搜索
Draft-ConfidentialTheProductivityofHealthCareandPharmaceuticals:QualityofLife,CauseofDeathandtheRoleofObesity*July24,2002RichardD.Miller,Jr.CenterforNavalAnalysesandH.E.Frech,IIIEconomicsDepartmentUniversityofCalifornia,SantaBarbara*EarlierversionswerepresentedinseminarsatCurtinUniversityinPerthAustralia,June21,2001andattheInternationalHealthEconomicsAssociationmeetingsinYork,England,July24,2001.Thanksareduetotheparticipantsofthoseseminars,especiallyHarryBloch,GavinMooney,JeffPetchy,andSandraHopkins.Thisworkextendsourearlierresearch,whichwassupportedbyagrantfromSciencesPodeParis.WearegratefulforsupportforthisresearchfromPfizer,Merck,AztraZeneca,andBristolMeyers-Squibb.2I.IntroductionandSummaryInanearlierstudy,sponsoredbySciencesPodeParis,weuseddatafromtheOrganizationforEconomicCooperationandDevelopment(OECD)tostudytheeffectsofpercapitaconsumptionofpharmaceuticalsandotherhealthcareonobjective,butcrude,measuresofhealth(FrechandMiller1999;MillerandFrech2000).Ourcrudemeasuresofhealthwerelifeexpectanciesatbirth,atage40,andatage60,andinfantmortality.Wefoundthatincreasedpharmaceuticalconsumptionisbotheconomicallyandstatisticallysignificantlyrelatedtoincreasedlifeexpectancyattheagesof40and60,evenwhencontrollingforpercapitaincomeandlifestylefactorssuchassmoking,alcoholconsumption,anddiet.Incontrast,wefoundnoeconomicallyorstatisticallysignificanteffectofnon-pharmaceuticalhealthcareconsumption.Ouranalysisofinfantmortalitywasnotrobust.Amongotherproblems,infantmortalityisnotconsistentlymeasured,notevenintherichcountries.Ourearlierstudyhasraisedasmanyquestionsasithasanswered.Inthisanalysis,weextendthatearlierworktoanswersomeofthesequestions.Wealsoreplicatetheoriginalworkwithlaterdataandaslightlyimprovedmodelthatallowsustotakeaccountofthelargeinternationalvariationinobesity.Wegobeyondsimplemortalitytoincludethequalityoflifeasameasureofhealth.Atthetimeofourearlierwork,therewerenodataonthis.Butnow,thankstotheworkoftheWorldHealthOrganization(WHO),thereareconsistentdataondisability-adjustedlifeexpectancy(DALE)formanycountries,includingalltheOECDcountries.Inanaturalextensionoftheearlywork,weexaminetheproductionofDALE.WefindthatpharmaceuticalconsumptionismorepowerfulinextendingDALEthanlife3expectancy.Wealsofindthattheresultsonlifeexpectancyareevenmorerobustandstrongerthanintheoriginalresearch.Wealsoextendtheoriginalworkinadifferentdirectionbyinvestigatinghowtheeffectsofpharmaceuticalconsumptionvarybydisease(causeofdeath).Here,wecannotlookatqualityoflife.Wefindthattheproductivityofpharmaceuticalconsumptionvariesgreatlybybothcauseofdeathandbyage.Forindividualsunder70yearsofage,pharmaceuticalconsumptionisveryhelpfulincirculatorydisease,buthaslittleeffectoncancerorrespiratorydisease.Atlaterages,pharmaceuticalconsumptionisgenerallyproductive.II.LiteratureReviewTherehavebeenmanyinternational,cross-countrystudiesofhealthcare.Mostlydrivenbybudgetaryconcerns,theyhavelargelyfocusedonthedeterminantsofhealthcareexpenditures.1Lessattentionhasbeenplacedonestimatingthedeterminantsofhealthitself.Scholarsfromanumberofdisciplineshavedoneworkinthisareaandtheresultshavegenerallybeenmixed.Wehavealreadyanalyzedmuchofthisliteratureinourearlierwork,especiallyFrechandMiller(1999,pp.2-21).Inthissectionwesummarizethemainconclusionsfromthatanalysis,thenreviewourownwork,andfinallyreviewotherstudies,particularlytheneweronesthataremostrelevant.ConclusionsfromourEarlierLiteratureReview.1See,forexample,JosephNewhouse(1977);DavidParkin,AllistarMcGuireandBrianYule(1987);Ulf-GGerdthamandB.Jonsson(1992)andUlf-GGerdtham(1991).4Certainoverallresultsfromearlierstudiesstandoutasrobustandsensible.Instudiesincludingpoorcountries,publichealthinfrastructuresuchaspotablewatersuppliesandsanitationarethemostpowerfuldeterminantsofhealth.Expansionofhealthcareserviceshasmuchlesseffect.Percapitaincomeandeducationarealsomorepowerfulthanmedicalcare.Dietandotherlifestylevariables,suchastobaccoandalcoholconsumption,havealsobeenfoundtobeimportant.Thesmallimpactofmedicalcarehasledtotheconventionalwisdominhealtheconomicsthatmedicalcaredoeslittletoimprovelifeexpectancy.But,afewgoodstudieshavefoundmedicalcaretomatter,includingacloselyrelatedoneusingOECDdata(ZweifelandFerrari1992).TheclosestprecursortoourearlierstudywasbyAkiraBabazonaandAlanHillman(1994).Unusualforitstime,itdisaggregatedhealthcaretoexamineeffectsonperinatalandinfantmortalityandlifeexpectancy.Incontrasttoourstudy,thisonefoundnoeffectofpharmaceuticalconsumptiononhealthmeasures.But,thisstudywasmarredbytheuseofinappropriatefunctionalformandseriouslyincorrectpharmaceuticalprices,aswasshownbyourownwork.Specifically,BabazonaandHillmanconverteddomesticpharmaceuticalprices,expressedindomesticcurrency,toacommoncurrencybyusingtheoverallpurchasingpowerparityexchangerate.Thiswouldbefineifrealpharmaceuticalpriceswerethesameinallmarkets,but,aswehaveshown,thisisfarfromthecase.Inamorerecentpaper,AndersAnellandMichaelWillis(2000,p.772)makethesamemistake,justifyingitbytheargumentthatthereisaworldmarketinpharmaceuticals.Whilethisistrue,cross