 2009).6 In these
circumstances,temporaryincentivesmayspeedadoptionbyhelpingtocompensateprovidersfor
theinitialfixedcostsofchangingtheirpracticepatternroutines.Thisamountstopayingprovidersa
time‐limitedperunitincentivefortheprovisionofacomponentoftheCPGsforaspecificcondition.7
Theuseoftemporaryincentivestoovercomeorganizationalinertiainfirmsissimilarinspirit
to theuseof temporary incentives to change individual and consumerbehavior. Firmsoftenuse
temporarypricediscounts,suchassalesandcoupons,tomarkettheirproducts(BlattbergandNeslin
1990;KirmaniandRao2000;andDupas2014).Discountsencourageindividualstopurchasegoods
thattheyarenotinthehabitofbuyingwhichinturnallowthemtoupdatetheirbeliefsaboutthe
product’sbenefits.Similarly,temporaryincentiveshavebeenusedtotrytohelpindividualsdevelop
betterhealthhabits suchas exercise andquitting smoking.8Recently, temporary incentiveshave
4 SeeDasandHammer (2005);DasandGertler (2007);Das,HammerandLeonard (2008);Barberand
Gertler(2009);LeonardandMasatu(2010);GertlerandVermeersch(2012);andMonahan,M.etal.(2015).5FormoreevidenceoforganizationalinertiaservingasabarriertoCPGcomplianceseeGrol(1990);Hudak,
O’DonnellandMazyrka(1995);Main,CohenandDiClemente(1995);andPathmanetal.(1996).6Weuseadifferentdefinitionofhabitsthanthebehavioraleconomicsliteraturewherehabitsarebasedon
theaddictionofmodelsofBeckerandMurphy(1988).Instead,werelyonthenotionsoffastandslowthinkingdiscussedinKahneman(2012)wheretasksperformedbasedonfastthinkingbecomehabits.
7Payinganupfrontlumpsumamountisanotheroption.However,itmaybehardertoensureandverifytheactual change in practice patterns. By paying based on actual performance the incentives also include acommitmentdeviceforcompliance.
8SeeforexampleVolppetal.(2008);Volppetal.(2009);CharnessandGneezy(2009);Johnetal.(2011);Royeretal.(2012);CawleyandPrice(2013);andAclandandLevy(2015).
3
beenusedtostimulatelong‐termsavingsintheformofinitiallyhighinterestratesandprice‐linked
savingsorlotteries(Gertleretal.2015,andSchaner2015).Toourknowledge,ourstudyisthefirst
to use a field experiment to examine the effects of temporary incentives on long‐run firm
performance.
Wetesttheeffectsoftemporaryincentivespaidtoclinicsforearlyinitiationofprenatalcare
usingafieldexperimentconductedwithPlanNacer,anArgentinegovernmentprogramthatprovides
health insurance tootherwiseuninsuredpregnantwomenandchildren.9Prenatal carebyskilled
healthprofessionalsbeginninginthefirsttrimesterofpregnancyisessentialforgoodmaternaland
newborn health outcomes, and is part of standardmedical training throughout theworld (WHO
2006).Throughearly initiationofcare,providersareable todetectandcorrect importanthealth
conditionssuchasinfectionsoranemiabeforetheyjeopardizematernalornewbornoutcomesas
wellasadvisemothersonproperprenatalnutritionandpreventionactivities(Schwarczetal.2001;
Carrolietal.2001aand2001b;CampbellandGraham2006).Despitetheserecommendationsand
the scientific evidence, take‐upof early initiationofprenatal care remains lowworldwide (WHO
2014).
Thefieldexperimentrandomizedtemporaryfinancialincentivestohealthcareclinicsinwhich
treatmentclinicswerepaida200%premiumforearlyinitiationofprenatalcare,i.e.beforeweek13.
Wefindthattherateofearlyinitiationofprenatalcarewas34%higherinthetreatmentgroupthan
in thecontrolgroup(0.42versus0.31)while the incentiveswerebeingpaid,andthat thehigher
levelsofearly initiationofprenatalcare inthetreatmentgrouppersistedat least15monthsand
likelymore than 24months after the incentives ended.We document that clinics changed their
routines by developing strategies to identify likely pregnant women and expanding the role of
communityhealthworkerstofindpregnantwomenandencouragethemtostartcareearly,andthat
thesechangesinroutinesalsopersistedatleast15monthsaftertheincentivesended.Despitethe
largeeffectof the incentivesonearly initiationof care,we findnoevidenceof aneffectonbirth
outcomes.
Ourresultsmayexplainthemechanismbehindrecentevidencethatpermanentperformance
incentives do indeed improve both quality and quantity of care.10 The standard neoclassical
9In2013,PlanNacerwasexpandedtootherpopulationsandrenamedProgramaSumar.10SeeforexampleBasingaetal.(2011);Floresetal.(2013);Bonfreretal.(2013);DeWalqueetal.(2015);
GertlerandVermeersch(2013);Gertleretal.(2014);andHuilleryandSeban(2014).MillerandBabiarz(2013)provideareview.
4
explanationisthatprovidersarereallocatingtheireffortacrossservicesinresponsetotheincreased
profit opportunities.11 However, previous studies have been unable to distinguish between this
mechanismandorganizational inertia.Onewaytodistinguishbetweenthetwomechanismsis to
observewhathappenswhenincentivesareremoved.Whiletheincentivesareinplaybothmodels
predict a positive response.However, once the incentives are removed, practice patterns should
reverttopriorlevelsinthestandardmodelsbutcontinueatthehigherlevelsunderorganizational
inertia.
Understanding themechanism bywhich financial incentiveswork is not only scientifically
interesting,butalsopolicyrelevant.Iftemporaryfinancialincentivesareabletoinduceprovidersto
adoptpermanentchangestotheirclinicalpracticepatterns,thentemporaryincentivescanachieve
aboostinperformanceatasubstantiallycheapercostthanpermanentincentives.Ourresultssuggest
that the mechanism behind positive provider responses to price increases is more related to
adjustmentcoststhantorespondingtohigherprofitmargins. Inthiscase, long‐termincreases in
productivitycanbeachievedmorecheaplythanthroughapermanentincreaseinfees.
2 CONCEPTUALFRAMEWORK
Wedevelopastylizedmodelofclinicalpracticepatternswhereclinicsincurafixedcosttochange
clinicalpracticeroutines.Weassumethatpatientsareidentical,thatclinicsprovidethesameservices
toallpatients,andthatdemandisexogenouslydetermined.
ObjectiveFunction:Clinicshaveapay‐offfunction ∝ ,where isprofits,Hishealth
oftherepresentativepatient,N isthenumberofpatients,and∝∈ 0,1 istheprovider’s intrinsic
valueofaunitofpatienthealth.12As∝risesthecliniciswillingtosacrificemoreincomeforpatient
health.When∝takesonvalue0,theclinicispurelyextrinsicallymotivated,andwhen∝is1theclinic
ispurelyintrinsicallymotivated.Whileweallowforbothextrinsicandintrinsicmotivationinthe
model,alloftheresultsfollowevenwithpureextrinsicmotivation.Allowingforintrinsicmotivation
does not change the direction of the predictions just the magnitude. Moreover, pure intrinsic
11SeeBakeretal.(1988);HolmstromandMilgrom(1991);Gibbons(1997);andLazear(2000).12Thereisevidencetosupportintrinsicmotivationasatleastpartiallymotivatingmedicalcareproviders.
SeeforexampleLeonardandMasatu(2010);Kolstad(2013);andClemenesandGotlieb(2014).
5
motivationby itselfdoesnotpredict that temporary incentiveswouldhave long termseffectson
productivity.13
HealthProductionFunction:Treatmenttechnology,asdefinedbyCPGs,involvestwoservices,
and where 1 if the clinic provides the service and 0 if not. If the clinic provides both
services,thenitisoperatingattheproductionpossibilitiesfrontier.Thehealthproductionfunction
fortherepresentativepatientis ,where isameanzerorandomshock.
ClinicalPracticeRoutine:Consideraclinicwhosecurrentclinicalpracticepatternroutineis
toprovide toallpatients.Inthiscase, istheclinic’sexistingclinicalpracticepatternroutine,and
isanadditionalservicethatthecliniccouldchoosetoaddtoitspracticeroutine.Iftheclinicwants
tointegratetheprovisionof intoitspracticepatternroutinethenitmustincuranupfrontfixed
costF.Thefixedcostincludesthecostofretoolingtobeabletoprovide ,thecostoflessreliable
serviceprovisionwhilethenewroutineisbeinglearned,andthecostofovercomingpsychological
resistancetochange.
Profits:Clinicsarepaid for andthemarginalcostofproviding toapatientis .Clinic
profitscanthenbeexpressedas:
∑ , (1)
where istheclinic’sdiscountrate.
Adoption:Theclinicadopts if
1 0 0 . (2)
Substitutionof(1)and(2) intothepay‐off functionandrearrangingtermsallowsustowritethe
conditionin(3)as:
∑ . (3)
13Withoutsomesortoffixedcostsofadjustment,bothintrinsicallyandextrinsicallymotivatedproviders
would still operate at the efficient frontier. Moreover, the intrinsic motivation literature suggests thatincentivescannegatively impactperformance.Thepsychology literature inparticularhas longargued thatperformance‐contingentincentivescanbedemotivatingforintrinsicallymotivatedworkers.ForexampleseeDeci(1971);PittmanandHeller(1987);Decietal.(1999);Deci(2001);EcclesandWigfiel(2002);DeciandRyan (2010). Benabou and Tirole (2003) embed these ideas in principle‐agent models that they use todemonstrate themechanisms throughwhich financial incentives can “crowd‐out” intrinsicmotivation andtherebynegativelyaffectperformance.Recentlaboratoryexperimentalevidenceonperformance‐contingentcontractsconfirmsthatincentivesinthepresenceofintrinsicmotivationcanresultinworseperformance.ForexampleseeFehrandFalk(1999);FehrandSchmidt(2000);GneezyandRuitichini(2000aand2000b);andArielyetal.(2009).
6
Clinicsaremorelikelytoadopt iftheprofitmarginfrom ishigher,theyaremoreintrinsically
motivated,theeffectof onpatienthealthishigher,theyhavehigherpatientvolumes,andthey
havelowerdiscountrates.
Organizational inertia: Inertia is defined as when the present value of the fixed costs of
changing organizational routine prevents the clinic from adopting a valuable improvement to
production. The conditions are 0 and ∑ , i.e. is
valuablebutnotadoptedbecauseofthefixedcostofadjustingorganizationalroutinetobeableto
provide .Clinicswhoaremoreintrinsicallymotivated(i.e.higher )arelesslikelytobefrozenby
organizationalinertiaandmaybeevenwillingtolosemoneyinordertoadopt ,especiallyif is
veryproductive(i.e.higher ).
TemporaryIncentives:Organizationalinertiacanbeovercomewithatemporaryincreasein
,thepriceof .14Consideranincreasetothepricepaidinperiod1thatdisappearsinsubsequent
periods.Withoutlossofgeneralitywecansimplifythemodelto2periodswith asthediscountrate.
Inthiscase,theincreaseof in inperiod1necessarytoinducetheprovidertoadopt is:
≧ 1 . (4)
Thetemporaryincentive, ,atminimumcoverstheremainderofthefixedcostofadjustmentthatis
not paid for the discounted present value of the future stream of surplus generated from the
provisionof .Theincentivegoesdownwithscale ,theprofitmargin ,theextenttowhich
clinicsareextrinsicallymotivatedtimesthemarginalproductof inthehealthproductionfunction
,andthediscountrate.
Cross‐Price Effects:One concern voiced in the literature is that price increases for some
servicesmightleadtoareallocationofeffortfromotherservicesthatremainunchangedleadingto
negativecross‐priceeffects.Theimplicitunderlyingmodelinthesepapersisanindividualphysician
allocating time between activitieswith a time budget constraint. In ourmodel of amedical care
organization that can hire more staff, cross‐price effects are generated based on the nature of
economies of scope in either the health care production function or cost function. If both the
productionandcostfunctionsareadditivelyseparable,thentherearenocross‐priceeffects.Ifthe
14Thealternativeisalumpsumpaymentthatisvulnerabletothepossibilityofnoncomplianceandmaybe
difficulttoverify.However,atemporaryincreasein requirestheclinictochangeroutinesandactuallyadoptinordertogetpaid.Inthissensethetemporarypriceincreasealsoincludesacommitmentdeviceandhence
isexantepreferable.
7
functionsarenotseparable,thenitispossibletohaveeithernegativeorpositivecross‐priceeffects
dependingthenatureofsubstitutabilityintheproductionandcostfunctions.
3 EXPERIMENTALDESIGN
ThefieldexperimentwasconductedbyPlanNacer,apublicinsuranceprogramthatbeganin2005
toimproveaccesstoqualityhealthcareforotherwiseuninsuredpregnantwomenandchildrenless
than6yearsold(Musgrove2010;Gertleretal.2014).LikeMedicaidintheU.S.andSeguroPopular
in Mexico, the national Plan Nacer program transfers funds to local governments, in this case
Provinces,whoarethenresponsibleforenrollingbeneficiaries,organizingtheprovisionofservices,
andpayingmedicalcareproviders.AninnovativefeatureoftheArgentineprogramisthatituses
financial incentives to ensure that beneficiaries receive high‐quality care. Financing from the
NationalleveltoProvincesisbasedfor60%onprogramenrollmentandfor40%onperformance.
Provincesthenusethosefundstopaypublichealthcarefacilitiesonafee‐for‐servicebasisforhealth
care provided to program beneficiaries. The national government determines the content of the
benefitspackage,whichisuniformacrossprovinces,whileprovincialgovernmentssetthepricethey
willpaytoprovidersforeachserviceinthatpackage.Healthfacilitiesarefreetochoosehowtouse
realizedrevenueswithinrelativelybroadguidelines.Some,thoughnotall,provincesallowhealth
facilitiestopaybonusestopersonnel.
PlanNacerscaledupbyfirstrecruitingandtrainingclinicsintheoperationsofitsprogram,
includingfeestructure,billing,andotherrules.Theprogramregularlyretrainstheclinicstokeep
themuptodateonanychangesandreinforceareasthatareperceivedtobeweak.Afterclinicsare
enrolled, clinic community outreach staff identify eligible women and children in the clinics’
catchmentareasinordertoenrollthemintotheprogram.Clinicoutreachstaffalsoregularlycontact
beneficiariestoencouragethemtotakeadvantageofprogrambenefits.
The field experiment was conducted with primary health care clinics in the Province of
Misiones,oneofthepoorestinthecountryandwithhighratesofmaternalandchildmortality.In
Misiones,theclinicisallowedtouseupto50%ofrevenuefromPlanNacerfeestopaybonusesto
facilitypersonnelatthediscretionofthefacilitydirector.TherolloutofPlanNacerinMisioneswas
completed in 2008 long before the pilot study. As such, both providers and beneficiaries were
knowledgeableoftheoperationofPlanNacerbeforetheexperimentbegan.
Theexperimentalinterventionwasdesignedtoencourageearlyinitiationofprenatalcarefor
PlanNacerbeneficiaries,therebyaligningtheincentivesinPlanNacerwithofficialArgentineclinical
8
practice guidelines, medical school training, and international scientific evidence. Before the
experiment, only one‐third of Plan Nacer beneficiarieswere initiating care in the first trimester
(National Ministry of Health, 2009 and 2010). The experiment randomized temporary financial
incentivestoprimaryhealthcareclinicsinwhichtreatmentclinicswerepaida200%premiumfor
earlyinitiationofprenatalcare,i.e.beforeweek13.
Table 1 presents the payment schedule for the periods before, during and after the
intervention.Priortotheinterventionperiod,theprovincepaidfacilities$40ARSforeachprenatal
visitregardlessofwhenitoccurredorwhetheritwasthefirstorasubsequentvisit.15Duringthe
interventionperiodthefeewasincreasedto$120ARSfor1stvisitsthatoccurredbeforeweek13but
remainedat$40ARSforsubsequentvisits.Afterthat,theinterventionperiodfeesrevertedtothe
originalpaymentof$40ARSforallvisits.Themodificationamountedtoa3‐foldincreaseinthefee
for1stvisitsbeforeweek13.Themodifiedfeestructurewasimplementedfor8months‐fromMay
2010 toDecember2010. Facilities selected to receive themodified fee structurewere invited to
participateandnotifiedofthetime‐limitedimplementationonApril14,2010.Facilitydirectorswere
requiredtosignaformalmodificationoftheirexistingcontractwithPlanNacerinordertoreceive
themodifiedfeestructure.
The studydesign included37clinicsoutof262primarycare facilitiesof theprovince,of
which 18 were randomly assigned to the treatment group and were offered the modified fee
schedule.Theother19 formed thecontrolgroup.Table2shows that compliancewith treatment
assignmentwasnotperfect:outof18facilitiesassignedtothetreatmentgroup,14wereactually
treatedasthreerefusedtosigntheagreementandafourthclosedbeforetheinterventionstarted.In
addition, oneof the facilitiesoriginally assigned to the control groupwasmistakenlyoffered the
treatmentandagreedtothemodifiedfeestructure.Intheend,therewere36facilitiesinthestudy
excludingtheonethatclosed.
4 DATA
The Province of Misiones maintains a well‐developed and long‐established automated medical
recordinformationsystemmanagedbytheprovincialauthorities.Personnelatpublicprimaryhealth
clinics and hospitals digitize a record of each service provided to each patient. The data are of
unusually high quality in that key outcomes such as dates of visits, services delivered, andbirth
weightarerecordedatthetimeofcarebytheprovider;thereforewedonotneedtorelyonmaternal
15Theexchangeratefor$1ARSwasaround$0.25USDbetween2009through2011.
9
recallofthesevariablescollectedinsurveyslongafterthevisit.Thedatausedintheanalysisare
extractedfromtheseclinicrecordsandcontaininformationontheuniverseofpatientsforthe36
clinicsinthestudy.Therecordsalsoincludetheindividual’snationalidentitynumber,whichisused
tolinktheindividualclinicmedicalrecordsfromprimaryhealthfacilitieswiththeregistryofhealth
insurancecoverage,theregistryofPlanNacerbeneficiaries,andhospitalmedicalrecords.Inall,97%
oftheprimaryclinicmedicalrecordsweremergedwiththedataoninsurancestatusandprogram
beneficiarystatus. Inaddition,75%of theseweresuccessfullymergedwithmedicalrecordsdata
fromhospitals.Thereforeouranalysis isabletoevaluatethe impactof the interventionforthose
womenwhoinitiatedtheirprenatalcareinoneoftheprimarycareclinicsofthesample.
4.1 ANALYSISSAMPLE
Figure2depictsthetimelineofthestudyandtheavailabilityofdatadividedinto4different
sub‐periods: (i) a16‐monthspre‐interventionperiod from January2009 toApril2010, (ii) an8‐
monthinterventionperiodfromMay2010toDecember2010,(iii)a15‐month“post‐intervention
periodI”fromJanuary2011toMarch2012and(iv)a9‐month“post‐interventionperiodII”from
April2012toDecember2012.
Prenatal care data was consistently collected for the first 3 periods from January 2009
throughMarch2012.StartinginApril2012,however,Misionesadoptedanewinformationsystem
andasaresultdata frompost‐interventionperiodIIcannoteasilybecomparedtodata fromthe
earlierperiods.Inparticular,thenewsystemchangedthecodesusedtoclassifythereasonforvisits
inordertofacilitatebilling.If inthefirstvisittheattendingphysicianrequestedanultrasoundto
confirmapregnancy,thisfirstvisitwaslabeledasa“carevisit”whilethesubsequent(second)visit,
waslabeledasthefirstprenatalvisit,ifindeedtheultrasoundconfirmedthepregnancy.Onaverage,
thiswouldledtoareductionintheshareofwomenwhohadavisitlabeledas“firstprenatalvisit”
beforeweek13andan increase intheweekspregnantat thetimeof thisvisit. If thenewcoding
system affected the treatment and control groups in the sameway, the differences between the
treatmentandcontrolgroupswouldstillcapturetheimpactoftheincentives,albeitpossiblywith
some measurement error. Therefore, we analyze the data from post‐intervention period II
separately,andinterprettheresultswithcaution.
10
TheanalysissampleincludespregnantwomenwhowerebeneficiariesofPlanNaceratthe
timeoftheirfirstprenatalvisit.16Whileinformationonprenatalcareutilizationisavailableforthe
fullsampleperiod,informationrelatedtobirthoutcomesisonlyavailableforwomenwhogavebirth
inapublichospitalthrough2011,i.e.womenwhobecamepregnantbeforeMay2011.
4.2 MEASUREMENTOFWEEKSPREGNANTAT1STPRENATALVISIT
Weconstructthenumberofweeksofpregnancyatthetimeofthefirstprenatalvisitasthe
differencebetweenthedateofthefirstvisitandthelastmenstrualdate(LMD).TheLMDisroutinely
collected at the time of the visit to calculate the estimated date of delivery (EDD) and both are
routinelyrecordedinthepatient’smedicalrecordattheclinic.17
Onepotentialproblemisthatmedicalpersonnelintreatmentfacilitiesmightmisreportthe
dateoflatefirstvisitasoccurringbeforeweek13sothattheycouldbilltotheprogram.Wethink
thisisunlikelyforthefollowingreasons.First,theweekofvisitisconstructedfromthedateofthe
firstprenatalvisitandtheLMD,bothofwhichalongwiththeEDDarerecordedinrealtimeinthe
medicalrecord.Inordertofalselyreportthatafirstvisitoccurredinthefirst12weeks,theprovider
wouldhavetoalterthedateofthefirstvisitrelativetoeithertheLMDortheEDDinthemedical
record.Thiswouldrequiresomeeffortifdoneinrealtimeandwouldbenoticeablebyauditorsif
alteredexpost.Second,PlanNacerusesexternalauditorstoverifytheaccuracyofclinicbilling.The
auditorscomparethedetailedclinicalrecordstothebillingrequeststofindinconsistenciesandthe
lattercanleadtosubstantialfinancialpenaltiesfortheprovinces.Finally,clinicalrecordsarelegal
documentsinArgentinaandpractitionerscouldlosetheirmedicallicenseifcaughtsystematically
misreportingforfinancialgain.
Tocorroborateourbeliefthatfalsereportingintheclinicrecordsisunlikely,weempirically
testwhetherthereisanyevidenceofsystematicmisreportingusingdatafromanalternativesource.
Specifically,weusegestationalageatbirthmeasuredbyphysicalexaminationobtainedfromhospital
recordstoconstructasecondestimateoftheLMDandweekspregnantatthetimeofthefirstprenatal
visit.Thehospitalpersonnelthatattendthebirthdonothaveanyincentivetomisreporthospital
records.Wethencomparetheestimatedweekoffirstvisitbasedongestationalageatbirthtothe
16Weexcludednon‐beneficiariesbecausemostofthemhaveprivatehealthinsuranceandassucharelikely
toreceivesomeofcareanddeliveratprivatefacilities.Sincewedonothavedatafromprivatefacilities,theoutcomesofmostoftheseobservationsarecensored.
17For10%ofthesampleLDMwasnotrecorded.Forthosecases,weusetheEDDtorecovertheLMD.
11
weekoffirstvisitreportedbythehealthfacilities.Theresultsdonotshowanyevidenceofsystematic
misreportingduetoincentives.AppendixAprovidesadetaileddiscussionoftheanalysisandresults.
4.3 DESCRIPTIVESTATISTICSANDBASELINEBALANCE
Table3reportsthedescriptivestatisticsforthekeyoutcomesofinterestanddemographic
characteristics at baseline, i.e. in the 16‐month pre‐intervention period (Jan 2009 –April 2010).
Outcomes are balanced at baseline in that there are no statistically significant differences in the
meansofvariablesbetween the treatmentandcontrolgroups.Onaveragewomenhad their first
prenatalvisitabout17.5weeksintotheirpregnancywithaboutone‐thirdofwomenhavingthatvisit
beforeweek13.Womencompletedabout4.7prenatalvisitsoverthecourseoftheirpregnancyand
morethan80%ofthemreceivedatetanusvaccine.Newbornsweighedapproximately3,300grams
onaverage,whileabout6%ofthemwerebornwithlowbirthweight(i.e.lessthan2,500grams),and
slightlymorethan9%ofbirthswerebornprematurely.
5 IDENTIFICATIONANDESTIMATION
We estimate both the intent‐to‐treat (ITT) and local average treatment (LATE) effects of the
incentives on outcomes. The ITT is the effect of assigning a clinic to treatment on outcomes,
regardlessofcompliance.Itcomparesthemeanoutcomeofthegroupassignedtotreatmenttothe
meanoutcomeofthegroupassignedtocontrolandisestimatedbyregressingtheoutcomeagainst
anindicatorofwhethertheclinicwasassignedtotreatment.TheLATEistheeffectofaclinicactually
receiving the incentives and is estimated regressing the outcome againstwhether the clinicwas
actually treated, using the clinic’s randomized assignment status as an instrumental variable for
actualtreatment(ImbensandAngrist1994).Inbothcases,thetreatmenteffectisidentifiedoffthe
variation induced by the randomized assignment status. In the discussion of results in the next
section,wereporttheLATEestimates.18
Oursampleisclusteredwithin36healthclinicssincetherandomassignmentoftreatment
occurredatthecliniclevel.Assuch,theremaybeintra‐clustercorrelationthatmustbeconsidered
forstatisticalinference.Standardmethodsofcorrectingstandarderrorsrelyonlargesampletheory
bothinthenumberofobservationsandinthenumberofclusters.Giventhesmallnumberofclusters
inoursample,weinsteadusestatisticalinferencemethodsthatarerobusttorandomizedassignment
oftreatmentamongasmallnumberofclusters.Specifically,weusetheWildbootstrapmethodto
18TheITTresultsarealmostidenticaltotheLATEestimates,whichisexpectedgiventherelativelyhigh
complianceratestotheoriginalassignment.TheITTresultsarepresentedinAppendixC.
12
generatep‐values for hypothesis testing in ITTmodels (Cameron et al. 2008) and an analogous
method for hypothesis testing in the LATE models (Gelbach et al. 2009). Our Wild bootstrap
procedureassignssymmetricweightsandequalprobabilityafterre‐samplingresiduals,anduses
999replications(DavidsonandFlachaire2008).
6 TIMINGOFFIRSTPRENATALVISIT
In thissectionwereport theresultsofanalysesof theeffectsof the temporary incentiveson the
timingofthefirstprenatalvisitandmechanismsbywhichclinicsachievedthoseresults.
6.1 DENSITIES
Figure3comparesthedensitiesofweekspregnantatthetimeofthefirstprenatalvisitsfor
theclinicsassignedtothetreatmentandcontrolgroups.PanelAshowsthatthereisnodifference
betweenthedensitiesofthetreatmentandcontrolgroupsinthepre‐interventionperiod.PanelB
shows that the treatment group density is to the left of the control group density during the
interventionperiod.Finally,PanelCandDshowthatthetreatmentgroupdensityisplacedtotheleft
ofthecontrolgroupdensityduringpost‐interventionperiodsIandII.Kolmogorov‐Smirnovtestsfor
equalityofthedistributionscannotberejectedforthepre‐interventionanalysis,butarerejectedfor
the intervention and both post‐intervention periods with p‐values of 0.031, 0.004, and 0.009
respectively.Theseresultsimplythatthetemporaryincentivesledtoearlierinitiationofcareinthe
treatmentgroupcompared to the control group in the interventionperiodand that thesehigher
levelsofcarepersistedforatleast15monthsandlikelyfor24monthsandmoreafterthehigherfees
wereremoved.
6.2 SHORT‐RUNEFFECTS
Table4reportstheestimatesoftheeffectsofthetemporaryfeesontheearlyinitiationof
care.PanelAreportstheresultsforweekspregnantatthetimeofthefirstprenatalvisitandPanelB
reportstheresultsforwhetherthefirstvisitoccurredbeforeweek13.Thefirstcolumnreportsthe
resultsfortheinterventionperiodandthesecondandthirdcolumnsreporttheresultsforthepost‐
interventionperiods.Duringtheinterventionperiod,onaveragewomeninthetreatmentgrouphad
their1stvisitabout1.5weeksearlierintheirpregnancythanwomeninthecontrolgroup.Theshare
ofwomeninthetreatmentgroupwhohadtheir1stvisitbeforeweek13 is11percentagepoints
higherthanthecontrolgroup;approximately35%higherthanthecontrolgroup.Bothestimatesare
significantlydifferentfromzeroatconventionalp‐values.
13
6.3 LONG‐RUNEFFECTS
Ourmodel of behavioral inertiaprovided clearpredictions aboutproviderbehavior once
temporary incentives disappear: i.e. if the fee increase is enough to overcome the fixed costs of
adaptinganewpractice,clinicsshouldmaintainhigherlevelsofprenatalcareafterincentivesare
removed. Column 2 of Table 4 reports estimated impact of the temporary fee increase on early
initiationofcareinthe15‐monthperiodafterthefeeswereremoved. Onaverage,womeninthe
treatmentgroupstartedtheircare1.6weeksearlierthanthoseinthecontrolgroup.Thedifference
betweenthetreatmentandcontrolgroupsintheshareofwomenwhohadtheir1stvisitbeforeweek
13was8percentagepoints.Bothestimatesarestatisticallydifferentfromzeroatconventionallevels.
Further,wecannotrejectthenullhypothesisthattheimpactisdifferentintheinterventionandpost‐
interventionperiods.Theseresultsareconsistentwiththehypothesisthattemporaryincentiveshelp
overcomebehavioralinertiaandmotivatelong‐runchangesinperformance.
Whilethereisnosignificantdifferentbetweentheeffectduringtheinterventionandthepost‐
interventionperiods,oneconcernmaybethattheeffectoftreatmentslowlytrendedtowardszero
aftertheincentivesended.Totestthishypothesis,weplotthemeannumberofweekspregnantat
the time of first prenatal visit for treatment and control groups, before, during and after the
intervention(Figure4).19Wesplitthepre‐interventionperiodintotwosub‐periodsof6‐monthseach
andthepost‐interventionperiodinto3sub‐periods:thefirsttwoare6monthsandthethirdis3
months.Thetreatmenteffectisthedifferencebetweenthetwolines.Whilethetreatmentandcontrol
groupshavesimilartrendsbeforetheintervention,thetreatmentgroupappearstoreceiveearlier
careduringtheintervention,andthechangepersistsaftertheendoftheintervention.Noticethat
thereis little ifanyfalloffoverthepost‐interventionperiod.Ratherthetreatmenteffectsremain
fairlyconstantoverthe15‐monthpost‐interventionperiodI.Figure5depictsthesamerelationship
fortheshareofwomenwhoreceivecarebeforeweek13ofpregnancy.20Again,theeffectsofthe
interventionappeartocontinueatasteadyrateafteritisdiscontinued.
6.4 LONGER‐RUNEFFECTS
TheperiodofanalysisinourmainresultsisrestrictedtoJanuary2009toMarch2012.Recall
thatstarting inApril2012, thevisitcodingsystemchanged.Hencestarting inApril2012what is
reportedasfirstvisitsinthedataisactuallyamixoffirstandsecondvisits.Asaresulttheaverageof
weekspregnantatfirstvisitincreasesandtheshareofpregnantwomenwhosefirstvisitwasbefore
19 As discussed above, the information from post‐intervention period II (April‐December 2012) uses a
differentmetricandisthereforenotincludedinthisfigure.20Ibidem.
14
week13fallsrelativetopreviousperiods.Column3inTable4showstheresultsforthislastperiod.
Themeanaverageofweekspregnantatthetimeofthefirstvisitforthecontrolgroupissubstantially
higherforthisperiodthanforpreviousperiodsandthemeansharethathadtheirfirstvisitbefore
week13issubstantiallylower,suggestingthatthereismeasurementerrorinourmainoutcomein
thisperiod.However,thisdifferenceincodingshouldhaveasimilareffectintreatmentandcontrol
clinics given the randomized assignment of the treatment. Therefore the difference between
treatmentandcontrolclinicsshouldcanceloutthemeasurementerrorandprovideuswithunbiased
estimatesoftheimpact.
The results in Table 4 show a statistically significant reduction in the number of weeks
pregnantatthetimeofthefirstvisitandastatisticallysignificantincreaseintheshareofpregnant
womenwhohadtheirfirstvisitbeforeweek13.Theseresultssuggestthatimprovedproductivity
fromthetemporaryfeeincreasepersistedatleast24monthsafterthefeeswereremoved.
6.5 ROBUSTNESS
Weimplementthreerobustnesschecks.First,themainsamplemayincludepregnanciesthat
startinoneperiodandendinanother,whichcouldcloudtheeffectoftheincentivesontimingofthe
firstvisit.Forexample,awomanwhois6monthspregnantandhasnothadaprenatalvisitwhenthe
intervention starts and subsequently receivesher firstprenatal checkupduring the intervention,
wouldbe counted as a third trimester first visit during the interventionperiod, even though the
intervention cannot affect whether she receives prenatal care before week 13. Hence, in this
robustnesstestwere‐estimatethemodelsonarestrictedsamplewherewomenarenomorethan
onemonthpregnantinthefirstmonthoftheperiodandnolessthan3‐monthspregnantinthelast
monthoftheperiod.Theresults,reportedinPanelsBofAppendixTablesB1andB2,areveryclose
inmagnitudeandstatisticalsignificancetothemainresultsinTable4.
Second,eventhoughtherewerenostatisticaldifferencesinbaselinemeans,itispossiblethat
randomizationwas not able to fully balance the treatment and control groups on unobservable
characteristicsgiventhesmallnumberofclinics.Inordertotestforthispossibility,weestimatethe
modelsusingdifference‐in‐differenceswithclinicandmonthfixedeffects.Theresults,reportedin
PanelsCofAppendixTableB1andB2,areverycloseinmagnitudeandstatisticalsignificancetothe
mainresultsinTable4.
Finally,instudiesinvolvingasmallsampleofclustersthereisaconcernthatafewoutliers
maydrivetheaverageeffectfoundintheprevioussections.Weexplorethispossibilitybyestimating
clinic‐specifictreatmenteffectswherebywecompareeachtreatedclinicindividuallytothecontrol
15
clinicsasagroup.AppendixFiguresB1andB2plottheseindividualclinictreatmenteffectsforthe
outcomesofweekspregnantatthetimeofthefirstprenatalvisit(B1)andfortheprobabilityofthat
thefirstvisitoccurredbeforeweek13(B2),respectively.Theresultsaresortedalongthex‐axisfrom
the lowest to thehighest estimated effect,while thedashedblue line is the intent‐to‐treat effect
calculatedbypoolingtheinterventionandthefirstpost interventionperiod. Thesolidblackline
representsazerotreatmenteffect.Theverticallinesare95%confidenceintervalsconstructedusing
standarderrorsobtainedfromtheWildbootstrapprocedure.Thefiguresshowthatthehypothesis
ofnotreatmenteffectisrejectedfor11outof17clinicsinFigureB1and12outof17clinicsinFigure
B2.Inaddition,thetreatmenteffectshavetheexpectedsignin15out17clinicsinFigureB1and14
outofclinicsinFigureB2.Thisprovidesevidencethatourresultsarenotdrivenbyafewlarge‐effect
clinics.
6.6 MECHANISMS
Inordertobetterunderstandhowclinicswereabletoachievesuchlargeincreasesinthe
share ofwomenwho initiated prenatal care beforeweek 13,we conducted a series of in‐depth
interviewswithprofessionalsinasub‐sampleof5treatmentclinicsand3comparisonclinics.21We
findthattreatmentclinicsadoptednewpracticesandchangedroutinesinordertoincreaseearly
initiationof prenatal care.After the initial invitation toparticipate in the pilot, all 5 interviewed
treatmentclinicsorganizedateammeetingwiththestaffinordertodiscussstrategiestorespondto
thenew incentive scheme.Various treatment clinics adopteddifferent strategies, but all of them
involvedexpandingthescopeofworkofcommunityhealthworkerstoidentifyandencouragenewly
pregnantPlanNacertoinitiatetheirprenatalcareearly.Insomeclinics,thedirectorsupportedthe
change in strategiesby changing theway the financial incentivesweredistributedbetween staff
members.22Inparticular,someofthemstartedallocatingtheincentivesconditionalonthenumber
21The clinics interviewedare located inPosadas, the capital ofMisionesProvince.Each interview took
approximately45minutes.TheinterviewswerecarriedoutinMay2015.22Upto2013,anyhealthfacilityparticipatinginPlanNacerinMisioneswasabletouseupto10%oftheir
PlanNacerfundstopayincentivestopersonnel.Ifthefacilityachievedasetofhealthtargetsmeasuringusingperformance indicators (tracers) set by the province, that facilitywas able to use up to 50%of funds formonetaryincentivestohealthprofessionals.Thebonusescouldbeassignedtoanypersonworkingatthehealthfacility, includingthehealthworkers,administrativepersonnel,volunteers,andeventopersonnelaffiliatedwithotherprogramsaslongastheywerenotabsentformorethan10workingdaysinamonth,theydidnotparticipate in a strike organized by the union, and they were not subjected to a disciplinary sanction(suspensionwithoutpayordismissal). Inallcases, the finaldecisionregardingassignmentof incentivestopersonnelwastheprerogativeoftheclinicdirector.
16
ofpregnantwomenthateachteammemberbroughttotheclinicinamonth.Thisallocationfurther
incentivizedhealthworkerstotestnewpractices.
The in‐depth interviews uncovered several innovative strategies that treatment clinics
developedtoidentifypregnanciesearly.Forinstance,healthworkersstartedtofollowupwomen
who used birth control pills.23 Specifically community health workers prioritized home visits to
womenwhohadnotpickeduptheirpills.Second,healthworkersstartedtargetingwomenathigh
risk of not coming in for an early checkup. According to the interviewed doctors, mothers who
already have children are less likely to initiate their prenatal visits early in a new pregnancy.
However,manyof thesewomenarealsoeligible forweekly freemilkdistribution for theirolder
children.Healthworkersmetthesemothersatthetimeofthemilkdistribution,enquiredabouttheir
last menstruation date, and offered an instant‐read pregnancy test to those women whose
menstruationwasoverdue.Third,healthworkersidentifieddifficultiesinprovidingearlyprenatal
care toadolescents,as theymightbeunwilling torevealapregnancy,especially to theirparents.
Communityhealthworkersthereforedecidedtochangethetimingofhomevisits,soastoincrease
thechanceoffindingadolescentsbythemselves.Inoneoftheinterviewedclinics,theworkflowwas
modifiedsoastoensurepredictableavailabilityofagynecologistoncertaindaysoftheweek.This
inturnprovidedaneasywayforcommunityhealthworkersandadministrativestaff toschedule
patientappointments.Otherclinicsintroducednewwaysofkeepingtrackof“atrisk”patients,such
asanotebookthatkepttrackofanyvisitstothehomesofwomenthatwereatrisk,oramapthat
identified catchment areas of community health workers with corresponding (potential)
pregnancies.
We are able to substantiate the claims of increased outreach using clinic administrative
records on the number of community outreach activities that resulted in actual maternal‐child
serviceattheclinic.24Figure6displaystheaverageandmediannumberofoutreachactivitiesthat
resulted in actual maternal‐child services for the pre‐intervention, intervention, and post‐
interventionIperiods.25Theresultsshowthatthereislittledifferenceinoutreachactivitiesbetween
treatmentandcontrolclinicsinthepre‐interventionperiod.Intheinterventionperiodthetreatment
23Birthcontrolpillsaredispensedfreeofchargebyeachhealthfacility’spharmacyunit,thoughwomen
cannotcollectmorethanamonthlysupplyatanyonetime.Thepharmacyunitkeepsrecordsofallbirthcontrolpillcollections.
24 Plan Nacer finances clinic outreach activities on a fee‐for‐service basis and employs an externalindependentauditortoauditclinicactivityreports.Treatmentandcomparisonclinicswerepaidthesamefeefortheseactivitiesbefore,duringandaftertheexperiment.
25Themediansarebettermeasuresofcentraltendencyasthedensitiesofbothactivitiesareasymmetricheavilyskewedtotheright.
17
groupevidencedsubstantiallymoreactivitiesthanthecontrolgroup,andthisdifferenceissustained
throughthepost‐interventionperiod.
Weusethedatatoestimatethedifferencesinlognumberofactivitiesbetweenthetreatment
andcontrolgroups.Theresultsshownodifferencesinactivitiesinthepre‐interventionperiodand
positive and statistically significant higher levels of activities in the treatment clinics in the
interventionandpost‐interventionIperiods(Table5).Again,wecannotrejectthatthehypothesis
that the effects are different in the intervention and post‐intervention periods implying that the
increaseinsuccessfuloutreachactivitiespersistedafterthetemporaryincentiveswereremoved.
6.7 PSYCHOLOGICALBARRIERS
In theprevious subsectionwedocumented tangible costsof adjustment to increase early
initiationof prenatal care.An additionalpotential costof adjustment is psychologicalbarriers to
change.Onewaytoovercomepsychologicalresistanceistomaketheguidelineortaskmoresalient
inthemindsoftheclinicstaff.26Theissueisnotoneoflackofknowledgeorinformationasinitiating
care in the first trimesterhasbeen inCPGssince the1970sandhasbeena long‐standingpartof
standardmedicaleducation. Rathertheissueistheimportanceorprioritythatstaffplaceonthe
task.
Thetemporaryincentivesmighthaveincreasedtheimportanceofearlyinitiationofcarein
thestaff’sminds,therebymakingitahigherpriorityforaction.Thehigherthepriorityofatask,the
lesslikelypsychologicalbarrierswouldstandinthewayofadoption.Kahneman(2012,pp8)states
that“…frequentlymentionedtopicspopulatethemind…”morethanothersand“…peopletendto
assesstherelativeimportanceofissuesbytheeasewithwhichtheyareretrievedfrommemory”.As
such,salience“…isenhancedbymerementionofanevent”(Kahneman2012,pp331).Ifincomplete
ornon‐adoptionofataskisamatterofsaliencethentheobservedtreatmenteffectsmaybeexplained
bythefactthattemporaryincentiveshelptoovercomethistypeofpsychologicalbarriertochange.
Whilewedonothaveinformationonthesalienceofearlyinitiationofcareduringorshortly
aftertheexperiment,weexplorewhetherthetemporaryfeeincreasemadeearlyinitiationofcare
moreimportantinthemindsoftheclinicstaffaftertheendoftheexperiment,usinganonlinesurvey
administeredtothechiefmedicalofficerofeachclinicabouttheabsoluteandrelativeimportanceof
26 Taylor and Thompson (1982) define salience as, “…the phenomenon that when one's attention is
differentiallydirectedtooneportionoftheenvironmentratherthantoothers,theinformationcontainedinthatportionwillreceivedisproportionateweightinginsubsequentjudgments”.SeeBordaloetal.(2012,2013)foramorerecentdiscussionofsalienceandchoicetheory.SeeDeMeletal.(2013),andKarlanetal.(2015)forempiricalanalysisofsalienceeffectsthroughinformationalreminders.
18
seven different prenatal care procedures including initiating prenatal care prior to week 13 of
pregnancy(seeAppendixD).
Figures8and9comparetheabsolutescoreandrelativerankingoftheproceduresinterms
of importanceforprenatalcare.Theabsolutescoresranges from0to5,with5beingthehighest
whiletherelativerankingsortsthesevenpracticesfrom1to7,with1beingthehighestranking.Our
outcomes of interest are the absolute score and relative ranking assigned to early initiation of
prenatalcare.Figure8showsthattheabsolutescoreassignedtoearlyprenatalcareisonaverage
4.8inthetreatmentgroupand4.7inthecontrolgroup.Figure9showsthatonaveragetherelative
rankingforthispracticeisalsosimilarbetweenthetwogroups,2.0forthetreatmentgroupand1.9
for the control group.Moreover, thesedifferences arenot statistically significant at conventional
levels(seeAppendixD).Theseresultssuggestthattheearlyinitiationofprenatalcareisofsimilar
highabsoluteandrelativeimportanceandthattemporaryfeesdidnothavealastingeffectoneither
theabsolutenorrelativeimportance.
6.8 ALTERNATIVEEXPLANATIONS
One alternative explanation for the short‐term treatment effects is that the incentives are
causingtreatmentclinicstotrytoattractpregnantwomenwhootherwisewouldhaveusedother
clinics.Thisisunlikelytobetrueasbeneficiarywomenareassignedtospecificclinicswhenenrolled
inPlanNacer.Moreover,thenumberofpatientspermonthandthesharethatinitiatecarebefore
week13arethesameinthepre‐andpost‐interventionperiodsforcontrolclinics,andtheaverage
monthly number of patients is also the same in the pre‐ and post‐intervention periods for the
treatmentclinics.
Analternativeexplanationforlong‐runresultsisthatafterthetemporaryincentivesended,
womenwhowerepregnantduringtheinterventionperiodspassedthemessageoftheimportance
of early initiation of care onto other beneficiarywomenwho became pregnant during the post‐
interventionperiod.Hence,thepersistenceoftheeffectoftheincentivesaftertheincentivesmight
becausedbyan informationalspillover.However, thehigheramountof thecommunityoutreach
activities in treatment clinics, the mechanism used to generate higher early initiation of care,
continuedintothepost‐experimentalperiodatthesamelevelasintheinterventionperiod.Hence,if
therewere information spillovers in the post‐intervention period, then onewould expect to see
highertreatmenteffectsinthepost‐interventionperiodthanintheinterventionperiod.
Finally,onemightarguethattheclinicscontinuedthenewroutinesafterthetemporaryfees
wereeliminatedbecausetheyfacedalargefixedcostofrevertingtotheoldroutinesandnotbecause
19
thenewroutinesaddednetvalue.However,inthiscase,wethinkthatthefixedcostsofreversingthe
routinesweresmall,becausethecommunityhealthworkerscouldsimplyhavereturnedtotheirold
patternsofactivities.
7 CROSS‐PRICEEFFECTS(SPILLOVER)
Whilethemodifiedfeeschedulewasdesignedtoaffectthetimingofthefirstprenatalvisit,
wemightexpectproviderstoreduceeffortsuppliedtootherservices,resultinginalowerprovision
of such services to patients. We test for this by estimating the effect of the incentives on the
probabilityofpregnantwomenhavingavalidtetanusvaccine,andthenumberofprenatalvisits.The
resultspresentedinTable6reportnoevidenceofcross‐priceeffects,positiveornegative,ineither
theinterventionperiodorinpost‐interventionperiodI.Infact,thelevelsoftheseservicesappearto
beconstantovertime.Whiletheconcernaboutcrowding‐outistypicallyforacontextofindividual
providersfacingtimeandeffortconstraints,ourresultsareconsistentwithafirmsettingwherethere
arenooveralleffortortimeconstraints.
8 BIRTHOUTCOMES
Nextwe address thequestionofwhether the effect of the incentives for early initiationof
prenatal care translated into improved birth outcomes as measured by birth weight, low birth
weight,andprematurebirth.AsshowninFigure7andreportedinTable7wefindnoeffectofthe
incentivesonbirthoutcomesineithertheinterventionperiodorinthepost‐interventionperiod.
Thereareanumberofpossiblereasonsforthis.First,thesamplecouldbetoosmalltobeable
todetectastatisticallysignificanteffectonoutcomes.However,thepointestimatesareverysmall,
halfofthemarenegativeandtheyareofsimilarmagnitudetodifferencesbetweentreatmentand
controlgroupsinthepre‐interventionperiod.Second,giventhattheresultsonbirthoutcomesare
obtainedfromananalysisofasubsampleofbeneficiariesforwhomwewereabletomergeprenatal
carerecordswithhospitalmedicalrecords,itispossiblethattheresultsinTable4donotholdfor
thissubsample.Wethereforereplicatetheprenatalcareanalysisusingonlythesubsampleofwomen
forwhomhospitalmedicalrecordsareavailable.Overall,weobtainsimilarresultstothoseobtained
withthefullsample.27Third,despitethemedicalliteratureandCPGrecommendation,itispossible
that early initiation of carematters only a small amount for the general population of pregnant
27Resultsofthisanalysisareavailableuponrequest.
20
women,evenifearlyinitiationofcaremattersagreatdealforhigh‐riskpatients.Highriskpatients
include, among others, smokers, substance abusers, those with poor medical and pregnancy
histories,andthosewhostartprenatalcareverylateintheirthirdtrimesteroronlywhenaproblem
occurs.Itmaybethattheincreaseinearlyinitiationofcarecomesfromprimarilylow‐riskmothers
whoarelesslikelytobenefitfromearlyinitiationofcare.Onewouldthinkthatitwouldbeeasierto
persuade low‐riskmothers to come a littler earlier than to convince high‐riskmotherswho are
reluctanttocomeforanycareatall.
Infact,thisisconsistentwiththesmallreductionintheaverageweekspregnantatthetime
ofthefirstprenatalvisit.Onaverage,womeninthetreatmentgroupinitiatedprenatalcareabout
1.5weeks earlier thanwomen in the control group. Prenatal caremay affect birth outcomes by
diagnosingandtreating illnesssuchashypertensionandgestationaldiabetesaswellas trying to
changematernalbehaviorthroughpromotingactivitiessuchasgoodnutrition,notsmokingandnot
consumingalcohol.Iftheinterventionhadinducedhigh‐riskwomenwhootherwisewouldhavehad
1stvisitmuchlaterinthepregnancy,thentheincentivesmayhavehadameasurableimpactonbirth
outcomes.Hence,whiletheincentiveswereeffectiveinincreasingearlyinitiationofcare,theydid
notmanagetosufficientlyaffectthegroupmostlikelytobenefit.Thesolutionmightbetocondition
incentivesonattendinghigh‐riskwomen,butriskisdifficultandexpensivetoidentifyandverifyand
thereforemaynotbecontractible.
9 DISCUSSION
Weexaminetheeffectsoftemporaryfinancialincentivesformedicalcareproviderstoincreaseearly
initiationofprenatalcareforpregnantwomenusingarandomizedcontrolledtrialinArgentina.The
interventionrandomlyallocatesathree‐foldincreaseinthefeepaidtohealthfacilitiesforeachinitial
prenatalvisitthatoccursbeforeweek13ofpregnancy.Thispremiumwasimplementedforaperiod
of8monthsandthenended.Usingdataonhealthservicesandbirthoutcomesfrommedicalrecords,
weinvestigateboththeshort‐termeffectsoftheincentiveandwhethertheeffectspersistoncethe
directmonetarycompensationdisappears.
Our results suggest that the temporary incentives motivated long‐run changes in
performance.Wefindthattheincentivesledtopregnantwomenbeing35%morelikelytoinitiate
prenatalcarebeforeweek13andthatthehigherlevelsofearlyinitiationofcarepersistedforatleast
15monthsandlikelymorethan24monthsaftertheincentivesended.Theseresultsareconsistent
with a model of providers who face a fixed cost to changing their clinical practice routines, i.e.
21
organizational inertia.Temporary incentives inducedproviders to adopt changes to their clinical
practicepatternsbyhelpingthemtoovercomeinertia.Oncetheyadoptchangestopracticepatterns
thattheybelievearebeneficialtopatients,thechangespersistevenafterthemonetaryincentives
disappear.Theseresultsareconsistentwiththefindingsfromin‐depthinterviewsthatevidenced
thattreatmentclinicsadoptedinnovativepracticesandchangedroutinesinordertoincreaseearly
initiationofprenatalcare.
Ourstudyaddstothegrowingbodyofevidencethatincentivesareeffectiveinimproving
providerperformance.Ourresultsalsohaveanumberofimportantpolicyimplications.First,our
results suggest that temporary incentives may be effective in motivating long‐term provider
performanceat a substantially lower cost thanpermanent incentives. Second,whilewe find that
incentivesareabletomotivatechangesinclinicalpracticepatterns,wedidnotfindimprovements
inhealthoutcomes.Themonetaryincentivesthatwereimplementedwerenotabletosufficiently
reachthosewomenforwhomearlyinitiationofprenatalcarewouldhavethelargesthealthimpact.
Therefore,incentivesmaybemademoreeffectivebydefiningex‐antethepopulationmostlikelyto
benefit,andtailoringincentivestowardsthispopulation.However,tailoringincentivestohighrisk
populations or those most likely to benefit from the services may not be contractible as these
characteristics are typically not observable. This is maybe a major limitation of using incentive
contractstoimprovehealthoutcomes.
22
REFERENCES
Acland, D., & Levy, M. R. (2015). “Naiveté, projection bias, and habit formation in gymattendance,”ManagementScience,61(1),146‐160.
Ariely,D.,Gneezy,U.,Loewenstein,G.,&Mazar,N.(2009).“Largestakesandbigmistakes,”TheReviewofEconomicStudies,76(2),451‐469.
Baker, G. P., Jensen, M. C., & Murphy, K. J. (1988). “Compensation and incentives: practice vs.theory,”TheJournalofFinance,43(3),593‐616.
Barber,S.L.,&Gertler,P.J.(2009).“Empoweringwomentoobtainhighqualitycare:evidencefroman evaluation of Mexico's conditional cash transfer programme,” Health Policy andPlanning,24(1),18‐25.
Basinga,P.,Gertler,P.J.,Binagwaho,A.,Soucat,A.L.,Sturdy,J.,&Vermeersch,C.M.(2011).“EffectonmaternalandchildhealthservicesinRwandaofpaymenttoprimaryhealth‐careprovidersforperformance:animpactevaluation,”TheLancet,377(9775),1421‐1428.
Becker,G.S.&Murphy,K.M.(1988).“Atheoryofrationaladdiction,”TheJournalofPoliticalEconomy,96(4),675‐700.
Becker,M.C.(2004).“Organizationalroutines:areviewoftheliterature,”IndustrialandCorporateChange,13(4),643‐678.
Benabou, R. & Tirole, J. (2003). “Intrinsic and extrinsic motivation,” The Review of EconomicStudies,70(3),489‐520.
Blattberg, R. C. & Neslin, S. A. (1990). “Sales promotion: concepts, methods, and strategies,”EnglewoodCliffs,PrenticeHall,NewJersey.
Bloom,N.,Propper,C.,Siler,S.,&VanReenan,J.(2015).“Theimpactofcompetitiononmanagementquality:Evidencefrompublichospitals,”TheReviewofEconomicStudies,82(2),457‐489.
Bonfrer, I.,Soeters,R.,vandePoel,E.,Basenya,O.,Longin,G.,vandeLooij,F.,&vanDoorslaer,E.(2013).“Theeffectsofperformance‐basedfinancingontheuseandqualityofhealthcareinBurundi:animpactevaluation,”TheLancet,381,S19.
Bordalo,P.,Gennaioli,N.&Shleifer,A.(2012).“Saliencetheoryofchoiceunderrisk,”TheQuarterlyJournalofEconomics,127(3):1243‐1285.
Bordalo,P.,Gennaioli,N.&Shleifer,A.(2013).“Salienceandconsumerchoice,”TheJournalofPoliticalEconomy,121(5),803‐843.
Cabana,M.D.,Rand,C.S.,Powe,N.R.,Wu,A.W.,Wilson,M.H.,Abboud,P.A.C.,&Rubin,H.R.(1999).“Why don't physicians follow clinical practice guidelines?: A framework forimprovement,”JAMA,282(15),1458‐1465.
Cameron,A.C.,Gelbach, J.B.,&Miller,D.L.(2008).“Bootstrap‐basedimprovements for inferencewithclusterederrors,”TheReviewofEconomicsandStatistics,90(3),414‐427.
Campbell,O.M.&Graham,W.J.(2006).“Strategiesforreducingmaternalmortality:Gettingonwithwhatworks,”TheLancet,368(9543),1284‐1299.
Campbell,S.,Reeves,D.,Kontopantelis,E.,Middleton,E.,Sibbald,B.,&Roland,M.(2007).“Qualityofprimarycare inEnglandwith the introductionofpay forperformance,”TheNewEnglandJournalofMedicine,357(2),181‐190.
23
Carroll,G.R.,&Hannan,M.T.(2000).“Thedemographyofcorporationsandindustries,”PrincetonUniversityPress.
Carroli,G.,Villar,J.,Piaggio,G.,Khan‐Neelofur,D.,Gülmezoglu,M.,Mugford,M.,&Bersgjø,P.(2001).“WHO systematic review of randomized controlled trials of routine antenatal care,” TheLancet,357(9268),1565‐1570.
Carroli,G.,Rooney,C.,&Villar, J. (2001). “Howeffective is antenatal care inpreventingmaternalmortality and seriousmorbidity? AnOverview of the Evidence,”PaediatricandPerinatalEpidemiology,15(s1),1‐42.
Cawley,J.,&Price,J.A.(2013).“Acasestudyofaworkplacewellnessprogramthatoffersfinancialincentivesforweightloss,”JournalofHealthEconomics,32(5),794‐803.
Charness,G.&Gneezy,U.(2009).“Incentivestoexercise,”Econometrica,77(3),909‐931.
Clemens,J.&Gottlieb,J.D.(2014).“Dophysicians'financialincentivesaffectmedicaltreatmentandpatienthealth?”TheAmericanEconomicReview,104(4),1320‐1349.
Das,J.,&Gertler,P.J.(2007).“Variationsinpracticequalityinfivelow‐incomecountries:aconceptualoverview,”HealthAffairs,26(3),w296‐w309.
Das, J.&Hammer, J. (2005). “WhichDoctor? Combining vignettes and item response tomeasureclinicalcompetence,”JournalofDevelopmentEconomics,78(2),348‐383.
Das,J.,Hammer,J.,&Leonard,K.(2008).“Thequalityofmedicaladviceinlow‐incomecountries,”TheJournalofEconomicPerspectives,22(2),93‐114.
Davidson,R.&Flachaire, E. (2008). "TheWildbootstrap, tamedat last," JournalofEconometrics,146(1),162‐169.
deMel,S.,McIntosh,C.,&Woodruff,C.(2013).“Depositcollecting:Unbundlingtheroleoffrequency,salience,andhabitformationingeneratingsavings,”TheAmericanEconomicReview,103(3),387‐92.
DeWalque,D.,Gertler,P.J.,Bautista‐Arredondo,S.,Kwan,A.,Vermeersch,C.,deDieuBizimana,J.,&Condo, J. (2015). “Using provider performance incentives to increase HIV testing andcounselingservicesinRwanda,”JournalofHealthEconomics,40(2),1‐9.
Deci, E. L. (1971). “Effecs of eternally mediated rewards on intrinsic motivation,” Journal ofPersonalityandSocialPsychology,18,105‐115.
Deci,E.L.,Koestner,R.,&Ryan,R.M.(1999).“Ameta‐analyticreviewofexperimentsexaminingtheeffectsofextrinsicrewardsonintrinsicmotivation,”PsychologicalBulletin,125(6),627.
Deci, E. L., Koestner, R., & Ryan, R. M. (2001). “Extrinsic rewards and intrinsic motivation ineducation:Reconsideredonceagain,”ReviewofEducationalResearch,71(1),1‐27.
Deci,E.L.andRyan,R.M.(2010).“Self‐determination,”JohnWiley&Sons,Inc.
DellaVigna, S. (2009). “Psychology and economics: evidence from the field,” Journal ofEconomicLiterature,47(2),315‐372.
Dupas,P.(2014).“Short‐runsubsidiesandlong‐runadoptionofnewhealthproducts:Evidencefromafieldexperiment,”Econometrica,82(1),197‐28.
Eccles, J. S. & Wigfield, A. (2002). “Motivational beliefs, values, and goals,” Annual Review ofPsychology,53(1),109‐132.
24
Fehr,E.&Falk,A.(1999).“Wagerigidity inacompetitiveincompletecontractmarket,” JournalofPoliticalEconomy,107(1),106‐134.
Fehr,E.&Schmidt,K.M.(2000).“Fairness,incentives,andcontractualchoices,”EuropeanEconomicReview,44(4),1057‐1068.
Flores,G.,Ir,P.,Men,C.R.,O’Donnell,O.,&vanDoorslaer,E.(2013).“Financialprotectionofpatientsthroughcompensationofproviders:TheimpactofhealthequityfundsinCambodia,”JournalofHealthEconomics,32(6),1180‐1193.
Gelbach,J.B.,Klick,J.,&Stratmann,T.(2009).“Cheapdonutsandexpensivebroccoli:theeffectofrelativepricesonobesity,”WorkingPaper.
Gertler,P.,Giovagnoli,P.I.,&Martinez,S.W.(2014).“Rewardingproviderperformancetoenableahealthy start to life: evidence from Argentina's Plan Nacer,”World Bank Policy ResearchWorkingPaper,6884,WorldBank,Washington,DC.
Gertler,P.SeiraE.,andScottA.(2015).“Long‐termeffectsoftemporaryprize‐linkedsavingslotteriesonaccountsopeningsandbalances,”UCBerkeleyWorkingPaper,BerkeleyCalifornia.
Gertler, P., & Vermeersch, C. (2012). “Using performance incentives to improve healthoutcomes,”WorldBankPolicyResearchWorkingPaper.
Gertler, P. & Vermeersch, C. (2013). “Using performance incentives to improve medical careproductivity and health outcomes,” NBER Working Papers 19046, National Bureau ofEconomicResearch,Cambridge,MA.
Gibbons,R.(1997).“Anintroductiontoapplicablegametheory,”JournalofEconomicPerspectives,11(1),127‐149.
Gibbons, R., & Henderson, R. (2012). “Relational contracts and organizationalcapabilities,”OrganizationScience,23(5),1350‐1364.
Gibbons, R., & Henderson, R. (2013). “What do managers do? Exploring persistent performancedifferences amongst seemingly similar enterprises,” The Handbook of OrganizationalEconomics,Chapter17,pages680‐731,RobertGibbonsandJohnRoberts,Editors,PrincetonUniversityPress,PrincetonandOxford.
Gneezy, U., & Rustichini, A. (2000a). “Pay enough or don't pay at all,” The Quarterly Journal ofEconomics,115(3),791‐810.
Gneezy,U.,&Rustichini,A.(2000b).“Afineprice,”TheJournalofLegalStudies,29(1),1‐17.
Grol,R.P.T.M.(1990).“Nationalstandardsettingforqualityofcareingeneralpractice:attitudesofgeneral practitioners and response to a set of standards,” British Journal of GeneralPractice,40(338),361‐364.
Grol, R. (2001). “Successes and failures in the implementation of evidence‐based guidelines forclinicalpractice,”MedicalCare,39(8),11‐46.
Grol,R.,&Grimshaw, J. (2003). “Frombestevidencetobestpractice:effective implementationofchangeinpatients'care,”TheLancet,362(9391),1225‐1230.
Hannan, M. T., & Freeman, J. (1984). “Structural inertia and organizational change,” AmericanSociologicalReview,149‐164.
Hoff, T. (2014). “When routines support or stifle innovation: Evidence from primary carepractices,”AcademyofManagementProceedings,Vol.2014,No.1,p.11116.
25
Holmstrom,B.&Milgrom,P.(1991).“Multitaskprincipal‐agentanalyses:Incentivecontracts,assetownership,andjobDesign,”JournalofLaw,Economics,&Organization,7(SpecialIssue),24‐52.
Hudak, B. B., O'Donnell, J., & Mazyrka, N. (1995). “Infant sleep position: pediatricians' advice toparents,”Pediatrics,95(1),55‐58.
Huillery,E.&Seban,J.(2014).“Pay‐for‐Performance,motivationandfinaloutputinthehealthsector:ExperimentalevidencefromtheDemocraticRepublicofCongo,”WorkingPaper,DepartmentofEconomics,SciencesPo,Paris.
Imbens, G.W. & Angrist, J. D. (1994). “Identification and estimation of Local Average TreatmentEffects,”Econometrica,62(2),467‐475.
John,L.K.,Loewenstein,G.,Troxel,A.B.,Norton,L.,Fassbender,J.E.,&Volpp,K.G.(2011).“Financialincentives for extended weight loss: a randomized, controlled trial,” Journal of GeneralInternalMedicine,26(6),621‐626.
Kahneman,D.(2012).“Thinking,fastandslow,”Farrar,StrausandGiroux,NewYork.
Karlan,D.,M.McConnell,S.Mullainathan&JonathanZinman(2015).“Gettingtothetopofmind:Howremindersincreasesavings,”ManagementScience,forthcoming.
Kirmani,A.&Rao,A.R. (2000). “Nopain,nogain:A critical reviewof the literatureon signalingunobservedproductquality,”JournalofMarketing,64(2),66–79.
Kolstad,J.T.(2013).“Informationandqualitywhenmotivationisintrinsic:Evidencefromsurgeonreportcards,”TheAmericanEconomicReview,103(7),2875‐2910.
Lazear, E. P. (2000). “Performance pay and productivity,”TheAmericanEconomicReview,90(5),1346‐1361.
Leonard,K.L.&Masatu,M.C. (2010), “Professionalismand theknow‐dogap:Exploring intrinsicmotivationamonghealthworkersinTanzania,”HealthEconomics,19(12),1461‐1477.
Main,D.S.,Cohen,S.J.,&DiClemente,C.C.(1995).“Measuringphysicianreadinesstochangecancerscreening:preliminaryresults,”AmericanJournalofPreventiveMedicine.
Miller,G.&Babiarz,K.S.(2013).“Pay‐for‐performanceincentivesinlow‐andmiddle‐incomecountryhealthprograms,”NBERWorkingPapers18932,NationalBureauofEconomicResearch,Inc.
Mohanan,M.,Vera‐Hernández,M.,Das,V.,Giardili,S.,Goldhaber‐Fiebert,J.D.,Rabin,T.L.,&Seth,A.(2015).“Theknow‐dogapinqualityofhealthcareforchildhooddiarrheaandpneumoniainruralIndia,”JAMAPediatrics.
Musgrove,P.(2010).“PlanNacer,Argentina:ProvincialmaternalandchildhealthinsuranceusingResults‐BasedFinancing(RBF),”Mimeo.
NationalMinistryofHealth(2009)."InformedegestiónPlanNacer,"ÁreaTécnica,UnidadEjecutoraCentral.BuenosAires,Argentina.
NationalMinistryofHealth(2010)."InformedegestiónPlanNacer,"ÁreaTécnica,UnidadEjecutoraCentral.RevisedversionMarch.BuenosAires,Argentina.
National Ministry of Health (2010b). ”Nomenclador único 2010,” Plan Nacer, Buenos Aires,Argentina.
Nelson,R.& S.Winter (1982). “An evolutionary theory of economic change,”HarvardUniversityPress.
26
Pathman,D.E.,Konrad,T.R.,Freed,G.L.,Freeman,V.A.,&Koch,G.G.(1996).“Theawareness‐to‐adherencemodelofthestepstoclinicalguidelinecompliance:thecaseofpediatricvaccinerecommendations,”MedicalCare,34(9),873‐889.
Pittman,T.S.&Heller,J.F.(1987).“Socialmotivation,”AnnualReviewofPsychology,38(1),461‐490.
Royer,H.M.Stehr,andJ.Sydnor(2012).“Incentives,commitmentsandhabitformationinexercise:evidencefromafieldexperimentwithworkersataFortune‐500company”NBERWorkingPaper18580,forthcominginAmericanJournalofEconomics:AppliedEconomics.
Schaner, S., (2015). “Thepersistent power of behavioral change: Long run impacts of temporarysavings subsidies for the poor.” Department of Economics, Dartmouth University,http://www.dartmouth.edu/~sschaner/main_files/Schaner_LongRun.pdf
Schuster,M.A.,McGlynn,E.A.,&Brook,R.H.(1998).“HowgoodisthequalityofhealthcareintheUnitedStates?,”MilbankQuarterly,76(4),517‐563.
Schwarcz,R.,Uranga,A.,Lomuto,C.,Martinez, I.,Galimberti,D.,García,O.M.,Etcheverry,M.E.,&Queiruga,M.(2001)."Elcuidadoprenatal:Guíaparalaprácticadelcuidadopreconcepcionalydelcontrolprenatal."NationalMinistryofHealth,Argentina.
Taylor,S.E.,&Thompson,S.C.(1982).“Stalkingtheelusive‘vividness’effect,”PsychologicalReview,89(2),155.
Thaler, R. H. & Sunstein C.R. (2009). “Nudge: Improving decisions about health, wealth, andhappiness,”PenguinBooks,NewYork.
Volpp,K.G.,John,L.K.,Troxel,A.B.,Norton,L.,Fassbender,J.,&Loewenstein,G.(2008).“Financialincentive–basedapproachesforweightloss:arandomizedtrial,”JAMA,300(22),2631‐2637.
Volpp,K.G.,Troxel,A.B.,Pauly,M.V.,Glick,H.A.,Puig,A.,Asch,D.A.,...&Audrain‐McGovern,J.(2009).“A randomized, controlled trial of financial incentives for smoking cessation,” The NewEnglandJournalofMedicine,360(7),699‐709.
Wooldridge, J. M. (2007). “Inverse probability weighted estimation for general missing dataproblems,”JournalofEconometrics,141(2),1281‐1301.
WorldHealthOrganization(2006).“Standardsformaternalandneonatalcare:Provisionofeffectiveantenatalcare,”WorldHealthOrganization,Geneva.
WorldHealthOrganization(2014).“WorldHealthStatistics:Healthrelatedmillenniumdevelopmentgoals,”WorldHealthOrganization,Geneva.
27
FIGURESANDTABLES
Figure1:ProviderCompliancewithClinicalPracticeGuidelines
Source:Authors’elaborationbasedon(‐)Schusteretal.(1998);(+)Grol(2001);(++)Campbelletal.(2007);(*)DasandGertler(2007);and(#)GertlerandVermeersch(2012).
28
Figure2:TimelineandDataAvailability
29
Figure3:DensitiesofWeeksPregnantat1stPrenatalVisit
Notes:Densities estimatedusinganEpanechnikovkernelwithoptimalbandwidth.P‐valesofKolmogorov‐Smirnovtestsofequalityofdistributionsbetweengroupsreportedbelowfigure.Thetwoverticallinesindicateweeks13and20ofpregnancy.Source:Authors’ownelaborationbasedondatafromtheprovincialmedicalrecordinformationsystem.
30
Figure4:MeanNumberofWeeksPregnantat1stPrenatalVisit
Notes:Thefirsttwopoints(circles)aremeansfor6‐monthperiodspriortotheinterventionperiod.Thethirdpoint(Diamond)correspondstotheinterventionperiod.Thefourthandfifthpoints(triangles)correspondto6‐monthsperiodsaftertheinterventionperiod,whilethelastpoint(triangle)isfora3‐monthperiod.
31
Figure5:ProportionofMotherswith1stPrenatalVisitbeforeWeek13ofPregnancy
Notes:Thefirsttwopoints(circles)aremeansfor6‐monthperiodspriortotheinterventionperiod.Thethirdpoint(Diamond)correspondstotheinterventionperiod.Thefourthandfifthpoints(triangles)correspondto6‐monthsperiodsaftertheinterventionperiod,whilethelastpoint(triangle)isfora3‐monthperiod.
32
Figure6:NumberofClinicOutreachActivities
Notes:Theheightofthebarsreportthemeanandmediannumberofoutreachactivitiesthatresultedinactualmaternal‐childserviceattheclinic,pertrimesterforthepre‐interventionperiod(January2009‐April2010),theinterventionperiod(May‐December2010),andpost‐interventionperiodI(January2011‐March2012)
33
Figure7:BirthWeightDensities
Notes:Densities estimatedusinganEpanechnikovkernelwithoptimalbandwidth.P‐valesofKolmogorov‐Smirnov tests of equality of distributions between groups reported below figure. Source: Authors’ ownelaborationbasedonmedicalrecordinformationsystem.
34
Figure8:AbsoluteScoreofImportanceofPrenatalCareServices
Notes:Thisgraphreportstheaverageoftheabsolutescorethatmeasurestheimportancegivenbyclinicstosevendifferentprenatalcareproceduresincludinginitiatingprenatalcarepriortoweek13ofpregnancy.Thedatawerecollectedusingashortonlinesurveyconductedintheclinicsthatparticipatedintheexperiment.(seeAppendixD)Theabsolutescoresrangefrom1to5,with5beingthehighestscoreintermsofimportance.Therespondwascodedzeroiftherespondentreportedthatthisprocedureisinappropriateforapregnantwoman.
35
Figure9:RelativeRankingofImportanceofPrenatalCareServices
Notes:Thisgraphreportstheaverageof therelativerankingthatmeasuresthedegreeofprioritygivenbyclinics to seven different prenatal care procedures including initiating prenatal care prior to week 13 ofpregnancy.Thedatawerecollectedusingashortonlinesurveyconductedintheclinicsthatparticipatedintheexperiment.(seeAppendixD)Therelativescoresaimedtorankthesevenpracticesfrom1to7,with1beingthehighestranking.Inpracticehowever,thesurveyinstrumentallowedtherespondenttorepeatnumbers.
36
Table1:Paymentsfor1stPrenatalVisit
TimePeriod Dates Paymentfor1stPrenatalVisit
Begin End BeforeWeek
13ofpregnancy
Atweek13ofpregnancyor
after
Pre‐Intervention January2009 April2010 $40ARS $40ARS
Intervention May2010 December2010 $120ARS $40ARS
PostIntervention January2011 December2012 $40ARS $40ARS
Source:NationalMinistryofHealth,Argentina(2010b)
Table2:ClinicAssignmentandComplianceStatus
AssignedtoTreatment
ActuallyTreated
Total Yes No
Yes 14 4 18
No 1 18 19
Total 15 22 37
Source:Authors’elaboration.
37
Table3:BaselineDescriptiveStatistics
AssignedTreatmentGroup
AssignedControlGroup
p‐Valuefortestofequalityofmeans
Mean(s.d.) N
Mean(s.d.) N
Largesample
WildBoot‐
Strapped
WeeksPregnantat1stPrenatalVisit 17.5 743 17.6 497 0.89 0.84 (7.48) (7.74)
1stVisitbeforeWeek13ofPregnancy 0.35 743 0.33 497 0.57 0.56 (0.48) (0.47)
TetanusVaccineDuringPrenatalVisit 0.80 743 0.84 497 0.34 0.41 (0.40) (0.37)
NumberofPrenatalVisits 4.68 743 4.28 497 0.39 0.45 (2.94) (2.77)
BirthWeight(grams) 3,328 552 3,291 379 0.36 0.37 (519) (558)
LowBirthWeight(<2500grams) 0.06 552 0.06 379 0.96 0.98 (0.23) (0.23)
Premature(gestationalage<37weeks) 0.09 319 0.10 249 0.83 0.82 (0.29) 0.30
MaternalAge 25.36 354 25.75 270 0.47 0.48 (6.49) 6.10
NumberofPreviousPregnancies 2.31 354 2.10 273 0.29 0.32 (2.39) (2.10)
FirstPregnancy 0.25 354 0.26 273 0.70 0.77 (0.43) (0.44)
Notes:Thistablepresentsmeansandstandarddeviationsinparenthesesforthetreatmentandcontrolgroupsduringthe16‐monthpre‐interventionperiodfromJanuary2009throughApril2010.P‐valuesfortestsequalityof treatment and control groupsmeans are presented in the last 2 columns.Wepresent both thep‐valuecomputedforlargesamplesandaWildbootstrappedp‐valuethatisrobustinsampleswithsmallnumbersofclusters(Cameronetal.2008).OurWildbootstrapprocedureassignssymmetricweightsandequalprobabilityafterre‐samplingresiduals(DavidsonandFlachaire2008)anduses999replications.
38
Table4:EffectsonTemporaryIncentivesonTimingof1stPrenatalVisit
(1) (2) (3)
InterventionPeriodPost‐Intervention
PeriodI(Jan2011–March2012)
Post‐InterventionPeriodII
(April–Dec2012)
A.WeeksPregnantat1stPrenatalVisit
Treatment ‐1.47** ‐1.63** ‐2.47**
(0.71) (0.75) (1.02)
LargeSamplep‐value 0.04 0.03 0.02
WildBootstrappedp‐value 0.08 0.03 0.03
ControlGroupMean 17.80 17.90 20.10
SampleSize 769 1,296 710
B.FirstPrenatalVisitBeforeWeek13ofPregnancy
Treatment 0.11** 0.08** 0.08**
(0.04) (0.04) (0.04)
LargeSamplep‐value 0.01 0.02 0.04
WildBootstrappedp‐value 0.03 0.05 0.06
ControlGroupMean 0.31 0.34 0.27
SampleSize 769 1,296 710
Notes:ThistablereportsLATEestimatesofthetreatmenteffectofthemodifiedfeescheduleonindicatorsofthe timing of the 1st prenatal visit. The differences are estimated from2SLS regressions of the dependentvariableonactualtreatmentstatusinstrumentedwithclinictreatmentassignmenttype.Thep‐valuesarefor2‐sidedhypothesistestsofthenullthatthedifferenceisequaltozero.Wepresentboththep‐valuecomputedforlargesamplesandaWildbootstrappedp‐valuethatisrobustinsampleswithsmallnumbersofclusters(Cameronetal.2008).OurWildbootstrapprocedureassignssymmetricweightsandequalprobabilityafterre‐samplingresiduals(DavidsonandFlachaire2008)anduses999replications.Column(1)reportstheresultsforthesampleobservedinan8‐monthinterventionperiod(May2010–December2010).Column(2)reportstheresultsforthesampleobservedinthe15‐monthperiodfollowingtheendoftheintervention(January2011–March2012).Column(3)reportstheresultsforthe9‐monthperiodafterthechangeinthecodingofthefirstprenatalvisit(April2012–December2012).Standarderrorsareinparentheses.*p<0.10,**p<0.05,***p<0.01.
39
Table5:ImpactonLogNumberofOutreachActivities
(1) (2)
InterventionPeriodPost‐InterventionPeriodI(Jan2011–March2012)
Treatment 0.47** 0.56** (0.23) (0.22)
LargeSamplep‐value 0.04 0.01
WildBootstrappedp‐value 0.04 0.02
Log(ControlGroupMean) 1.93 1.93
SampleSize 324 324
Notes:ThistablereportsLATEestimatesofthetreatmenteffectofthemodifiedfeeschedule.Thedependentvariableisthelogofthenumberofclinicoutreachactivitiesthatresultedinactualmaternal‐childserviceattheclinicpertrimester.Thep‐valuesarefor2‐sidedhypothesistestsofthenullthatthedifferenceisequaltozero.Wepresentboththep‐valuecomputedforlargesamplesandaWildbootstrappedp‐valuethatisrobustin samples with small numbers of clusters (Cameron et al. 2008). OurWild bootstrap assigns symmetricweights and equal probability after re‐sampling residuals (Davidson and Flachaire 2008) and uses 999replications.Theseareonlycomputedforthecoefficientsoftreatmentinteractedwitheachperiod.Standarderrorsareinparentheses.*p<0.10,**p<0.05,***p<0.01.
40
Table6:Cross‐PriceEffects(Spillover)
(1) (2)
InterventionPeriodPost‐InterventionPeriodI
(Jan–Dec2011)
A.TetanusVaccine
Treatment
0.02 ‐0.02 (0.08) (0.05)
LargeSamplep‐value 0.76 0.62
WildBootstrappedp‐value 0.75 0.67
ControlGroupMean 0.79 0.84
SampleSize 769 1,053
A.Numberofvisits
Treatment
0.39 0.51 (0.33) (0.58)
LargeSamplep‐value 0.24 0.38
WildBootstrappedp‐value 0.27 0.41
ControlGroupMean 4.05 4.40
SampleSize 769 1,053
Notes:ThistablereportsLATEestimatesofthetreatmenteffectofthemodifiedfeescheduleonindicatorsofother services. The differences are estimated from 2SLS regressions of the dependent variable on actualtreatmentstatusinstrumentedwithclinictreatmentassignmenttype.Thep‐valuesarefor2‐sidedhypothesistestsofthenullthatthedifferenceisequaltozero.Wepresentboththep‐valuecomputedforlargesamplesandaWildbootstrappedp‐valuethatisrobustinsampleswithsmallnumbersofclusters(Cameronetal.2008).OurWildbootstrapprocedureassignssymmetricweightsandequalprobabilityafterre‐samplingresiduals(Davidson and Flachaire 2008) and uses 999 replications. Column (1) reports the results for the sampleobservedinan8‐monthinterventionperiod(May2010–December2010).Column(3)reportstheresultsforthesampleobservedinthe12‐monthperiodfollowingtheendoftheintervention(January2011–December2011).Standarderrorsareinparentheses.*p<0.10,**p<0.05,***p<0.01.
41
Table7:ImpactofIncentivesonBirthOutcomes
(1) (2)
InterventionPeriod Post‐InterventionPeriodI(Jan–Dec2011)
A.BirthWeight
Treatment
‐37.34 25.109 (48.61) (40.67)
LargeSamplep‐value 0.44 0.54
WildBootstrappedp‐value 0.49 0.51
ControlGroupMean 3,304 3,279
SampleSize 555 802
B.LowBirthWeight
Treatment
0.01 ‐0.01 (0.02) (0.02)
LargeSamplep‐value 0.63 0.60
WildBootstrappedp‐value 0.61 0.56
ControlGroupMean 0.05 0.06
SampleSize 555 802
B.Premature
Treatment
0.03 ‐0.04 (0.03) (0.02)
LargeSamplep‐value 0.31 0.08
WildBootstrappedp‐value 0.28 0.12
ControlGroupMean 0.09 0.12
SampleSize 414 708
Notes:ThistablereportsLATEestimatesofthetreatmenteffectofthemodifiedfeescheduleforonindicatorsof birthoutcomes. The observations includewoman forwhomweare able toobtain informationonbirthoutcomesprovidedinpublichospitalbirthrecords.Thedifferencesareestimatedfrom2SLSregressionsofthedependentvariableonactual treatment status instrumentedwithclinic treatmentassignment type.Thep‐valuesarefor2‐sidedhypothesistestsofthenullthatthedifferenceisequaltozero.Wepresentboththep‐value computed for large samples and a Wild bootstrapped p‐value that is robust in samples with smallnumbersofclusters(Cameronetal.2008).OurWildbootstrapprocedureassignssymmetricweightsandequalprobabilityafterre‐samplingresiduals(DavidsonandFlachaire2008)anduses999replications.Column(1)reportstheresultsforthesampleobservedinan8‐monthinterventionperiod(May2010–December2010).Column (2) reports the results for the sample observed in the 12‐month period following the end of theintervention(January2011–December2011).Standarderrorsareinparentheses.*p<0.10,**p<0.05,***p<0.01.
42
APPENDIXA:TESTOFMISREPORTINGWEEKSPREGNANTAT1STPRENATALVISIT
Oneconcernisthatthefinancialincentivesmaycauseclinicstomisreporttheweekofpregnancyat
thefirstvisit.Inthisappendixwereporttheresultsoftestforthisbehavior.Recallthatinourmain
analysisweconstructtheweekofpregnancyatthefirstvisitusingthedateofthefirstvisitandthe
last menstrual date (LMD) as reported by the women. If the latter is not available we use the
estimateddateofbirth(EDD)asrecordedbythephysicianinthefirstvisit.TheEDDiscalculatedoff
theLMDasreportedbythewomenduringherfirstvisit.Whileclinicmedicalrecordsshouldcontain
bothdates,about10%ofrecordsaremissingtheLMD.
OnepossiblewayofmisreportingtheweekofpregnancyatthefirstvisitistochangetheLMD
andtheEDDinthepatient’sclinicalmedicalrecord.Forinstance,ifawomanisinher21stweekof
pregnancyatthefirstvisit,thephysiciancouldadd7daystotheLMDandEDDsothatthevisitfalls
intothe20thweekofpregnancy.Bothwouldhavetobechangedinordertodeceivetheauditors.
To test for this possibility we use gestational age at birth (GAB) in weeks measured by
physicalexaminationatthetimeofbirth,registeredinthehospitalmedicalrecord.Wethencompare
theweekselapsedfromthefirstprenatalvisittothedeliverydatebasedonGABtoweekselapsed
fromfirstvisittothedeliverydatebasedonEDD.WhileEDDiscollectedbytheclinicwhohasan
incentive tomisreport, theGAB is collectedby thehospital at timeofdeliverywhere there isno
incentivetomisreport.
FigureA1plotsthenumberofweekstodeliveryfromthetimeofthe1stvisitbasedonGAB
(y‐axis)totheonebasedonEDD(x‐axis).Ifthereisnodifferencebetweenthetwomeasures,then
allofthedatesshouldfallonthe45‐degreeblueline.ThereshouldbesomedifferencesasEDDisan
estimatethatassumesnoprematurityatbirth,andtherecouldbedataentryinGABandEDDand
recallerrorsinEDD.FigureA1showsthatalmostallofthedataembracetheblue45‐degreelineand
mostoftheobservationsoffthelinearesituatedaboveit,consistentwithprematurityexplainingthe
differences.
IftheclinicchangestheEDDinordertocapturehigherpayments,wewouldexpectgreater
differences, for the treatment group, betweenGAB andEDDbelow the 12‐week thresholds than
aboveitduringtheinterventionperiodwhentheincentivesareinforce,butnodifferencesinthe
pre‐intervention period. In order to test this, we estimate the following difference in difference
regression:
13 13 (A1)
43
where isweeksofpregnantatthefirstvisitbasedonEDDforindividualigettingcareinclinic
j, isthenumberofweeksatthefirstvisitbasedonGABforindividualigettingcareinclinicj,
isaclinicfixedeffect, 13 isanindicatorofwhethertheclinicreportedthefirstvisitto
beinthefirst12weeksbasedonEDD, isanindicatorofwhethertheclinicwasactuallytreated,
and isanerrorterm.
Intheabsenceofmisreportingandnoprematuritythereshouldbenodifferencebetweenthe
twomeasuresand wouldhaveacoefficientof1.However,becauseprematurebirthsoccurbefore
EDD,weexpect tobeclosetobutlessthanone.Thenwecaninterprettheothercoefficientsasthe
effecton accountingforaverageweeksofprematurity.Sothedependentvariableis
theerrorinEDDinforecastingactualdeliverydate.Equation(A1)takesonadifferenceindifference
interpretationinthesensethewearedifferencingthechangeintheforecasterrorbetweenthepre‐
interventionandinterventionperiodsforthegroupofpregnantwomenforwhichaclinicreportsas
havingtheirfirstvisitbefore13weeksandthegroupofpregnantwomenforwhichaclinicreports
havingthefirstvisitinweek13orlater.Ifthereisnodifferenceintheerrorforthetreatmentgroup
inthepostperiodthen , theinteractionbetweentreatmentandreportedhavingthefirstperiod
beforeweek13,willbezero.Wefindnoevidenceofmisclassificationbytreatedclinics(SeeTable
A1).
44
FigureA1
ComparisonofWeeksPregnantat1stPrenatalVisitBasedonGestationalAgeatBirthandBasedonDateofLastMenstruation
Source:Authors’ownelaborationbasedondatafromtheprovincialmedicalrecordinformationsystem.
45
TableA1:TestforMisreportingWeeksPregnantat1stPrenatalVisit
DependentVariable:WeeksPregnantat1stPrenatalVisit,byGestationalAgeatBirth
WeeksPregnantbyEDD 0.90*** (0.02)
1(WeeksPregnantbyEDD<13) ‐0.13 (0.31)
1(WeeksPregnantbyEDD<13)x1(Treated=1) ‐0.03 (0.44)
Constant 1.33*** (0.39)
Observations 1730
AdjustedR2 0.82
Thedependentvariableisweekspregnantatthefirstprenatalvisitconstructedusinggestationalageatbirth.Theindependentvariableisweekspregnantatthefirstvisitconstructedbyusingthelastdayofmenstruationorestimateddeliverydate(EDD).Theinteractionterminteractsadichotomousindicatorforwhetherthevisitwasbeforeweek13andadichotomousindicatorforwhethertheclinicwasactuallytreated.Theregressioncontrolsforclinicfixedeffectsbyaddingabinaryindicatorforeachclinicinthesample.Standarderrorsareinparentheses.*p<0.10,**p<0.05,***p<0.01.
46
APPENDIXB:ROBUSTNESSTESTRESULTS
FigureB1:IndividualClinicTreatmentEffectsforWeeksPregnantat1stPrenatalVisit
Notes:Thisfigureplotsindividualclinictreatmenteffectsfortheoutcomeofweekspregnantatfirstprenatalvisit.WerunOLSregressionoftheoutcomecomparingeachclinicassignedtothetreatmentgrouptoallclinicsassignedtothecontrolgrouppoolingtheinterventionperiodandthepost‐interventionperiodI(henceMay2010‐March 2012). One treatment clinic is not included because of its insufficient sample size. This cliniccorrespondstooneofthetwothatdidnottakeuptreatment.Thetrianglesymbolreferstotheclinicthatwasassignedtotreatmentbutdidnottakeupthetreatment.Thex‐axisissortedfromthelowesttothehighestclinic‐specificimpact.Thedashedbluelineistheintent‐to‐treateffectcalculatedbypoolingtheinterventionand the first post intervention period. The vertical lines are 95% confidence intervals constructed usingstandarderrorsobtainedfromtheWildbootstrapprocedure.
47
FigureB2:IndividualClinicTreatmentEffectsfor1stPrenatalVisitbeforeWeek13ofPregnancy
Notes:Thisfigureplotsindividualclinictreatmenteffectsfortheoutcomeoffirstprenatalvisitbeforeweek13.WerunOLSregressionoftheoutcomecomparingeachclinicassignedtothetreatmentgrouptoallclinicsassignedtothecontrolgrouppoolingtheinterventionperiodandpostinterventionperiodI(henceMay2010‐March2012).Onetreatmentclinicisnotincludedbecauseofitsinsufficientsamplesize.Thiscliniccorrespondstooneofthetwothatdidnottakeuptreatment.Thetrianglesymbolreferstotheclinicthatwasassignedtotreatmentbutdidnottakeupthetreatment.Thex‐axisissortedfromthelowesttothehighestclinic‐specificimpact.Thedashedbluelineistheintent‐to‐treateffectcalculatedbypoolingtheinterventionandthefirstpost interventionperiod.Thevertical linesare95%confidenceintervalsconstructedusingstandarderrorsobtainedfromtheWildbootstrapprocedure.
48
TableB1:RobustnessTestsforWeeksPregnantat1stPrenatalVisit
(1) (2) (3)
InterventionPeriodPost‐Intervention
PeriodI(Jan2011–March2012)
Post‐InterventionPeriodII
(April–Dec2012)
A.ResultsfromTable4
Treatment ‐1.47** ‐1.63** ‐2.47**
(0.71) (0.75) (1.02)
LargeSamplep‐value 0.04 0.03 0.02
WildBootstrappedp‐value 0.08 0.03 0.03
ControlGroupMean 17.80 17.90 20.10
SampleSize 769 1,296 710
B.EstimatesUsingRestrictedSample
Treatment ‐1.47* ‐2.01*** ‐2.01* (0.77) (0.70) (1.11)
LargeSamplep‐value 0.06 0.00 0.07
WildBootstrappedp‐value 0.09 0.02 0.12
ControlGroupMean 17.96 18.32 17.01
SampleSize 760 1,326 425
C.Difference‐in‐DifferencesEstimates
Treatment ‐1.35** ‐1.74*** ‐2.35* (0.64) (0.63) (1.31)
LargeSamplep‐value 0.036 0.005 0.072
WildBootstrappedp‐value 0.060 0.014 0.144
ControlGroupMean 17.80 17.90 20.10
SampleSize 4,015 4,015 4,015
Notes:ThistablereportsLATEestimatesofthetreatmenteffectofthemodifiedfeescheduleonweekspregnantat1stprenatalvisit.Thep‐valuesarefor2‐sidedhypothesistestsofthenullthatthedifferenceisequaltozero.Wepresentboththep‐valuecomputedfor largesamplesandaWildbootstrappedp‐valuethat isrobust insamples with small numbers of clusters (Cameron et al. 2008). Our Wild bootstrap procedure assignssymmetricweightsandequalprobabilityafterre‐samplingresiduals(DavidsonandFlachaire2008)anduses999replications.Column(1)reportstheresultsforthesampleobservedinan8‐monthinterventionperiod(May2010–December2010).Column(2)reportstheresultsforthesampleobservedinthe15‐monthperiodfollowingtheendoftheintervention(January2011–March2012).Column(3)reportstheresultsforthe9‐monthperiodafterthechangeinthecodingofthefirstprenatalvisit(April2012–December2012).Standarderrorsareinparentheses.*p<0.10,**p<0.05,***p<0.01.
49
TableB2:RobustnessTestsfor1stPrenatalVisitbeforeWeek13
(1) (2) (3)
InterventionPeriodPost‐Intervention
PeriodI(Jan2011–March2012)
Post‐InterventionPeriodII
(April–Dec2012)
A.ResultsfromTable4
Treatment 0.11** 0.08** 0.08**
(0.04) (0.04) (0.04)
LargeSamplep‐value 0.01 0.02 0.04
WildBootstrappedp‐value 0.03 0.05 0.06
ControlGroupMean 0.31 0.34 0.27
SampleSize 769 1,296 710
B.EstimatesUsingRestrictedSample
Treatment 0.09** 0.10** 0.10* (0.04) (0.04) (0.06)
LargeSamplep‐value 0.03 0.01 0.08
WildBootstrappedp‐value 0.08 0.02 0.11
ControlGroupMean 0.31 0.33 0.36
SampleSize 760 1,326 425
C.Difference‐in‐DifferencesEstimates
Treatment 0.09* 0.07 0.07 (0.05) (0.05) (0.06)
LargeSamplep‐value 0.08 0.11 0.23
WildBootstrappedp‐value 0.13 0.17 0.24
ControlGroupMean 0.31 0.34 0.27
SampleSize 4,015 4,015 4,015
Notes:ThistablereportsLATEestimatesofthetreatmenteffectofthemodifiedfeescheduleanindicatorofwhetherthe1stprenatalvisitoccurredbeforeweek13ofpregnancy.Thep‐valuesarefor2‐sidedhypothesistestsofthenullthatthedifferenceisequaltozero.Wepresentboththep‐valuecomputedforlargesamplesandaWildbootstrappedp‐valuethatisrobustinsampleswithsmallnumbersofclusters(Cameronetal.2008).OurWildbootstrapprocedureassignssymmetricweightsandequalprobabilityafterre‐samplingresiduals(Davidson and Flachaire 2008) and uses 999 replications. Column (1) reports the results for the sampleobservedinan8‐monthinterventionperiod(May2010–December2010).Column(2)reportstheresultsforthe sampleobserved in the15‐monthperiod following theendof the intervention (January2011–March2012).Column(3)reportstheresultsforthe9‐monthperiodafterthechangeincodingofthefirstprenatalvisit(April2012–December2012).Standarderrorsinparentheses.*p<0.10,**p<0.05,***p<0.01.
50
APPENDIXC:ITTRESULTS
TableC1:ITTEstimatesoftheEffectofTemporaryIncentivesonTimingof1stPrenatalVisit
(1) (2) (3)
Intervention
Period
Post‐InterventionPeriodI
(Jan2011–March2012)
Post‐InterventionPeriodII
(April–Dec2012)
A.WeeksPregnantat1stPrenatalVisit
Treatment ‐1.39** ‐1.59** ‐2.47** (0.67) (0.73) (1.02)
LargeSamplep‐value 0.04 0.03 0.02
WildBootstrappedp‐value 0.09 0.03 0.03
ControlGroupMean 17.80 17.90 20.10
SampleSize 769 1,296 710
B.FirstPrenatalVisitBeforeWeek13ofPregnancy
Treatment 0.10*** 0.08** 0.08** (0.04) (0.04) (0.04)
LargeSamplep‐value 0.01 0.02 0.04
WildBootstrappedp‐value 0.03 0.05 0.08
ControlGroupMean 0.31 0.34 0.27
SampleSize 769 1,296 710
Notes:ThistablereportsITTestimatesofthetreatmenteffectofthemodifiedfeescheduleonindicatorsofthetimingofthe1stprenatalvisit.TheLATEestimatesarereportedinTable4.ThedifferencesareestimatedfromOLS regressionsof thedependentvariableonan indicator for clinic treatment randomassignment.Thep‐valuesarefor2‐sidedhypothesistestsofthenullthatthedifferenceisequaltozero.Wepresentboththep‐value computed for large samples and a Wild bootstrapped p‐value that is robust in samples with smallnumbersofclusters(Cameronetal.2008).OurWildbootstrapprocedureassignssymmetricweightsandequalprobabilityafterre‐samplingresiduals(DavidsonandFlachaire2008)anduses999replications.Column(1)reportstheresultsforthesampleobservedinan8‐monthinterventionperiod(May2010–December2010).Column (2) reports the results for the sample observed in the 15‐month period following the end of theintervention(January2011–March2012).Column(3)reportstheresultsforthe9‐monthperiodafterthechange in the coding of the first prenatal visit (April 2012 – December 2012). Standard errors are inparentheses.*p<0.10,**p<0.05,***p<0.01.
51
TableC2:ITTofCross‐PriceEffects(Spillover)
(1) (2)
InterventionPeriodPost‐InterventionPeriod
(Jan–Dec2011)
A.TetanusVaccine
Treatment 0.02 ‐0.02 (0.07) (0.05)
LargeSamplep‐value 0.76 0.62
WildBootstrappedp‐value 0.80 0.59
ControlGroupMean 0.79 0.84
SampleSize 769 1,053
A.Numberofvisits
Treatment 0.37 0.50 (0.32) (0.57)
LargeSamplep‐value 0.24 0.38
WildBootstrappedp‐value 0.27 0.40
ControlGroupMean 4.05 4.40
SampleSize 769 1,053
Notes:ThistablereportsITTestimatesofthetreatmenteffectofthemodifiedfeescheduleonindicatorsofotherservices.TheLATEestimatesarereportedinTable5.ThedifferencesareestimatedfromOLSregressionsofthedependentvariableonanindicatorforclinictreatmentrandomassignment.Thep‐valuesarefor2‐sidedhypothesistestsofthenullthatthedifferenceisequaltozero.Wepresentboththep‐valuecomputedforlargesamplesandaWildbootstrappedp‐valuethatisrobustinsampleswithsmallnumbersofclusters(Cameronetal.2008).OurWildbootstrapprocedureassignssymmetricweightsandequalprobabilityafterre‐samplingresiduals (DavidsonandFlachaire2008) anduses999 replications. Column (1) reports the results for thesampleobserved inan8‐month interventionperiod(May2010–December2010).Column(3)reports theresultsforthesampleobservedinthe12‐monthperiodfollowingtheendoftheintervention(January2011–December2011).Standarderrorsareinparentheses.*p<0.10,**p<0.05,***p<0.01.
52
TableC3:ITTEffectsofIncentivesonBirthOutcomes
(1) (2)
InterventionPeriod Post‐InterventionPeriod(Jan–Dec2011)
A.BirthWeight
Treatment ‐34.88 24.48 (45.38) (39.63)
LargeSamplep‐value 0.44 0.54
WildBootstrappedp‐value 0.46 0.57
ControlGroupMean 3304.82 3279.13
SampleSize 555 802
B.LowBirthWeight
Treatment 0.01 ‐0.01 (0.02) (0.01)
LargeSamplep‐value 0.63 0.60
WildBootstrappedp‐value 0.61 0.63
ControlGroupMean 0.05 0.06
SampleSize 555 802
B.Premature
Treatment 0.03 ‐0.04* (0.03) (0.02)
LargeSamplep‐value 0.31 0.08
WildBootstrappedp‐value 0.32 0.09
ControlGroupMean 0.09 0.12
SampleSize 414 708
Notes:ThistablereportsITTestimatesofthetreatmenteffectofthemodifiedfeescheduleforonindicatorsofbirthoutcomes.TheLATEestimatesarereportedinTable6.Theobservationsincludewomanforwhomweareabletoobtaininformationonbirthoutcomesprovidedinpublichospitalbirthrecords.Thedifferencesareestimated from OLS regressions of the dependent variable on an indicator for clinic treatment randomassignment.Thep‐valuesarefor2‐sidedhypothesistestsofthenullthatthedifferenceisequaltozero.Wepresentboththep‐valuecomputedforlargesamplesandaWildbootstrappedp‐valuethatisrobustinsampleswith small numbers of clusters (Cameron et al. 2008). Our Wild bootstrap procedure assigns symmetricweights and equal probability after re‐sampling residuals (Davidson and Flachaire 2008) and uses 999replications.Column(1)reportstheresultsforthesampleobservedinan8‐monthinterventionperiod(May2010 – December 2010). Column (2) reports the results for the sample observed in the 12‐month periodfollowingtheendoftheintervention(January2011–December2011).Standarderrorsareinparentheses.*p<0.10,**p<0.05,***p<0.01.
53
APPENDIXD:ONLINESURVEYOFCLINICS
IncollaborationwiththeProvincialManagementUnitoftheprogram(UGPS),inMay2015
weconductedashortonlinesurvey(usingSurveyMonkey®)inthoseclinicsthatparticipatedinthe
pilot.Thesurveyaimstomeasuretheabsoluteandrelativeimportanceofsevendifferentprenatal
careproceduresincludinginitiatingprenatalcarepriortoweek13ofpregnancy.Theabsolutescores
rangefrom1to5,with5beingthehighestscoreintermsofimportance,andanadditionaloptionof
zero indicating that theprocedure isnotappropriate forapregnantwoman.Hence, theabsolute
scorerangesfrom0to5points.Therelativerankingaimedtosortthesevenpracticesfrom1to7,
with1beingthehighestranking.Inpracticehowever,thesurveyinstrumentallowedtherespondent
torepeatnumbers.
Thesurveywassentouttobyemailtoclinicsdirectors(orthenextpersoninrank).Wewere
unable to obtain current email addresses for8 out of the 36 clinics.Another4 clinics confirmed
havingreceivedtheemailbutrefusedtoanswerit.Outofthe24clinicsthatdidrespondtothesurvey,
21 fully completed it while 3 only partially completed it. Out of the 21 clinics with complete
responses,13belongtothetreatmentgroupand8tothecontrolgroup.AppendixTableD1shows
thattherearenosignificantdifferencesinbaselinecharacteristicsbetweenclinicsthatrespondedto
thesurveyandclinicsthatdidnotrespond.Inaddition,weaccountforsurveynon‐responseusing
InverseProbabilityWeightingbasedon the logistic regressionreported inTableD2(Wooldridge
2007).WereportresultsforbothIPWandnon‐IPWregressions.
Figures8and9donotsuggestanydifferenceintheabsolutescoreandrelativerankingofthe
procedures between treatment and control clinics. To test for the significance of the differences
betweenthetwogroups,werunanOLSregressionoftheabsolutescoreandtherelativeranking
againstabinaryindicatorfortreatment.Toaccountforthesmallsamplesizewealsocomputethep‐
value for the differences in means permuting our data and using a random sample of 10,000
permutations.TheresultsareshowninTablesD3andD4.
54
OnlineSurveyQuestionnaire
Weaskforyourcollaborationincompletingabriefsurveyaboutprenatalcareservicesprovidedatyourhealthfacility.
Important:Whenansweringthesurvey,pleasethinkofahypotheticalcaseofawomanwiththefollowingcharacteristics:
25yearsold Livinginthesameneighborhoodwhereyourhealthfacilityislocated Withoutanyapparentsignofdisease 6weekspregnant Hadapreviouslow‐riskpregnancy
1. Pleaseassignascorebetween1to5toeachofthefollowingservicesthatcouldbe
deliveredtothepregnantwomanpresentedinthehypotheticalcase.
1correspondstoaservicetowhichyouassignthelowestimportance5correspondstoaservicetowhichyouassignthehighestimportance
1 2 3 4 5
Notappropriateforapregnantwoman
Prenatalultrasound
ThoraxX‐Ray
Firstprenatalvisitbeforeweek13ofpregnancy
Bio‐psycho‐socialpregnancycounselingvisit
CombinedDiphtheria/Tetanusvaccine
Bloodtestwithserology
Bloodtestwithoutserology
55
Pleaserankinorderofpriority(from1to7)thefollowing7healthservicesthatcouldbedeliveredtothepregnantwomanofthehypotheticalcase.
1correspondstotheserviceyouwouldprioritizethemost7correspondstotheserviceyouwouldprioritizetheleast
Prenatalultrasound
ThoraxX‐Ray
Firstprenatalvisitbeforeweek13ofpregnancy
Bio‐psycho‐socialpregnancycounselingvisit
CombinedDiphtheria/Tetanusvaccine
Bloodtestwithserology
Bloodtestwithoutserology
56
TableD1:BaselineCharacteristicsofClinics,byOnlineSurveyResponseStatus
Non‐
respondent Respondent P‐value Obs.
NumberofPregnantWomenAttendedperYear 48.60 64.90 0.33 36
WeeksPregnantat1stPrenatalVisit 17.44 16.77 0.15 36
1stVisitbeforeWeek13ofPregnancy 0.34 0.38 0.27 36
%ofPregnantWomenwhoarePlanNacerBeneficiaries 0.61 0.64 0.59 36
TetanusVaccineDuringPrenatalVisit 0.74 0.81 0.22 36
NumberofPrenatalVisits 4.26 4.42 0.72 36
BirthWeight(Grams) 3,283 3,320 0.33 36
GestationalAge(Weeks) 38.65 38.47 0.57 31
LowBirthWeight(<2500Grams) 0.06 0.07 0.73 31
Premature(GestationalAge<37Weeks) 0.10 0.13 0.60 31
Notes:This table reports themeansofbaselinecharacteristics for clinics that responded to theMay2015online survey and for clinics thatdidnot respond.The characteristics are taken from themedical recordsinformationsystem(2009).Thep‐valuesforthetestsofdifferencesinmeansarecomputedusingpermutationteststhatarerobustforsmallsamplesizes.
57
TableD2:ProbabilityofRespondingtotheOnlineSurvey,LogitCoefficientsandMarginalEffects
Coeff. Marg.Eff.
TreatmentGroup 1.498 0.274 (1.111) (0.180)
BirthWeight(grams) 0.100 0.018 (1.076) (0.196)
WeeksPregnantat1stPrenatalVisit ‐0.594 ‐0.109 (0.648) (0.121)
1stVisitbeforeWeek13ofPregnancy ‐3.590 ‐0.657 (9.026) (1.670)
%ofPregnantWomenwhoarePlanNacerBeneficiaries 1.620 0.296 (4.359) (0.774)
TetanusVaccineDuringPrenatalVisit 3.350 0.613 (3.817) (0.646)
NumberofPrenatalVisits ‐0.099 ‐0.018 (0.559) (0.101)
Constant 7.644 (18.248)
Observations 36 36
Notes: This table reports the coefficients and marginal effects from a logit regression that estimates theprobabilitythataclinicrespondedtotheMay2015onlinesurvey.
58
TableD3:DifferencesinAbsoluteScoreandRelativeRankingofEarlyPrenatalCare
AbsoluteScore RelativeRanking
(1)OLS
(2)OLS‐IPW
(3)OLS
(4)OLS‐IPW
Difference(Treatment–Control) 0.20 0.13 0.10 0.14 (0.22) (0.92) (0.21) (0.89)
LargeSamplep‐value 0.38 0.89 0.65 0.88
Permutationp‐value 0.35 1.00 0.46 0.99
Observations 20 20 20 20
Controlgroupmean 4.57 1.88 4.66 1.88
Notes:Column(1)showsthedifferencesbetweentreatmentandcontrolclinicsintheabsolutescoreassignedtothepracticeofearlyprenatalcarewithoutanyadjustmentofsampleloss.Column(2)adjustsforsamplelossbyInverseProbabilityWeighting.Column(3)showsthedifferencesbetweentreatmentandcontrolclinicsintherelativerankingassignedtoearlyprenatalcareamongsevendifferentpractices.Column(4)isthesameasColumn(3)butadjustsforsamplelossbyInverseProbabilityWeighting.(Wooldridge2007)ThecoefficientsareobtainedfromanOLSregressionofeachoutcomeagainstatreatmentbinaryindicator.ThethirdrowshowstheP‐valueobtainedfrompermutingthedatausingarandomsampleof10,000permutations.Standarderrorsareinparentheses.Weloseoneobservationineachcasebecauseofmissingdataineachspecificquestion.