Prices of pharmaceuticals : a comparison of prescription drug prices in Sweden with nine European countries

S A M F U N N S - O G

N Æ R I N G S L I V S F O R S K N I N G A S I n s t i t u t e f o r R e s e a r c h i n E c o n o m i c s a n d B u s i n e s s A d m i n i s t r a t i o n

R e p o r t

Prices of Pharmaceuticals:

A Comparison of Prescription Drug Prices in Sweden with Nine European Countries

Kurt R. Brekke

Tor Helge Holmås

næringslivsforskning AS

- er et selskap i NHH-miljøet med oppgave å initiere, organisere og utføre eksternfinansiert forskning. Norges Handelshøyskole, Universitetet i Bergen og Stiftelsen SNF er aksjonærer.

Virksomheten drives med basis i egen stab og fagmiljøene ved NHH og Institutt for økonomi (UiB).

SNF er Norges største og tyngste forsk- ningsmiljø innen anvendt økonomisk- administrativ forskning, og har gode samarbeidsrelasjoner til andre forsk- ningsmiljøer i Norge og utlandet. SNF utfører forskning og forskningsbaserte utredninger for sentrale beslutnings- takere i privat og offentlig sektor.

Forskningen organiseres i programmer og prosjekter av langsiktig og mer kortsiktig karakter. Alle publikasjoner er offentlig tilgjengelig.

in Economics and Business Administration

- is a company within the NHH group.

Its objective is to initiate, organize and conduct externally financed research.

The company shareholders are the Norwegian School of Economics and Business Administration (NHH), the University of Bergen (UiB) and the SNF Foundation. Research is carried out by SNF´s own staff as well as faculty members at NHH and the Department of Economics at UiB.

SNF is Norway´s largest and leading research environment within applied economic administrative research. It has excellent working relations with other research environments in Norway as well as abroad. SNF conducts research and prepares research-based reports for major decision-makers both in the private and the public sector. Research is organized in programmes and projects on a long-term as well as a short-term basis. All our publications are publicly available.

SNF ^REPORT N ^O . 01/12

P ^RICESOF P HARMACEUTICALS :

A C OMPARISONOFPRESCRIPTIONDRUGPRICESIN

S WEDENWITHNINE E ^UROPEAN C ^OUNTRIES

—”–Ǥ”‡‡ƒ†‘”‡Ž‰‡‘Ž¤•

’”‘Œ‡…–‘ǤʹͶͻͻ

Š‡’”‘Œ‡…–‹•ˆ—†‡†„›

Ȃ†‡ˆ‘”•ƒ†‡Ž¡‡‡†‡Ž•ˆÚ”‡–ƒ‰‡

ȋ Ǧ( Ȍ

ǡ ʹͲͳʹ

©Materialetervernetetteråndsverkloven.Uten uttrykkeligsamtykkeereksemplarfremstillingsom utskriftogannenkopieringbaretillattnårdeter hjemletilov(kopieringtilprivatbruk,sitato.l.)eller avtalemedKopinor(www.kopinor.no)

Utnyttelseistridmedlovelleravtalekanmedføre erstatningsͲogstraffeansvar.

ISBN978Ͳ82Ͳ491Ͳ0787Ͳ2Printedversion

ISBN978Ͳ82Ͳ491Ͳ0788Ͳ9Onlineversion ISSN0803Ͳ4036

Preface

ThisreporthasbeenconductedonbehalfofLIF–theresearchͲbasedpharmaceuticalindustryin Sweden.ThepurposeoftheprojecthasbeentostudythepricelevelofpharmaceuticalsinSweden relativetoasetofEuropeanreferencecountries.TheprojectwasinitiatedinDecember2011and finalizedinFebruary2012.DuringtheprojectperiodwehadameetingwithLIFinJanuary2012.

TheprojectbuildsonfourpreviousreportswehavewrittenoncrossͲcountrypricecomparisonsin EuropeusingNorwayasthebasecountry.Thefirstreport(SNFreport05/08)writtenforthe NorwegianMinistryofHealth,whereasthethreesubsequentreports(SNFreport06/09,08/10,and 11/11)waswrittenfortheNorwegianPharmacyAssociation.Thecurrentreportdiffersfromthe previousonesalongtwoimportantdimensions:(i)weuseSwedenasthebasecountryforprice comparisons;(ii)weselectasampleofsubstancesthathasnogenericsalesinSweden.

TheprojecthasbeenundertakenbyProfessorKurtR.BrekkeattheNorwegianSchoolofEconomics (NHH)andSeniorResearcherTorHelgeHolmåsattheUniRokkanCentre.Theauthorsareaffiliated totheInstituteofResearchofEconomicsandBusinessAdministration(SNF)andtheCentrefor HealthEconomicsinBergen(HEB).

WearegratefultoIMSHealthforallowingustousethedatausedinSNFreport11/11inthisproject.

WearealsogratefulforcommentsandsuggestionsbyKarolinaAntonovfromLIF,althoughthisdoes notmakeherresponsibleforcontentandtheconclusionsinthereport.Possibleerrorsandmistakes areofcoursetheresponsibilityoftheauthors.

Bergen,February2012

KurtR.Brekke TorHelgeHolmås

Summary

WestudythepricelevelofpharmaceuticalsinSwedenrelativetothefollowingnineEuropean countries;Austria,Belgium,Denmark,Finland,Germany,Ireland,theNetherlands,Norway,and UnitedKingdom(UK).Oursampleconsistsofprescriptiondrugsthatdonothavegenericsalein Sweden.UsingIMSHealthdataonpricesandsalesvolumesforthefirsthalfof2010,wecompute severalpriceindicestodescribethepricedifferencesandpotentialcostsavingsinthenonͲgeneric marketsegment.OurresultsshowthattheSwedishpricelevelisslightlybelowaveragerelativeto theotherEuropeancountries.UK,NorwayandtheNetherlandstendtohavelowerpricesthan Sweden,whereasGermany,IrelandandDenmarktendtohavehigherprices.Finlandhaslower pricesthanSwedenonwholesalelevel,butslightlyhigherpricesatretaillevel.AustriaandBelgium haveaboutthesamepricelevelasSweden.

TableofContents

Preface Summary

Chapter1 Introduction...1

Chapter2 DataandSample...3

2.1 Thesampleofsubstances...3

Chapter3 MethodandAnalysis...8

3.1 Basicpriceindexanalysis...8

3.2 Matchingofproducts...10

Chapter4 Results...12

4.1 Priceindicesbasedonidenticalpacks...12

4.2 Priceindicesbasedonsubstance(dose)prices...15

4.3 Genericcompetition...18

4.4 Parallelimport...20

4.5 ComparisonwiththeNorwegianstudy...21

4.6 Priceindicesfornewandoldproducts...22

Chapter5 ConclusionsandRemarks...25

Referencelist... ...26

APPENDIX... ...27

1

Chapter1 Introduction

ThepurposeofthisstudyistoanalyzethepricelevelofpharmaceuticalsinSwedenrelativetoother EuropeancountriesinthenonͲgenericmarketsegment.WehaveobtaineddatafromIMSHealthon pricesandsalesvolumesforthefirsthalfof2010.ThesamplecontainstopsellingprescriptionͲbound substanceswithnogenericsaleontheSwedishmarket.Wecomparethepricesofthissetof(brandͲ name)productswiththefollowingnineEuropeancountries:Austria,Belgium,Denmark,Finland, Germany,Ireland,theNetherlands,Norway,andUnitedKingdom(UK).

Weusepriceindexanalysistostudythepricedifferencesacrosscountries.Thisusuallyentailsa tradeͲoff between precision and representativity, which is particularly present considering heterogeneousproductssuchaspharmaceuticals.Wethereforetaketwodifferentapproachesthat invariousdegreessatisfythesecriteria.First,wecomparepricesofidenticalpacksinSwedenandthe referencecountries.Thisapproachyieldsahighdegreeofprecision,butisnotlikelytoproduce representativesamplesofproducts,implyingthatpricedifferencesarepossiblyincorrect.Second, wecompareaveragesubstance(dose)pricesacrosscountries.Thisapproachgeneratesmore representativesamplesineachcountry,andisthereforelikelytoproducemorereliablemeasuresof pricedifferencesandpotentialcostsavingsacrosscountries.

WecomputeawidesetofpriceindicesusingSwedenasthebasecountry.First,wecompute bilateralpriceindicesforallmatchingproducts.TheseindicesshowthattheSwedishpricelevelis slightlybelowaverageatbothwholesaleandretaillevel.Second,wecomputeseparatepriceindices forprotectedandnonͲprotectedsubstances.Intheprotectedsegment,theSwedishpricelevelis fairlylow,whereasinthenonͲprotectedsegmenttheSwedishpricelevelismoreatthehigherend.

Thelattermightbeduetogenericsalesinthereferencecountries,sowecomputepriceindices whereweexcludesubstanceswithgenericsaleinthereferencecountries.Asexpected,theSwedish pricelevelbecomesmorefavorable,buttheeffectisrathersmallanddoesnotchangetherankingof countries.

Third,wecomputepriceindicesforthesubstanceswithparallelimportsinSweden.Forthissample ofproducts,theSwedishpricelevelisfairlyhigh.However,thisismostlikelyduetothefactthat parallelimportismoreprofitableforproductswithrelativelyhighprices.Finally,wecomparethe priceindicesderivedinthisstudywiththeonesobtainedinBrekke,HolmåsandStraume(2011) (BHSͲstudy)thatusedNorwayasthebasecountry.Asexpected,wefindthatSwedenbecomes

2

relativelycheaperwhenbeingthebasecountry.However,theeffectisnotverystronganddoesnot changetherankingofcountriesqualitatively.

Therestofthereportisorganizedasfollows.InChapter2weprovideanoverviewoverthedataand oursampleofproducts.InChapter3wepresentthemethodsthatareusedtoconstructtheprice indices.InChapter4wepresentanddiscusstheresultswederivefromtheanalysis,and,finally,in Chapter5wedrawsomeconclusionsandoffersomeremarks.

3

Chapter2 DataandSample

DataareprovidedbyIMSHealthandcontaindetailedinformationonprescriptionͲboundsalesthe firstsixmonthsof2010forthefollowingtenEuropeancountries:Austria,Belgium,Denmark, Finland,Germany,Ireland,theNetherlands,Norway,Sweden,andtheUK.Thedatasetisthesame asusedinBHS(2011)andcomprisesthe300mostsellingsubstancesinNorway.Inthecurrentstudy welimitthesampletothesetofsubstancesthatdonothavegenericsaleinSweden.Belowwefirst describeoursampleofsubstancesandproducts,andthenthedatawemakeuseofinouranalysis.

2.1 Thesampleofsubstances

Inourdatathereare169substanceswithoutgenericsalesinSweden.Weexclude16ofthese substances,becausetheylackinformationonpatentstatus,whichmakesitdifficulttocompute separatepriceindicesfortheprotectedandnonͲprotectedmarketsegment.¹Thisleavesuswitha sampleof153substances.TableA.1intheAppendixprovidesafulllistofthesesubstanceswith informationaboutpatentstatus,genericsale,andparallelimport.Inthetablebelowwereportthe numberofsubstancesineachcountryinoursample.

Table2.1.NumberofsubstancesinSwedenandreferencecountries,2010.

All substances

Substances withpatent

status

“Protected“

Substances withpatent status“NotͲ Protected“

Substances withgeneric

sales

Substances withparallel

import

Sweden 153 106 47 0 59

Norway 153 96 57 12 37

Denmark 148 109 39 14 89

Finland 142 85 57 15 17

UK 137 95 42 18 97

Germany 143 104 39 20 122

Netherlands 144 101 43 21 112

Belgium 130 91 39 12 6

Austria 139 101 38 11 2

Ireland 136 96 40 12 0

Globalsubstances 104 54 23 Ͳ Ͳ

1The16substanceswithoutinformationaboutpatentstatusarethefollowing:ALFACALCIDOL,

CALCIUM;COLECALCIFEROL,CINACALCET,CLOSTRIDIUMBOTULINUMTOXINTYPEA,CYANOCOBALAMIN, CYANOCOBALAMIN;FOLICACID;PYRIDOXINE,EPINEPHRINE,FOLLICLEͲSTIMULATINGHORMONE;LUTEINISIN, GLUCOSAMINE,HYDROXOCOBALAMIN,IMMUNOGLOBULINBASE,LANTHANUM,LITHIUM,PALIVIZUMAB, POLLEN,URSODEOXYCHOLICACID.

4

Thetableshowsthat106ofthe153substanceswithoutgenericsalesinSwedenareprotectedby patentregulation.However,theresidual47substancesarenonͲprotected,butgenericproducers haveabstainedfromenteringthemarket.Thetablealsoshowsthat59ofthe153substanceshave salesofparallelimportedproductsinSweden.

Thematchingofsubstancesacrosscountriesisfairlyhigh.Infact,allofthe153substancesinSweden arepresentontheNorwegianmarket.InDenmarkonly5ofthe153substancesarenotonthe market.Belgiumhasthelowestnumberwith130matchingsubstances.Thenumberofglobal substances–i.e.,substancesthatareonthemarketinallcountries–is104.Thisreflectsthatitnot thesamesubstancesthataremissinginthedifferentcountries.

Thesecondcolumninthetableshowsthenumberofsubstancesthatareprotectedbypatent regulation.InSweden106ofthe153substancesareprotected.Interestingly,thisnumbervaries acrossthecountriesinoursample.Forinstance,inNorwayonly96substancesareprotected, whereasinDenmark109substancesareprotected.Ifweconsiderthe104globalsubstances,only54 substancesareprotectedinallofthecountriesinoursample.Thesefiguresdemonstratethe variationinthenationalenforcementofpatentprotection,despitetheharmonizationacross EuropeancountriesthroughEUlegislation.

Protectedsubstancesdonothavegenericcompetitionbydefinition.However,thisisnotthecasefor thenonͲprotectedsubstances.Inthissegmentabsenceofgenericcompetitionisduetothegeneric producersnotfindingitprofitabletolaunchtheirgenericversionsonnationalmarkets.Whilenone ofthe153substanceshavegenericsalesinSweden,thisisnotthecaseinourreferencecountries.

ThehighestnumberofsubstanceswithgenericsalesinoursampleisintheNetherlands,where21 (of43nonͲprotected)substanceshavegenericsales.ThelowestnumberisinAustriawith11(of38 nonͲprotected)substanceshavegenericsales.Intable2.2belowpresentthefulllistofsubstances thathavegenericcompetitioninthereferencecountries.Thisdemonstratesthevariationacross countriesintheextentofgenericcompetitioninthenonͲprotectedsegment.

Parallelimportisanothersourceofcompetitionmainlyforprotectedsubstances,butpotentiallyalso fornonͲprotectedsubstances,especiallyintheabsenceofgenericcompetition.Weseefromthe tablethatparallelimportplaysalargeroleinGermanyandtheNetherlands.InGermany122of143 (matching)substanceshaveparallelimport.However,inIreland,AustriaandBelgiumparallelimport seemstoplayaveryminorrole.Infact,inIrelandwedonotobserveparallelimportofanyofthe productsinoursample.

5

Table2.2.Substanceswithgenericsalesinthereferencecountries,2010.

Norway DenͲ

mark

Finland UK GerͲ many

NetherͲ lands

Belgium Austria Ireland

AMITRIPTYLINE X X X X X

ATORVASTATIN X

BETAMETHASONE,SALICYLICACID X

BUDESONIDE,FORMOTEROL X

BUMETANIDE X X X X

C1INHIBITOR(HUMAN) X X

CALCIPOTRIOL X X X X

CLOBETASOL X X X X X X

DESOGESTREL,ETHINYLESTRADIOL X X X X

DIPYRIDAMOLE X X X X

DONEPEZIL X X X

DORZOLAMIDE,TIMOLOL X X X

EBASTINE X

EPINEPHRINE,LIDOCAINE X X X X X

ESCITALOPRAM X X X

ESOMEPRAZOLE X X X X

FLECAINIDE X X X X X X X

FLUVASTATIN X X X X X X X

FUSIDICACID X X

HYDROXYZINE X X X

LATANOPROST X

LERCANIDIPINE X X X X X X X X

LEVONORGESTREL X

METHENAMINE X X

METHYLPREDNISOLONE X X X

METOCLOPRAMIDE X X X X X X X

MONTELUKAST X X

NIFEDIPINE X X X X X X X X

NYSTATIN X X X

OLANZAPINE X X

PERPHENAZINE X X

PIVMECILLINAM X X X

PRAMIPEXOLE X

PROGESTERONE X X X X X X

QUETIAPINE X X X

SALMETEROL X

SIBUTRAMINE X

SILDENAFIL X

TERBUTALINE X X X X

THALIDOMIDE X

TIBOLONE X

VALSARTAN X

WARFARIN X X X

Letustakeacloserlookatthenumberofproductsinoursample.Thetablebelowprovidesthe numberof(unique)packsonthemarketforeachcountry.Weseethatthereare791uniquepacksof the153differentsubstancesonthemarketinSweden.Thisimpliesanaverageofslightlyabove5 differentpackspersubstance.SincetherearenogenericsalesinSwedenforthissampleof substances,the791packsareallbrandͲnameorparallelimported(brandͲname)products.

6

Table2.3.Numberofpacksineachcountrydependingonpatentstatusandparallelimports,2010.

Allsubstances Patentstatus

“Protected“

Patentstatus

“NotͲProtected“

Parallelimport

Sweden 791 566 225 170

Norway 630 377 253 53

Denmark 732 503 229 292

Finland 664 363 301 33

UK 661 436 225 354

Germany 1200 750 450 759

Netherlands 807 529 278 450

Belgium 516 340 176 9

Austria 606 383 223 4

Ireland 461 311 150 0

Thefirstcolumnshowsthatthereisquitesomevariationacrossthecountriesinthenumberofpacks onthemarketinoursampleofsubstances.Germanyhasthehighestnumberwith1200uniquepacks ofthe143substances.Irelandhasthelowestnumberofpacks,withonly461packsofthe136 substancesinthesample.Thenumberofpacksislikelytobehigherforsubstanceswithparallel importsand/orgenericsales,wherethebrandͲnameproducersfacecompetitionfromparallel tradersorgenericproducers.

Letusfinallytakealookatthenumberofdosesperpackineachcountry.Weseefromthetable belowthatthereissomevariationacrosscountriesintheaveragepacksize.Indeed,Swedenhasthe highestaveragenumberofdoses(45.1)perpack,whereasUKandtheNetherlandshavethelowest packsizeswith29.9and27.3dosesonaverage.

Table2.4.Averagenumberofdosesperpack,2010.

Allsubstances Substanceswith

patentstatus“Protected“

Sweden 45.1 39.6

Norway 44.5 38.0

Denmark 41.1 36.7

Finland 40.5 35.2

UK 29.9 26.7

Germany 40.4 39.1

Netherlands 36.7 27.2

Belgium 37.9 35.7

Austria 27.3 25.6

Ireland 32.5 27.9

7

Wealsoseethatthepacksizesareonaveragelowerforthesubstancesthatareprotectedbypatent regulation.Thispatternisconsistentacrosscountries,andimpliesthatthenonͲprotectedsegment hasahigheraveragenumberofdosesperpack,whichdrivesuptheaverageforallsubstances.This canbeduetogenericproducersenteringthenonͲprotectedmarketwithlargepacksand/orparallel importersenteringtheprotectedmarketwithsmallpacks.Alternatively,thehigheraveragepacksize inthenonͲprotectedsegmentisduetosampleselection,wheresomeofthesubstancesinthenonͲ protected(protected)segmentarecharacterizedbyrelativelylarge(small)packsizes.

Moreimportantly,therelativelylargevariationinthenumberofpacksandtheaveragepacksize acrossraisesaconcernregardingtherepresentativityofpricecomparisonsbasedonmatchingof identicalpacks.Theissueisrelatedtohowmanyofthe791uniquepacksontheSwedishmarketwe areabletofindinthereferencecountries.Indeed,ifalargenumberofpacksarenotpossibleto match,thentheresultingpricedifferencesmightbebiasedandpotentiallyincorrect.Wewillreturn tothisissueinthenextchapters.

8

Chapter3 MethodandAnalysis

Wewillusestandardpriceindexmethodtocomparepricesacrosscountries.²Priceindicesareoften sensitivetohowtheyarecalculated.Theresultsmightvaryaccordingtothesample,thematching procedure,andtheuseofweights.Wethereforecomputeseveralpriceindicesusingdifferent procedures.Inthischapterwefirstdescribesomegeneralaspectsofpriceindexanalysisandthen afterwardsthevariousmatchingproceduresthatweuse.

3.1 Basicpriceindexanalysis

Apriceindexisaweightedaverageofpricesfordifferentproducts,usuallycalculatedovertime,such astheconsumerpriceindex.Ifwehavetwotimeperiods,period0andt,andtwoproducts,product 1and2,wecanexpressapriceindexasfollows:

, 100

2 0 2 1 0 1

2 2 1

1 u

w p w p

w p w I p

t t

P

wherew1andw2areweightsappliedtotherespectiveprices p₁⁰,p₁^t,p₂⁰and p^t₂.Whencalculating priceindicesitiscommontousesoldquantitiesoftheproductsasweightstotakeaccountofthe relativeimportanceofthepricesofthevariousproducts.Wecanobtaintwodifferentpriceindices dependingonthechoiceofweights.Ifwechoosesoldquantitiesinthecomparisonperiod(periodt) asweights,weobtainthesoͲcalledPaaschepriceindex:

, 100

2 0 2 1 0 1

2 2 1

1 u

t t

t t t t

P p q p q

q p q P p

whereq₁^t andq₂^tarethequantitiesofproduct1and2soldinperiodt.Ifweusequantitiessoldin thebaseperiod(period0)asweights,weobtainthesoͲcalledLaspeyrespriceindex:

,

0 100

2 0 2 0 1 0 1

0 2 2 0 1

1 u

q p q p

q p q L p

t t

P

whereq₁⁰andq₂⁰arethequantitiesofproduct1and2soldinperiod0.Boththesepriceindiceswill expresschangesinaveragepricesovertime.Ifpricesareless(more)than100,thismeansthatthere hasbeenareduction(increase)inaveragepricesovertheperiod.

2SeeDanzon(1999)andDanzonandChao(2000)foradiscussionandanalysisofcrossͲcountryprice comparisonsofpharmaceuticals.

9

Inthisstudywecalculatedifferencesinaveragepricesacrosscountries(notovertime)tostudy whetherthepricesinSwedenarehigherorlowerthaninthereferencecountries.Toexplainhowthe priceindicesareconstructed,weassumetwocountries,SwedenandAbroad,whereproducts1and 2aresold(butinpotentiallydifferentquantities).Thegeneralpriceindexcanthenbeexpressedas

, 100

2 2 1 1

2 2 1

1 u

w p w p

w p w

I p_{S S}

A A

P

wherep₁^Aandp₂^Aarethepricesofproduct1and2abroad,p₁^Sandp₂^Sarethepricesofproducts 1and2inSweden,andw1andw2aretheweightstobeappliedtothesedifferentprices.Ifweuse quantitiessoldabroadasweights,wecalculatethePaaschepriceindex.However,forcrossͲcountry pricestudies,itismorecommontocomputetheLaspeyrespriceindex,wherethequantitysoldin thebasecountry (Sweden) is used as weights. The Laspeyres price index for crossͲcountry comparisonscanbeexpressedasfollows:

, 100

2 2 1 1

2 2 1

1 u

S S S S

S A S A

P p q p q

q p q

L p

whereq₁^Sandq₂^Sarequantitiessoldofproduct1and2inSweden.Ifthepriceindexismore(less) than100,thismeansthataveragepricesabroadarehigher(lower)thaninSweden.However,itdoes notmeanthatallpricesarehigherabroadthaninSweden.Wecanimaginethatproduct1hasa higherpriceabroadthaninSweden(p₁^A ! p₁^S),whiletheconverseistrueforproduct2(p₂^A p₂^S).

Whetherthepriceindexwillbehigherorlowerthan100dependsonthemagnitudeofprice differencesandthequantityweights.

UsingSwedishquantityweights,thepriceindicesprovideameasureofwhattheconsumptionof pharmaceuticalsinSwedenwouldcostwiththeforeignpricelevel.Apriceindexbelow100would showthecostsavingsthatcanbeobtainedifSwedenimportedtheforeignpricesgiventhatthe Swedishconsumptionremainedunchanged.Thisisastrongassumptionthatisonlyreasonableifthe demandisperfectlypriceinelastic.Ifthisisnotthecase,thendemandresponseswouldeither counteractorreinforcethecostsavingsreportedbythepriceindices.Wecanalsoimaginesupply sideresponsesduetocompetitionwhenpricesofrivalproductsarechanged.Inaddition,the implementationoflowerpricesisofcourseahardtask.Thus,thecostsavingsmeasuredbytheprice indicesshouldbetreatedwithsomecaution.

10

3.2 Matchingofproducts

Havingdecidedonthetypeofpriceindextocompute,thenextquestionishowtomatchthe productsinthebasecountry(Sweden)withtheproductsinthereferencecountries.Asmentioned before,thereisaninevitabletradeͲoffbetweenprecisionandrepresentativitywhenitcomesto comparingpricesofheterogeneousproductssuchaspharmaceuticals.

Precisionisattainedwhencomparingpricesofexactlythesameproductacrosscountries.Acommon approachistoselectthemostsoldpackwithagivensubstanceinthebasecountryandcomparethe priceofthispackwiththepriceofanidenticalpackinthereferencecountry.Theproblemwiththis procedureistwoͲfold.First,thispackisrepresentativeinthebasecountryduetobeingthemost soldone,butrarelyrepresentativeinthereferencecountry,whereotherpackswiththesame substancemighthavehighersales.Second,thispackmightnotbesoldatallinthereference country.Ifthisisthecaseforseveralpacks,thenthesampleofproductsthatformsthebasisfor pricecomparisonswouldbebiasedandnotrepresentativeforboththebaseandthereference country.Thepriceindiceswouldalsobeverysensitivetothesampleofproducts.

Inordertomitigatethisproblemsomewhat,wedonotmatchonlythemostsellingpacks,but insteadmatchthewholepopulationofpacksonthemarketinthebasecountry.Thisgivesamuch broadersampleandincreasestherepresentativityatthesametimeastheprecisionintheprice comparisonsispreserved.Incomputingthepriceindices,weusethenumberofdosessoldofagiven packinthebasecountry(Sweden)asweights.However,thematchingnumbersshowthateventhis procedureresultsinasignificantlossofproductsinboththebaseandthereferencecountries,which meansthatthereisstillaconcernthatthepricedifferencesarenotrepresentativeacrosscountries.

To ensure representativity we compute the (volumeͲweighted) average dose price for each substanceineachcountry.Toillustratehowthispriceiscomputed,wemayconsiderthefollowing example.AssumethatforactivesubstanceA(forexampleinSweden)wehavethreedifferentpacks withthefollowingpricesandsalesvolumes:

x Pack1A:thepriceisSEK10perdoseandthesalesvolumeis5doses x Pack2A:thepriceisSEK20perdoseandthesalesvolumeis10doses x Pack3A:thepriceisSEK30perdoseandthesalesvolumeis15doses ThevolumeͲweightedaveragedosepriceisthencomputedasfollows:

33 . 23 30 SEK

30 15 30 SEK

20 10 30 SEK

10 5

SEK u u u

11

Thearithmetic(unͲweighted)averagedosepriceintheexampleisSEK20.ThevolumeͲweighted averagedosepriceishigherbecausethemoreexpensivepacksalsohavehighersalesthanthe cheaperpack.Ifthecheaperpackhadahighersalesvolumethanthemoreexpensivepacks,thenthe volumeͲweightedaveragepricewouldbecomelowerthanthearithmeticaverageprice.Finally,ifwe selectedonlythetopͲsellingpack,wewouldinthisexamplereportapricelevelofSEK30perdose, andconsiderthistoberepresentativepriceforthissubstanceinthebasecountry(Sweden).

Thus,theadvantageofusingthevolumeͲweightedaveragedosepriceisthatitmakesuseofallsales informationineachcountryandproducesarepresentativepricelevelofeachcountry.Moreover, usingthismeasureimpliesthatwematchsubstancesandnotpacksacrosscountries.Thesetof matchingsubstancesismuchlargerthanthesetofmatchingpacks,whichmeansthatsampleisnot biasedandthepriceindicesarelikelytobemoreprecisemeasuresofpricedifferenceacross countries.

12

Chapter4 Results

Inthischapterwepresenttheresultsfromthepriceindexanalysis.Wehavecomputedawiderange ofpriceindicesforvariousmarketsegmentsusingdifferentmatchingprocedures.Thecompleteset ofpriceindicescanbefoundinTableA.2toA.5intheAppendix.Belowwepresenttheoneswefind moreinterestingandrelevant.

4.1 Priceindicesbasedonidenticalpacks

Wefirstcomparepricesofidenticalpacks(samesize,strengthandformulation).Astandard approachistoselectthetopsellingpackwithinagivensubstanceinthebasecountry(Sweden)and comparethepriceofthispackwithpricesofidenticalpacksinthereferencecountries.Theproblem withthisprocedureisthatthesamplebecomesverysmallandpotentiallybiased,sotheresulting priceindiceswillbeincorrectmeasuresofcrossͲcountrypricedifferences.

WethereforetakethesameapproachasinBHS(2010)andmatchall(notjustthetopselling) identicalpacksbetweenSwedenandeachreferencecountry.Thisgivesusamuchlargerandmore representativesampleofproducts.Quantityweightsarecomputedbydividingthenumberofdoses soldofagivenpackwiththetotalnumberofdosessoldofallthematchingpacks.Weusethe Swedishsalesvolumestocomputethequantityweightsandusethesetocomputethebilateralprice indices.Sincethesampleofmatchingpacksvariesacrosscountries,therewillbeauniquesetof quantityweightsforthedifferentpriceindices.

Achallengewithmatchingidenticalpacksisthattheremightbeseveralidenticalpackswithinagiven country.Thesamepack(size,strengthandformulation)maybeofferedbydifferentfirms,suchas brandͲnameproducersandparallelimporters.Insomecountriestherearealsogenericproducers.

WehandlethisissuebycomputingthevolumeͲweightedaveragepriceforidenticalpackswhen thereareseveralidenticalpacksinagivencountry.This(representative)packpriceisthenthebasis forcomputingthepriceindices.

ThefullsetofbilateralpriceindicesbasedonidenticalpackscanbefoundinTableA.2inthe Appendix.Belowwepresentsomeofthepriceindices.Wefirstconsiderthepriceindicesbasedon allsubstances(thefullsample)inSweden.Thefigurebelowshowsthesepriceindicesatbothretail andwholesalelevel.Thecountriesarerankedfromlowesttohighestpricesaccordingtothewhole pricelevels.

13

Figure4.1.Bilateralpriceindicesbasedonidenticalpacks,allsubstances,2010.

ThefigureshowsthattheSwedenisinthemiddleofthepriceranking.Irelandisthemostexpensive countrywith39percenthigherwholesalepricesthanSweden.AlsoGermanyandDenmarkarealso moreexpensivethanSwedenwith,respectively,19and16percenthigherwholesaleprices.Austria, BelgiumandtheNetherlandshaveaboutthesamepricelevelasSweden.Theresidualcountrieshave significantlylowerprices.UKisthecheapestcountrywith24percentlowerwholesalepricesthan Sweden.NorwayandFinlandhaveabout10percentlowerpricesatwholesalelevel.Theprice rankingatretaillevelismorefavorableforSweden.Thisisduetolowerpharmacymarginsthanthe restofthecountriesexceptforUK.

ThenextfigureshowstheseparatepriceindicespackswithprotectedandnonͲprotectedsubstances.

WematchonlypacksthathavethesamepatentstatusinSwedenandthereferencecountries,and computebilateralpriceindices.Wereportthepriceindicesbasedonpricesatwholesalelevel.The priceindicesforretaillevelcanbefoundintableA.2intheAppendix.Thecountriesarerankedfrom lowesttohighestpricesintheprotectedmarketsegment.

14

Figure4.2.Bilateralpriceindicesforidenticalpacks,protectedandnonͲprotectedsubstances, wholesaleprices,2010.

ThefigureshowsthattheSwedishpricelevelismorefavorablefortheprotectedmarketsegment.In thismarketsegmentonlyUKandtheNetherlandshavelowerwholesaleprices.TheUKpricelevelis 28percentlowerthaninSweden,whereastheNetherlandshas8percentlowerprices.Intheother end,wefindIrelandwith35percenthigherprices.AlsoDenmarkandGermanyhavesignificantly higherpricesthanSwedenintheprotectedmarketsegment.Theresidualcountrieshaveonly marginallyhigherwholesaleprices.

InthenonͲprotectedmarketsegmentthepictureisdifferent.Inthissegment,theSwedishpricelevel ismoreattheaverage.UK,Belgium,NorwayandtheNetherlandshavelowerprices.However,the costsavingsfromimportingtheUKpricelevelissmallerthanintheprotectedsegment.Irelandand GermanyareparticularlyexpensiveinthenonͲprotectedsegment.ImportingtheIrishpricelevel wouldresultina51percentincreaseinthepharmaceuticalexpendituresinSwedenassumingthat theconsumptionpatternisunchanged.

Theproblemwithcomparingpricesofidenticalpacksis,asmentionedbefore,thatthesample becomessmall,potentiallybiasedandthusnonͲrepresentativeforSwedenandthereference countries.Thetablebelowreportsthepopulationofpacksineachcountryandthenumberof identicalpacksthatcanbematchedwithSweden.

15

Table4.1.NumberofpacksineachcountryandnumberofidenticalpackswithSweden.

Allsubstances Protected NonͲprotected

Population Matched Population Matched Population Matched

Sweden 791 Ͳ 566 Ͳ 225 Ͳ

Norway 630 438 377 263 253 97

Denmark 732 493 503 364 229 116

Finland 664 448 363 274 301 85

UK 661 313 436 248 225 63

Germany 1200 434 750 337 450 82

Netherlands 807 337 529 262 278 64

Belgium 516 282 340 225 176 50

Austria 606 288 383 229 223 51

Ireland 461 272 311 213 150 48

Thetableshowsthatthesampleofmatchingpacksissignificantlylowerthanthetotalnumber (population)ofpacksinSwedenandthereferencecountries.Thebiggestreductioninthesample occurswhenwematchpackswithGermany.Thenumberofmatchingpacksbetweenthesetwo countriesis434,whereasGermanyandSwedenhaveapopulationofpacksequalto1200and791, respectively.Notice,however,thatthelossofobservationsisnotaslargeasitappears,sincethere mightbeseveralsuppliers(brandͲnameproducers,paralleltraders,genericproducers)ofthesame packineachcountry.OurprocedureofcomputingthevolumeͲweightedaveragepackpricewhen thereareseveralidenticalpacksinagivencountryimpliesthattheactuallossofinformationis somewhatlower.However,theproblemisstillsignificantasshowninthetable.Inthefollowing sections,wethereforefocusthepriceindicesbasedonthe(volumeͲweighted)averagedosepricesat substancelevel.

4.2 Priceindicesbasedonsubstance(dose)prices

Theaveragedosepricespersubstancearecomputedusingallsalesinformationineachcountry.

Sincewehaveinformationaboutthedosepriceandthenumberofdosessoldofeachpack,wecan computethevolumeͲweightedaveragedosepriceforeachsubstanceineachcountry.Wethen comparethesesubstance(dose)pricesacrosscountriesforthesetofmatchingsubstancesand constructpriceindicesusingtheSwedishquantityweights.Thequantityweightsaresimplythe numberofdosessoldofagivensubstancerelativetothetotalnumberofdosessoldofallmatching substances.Sincethesampleofmatchingsubstancesvariesacrosscountrieswhenwecompute bilateralpriceindices,therewillbeauniquesetofquantityweightsforeachcomparison.

16

ThefullsetofbilateralpriceindicesbasedonvolumeͲweightedaveragesubstance(dose)pricesis reportedinTableA.3intheAppendix.Here,wewillfocusontheoneswefindmostrelevant.Letus firstconsiderthebilateralpriceindicesforallmatchingsubstancesirrespectiveofwhethertheyare protectedornot.Inthefigurebelowwereportthesepriceindicesatwholesaleandretaillevel.

Figure4.3.Bilateralpriceindicesbasedonaveragesubstancepricesatretailandwholesalelevelfor allsubstances.

ThefigureshowsthattheSwedishpricelevelisonaverageforthisgroupofcountries.Germanyis themostexpensivecountry.Importingpricesatwholesalelevelwouldresultina29percentincrease intheSwedishpharmaceuticalexpendituresassumingtheconsumptionisunchanged.Thecheapest countryisUKwith19percentlowerpricesthaninSweden.Ifweconsiderretail(pharmacy)prices, theSwedishpricelevelbecomesslightlymorefavorable.Thisreflectsthatthepharmacymarginsin SwedenarelowerthaninthereferencecountriesexceptforUK.AustriaandparticularlyFinland appeartohavesubstantialpharmacymarginsastheirpriceindicesatretaillevelbecomemuch higherrelativetoSweden.

LetusnowdividethesampleintoprotectedandthenonͲprotectedsubstancesandcompareprices ofthematchingsubstances.Notethatwecomparepriceofsubstanceswiththesamepatentstatus inSwedenandthereferencecountries,implyingthatweexcludesubstancesthatareprotectedin

17

Swedenbutnotinthereferencecountry,andviceversa.Thefigurebelowshowsthebilateralprice indicesatwholesalelevelforthesetwomarketsegments.Therankingofcountriesisbasedonthe priceindicesfortheprotectedsegment,whichisalsothelargestmarketsegment.

Figure4.4.Bilateralpriceindicesbasedonaveragesubstancepricesatwholesalelevelforprotected andnonͲprotectedsubstances.

ThefigureshowsthatSwedenhasaslightlybelowaveragepricelevelfortheprotectedsubstances.

Germanyisthemostexpensivecountrywitha25percenthigherwholesalepricelevelthanSweden.

AlsoIrelandandDenmarkhavesignificantlyhigherwholesalepricesthanSwedenintheprotected segment.UKisthecheapestcountry.ImportingtheUKwholesalelevelwouldresultina19percent reductionintheSwedishpharmaceuticalexpendituresintheprotectedsegmentassumingthe consumptionisunchanged.TheNetherlandandNorwayalsohavelowerwholesalepricesthan Sweden,whereasFinland,AustriaandBelgiumhaveaboutthesamepricelevel.

ThepictureissomewhatdifferentforthenonͲprotectedsubstances.Inthissegment,theSwedish wholesalepricelevelisaboveaverage.GermanyandDenmarkarethetwomostexpensivecountries inthemarketsegmentwith,respectively,32and27percenthigherpricesthanSweden.Irelandis alsoslightlymoreexpensive.However,therestofthecountrieshavealowerwholesalepricelevelin thenonͲprotectedmarketsegment.UKisthecheapestcountry.ImportingtheUKpricelevelwould resultinan18percentreductionintheSwedishexpendituresonnonͲprotectedsubstances.

18

4.3 Genericcompetition

OursampleisdefinedbythesetofsubstanceswithoutgenericsalesinSweden.However,asshown inTable3.1,someofthesesubstancesdohavegenericsalesinthereferencecountries.Apotential issueisthatgenericcompetitionmightdrivedownthepriceindicesforthereferencecountries,so thatSwedenseemstobemoreexpensivethantheyactuallyarewhenaccountingforgeneric competition.Notice,however,thatthepriceindicesfortheprotectedsegmentdonotinclude substanceswithgenericsalesinthereferencecountriesaswematchonlysubstancesthatare protectedinbothcountries.Tolookmorecarefullyattheissueofgenericcompetition,wehave computedtodifferentsetsofpriceindices:(i)priceindicesforsubstanceswithoutgenericsale;(ii) priceindicesfornonͲprotectedsubstanceswithoutgenericsale.Inthefigurebelowwereport bilateralpriceindicesatwholesalelevelforallsubstancesirrespectiveofwhetherthereisgeneric competitionornotandbilateralpriceindicesforthesetofsubstancesthatdonothavegenericsales.

Figure4.5:Bilateralpriceindicesbasedonaveragesubstancepricesatwholesalelevelforall matchingsubstancesandmatchingsubstanceswithoutgenericcompetition.

ThefigureshowsthatgenericcompetitioninthereferencecountriesforsomenonͲprotected substancesdoesnotmattermuchforthepriceindicescomputedforallmatchingsubstances.

However,thereisatendencyforSwedenbecomingslightlycheaperwhenexcludingthesubstances inthereferencecountriesthathavegenericsales,aswewouldexpect,buttheimpactisvery

19

moderate.TheonlyexceptionisFinland,whichis11percentcheaperthanSwedenatwholesalelevel forallsubstances,butonly3percentcheaperthanSwedenwhenweexcludethesubstanceswith genericcompetition.

WedothesameexerciseforthenonͲprotectedsegment.InTableA.3wereportthebilateralprice indicesforthenonͲprotectedsubstancesthatdonothavegenericsale.

Figure4.6.BilateralpriceindicesbasedonaveragesubstancepricesatwholesalelevelfornonͲ protectedsubstanceswithandwithoutgenericsales.

ThefigureshowsmostlythesamepicturefornonͲprotectedsubstancesasforallsubstances.When excludingsubstanceswithgenericsaleinthereferencecountries,Swedenbecomescheaperrelative tothereferencecountries.Formostcountriestheeffectissmall,butforsomecountriesgeneric salesseemtomatterforthepricelevels.Inparticular,DenmarkandGermanybecome,respectively, 12and7percentmoreexpensiverelativetoSwedenwhenexcludingthesubstanceswithgeneric saleinthesecountries.Thus,genericcompetitionplaysaroleforthepriceindices,buttheimpactis verylowinmostcasesanddoesnotgenerateanysignificantupwarddistortionontheSwedishprice level.

20

4.4 Parallelimport

IntheEUparalleltradeisencouragedamongmembercountriesalsoforpharmaceuticals.Parallel importersconstituteacompetitivethreatforthebrandͲnameproducers.InTable3.1weshowed thattheextentofparallelimportvariesquitealotamongthecountriesinoursample.Germanyhas thehighestnumber(122)ofsubstanceswithparallelimportandthehighestnumberofparallel importedpacks(759),whereasIrelandhasnoparallelimportsatallforthesubstancesinoursample.

InSweden59of153substanceshasparallelimport,and170of791packsareparallelimported.The figurebelowcomparesthebilateralpriceindicesatwholesalelevelforallsubstancesandforthe substanceswithparallelimportsinSweden.Wedonotrestrictthecomparisontomatching substanceswithparallelimports,sincethisnumberistoolow.Thefiguresare,however,reportedin TableA.3intheAppendix.

Figure4.7:Bilateralpriceindicesbasedonsubstancepricesatwholesalelevelforallsubstancesand substanceswithparallelimportsinSweden.

Thefigureshowsaclearpattern.Swedenbecomesrelativelymoreexpensiveforthesetof substancesthathaveparallelimport.Thisisperhapsasexpected,sinceparallelimportisprofitable onlyifthepricelevelintheimportingcountryishigherthanintheexportingcountries.Thus,we wouldexpecttoobserveparallelimportforsubstanceswithrelativelyhighpricesinSweden.Thisis probablyalsothereasonwhythehighͲpricecountryGermanyhasalotofparallelimports.Onthe otherhand,onewouldexpectthatdefactoparallelimportwouldleadtolowerpricesdueto

21

competitionwiththeoriginalbrandͲnameproducer.Thisis,however,notconsistentwiththeresults fromthepriceindicesabove.

4.5 ComparisonwiththeNorwegianstudy

AninterestingquestioniswhethertheuseofSwedishconsumptionweightsinfluencestheprice indicesinanysignificantway.Itisusuallyarguedthateachcountrybecomesrelativelycheaperwhen beingthebasecountry.WethereforecomparetheresultsfromthisstudywiththeNorwegianstudy byBHS(2011).Wefocusontheglobalpriceindicesforsubstanceswithoutgenericsale(inall countries),sincethesetofmatchingsubstancesisthesameinbothstudiesfortheseindices.The globalpriceindicesarereportedintableA.3inthisreportandtable4.3inBHS(2011).Wemodifythe priceindicesinBHS(2011)byassumingSwedentobethebasecountry.Forillustrativepurposeswe assumethebaseindextobezero(ratherthan100).Thefigurebelowshowsthecomparisonofthe priceindicesusingSwedishandNorwegianconsumptionweights.

Figure4.8.Globalpriceindices,averagedoseprices,wholesalelevel,substanceswithoutgeneric salesinallcountries(N=73),NorwegianandSwedishweights.

Asexpected,weseethatthepriceindicesaremorefavorableforSwedenwhenusingSwedish consumptionweights.ThehighͲpricecountriesbecomemoreexpensive,whereasthelowͲprice countriesbecomelesscheaprelativetoSweden.Thesamepictureispresentatretaillevelaswell.

22

ThegainfromimportingtheUKwholesalepricelevelisreducedfrom19.7to14.4percent,whereas thecostofimportingtheGermanpricelevelisincreasedfrom27.1to31.7percent.Thisshowsthat onecannotsimplyusepriceindicesbasedonothercountriesconsumptionweightstoinfercost savingsinownhomecountry.Infact,itmightbereasonabletoassumethatthecostsavings (increases)generallyarelower(higher)usingconsumptionweightsfromhomecountry,asillustrated inthefigureabove.Theargumentisthateachcountrytendstoconsumemoreofpharmaceuticals thatarerelativelycheap.Whetherthisisduetoconsumerbehaviororregulationisanopen question.

WecouldhavecomparedthebilateralpriceindicesinthisstudywiththeNorwegianstudybyBHS (2011).Ifwedoso,theSwedishpricelevelwouldbereducedfrom12to6percenthigherthanthe Norwegianpricelevelatwholesalelevel.Theproblemis,however,thatthesetofsubstancesforthe residualcountriesvaryacrossthestudiesfortheseindices.Thesetofmatchingsubstancesbetween NorwayandthereferencecountriesisnotthesameasbetweenSwedenandthereference countries,exceptforbetweenNorwayandSwedenofcourse.Thismeansthatdifferencesinprice indicesinthetwostudiesmightbeduetodifferentsamplesofproductsratherthandifferent consumptionweights.

4.6 Priceindicesfornewandoldproducts

Finally,wehaveaddedinformationabouttheintroductiondatesontheSwedishmarketforthe substancesinoursample.Thisinformationcangiveusanideaofhowthepriceshavedevelopedin thedifferentcountriesovertime.Inparticular,wecanstudythepricelevelsofnewandold substances.WesplitthesampleintosubstanceslaunchedontheSwedishmarketbefore1990, between1990and2002,andafter2002until2010.Intheperiodbefore1990,thegrowthinthe pharmaceuticalexpenditureswasfairlystableandmoderate,andApoteketAB(thestateͲowned wholesaler)hadaroleinthepricesetting.Intheperiodbetween1990and2002severalblockbusters wereintroducedontheSwedishmarket,resultinginescalatingpharmaceuticalexpenditures.Inthis periodtheresponsibilityforpricesettingofpharmaceuticalswastransferredtothesocialinsurance body called Riksförsäkringsverket (RFV). In 2002 the responsibility for price setting of pharmaceuticalswastransferredtoanewregulatorybody,whichnowiscalledTandvårdsͲoch Läkemedelsförmånsverket (TLV). At the same time Sweden introduced reference pricing and mandatorygenericsubstitutionatpharmacylevel.

23

Thefigurebelowshowsthepriceindicesforthesethreeperiods.Wehaverankedthecountriesfrom lowesttohighestpricesinthelastperiod(after2002).Thefullsetofpriceindicescanbefoundin tableA.5intheAppendix.

Figure4.9.Bilateralpriceindicesforaveragesubstancepricesatwholesalelevelforallsubstances dependingontheintroductiondateinSweden.

Thefigureshowsquitesomevariationinthepricedifferencesovertime.Ifwecomparetheold substances(before1990)withthenewsubstances(after2002),weseethatsomecountries (Germany,Ireland,Denmark,andUK)haveahigherpricelevelthanSwedenontheoldsubstances thanonthenewsubstances.UK,whichisthecheapestcountry,isactually10percentmore expensivethanSwedenontheoldsubstances,but11percentcheaperthanSwedenonthenew substances.Denmarkis53percentmoreexpensivethanSwedenontheoldsubstances,whereasthe pricedifferenceisonly14percentonthenewsubstances.Fortherestofthecountries,thefigures arereversed,thoughthechangesaremoremoderate.Austriahas2percenthigherpricesthan Swedenonoldsubstances,but9percenthigherpricesonnewsubstances.Similarfiguresapplyto Belgium,Finland,theNetherlands,andNorway.

24

Ifwecomparethepricedifferencesofsubstanceslaunchedbetween1990and2002withthemore recentsubstanceslaunchedafter2002,weseeatendencytoSwedenbecominglessexpensivethan thereferencecountries.Forinstance,UKis24percentcheaperthanSwedenonthesubstances introducedbetween1990and2002,butonly11percentcheaperonthesubstancesintroducedafter 2002.ThesamefiguresapplytoNorway,theNetherlands,Finland,Belgium,Austria,andIreland.The exceptionsareDenmarkandGermany,whichtendstohaveslightlylowerpricesonthemorerecent substances.Thus,thereisatendencytoSwedenbeingmoreexpensiveontheoldsubstances launchedbefore1990,butlessexpensiveonthemorerecentsubstanceslaunchedafter2002.

25

Chapter5 ConclusionsandRemarks

WehaveanalyzedthepricelevelofpharmaceuticalswithoutgenericsaleinSwedenrelativetothe followingnineEuropeancountries:Austria,Belgium,Denmark,Finland,Germany,Ireland,the Netherlands,Norway,andUK.UsingdetailedproductͲleveldataonpricesandvolumesforall prescriptionͲboundsalesforalargesetofsubstanceoverthefirstsixmonthsin2010,wehave computedawidesetofdifferentpriceindices,andderivedthefollowingsetoffindings.

First,theSwedishpricelevelisslightlybelowaveragecomparedwiththereferencecountries consideringallmatchingsubstances.Germany,IrelandandDenmarktendtohavethehighestprices, whereasUK,NorwayandtheNetherlandstendtohavethelowestprices.Thepriceindicesatretail levelaremorefavorableforSwedenduetolowpharmacymargins.Second,theSwedishpricelevelis moreinthehigherͲendwhenconsideringthenonͲprotectedsubstances.Weshowthatthisispartly duetothepresenceofgenericcompetitioninsomeofthereferencecountries,thoughthiscannot accountforthefullpricedifference.Third,consideringthesubstanceswithparallelimportin Sweden,wefindthattheSwedishpricelevelislessfavorable.Thisislikelytobeduetothefactthat parallelimportismoreprofitableforproductswithrelativelyhighprices.

Fourth,wecomparethepriceindicesinthisstudywiththeBHS(2011)studythatusesthesamedata set,butcomputesthepriceindicesusingNorwegianratherthanSwedishconsumptionweights.As expected,wefindthatSwedenbecomesrelativelycheaperwhenbeingthebasecountry.Forthe bilateralpriceindices,theBHSstudyfindthatNorwayhas12percentlowerwholesaleprices, whereasinthecurrentstudywefindthatSwedenhasonly6percenthigherwholesaleprices.Both figuresarecorrect,butshowthatonecannotinfertheinversepricedifference,andthatthereisa tendencyforeachcountrytobecomerelativelycheaperwhenbeingthebasecountry.

Finally,wesplitthesampleintothreegroupsdependingontheirlaunchingdateontheSwedish market.WefindthatSwedentendstohavehigherpricesontheoldsubstancesintroducedbefore 1990,butrelativelylowerpricesonthemorerecentsubstanceslaunchedafter2002.

26

Referencelist

Brekke,K.R.,Holmås,T.H.,Straume,O.R.,2009.ArepharmaceuticalsinexpensiveinNorway?A comparisonofpricesofprescriptiondrugsintenEuropeancountries.Reportno.06/09,Institutefor ResearchinEconomicsandBusiness(SNF),Bergen.

Brekke,K.R.,Holmås,T.H.,Straume,O.R.,2010.ArepharmaceuticalsstillinexpensiveinNorway?A comparisonofpricesofprescriptiondrugsintenEuropeancountries.Reportno.08/10,Institutefor ResearchinEconomicsandBusiness(SNF),Bergen.

Brekke,K.R.,Holmås,T.H.,Straume,O.R.,2011.ComparingpharmaceuticalpricesinEurope.Report no.11/11,InstituteforResearchinEconomicsandBusiness(SNF),Bergen.

Danzon,P.M.,1999.PriceComparisonsforPharmaceuticals.AReviewofU.S.andCrossͲNational Studies.AmericanEnterpriseInstituteforPublicPolicyResearch.

DanzonPM,ChaoLͲW,2000.CrossͲnationalpricedifferencesforpharmaceuticals:howlarge,and why?JournalofHealthEconomics19,159Ͳ195.

27

APPENDIX

TableA.1.Listofsubstanceswithinformationonrang,genericsales,patentprotection,and parallelimport

Molecule Rang:Sales

inSweden

Sale Generics.

Numberof countries

Protected Sweden

Protected.

Numberof countries

PͲimport Sweden

PͲimport.

Numberof countries

ETANERCEPT 1 0 1 8 0 2

ADALIMUMAB 2 0 1 8 0 1

BUDESONIDE,FORMOTEROL 3 1 0 1 1 4

OCTOCOGALFA 4 0 1 7 0 1

ATORVASTATIN 5 1 1 7 1 6

OLANZAPINE 6 2 1 6 1 5

CANDESARTANCILEXETIL 7 0 1 9 1 4

PREGABALIN 8 0 1 9 1 6

ESOMEPRAZOLE 9 4 0 1 1 3

TIOTROPIUMBROMIDE 10 0 1 9 0 4

FLUTICASONE,SALMETEROL 11 0 1 8 1 6

INSULINGLARGINE 12 0 1 9 1 4

INSULINASPART 13 0 1 9 0 3

INSULINASPARTPROTAMINECRYSTALLINE 14 0 1 9 0 4

DALTEPARINSODIUM 15 0 0 0 0 4

DARBEPOETINALFA 16 0 1 8 0 2

RANIBIZUMAB 17 0 1 7 0 2

QUETIAPINE 18 3 1 6 1 5

DONEPEZIL 19 3 1 5 1 4

SILDENAFIL 20 1 1 8 0 5

ESCITALOPRAM 21 3 1 5 1 5

MOROCTOCOGALFA 22 0 1 8 0 0

INSULINHUMANISOPHANE 23 0 1 9 0 4

LATANOPROST 24 1 1 8 1 3

ARIPIPRAZOLE 25 0 1 9 1 5

DULOXETINE 26 0 1 9 1 4

EPTACOGALFA 27 0 1 4 0 0

PRAMIPEXOLE 28 1 0 0 0 4

TOLTERODINE 29 0 1 8 0 4

EPOETINBETA 30 0 1 8 0 2

TADALAFIL 31 0 1 9 0 5

CANDESARTANCILEXETIL,HYDROCHLOROTHIAZIDE 32 0 1 8 1 3

MONTELUKAST 33 2 1 6 0 4

TERBUTALINE 34 4 0 0 0 4

LEVETIRACETAM 35 0 0 0 1 4

WARFARIN 36 3 0 0 0 0

DESLORATADINE 37 0 1 9 0 5

DESOGESTREL 38 0 0 8 0 4

INSULINLISPRO 39 0 1 9 0 3

EZETIMIBE 40 0 1 9 0 5

ZOLMITRIPTAN 41 0 1 9 1 6

BOSENTAN 42 0 1 7 0 1

INSULINDETEMIR 43 0 1 9 0 4

MEMANTINE 44 0 1 9 1 4

GALANTAMINE 45 0 1 8 1 4

ATOMOXETINE 46 0 1 9 0 3

SOLIFENACIN 47 0 1 9 0 4

FOLLITROPINALFA 48 0 1 9 1 3

RIVASTIGMINE 49 0 1 8 1 6

FOLLITROPINBETA 50 0 1 9 1 3

BETAMETHASONE,CALCIPOTRIOL 51 0 1 9 1 4

PEGINTERFERONALFAͲ2A 52 0 1 8 0 2

LEVONORGESTREL 53 1 0 3 0 3

IRBESARTAN 54 0 1 7 1 4