Does independent needs assessment limit use of publicly financed long-term care?

(1)

ContentslistsavailableatScienceDirect

Health

Policy

j o u r n al ho me p ag e :w w w . e l s e v i e r . c o m / l o c a t e / h e a l t h p o l

Does

independent

needs

assessment

limit

use

of

publicly

ﬁnanced

long-term

care?

Pieter

Bakx

a,∗

_,

_Rudy

_Douven

a,b

_,

_Frederik

_T.

_Schut

a

a_Erasmus_School_of_Health_Policy_and_Management,_Erasmus_University_Rotterdam,_the_Netherlands

b_CPB_Netherlands_Bureau_for_Economic_Policy_Analysis,_the_Netherlands

a

r

t

i

c

l

e

i

n

f

o

Articlehistory:

Received16April2020

Receivedinrevisedform1September2020

Accepted3September2020

Keywords:

Long-termcare

Long-termcareinsurance

Independentneedsassessment

Moralhazard

a

b

s

t

r

a

c

t

Inhealthcaretheassessmentofpatients’needsistypicallyentrustedtohealthcareproviders.Bycontrast,

inpubliclyﬁnancedlong-termcare(LTC)needsassessmentisoftendelegatedtoanindependentassessor.

OnerationaleofferedforindependentneedsassessmentinLTCistolimitthescopeformoralhazardand

supplier-induceddemand,whichmaybeparticularlystrongincaseofpublicLTCinsurance.Westudy

whetherindependentneedsassessmentrestrictsuseofpubliclyﬁnancedLTCattheintensivemargin

(i.e.afterpeoplearebeingassessedtobeeligibleforreceivingcare).Therefore,welinknationwideDutch

administrativedatasetsaboutindividualLTCuseandeligibilitydecisionsbytheindependentassessment

agencyin2012.Weﬁndforvirtuallyalltypesofcare,allpopulationsubgroups,andallregionsthatLTC

usebypatientswassubstantiallylessthanthemaximumamountofcareallowedbytheindependent

assessor.ThissuggeststhatintheNetherlandsindependentneedsassessmentinLTCdoesnotimposea

bindingconstraintonuseonceapersonisconsideredeligibleforcare.Still,independentneedsassessment

mayhavereducedLTCuseattheextensivemargin.Asigniﬁcantproportionoftheapplicationsforcare

(16%)wasrejected.Inaddition,theindependentassessmentmaydetersomepeoplefromapplying.

(http://creativecommons.org/licenses/by/4.0/).

1. Introduction

Traditionally,theneedsofpatientsareassessedbytheirhealth careproviders.Buttheseprovidersmaybeself-interestedandhave superiorinformationaboutpatients’needs,resultingin principal-agentproblemsforthepatientandthethird-partypayer[1].In thepresenceofcomprehensivehealthinsurance,thismayresultin moralhazardandsupplier-induceddemand.

Independentneedsassessmentmaybeespeciallypopularinthe presenceofpubliclong-termcare(LTC)ﬁnancingbecausemoral hazardandsupplier-induceddemandareparticularlylikelyinthis context.LTCenablestheelderlyandthedisabledtocopewiththeir limitations.Receivingmorecareandsupportthanstrictlyneeded islikelytogeneratepositivemarginalbeneﬁtsforpatientsbecause it is offeringadditional comfort.Moreover,if LTCprovidersare paidfee-for-serviceandareallowedtoperformneedsassessment themselves,theymaybeinclinedtoinducemoredemandfortheir servicesthanstrictlynecessary.Hence,ifduetothepresenceof comprehensiveLTCinsurancethemarginalcostsforpatientsare

∗ Correspondingauthorat:POBox1738,3000DRRotterdam,theNetherlands.

E-mailaddress:bakx@eshpm.eur.nl(P.Bakx).

low,theriskofmoralhazardandsupplier-induceddemandmay beparticularlyhighforLTCservices.

Theseproblemsmaybereducedbydelegatingtheassessment ofpatients’needstoanindependentassessor.Independentneeds assessmentisuncommoninhealthcare,butfrequentlyusedinLTC, particularlyincountrieswithacomprehensivepublicLTCﬁnancing scheme,suchastheNetherlands,Belgium,Germany,Japan,Norway andSwitzerland[2–5].

We willexamine whetherindependent needsassessment is indeedlikelytoconstrainLTCdemandinthecontextoftheDutch publicLTCinsurancescheme.Independentneedsassessmenthas tworoles. First,it determineswhethera person iseligible (the extensivemargin).Second,conditionalonbeingeligiblethe asses-sorrestrictstheuseofcarebyspecifyingamaximumamountof care(theintensivemargin).Inthispaperwestudytheintensive margin,i.e.doesthemaximumamountofcaresetbytheassessor limittheamountofcareusedbypeople?Weareabletostudythis becausetheassessmentonlylimitsuseiftheactualcareusedby peopleisclosetothemaximumamountofcaresetbytheassessor. Ifthismaximumisnotbinding,implyingthatmostpeopleuseless carethantheyareentitledto,thenitisunlikelythatthisrestriction limitsdemand.Bycontrast,ifmosteligiblepeopleusethe max-imumadmittedamountofcare,thisimpliesthattherestriction isbindinganditmayindeedhaveservedasademandconstraint. https://doi.org/10.1016/j.healthpol.2020.09.003

(2)

TheDutchinstitutionalsettingallowsustostudythisphenomenon withprecisionbecauseboththeeligibilitydecisionbytheassessor andtheactualuseofLTCareadministeredforeachpatientandcan belinkedattheindividuallevel.

Weusethesenationwidedatatoexaminewhethertheamount of homecareused bya patientequals themaximumadmitted amountbytheindependentagency.Wefindforvirtuallyalltypes ofcare,allpopulationsubgroups,andallregionsthatLTCuseby patientswassubstantiallylessthanthemaximumamountofcare admitted bytheindependentassessor.Thissuggeststhatinthe Dutchcontextattheintensivemarginindependentneeds assess-mentinLTChardlylimitsLTCuse.Duetoalackofdataweleave forfutureresearchwhetherindependentneedsassessmentmay havereducedLTCuseattheextensivemargin.In2009asignificant proportionoftheapplicationsforcare(16%)wererejectedbythe assessmentagency[6].Rejectionsoccurbecausetheapplicantdoes notfulfilltheeligibilityrequirementsorquitstheassessment pro-cedure,orbecausetheassessorreferstheapplicanttocarefinanced throughotherfinancingschemes.

Ourpapercontributestotheliteratureintwoways.First,toour knowledgeitisthefirststudythatanalyzestheroleofan inde-pendentneedsassessorinrationingaccesstocareindetailusing quantitativedata.Second,usingpopulationdataattheindividual level,ourstudyisthefirsttoanalyzethenon-take-up,i.e.the max-imumamountofcareadmittedbytheassessorminustheactual useofcarebythepatient.Ourpaperthereforecontributestothe understandingofnon-take-upofhealthinsurancebenefitsandits determinants(seee.g.[7,8]).Ourpaperisorganizedasfollows.First, wediscusstherationaleforindependentneedsassessment.Next, webrieflydescribetheDutchcontextofpublicLTCinsurance.Then weexplainhowweanalyzethepotentialimpactoftheassessment agencyonconstrainingLTCuse, followedbyapresentationand discussionoftheempiricalresults.

2. Rationaleforindependentneedsassessment

The choiceforindependentneedsassessment inhealth care dependsonwhetheritspotentialbenefitsoutweighitspotential costs.Themainbenefitisthatitreducespotentialbiasinthe assess-mentresultingfromproviderinterestsorpatientdemand.When providersandrecipientshaveaninterestinprovidingorobtaining more ormoreexpensivecarethanstrictlyneeded,independent needsassessmentmayreduceoverprovisionandinefficientlyhigh expenditures.Thiseffectmaybeamplifiedifinsurancereducesthe marginalcostofconsumingLTCfortherecipient.Thesepotential benefits,however,havetobeweighedagainstthemonetarycosts of assessmentsandauditingaswellasthetime costsinvolved. Inaddition,theindependentassessormayalsobebiaseddueto pressurebystakeholders,financialrestrictionsorregulations,and maybelessabletomakeanappropriateneedsassessmentbecause theassessormaybelessinformedaboutthespecificneedsofthe patientthanaproviderduetoalesspersonalandfrequentcontact withthepatient.

The choiceforanindependentassessment oranassessment donebyaproviderdependsonthecharacteristicsofthehealth carethattheassessmentisneededforandonotherfactors.Atleast ﬁvefactorscanbedistinguishedthatarerelevantfordetermining thefeasibilityanddesirabilityofindependentneedsassessment. First, sinceindependentassessmenttakestime, itisonly possi-blewhencareisnoturgent(i.e.forelectivecare).Second,ifthe needforcarecannotbedeﬁnedpreciselythereismoreroomfor moralhazardandinducingdemand,inwhichcasethebiasinthe providers’decisionsmaybelarger.Anindependentassessormay createmoretransparencyandequityintheeligibilitycriteriafor obtainingcare.Third,anindependentassessorismoreattractive

iftherearebenefitsfromoveruseforprovidersandpatients,e.g. underfee-for-servicecontracts,positivemarginalbenefitsforthe patientfromoveruseandthegovernmentoraninsurerpaysalarge shareofthecosts.Fourth,anindependentassessorislessbeneficial ifthereareotherrestrictionsonsupply(e.g.providerbudgets)and demand(e.g.co-payments)astheserestrictionsalsoworktolimit supplier-induceddemandandmoralhazard.Fifth,thevalueofan independentassessmentishigherwhentheassessorhasthe abil-ityandtheincentivestoactintheinterestofitsprincipal.External pressuree.g.becauseapplicantsmaychallengeeligibilitydecisions incourt,maycausearationalindependentassessortobemore lenientthanitsprincipaldesires[9].

Thesefivefactorsmeanthatindependentneedsassessmentin thecontextofLTCisfeasible,becauseLTCisoftenforpersisting ratherthanurgentproblems,andthatisdesirablebecausethereis likelymuchroomformoralhazardandsupplierinduceddemand. Thereare tworeasonsforthis.First, thedemandfor LTCis not onlyafunctionofsomeone’sfunctionallimitations,butperceptions andpersonalcircumstances,e.g.theavailabilityofinformal sup-port,mattertoo.Hence,theneedforcareoftencannotbedefined precisely.Second,inthecaseofLTCinsurance,servicesforwhich marginalcostsarehigherthanthebenefitsforsocietyorthe third-partypayermaybewelfare-increasingforindividualprovidersand patients.

3. Long-termcareintheNetherlands

PublicLTCinsurancepaysfor94%(2012figure)ofallLTC expen-dituresintheNetherlands;theremainderisfinancedthroughthe SocialSupportAct[10].PeoplewhoareeligibleforpublicLTC insur-ancebenefitscanchoosebetweenreceivingtheseservicesinkind ortotakeoutacashbenefitthatisequaltoroughly75%ofthecosts ifthecarewereprovidedinkind[11].

Until1998,LTCproviderswereresponsiblefortheassessments ofpeople’s needsfor carecovered by thepublicLTCinsurance scheme.ToreducetheinﬂuenceofprovidersonLTCuse,in1998 thistaskwasentrustedtoregionalindependentassessment agen-cies.In2005,allregionalassessmentagenciesweremergedintoa centralagencyforneedsassessment(CIZ)toreducetheprevailing regionalvariationinneedsassessment.SincethenCIZhascarried outorauditedtheneedsassessmentforpublicLTCinsurance ben-eﬁts.ItdoessoaccordingtorulessetbytheMinistryofHealth [12,13].

Eligibility for LTC covered through public LTC insurance is usually requested by the applicant or someone who does the applicationontheirbehalf.Eligibilitydependsonthehealthand health-relatedlimitationsoftheapplicant.Until2015,manyother aspectssuchaslivingconditions,socialenvironment,psychicand socialfunctioningoftheapplicant,thepresenceofother profes-sionalservicesand informalcarethatthepatientreceiveswere alsotakenintoaccount[14–16].Togetinsightintheseaspectsof theapplication,theassessorusestheinformationprovidedonthe applicationformandmaygatherinformationonthecriterialisted about,e.g.viahealthcareprovidersorthroughahomevisit.Insome cases(e.g.afterahospitalization)inwhichtheneedforhomecare isstraightforwardtoestablish,theassessmentissometimes dele-gatedtoahomecareprovider.Inthesecases,CIZhastheroleofan auditor[13].

Homecareprovidersareprivateentities,whichareeither for-profitornot-for-profit[11].Homecareprovidersarepaidforevery hourofcareprovided.Thesefee-for-servicecontractsmeanthat theyhave incentives for overprovision. The providersare con-tractedbyregionalsinglepayers,eachofwhichisconstrainedby anannualbudgetthatissetatthenationallevel.Regionalpayersdo notbearanyfinancialriskforLTCexpensesanddonotcompetefor

(3)

consumers,andthereforehavelittleincentivetomonitorthe efﬁ-ciencyofcare.Theregionalproviderbudgets,however,maylimit LTCusebecausetheyconstrainthetotalamountspentonLTC. Non-publicdatafromtheDutchHealthCareAuthorityshowthatfrom 2011to2013all32singlepayersspentatleast98.3%oftheregional budgetperyear,whichsuggestthattheregionalbudgetsmayact asademandconstraint.Nevertheless,onlyafewindividualswere onwaitinglistsduringtheperiod(2012)westudied[17].

In additiontoregionalbudgets,income-relatedco-payments mayalsorestrictpubliclyﬁnancedLTCdemand,althoughDutch co-paymentsarerelativelylowincomparisontoothercountries [2,18].

The key question addressed in this paper is whether inde-pendentneedsassessment effectivelyrestrictspubliclyﬁnanced LTC demand at the intensive margin in the presence of other potentialdemandconstraints(i.e.regionalprovidersbudgetsand co-payments).

4. Empiricalanalysis

4.1. Methodsanddata

AttheintensivemargintheassessorcanreduceLTCdemand foreachofthetypesofLTCthatisused.Inhomecaretheassessor speciﬁesarangeofhourstoeachclient.Thisrangecontainsalower andanupperboundofhourscareperweek.Theupperboundisthe maximumadmittedamountofpubliclyﬁnancedcaretheclientis allowedtoreceive.Ifclientsuselesshoursofcare,thenthe maxi-mumadmittedamountthenweconcludethattheassessmentdoes notlimitLTCuse.

Toinvestigatethis,weuseanationwideadministrativedataset that encompasses about 600,000 individuals being eligible for receiving homecarein theNetherlands for 13 four-week peri-odsin2012.Welinkthreesetsinformationattheindividuallevel throughauniquepersonIDcreatedbyStatisticsNetherlands.The ﬁrstdatasetcontainstheeligibilitydecisionsmadebythe indepen-dentassessor(CIZ). Thesedecisionsspecifyforwhich typesand amounts(inhours)ofhomecaretheindividualiseligiblefor.We studyfourtypesofhomecare:personalcare,nursing,group assis-tance andindividual assistance.Theseconddatasetrecordsthe

Table1

Eligibilityforhomecareanduseofcareinkindin2012,uniqueindividuals.

Number %careinkinda _%_cash_beneﬁt _%_no_care

Typeofhomecare

Personalcare 404,508 79.3 14.3 6.4

Nursing 187,823 81.4 10.0 8.6

Individualassistance 194,533 58.7 32.2 9.1

Groupassistance 106,934 62.6 24.9 12.5

a_CBS₍₂₀₁₈₎_[19]_.

actualuseofcare.ThisdatacomesfromtheCentralAdministration Ofﬁce(CAK)ofthepublicLTCinsurancescheme.Itcontainsthe numberofhoursofhomecarethatwasprovidedin-kindasitwas billedbytheproviders.Third,welinkbackgroundcharacteristics fromotheradministrativesources.Thesebackground character-isticsaretheage,gender,theregionofresidence,thehousehold composition,householdincomeandpriorhealthcareexpenditures ofallusers.

Abouttwo-thirdsoftheindividualswhoareeligibleforhome careareelderlywhoeitherhaveasomaticorapsychogeriatric con-ditionthatcausesfunctionallimitationsrequiringhomecare.The otherindividualsareyoungdisabledwhohaveapsychiatric disor-der,orasensory,physicalormentalhandicap.Ofallindividuals,55 percentliveswithaspouseorachildwhile45percentlivesalone; 58percentisfemale.Finally,theeligiblepopulationisratherpoor onaverage:morethanhalfisintheﬁrstthreeincomedeciles[19]. Weanalyzeif therestrictionsthat areimposed by indepen-dentassessorslimittheirusebycalculatingthetake-upratio:the amountofcarethatapersonuseddividedbytheamountthatthis personwaseligiblefor.Ifthistake-upratioislowerthan1,the restrictionsthattheindependentassessorimposedwerenot bind-ingforthisperson.Wecalculatethisratioforeachtypeofhome care.Subsequently,weuseordinaryleastsquaresregressionsto regressthetake-upratioonbackgroundcharacteristicstoﬁndout iftheratiovariessystematicallyacrosssubgroupsinthepopulation. 4.2. Results

LTCuseisinmostcaseslowerthanthemaximumamountof careasadmittedbytheassessor.AsshowninTable1,asubstantial

Table2

Useofpersonalcareconditionaloneligibilityin2012.

n(%oftotal)b _%_who_use_care_in_kindc _Median_number_of_hours_used_(%)d _Take-up_share_between _Take-up_share_>_100%f

90%–100%e

Eligibility(hoursperweek)a

0−2 622,261(18.3%) 61.0% 1.0 18.2% 3.3% 2−4 957,594(28.1%) 64.3% 2.3 17.6% 4.0% 4−7 934,245(27.4%) 71.2% 4.0 10.9% 3.0% 7−10 423,490(12.4%) 74.7% 6.3 14.0% 3.2% 10−13 179,077(5.3%) 71.6% 8.8 16.1% 4.1% 13−16 106,154(3.1%) 64.8% 11.3 18.3% 4.7% 16−20 57,818(1.7%) 58.9% 14.6 19.2% 3.8% 20−25 112,454(3.3%) 70.0% 16.6 20.2% 4.7% >25 14,486(0.4%) 28.2% 23.2 – –

Note:resultsforothertypesofhomecareareavailableuponrequest.

a_Categories_of_hours_the_client_is_eligible_for._If_the_individual_changed_from_one_category_to_another_during_a_four-week_period,_the_individual_is_assigned_to_the_highest

category.

b_n_refers_to_number_of_observations_in_a_category_where_every_observation_is_for_a_four-week_period,_hence_there_are₁₃_periods_in_2012._The_percentage_between_brackets

representstheshare:thenumberofobservationspercategorydividedbyallobservationsforthistypeofcare.

c_The_percentages_in_this_table_are_calculated_using_monthly_{observations.}_Hence,_they_are_all_lower_than_the_average_reported_in_Table₁_(79.3_%),_which_is_calculated_over

theentireyear.

d_Individuals_who_used_at_least_some_care_provided_in_kind_only.

e_Refers_to_the_number_of_observations_where_clients_have_take-up_shares_between₉₀_%–100_%_of_the_maximum_admitted_amount_of_eligible_hours._For_example,_if_the

maximumadmittedamountofhoursis16hperfourweek-period(category2−4)wecalculatedthetake-upsharebetween90–100%asthenumberofobservationswhere

clientsuse15and16h,and60%oftheobservationswhereclientsuse14h,anddivideitbythetotalnumberofobservationsforthatcategory.Forthelastcategorythereis

noﬁxedmaximumadmittedamountofhours.

f_Number_of_observations_where_clients_use_more_than₁₀₀_%_of_the_maximum_admitted_amount_of_eligible_hours_divided_by_the_total_number_of_{observations.}_For_the_last

(4)

Fig.1.Useofpersonalcarebyindividualseligiblefor0to8h(upperleft),8-16h(upperright),16-28h(lowerleft)and28-40h(right)perfour-weekperiod.

Note:Thegraphsdepictonlyindividualswhousedsomecarethatwasprovidedinkindandwhousedlessthan125%ofthemaximumnumberofhours.Themedianis

calculatedusingallobservationsofindividualswhousedsomecarethatwasprovidedinkind.Theresultsforindividualseligibleformorethan40hofpersonalcareper

four-weekperiodandforothertypesofhomecareareavailableuponrequest.

shareoftheeligibleindividualsdoesnotuseLTCatall,varyingfrom 6.4percentforpersonalcareto12.5percentforgroupassistance. Furthermore,asmallgroupoptsforacashbeneﬁtwhilethelargest groupusescarethatisprovidedinkindatsomepointduringthe year(seeTable1).

Oftheindividualsusingcareinkindatsomepoint,onlyveryfew individualstakeupanamountofcarethatisclosetothemaximum amounttheyareeligiblefor.Thisisillustratedforpersonalcare inTable2andFig.1.Forexample,asshowninFig.1(lowerright panel),themajorityofpeopleentitledtoreceiving28−40hours ofpersonalcaretookuplesscarethanthelowerboundof28h (themedianindividualusingabout25h),andthesameobservation holdsforpatientsentitledtolesspersonalcare(otherpanels)and othertypesofhomecare.

Thereismuch variationinhow muchofthehomecarethat someoneiseligibleforisused(Table2).Whatshareisusedvaries acrosstypesofcareandbytheamountthatthepersoniseligible for.Forinstance,medianuseis1hforindividualswhoare eligi-blefor0−2hofpersonalcareperweek.Thatis,forthisgroup,use isequalto100%oftheaverageof0and2(theupperandlower boundofthecategory).Yet,medianusebyindividuals whoare eligiblefor10−13hisonly8.75hofcare(76%of11.5). Further-more,theproportionofthesampleforwhomeligibilityisbinding, i.e.whousebetween90–100%ofthetotalamountofhoursthey areeligiblefor,rangesbetweenonly10−20%ofthepopulation.A smallproportion(3–5%)ofthehomecareusersreceivesmorecare thanthemaximumadmittedamount,whichsuggeststhatin prac-ticethemaximumisnotalwaysstrictlyapplied.Figuresforthe othertypesofhomecare(i.e.nursing,individualassistanceand groupassistance)showhighlysimilarpatternsastheones

pre-sentedhereforpersonalcareandareavailablethroughtheonline supplement.

Theregressionanalysesrevealthatthetake-uprationotonly differsbythetypeofhomecareandthenumberofthehoursoneis eligiblefor,butalsoacrosssubgroupsofusers,andthattheoverall meanshidesubstantialheterogeneity(Table3).Forexample,the take-upratioincreaseswithage,islowerformen,andfor individ-ualswithchildren,withaspouse,andwithsomaticcondition,and forthosewhoarelivinginruralareasandarenotofforeigndescent. Thereisalsoregionalvariation.Forexample,forpersonalcarethere isadifferenceof19percentagepointsbetweentheregionswiththe highestandthelowestaveragetake-upratio,i.e.an11-minute dif-ferenceperhourofcareforwhichsomeoneiseligible.Theseresults suggestthatvarioussubgroupsinthepopulationexperience differ-entbarrierstousehomecare,orthatthesesubgroupsaretreated differentlybyindependentassessors.

Thereareonlyafewsubgroupsforwhichthepredicted take-upratioiscloseto1andhencetherestrictionsfromtheeligibility assessmentmaybebinding(e.g.thosefromTurkishdescent)and thosearesmallinsize.Insum, forvirtuallyalltypesofcare,all populationsubgroups,andallregionspatientsuselesshomecare thantheyareeligiblefor,indicatingthatforthemajorityofhome careuserstherestrictionsonthenumberofhoursimposedbythe assessormaynotbeeffectiveinconstrainingdemandforpublicly ﬁnancedhomecare.

5. Discussion

Thereareseveralpotentialexplanationsforthelowuptakeof publicLTCbeneﬁtsamongeligiblepeople.Firstofall,otherpublic

(5)

Table3

Regressionresults.

Take-upratiopersonalcarea

Age18−64 −0.051(0.001)**

Age65−69 −0.036(0.001)**

Age70−74 −0.021(0.001)**

Age75−79 −0.013(0.001)**

Age80−84 Referencecategory

Age85−89 0.007(0.001)**

Age90−94 0.019(0.001)**

Man −0.016(0.000)**

Woman Referencecategory

Deceasedbefore1January2014 0.009(0.001)**

Surviveduntil1January2014 Referencecategory

Timetodeath,inmonths 0.000(0.000)

Notofforeigndescent Referencecategory

Foreigndescent:Westerncountries 0.090(0.003)**

Foreigndescent:Turkey 0.135(0.003)**

Foreigndescent:Morocco 0.028(0.002)**

Foreigndescent:Suriname 0.015(0.004)**

Foreigndescent:NetherlandsAntillesand

Aruba

0.001(0.001)

NoChildren Referencecategory

Children −0.008(0.001)**

Numberofchildren 0.001(0.000)**

Householdsize 0.000(0.000)

Nospouse/spouselivesinanother

household

Referencecategory

Spouselivesinthesamehousehold −0.038(0.000)**

Municipality:verystronglyurbanized Referencecategory

Municipality:stronglyurbanized −0.024(0.001)**

Municipality:moderatelyurbanized −0.029(0.001)**

Municipality:littleurbanized −0.031(0.001)**

Municipality:rural −0.032(0.001)**

Somaticcondition Referencecategory

Psychogeriatriccondition 0.042(0.001)**

Psychiatriccondition 0.030(0.001)**

Physicaldisability 0.051(0.001)**

Mentaldisability 0.036(0.002)**

Sensorydisability 0.017(0.002)**

Healthcareexpendituresinpreviousyear

(in1000euro)

0.000(0.000)**

Intercept 0.890(0.002)**

Regionsb _Yes

Timeb _Yes

Amountofcareb _Yes

Income:5%categoriesb _Yes

Numberofobservations 2,050,715

Note:resultsforothertypesofhomecareareavailableuponrequest.

a_Individuals_who_used_at_least_some_care_provided_in_kind_only.

b_Both_regressions_included₃₁_indicators_for_single-payer_regions_and₁₂

indica-torsforthesecondthroughthe13thfour-weekperiod.Inaddition,theregression

forpersonalcarecontained9indicatorstocontrolforthenumberofhoursthe

indi-vidualiseligibleforand19indicatorsforthelowest195%categoriesbasedon

standardizedhouseholdincome.

policiessuchascopaymentsandregionalbudgetrestrictionsmay partlyexplainthis.Apotentialconsequenceofalongerperiodof tightregionalbudgetsisaconstrainedsupplyside,resultingina lackofaccesstoservicesbypatients.Second,theassessmentrules ortheassessor’sinterpretationmaybetoogenerous.Third, infor-malcare,whichisaclosesubstituteforformalhomecareinsome cases,mayplayarole.Furtherresearchisneededtounderstandthe roleofeachofthecausesofthediscrepancybetweentheamount ofcarepeopleactuallyuseandareeligiblefor.

ThelimitedimpactofindependentneedsassessmentonLTCuse raisesthequestionaboutitseffectiveness,atleastwithintheDutch context.Ifmoralhazardandsupplierinduceddemandare effec-tively counteracted by budgetary restrictionsand co-payments, independent needs assessment may only be necessary on the extensivemargintopreventnon-eligiblepeoplefromusingLTC altogether,butnotontheintensivemargin.Thus,giventhefew userswhoseLTCuseisclosetothemaximum,itmaybesufﬁcient toassessthetypeofcareneededbutnotthenumberofhoursper

week,whileleavingtheallocationofthebudgetamonguserstothe payersortheproviders.Attheextensivemarginindependentneeds assessmentintheDutchLTCsystemmayeffectivelyblocksome peoplefromusingcarealtogether.Itisnotclear,however,towhat extentandforwhich peopletheindependentneedsassessment posesbarrierstoaccesscare.Toinvestigatethis,dataarerequired onwhichapplicationsarerejected,butthesedataarecurrentlynot availableforresearch.

AreformofDutchsystemLTCﬁnancingthatwasimplemented in2015mayshedfurtherlightontheimpactofindependentneeds assessmentonLTCuse. Aspartof thereform, mosthomecare wastransferredfromthepublicLTCinsuranceschemetoeither thepublichealthinsurancescheme(personaland nursingcare) orthemunicipalities(socialsupportandpersonalassistancewith activitiesofdailylife).Consequently,forpersonalandnursingcare independent needsassessors have been replaced by providers, whereasforsocialsupportandpersonalassistancetheyhavebeen replacedbymunicipalities.Theeffectofabolishingtheindependent needsassessmentcannotbeseparatedfromtheeffectsofother partsofthe2015reform[20].Yet,ourﬁndingthatindependent needsassessmenthadalimitedimpactontheintensivemarginof homecareusebeforethereformsuggeststhatabolishingthismay nothavehadasubstantialeffectontheusebyexistingusers.

Finally,thelimitedeffectofneedsassessmentontheintensive marginofhomecareuseintheNetherlandsalsoraisesquestions abouttheeffectivenessoftheindependentassessment inother countries,wherethedemandandsupplyofLTCareoftenmore restrictedthroughothermeasures[18].Furthermore,ourﬁndings suggestthatsettingobjectiveeligibilitycriteriaisnotenoughfor ensuringthattheactualusefollowsthesamerules,asweﬁndthat non-take-upisconsiderableandtheextentofnon-take-upvaries acrosssubgroupsinthepopulation.

6. Conclusion

ComprehensiveLTCinsurancemaygiverisetomoralhazardand supplier-induceddemand.Astrategythatoftenusedforpublicly ﬁnancedLTC–butnotforothertypesofhealthcare–istoorganize anindependentneedsassessment.Suchanindependent assess-mentmaycountermoralhazardandsupplier-induceddemandby determiningwhichtypesandwhichamountofLTCapersonreally needs.

Although the extent of moral hazard and supplier-induced demandinpubliclyﬁnancedLTCcannotbeestablished,weareable toinvestigatewhetherindependentneedsassessmenteffectively restrictsLTCuseat theintensivemargin. Ifindependentneeds assessmentdoesnotimposeabindingconstraintonLTCuse,itis highlyunlikelythatitreducesmoralhazardandsupplier-induced demandamongthosewhoareeligibleforcare.Hence,thiswould removeanimportantreasonfororganizingtheindependentneeds assessment.

Wefindthatindependentneedsassessmentdoesnotseemtobe effectiveinconstrainingpubliclyfinancedLTCuseoncepeopleare consideredeligibleforreceivingcare.TheuptakeoftheLTCbenefits variesacrosssubgroups,butvirtuallyallsubgroupsuseonlypartof theLTCforwhichtheyareeligible.Thevariationinuptakeis associ-atedwiththepatient’spersonalandhouseholdcharacteristicsand hisorherregionofresidence.

Declarationofcompetinginterest

PartoftheresearchwasdonewhilePieterBakxperformed con-tractresearchfortheCPB.PieterBakxandErikSchutacknowledge supportthroughtheNETSPARgranton“Optimalsavingand insur-anceforoldage:theroleofpubliclong-termcareinsurance”.These fundingsourceswerenotinvolvedintheresearch.

(6)

Acknowledgement

The authorsthankPeterAlders,AdamElbourne,Esther Mot, Marielle Non, Johan Polder and Wouter Vermeulen for com-ments.Inthisarticleweusenon-publicmicrodatafromStatistics Netherlands.ApreviousversionofthepaperispublishedasCPB DiscussionPaper327.

References

[1]ArrowK.Uncertaintyandthewelfareeconomicsofmedicalcare.American EconomicReview1963;53(5):941–73.

[2]BakxP,ChernichovskyD,PaolucciF,SchokkaertE,TrottmannM,WasemJ,Schut F.Demand-sidestrategiestodealwithmoralhazardinpublicinsurancefor long-termcare.JournalofHealthServicesResearch&Policy2015;20(3):170–6.

[3]GoncalvesJ,WeaverF.Homecare,hospitalizationsanddoctorvisits.University ofGenevaworkingpaperseries,no.14-09-5.Geneva:UniversityofGeneva; 2014.

[4]TamiyaN,NoguchiH,NishiA,ReichM,IkegamiN,HashimotoH,ShibuyaK, KawachiI,CampbellJ.Populationageingandwellbeing:lessonsfromJapan’s long-termcareinsurancepolicy.Lancet2011;378(9797):1183–92.

[5]WilleméP,GeertsJ, CantillonB,MusscheN.Long-termcareﬁnancingin Belgium.In:Costa-FontJ,CourbageC,editors.Financinglong-termcarein Europe.Institutions,marketsandmodels.Houndmills,Basingstoke:Palgrave Macmillan;2012.

[6]CIZ.TrendrapportagelandelijkeindicatiestellingAWBZ2009.Driebergen:CIZ; 2010.

[7]ArrighiY,DavinB,TrannoyA,VentelouB.Thenon-takeupoflong-termcare

beneﬁtinFrance:apecuniarymotive?HealthPolicy2015,http://dx.doi.org/

10.1016/j.healthpol.2015.07.003.

[8]Guthmuller S, JusotF, Wittwer J. Improvingtakeup ofhealth insurance program: a social experiment in France. Journal of Human Resources 2014;49(1):167–94.

[9]PrendergastC.Thelimitsofbureaucraticefﬁciency.JournalofPoliticalEconomy 2003;111(5):929–58.

[10]CBS.Gezondheidenzorgincijfers2014.TheHague:CBS;2015.

[11]MotE.TheDutchsystemoflong-termcare.Centraalplanbureaudocument, no.204.TheHague:CPB;2010.

[12]RMO.Indicatiestelling:omstredentoegangtotzorg.TheHague:RMO;2010.

[13]LindeboomM,vanderKlaauwB,VriendS.Auditratesandcompliance:aﬁeld experimentincareprovision.JournalofEconomicBehavior&Organization 2016;131:160–73.

[14]Rijksoverheid. Beleidsregelsindicatiestelling AWBZ 2012;2011[Accessed

on12June2019]http://wetten.overheid.nl/BWBR0030849/geldigheidsdatum

02-01-2012.

[15]Rijksoverheid.Beleidsregelsindicatiestellingwlz 2015;2014[Accessedon

12 June 2019] http://wetten.overheid.nl/BWBR0036073/geldigheidsdatum

23-04-2015.

[16]DiepstratenM,DouvenR,WouterseB.Canyourhousekeepyououtofanursing home?HealthEconomics2020;29(5):540–53.

[17]CVZ.WachtlijstonderzoekAWBZ.Factorendievaninvloedzijnopde betrouw-baarheidvanwachtlijstinformatie.Diemen:CVZ;2013.

[18]OECD.Helpwanted?Providingandpayingforlong-termcare.OECDhealth policystudies.Paris:OECD;2011.

[19]CBS.Monitorlangdurigezorg;2018[Accessedon12June2019] mlzstat-line.cbs.nl.

[20]BakxP,Garcia-GomezP,RellstabS,SchutF,vanDoorslaerE.Hervorming lang-durigezorg:trendsinhetgebruikvanverplegingenverzorging.NetsparDesign Paper141;2020.