|

 

The welfare effects of innovative pharmaceuticals

[Save eerst dit bestand als XXX(titel).doc] The welfare effects A pilot study for the Netherlands Apostolos Tsiachristas
Research commissioned by the American Chamber of Commerce Pharma-
ceutical Committee
Aarts De Jong Wilms Goudriaan Public Economics bv (APE) and
Maastricht University
Den Haag, January 2008

The welfare effects of innovative pharmaceuticals: A pilot study for the
Netherlands
Apostolos Tsiachristas, René Goudriaan & Wim Groot
Ape rapport nr. 511
2008 Aarts De Jong Wilms Goudriaan Public Economics bv (APE) and
Maastricht University Website: Cover: Brordus Bunder, Amsterdam Niets uit deze uitgave mag worden verveelvoudigd en/of openbaar gemaakt via druk, fotoko- pie of op welke andere wijze dan ook, zonder voorafgaande schriftelijke toestemming. TABLE OF CONTENTS
EXECUTIVE SUMMARY 1.2 Research aim and objectives 1.4 Overview of the paper 2 INTERNATIONAL POLICIES TOWARDS NEW DRUGS 2.1 Guidelines and thresholds 2.1.1 United Kingdom 2.1.3 New Zealand 2.1.6 The Netherlands 2.1.7 Summary of guidelines and implied thresholds 2.2 Institutional controls and regulations 2.2.1 United Kingdom 2.2.3 New Zealand 2.2.6 The Netherlands 2.2.7 Summary of institutional controls and regulations 3 WELFARE EFFECTS: EMPIRICAL EVIDENCE AND THEORETICAL CONCERNS 3.1 The welfare impact of pharmaceuticals 3.2 Issues of concern 4 STUDY DESIGN AND RESEARCH PREPARATION 4.1 Research steps Ape
4.2 Selection of studies and data collection 4.2.2 Policy orientation 4.2.3 Generalizability and international comparability 4.2.5 Disease variety 4.2.6 Producers variability 4.2.7 Economic importance 4.3 Data analysis 4.3.1 Adjustment of cost-effectiveness data 4.3.2 The formula used in the estimation of welfare gains 4.3.3 Value of a QALY used in our benchmark scenario 4.3.4 Other parameters taken into account 4.4 Alternative scenarios 5 INTERNATIONAL COMPARISONS AND ESTIMATES OF THE WELFARE GAINS 5.1 International comparison of input parameters 5.1.1 Discount rates 5.1.2 Time horizons 5.1.3 Costs included 5.1.4 Comparative medicine 5.2 Welfare gains in the Netherlands 6 INSTITUTIONAL BARRIERS 6.1 Introduction 6.2 Boundaries in the financing of the different sectors 6.3 Authorization and reimbursement 6.4 Imperfect incentives for health insurers 7 DISCUSSION AND CONCLUSIONS 7.2.1 International differences 7.2.2 Welfare effects in the Netherlands 7.2.3 Institutional barriers 7.3 Limitations of the study 7.4 Suggestions for future research APPENDIX A: LIST OF DEFINITIONS Ape
APPENDIX B: RESULTS FROM LITERATURE REVIEW APPENDIX C: STATISTICS OF DISCOUNTING RATES FOR EFFCTS AND COSTS APPENDIX D: STATISTICS OF TIME-HORIZONS APPENDIX E: WELFARE GAINS RESULTS PER SAMPLE APPENDIX F: DIFFERENCES CONCERNING COMMON DRUGS APPENDIX G: PRICE REDUCTIONS AND INTRODUCTION DELAYS Ape
VOORWOORD
Tot onze verbazing is de over de omvang van de welvaartswinsten van in-novatieve geneesmiddelen tot nu toe vrijwel niets bekend. De vaklitera-tuur biedt nauwelijks aangrijpingspunten over de kwantitatieve betekenis van de welvaartswinsten van geneesmiddelen; dit in tegenstelling tot veel andere sectoren van de economie. APE en de Universiteit Maastricht stel-len het zeer op prijs dat het Pharmaceutical Committee van de Amerikaan-se Kamer van Koophandel ons financieel in staat heeft gesteld om de wel-vaartseffecten van innovatieve geneesmiddelen te onderzoeken. De voorliggende studie heeft nadrukkelijk het karakter van een pilot study, die vraagt om verder onderzoek. In het onderzoek is gebruik gemaakt van de resultaten van farma-economische studies, die in de meeste gevallen door overheidsinstanties of adviesorganen zijn gepubliceerd. Het gebruik hiervan houdt geen kwaliteitsoordeel over de betreffende studies in. Ge-zien het explorerende karakter van het onderzoek en uit oogpunt van transparantie hebben wij ervoor gekozen om een uniforme methode te hanteren voor de bepaling van de welvaartseffecten van alle onderzochte geneesmiddelen. Het is mogelijk dat daardoor onvoldoende recht wordt gedaan aan eigenschappen van afzonderlijke geneesmiddelen. Daardoor kan in individuele gevallen sprake zijn van een onder- of overschatting van de welvaartseffecten. Veel personen hebben belangrijke bijdrage geleverd aan het project. In de eerste plaats fungeerde tijdens het onderzoek een begeleidingscommissie bestaande uit: Robert Berkelbach van der Sprenkel (MSD), Werner Brou-wer (Erasmus Universiteit Rotterdam), Martin van der Graaff (Nefarma), Willem Jan Meerding (Pfizer) en Marc Pomp (Centraal Planbureau). Bij de uitvoering van het project hebben wij dankbaar gebruik gemaakt van hun commentaren op eerdere concepten van het rapport. In de tweede plaats heeft een groot aantal personen ons van gegevens of advies voorzien. Het betreft: Stan van Belkum (CBG), Kim Blaas (Wyeth), Michel Boucher (CADTH, Canada), Jolanda Crombach (Wyeth), Gepke Delwel (CVZ), Fred van Doorn (MSD), Dirk-Jan Griffioen (RIVM), Rachel Grocott (PHARMAC, Australië), Geke Hage (Z-Index), Jacqueline Janssen (SFK), Anne-Janine Janszen (Eli Lilly), Kathleen Heppell (NICE, Verenigd Koninkrijk), Bart van der Lelie (Schering-Plough), Ulf Persson (IHE, Zweden), Hans Piepenbrink (CVZ), Joakim Ramsberg (LFN, Zweden), Carla Vos (MSD), Aglae Weijers (Bristol-Myers Squibb) en Ilse van Woudenberg, (Janssen-Cilag). Ape
Wij danken alle genoemde personen en de opdrachtgever voor hun bijdra-ge aan het project. De verantwoordelijkheid voor het onderzoek berust echter bij APE en Universiteit Maastricht. De auteurs Ape
Aanleiding In de beleidsdiscussie over nieuwe geneesmiddelen ligt de nadruk vaak op kostenbeheersing. De baten van nieuwe geneesmiddelen, zoals gezond-heidswinst, een betere kwaliteit van leven en besparingen op de kosten in andere delen van de zorg of de sociale zekerheid, krijgen in deze discussie meestal weinig of geen aandacht. Zowel in Nederland als in het buitenland zijn weinig markten zo streng gereguleerd als de geneesmiddelenmarkt. Onderzoeksobject en gegevens In de voorliggende studie doen we een eerste poging om meer inzicht te krijgen in de welvaartswinsten van innovatieve geneesmiddelen, die na 1997 op de Nederlandse markt zijn toegelaten. De welvaartswinsten zijn gedefinieerd als het verschil tussen de waardering van de voor kwaliteit gecorrigeerde gewonnen levensjaren (QALY's) en de daarmee gemoeide kosten. Over de omvang van de welvaartswinsten van innovatieve ge-neesmiddelen is vrijwel niets bekend. Voor de bepaling van de welvaartswinsten maken wij – waar mogelijk - gebruik van farmaco-economische studies die door overheidsinstanties of adviesorganen zijn gepubliceerd. De studies zijn afkomstig van het College voor Zorgverzekeringen (CVZ) en de Engelse en Canadese tegenhangers (respectievelijk NICE en CADTH). Deze zijn aangevuld met een beperkt aantal studies uit de literatuur om het scala van behandelde aandoeningen te verbreden. Het gebruik van de 52 farmaco-economische studies houdt geen kwaliteitsoordeel over de betreffende studies in. Aanpak We hebben de resultaten van de buitenlandse farmaco-economische stu-dies zo goed mogelijk ‘vertaald' naar de Nederlandse situatie. Gezien de uiteenlopende uitgangspunten van de verschillende farmaco-economische studies is het echter ondoenlijk om een exacte vertaling te maken. Daar-door zijn onze resultaten uiteraard met enige marges omgeven. Ons on-derzoek heeft nadrukkelijk het karakter van een verkennende studie. De geldelijke waardering van een QALY loopt uiteen tussen de verschillen-de landen. Naar internationale maatstaven is de in Nederland doorgaans gebruikte waarde van € 20.000 per QALY laag. Daarom hanteren we in het rapport drie scenario's voor de waardering van een QALY: € 20.000, € 50.000 en € 80.000. Het middenscenario vormt de benchmark, mede Ape
omdat de waardering van een QALY ruwweg overeenkomt met de interna-tionaal gebruikelijke waarde. Uit oogpunt van eenvoud en transparantie maken we voor alle geneesmiddelen gebruik van dezelfde waardering van een QALY. Resultaten Het gebruik van de 52 onderzochte innovatieve geneesmiddelen leidt in het midden- en het hoge scenario tot substantiële welvaartswinsten: res-pectievelijk € 1,7 mld. (0,3% van het BBP) en € 3,6 mld. (0,7% van het BBP). In het lage scenario zijn de welvaartswinsten negatief (€ -0,2 mld.). De welvaartswinsten van de gezamenlijke CVZ-studies zijn in alle drie sce-nario's gemiddeld positief. Alleen de resultaten op basis van de studies van NICE/CADTH en de literatuur pakken in het lage scenario negatief uit. Dat is waarschijnlijk mede het gevolg van de verschillen in kostentoerekening in de buitenlandse studies. Deze studies salderen de kosten niet met be-paalde opbrengsten in de vorm van productiviteitswinsten en een geringer beroep op de sociale zekerheid. In plaats daarvan hanteren zij een hogere waardering voor een QALY. Institutionele belemmeringen De welvaartswinsten zijn aanzienlijk, maar kunnen nog toenemen als insti-tutionele belemmeringen voor de efficiënte en effectieve inzet van innova-tieve geneesmiddelen worden weggenomen. De maatschappelijke kosten van de huidige regulering van de geneesmiddelenmarkt lijken groter dan de maatschappelijke baten. Vertragingen in de toelating van de 52 onder-zochte geneesmiddelen zijn bijvoorbeeld goed voor € 970 mln. aan gemis-te welvaartswinsten. De belangrijkste belemmeringen zijn: ƒ schotten in de financiering van extramurale farmaceutische zorg versus intramurale zorg en sociale zekerheid; ƒ het proces van toelating en vergoeding van nieuwe geneesmiddelen; ƒ verkeerde prikkels in de financiering van de zorgverzekeraars. Verder onderzoek Onze unieke pilot study is onvermijdelijk aan beperkingen onderhevig. Ten eerste berust de studie op gegevens van slechts 52 nieuwe geneesmidde-len, zodat het lastig is om te beoordelen wat de zeggingskracht is van de resultaten voor andere geneesmiddelen en andere aandoeningen. Ten tweede is de studie gebaseerd op heterogene farmaco-economische evalu-aties uit verschillende bronnen. Ten derde hebben we noodgedwongen een aantal ruwe vertaalslagen moeten maken om de buitenlandse studies ver-gelijkbaar te maken met de Nederlandse situatie. Ape
Deze beperkingen moeten worden weggenomen in verder onderzoek. Dat onderzoek moet zich naar onze mening richten op: 1. een verbreding van de set van onderzochte geneesmiddelen; 2. een splitsing van de onderzochte geneesmiddelen in verschillende aan- doeningengroepen om inzicht te krijgen in de verschillen tussen deze groepen en - waar noodzakelijk – verschillende waarderingen voor de betreffende QALY's te hanteren; 3. de opzet van een raamwerk waarmee de resultaten van de internatio- nale farmaco-economische studies beter kunnen worden vergeleken; 4. een analyse van de mogelijke systematische verschillen tussen farma- co-economische studies waarin de resultaten worden gerapporteerd in QALY's dan wel in gewonnen levensjaren. 5. een empirische vergelijking van de verschillende methoden voor de bepaling van de welvaartseffecten van nieuwe geneesmiddelen. Ape
EXECUTIVE SUMMARY
Motivation Policy discussions on new medicines are often focused on cost contain-ment. In general, little attention is paid to the benefits of new medicines, such as health gains, higher quality of life, and cost savings in other health care sectors or social security. In the Netherlands, as well as in other countries, few sectors are so heavily regulated as the pharmaceutical mar-ket. Research aim and data In the current study, we present a first attempt to gain some more insight in the welfare gains of innovative medicines introduced in the Dutch mar-ket after 1997. Welfare gains are defined as the difference between the estimated value of a quality adjusted life year (QALY) gained by innovative pharmaceuticals and their costs. Very little is known about the magnitude of the welfare gains of innovative pharmaceuticals. In the estimation of the welfare gains we have used - where possible - pharmaco-economic studies published by governmental institutions or quasi-autonomous non-governmental organizations. The studies have been collected by the Dutch Health Care Insurance Board (CVZ) and its British and Canadian counterparts (NICE and CADTH respectively). In or-der to broaden the spectrum of diseases, the sample has been augmented with a limited number of pharmaco-economic studies collected from the scientific literature. We have selected 52 studies. The selection does not imply any quality judgment over the methodology and the results of the studies concerned. Approach We have ‘translated' the results of the pharmaco-economic studies col-lected outside CVZ - as much as possible - to the Dutch situation. In view of the diverging assumptions of the different pharmaco-economic studies, an exact ‘translation' of the foreign studies is, however, infeasible. There-fore, our results contain some margin. Our research is primarily an ex-ploratory study. The monetary valuation of a QALY varies among countries. Compared to international standards, the €20,000 value of a QALY generally used in the Netherlands is relatively low. Therefore, in this report, we employ three scenarios with a different valuation of a QALY: € 20,000, € 50,000, and Ape
€ 80,000. The middle scenario is the benchmark scenario in our analysis, because it roughly corresponds to the valuation that is used internation-ally. For the sake of simplicity and transparency, we have used the same value of a QALY for all pharmaceutical products. Results The analysis of the 52 examined new pharmaceuticals show that the wel-fare gains in the middle and high scenarios are substantial. They amount to € 1.7 billion (0.3% of the Dutch GDP) and € 3.6 billion (0.7% of the Dutch GDP), respectively. In the low scenario the welfare gains are nega-tive (€ -0.2 billion). The welfare gains of all CVZ studies taken together are positive in all three scenarios. Only the results based on the NICE/CADTH and literature samples are negative in the low scenario. It is likely that this result is due to differences in the cost-accounting methods used in the non-Dutch studies. In the foreign studies, cost savings in the form of productivity gains and lower utilization of social security are usu-ally ignored. In stead, these countries use a higher valuation of a QALY than in the Netherlands. Institutional barriers The welfare gains are apparently substantial but they could be higher if institutional barriers for an efficient and effective utilization of innovative drugs are removed. The social costs of the current regulation of pharma-ceutical markets in the Netherlands appear to be larger than their social benefits. For instance, in our sample the delay in the introduction of new medicines approximately results in € 970 million foregone welfare gains. The most important barriers concern: ƒ boundaries in the financing of the outpatient pharmaceutical care ver- sus inpatient care and social security; ƒ the process of authorization and reimbursement of new medicines; ƒ imperfect incentives in the financing system of the health insurers. Further research Our unique pilot study is unavoidably subject to limitations. First, the study is based on only 52 new medicines. Consequently, it is difficult to judge whether these results can be generalized to other medicines and diseases. Second, our study is based on heterogeneous pharmaco-eco-nomic evaluations derived from different sources. Third, we have been forced to make some rough adjustments in order to make the non-Dutch studies as comparable to the Dutch situation as possible. Ape
These limitations have to be overcome in further research. In our opinion that research has to focus on: 1. an extension of the sample of the examined medicines; 2. a breakdown of the analyzed medicines in different disease groups in order to get an insight in the differences between these groups and to adopt - where necessary - different values of a QALY for these groups; 3. the development of a framework in which the results of the interna- tional pharmaco-economic studies can be better compared; 4. an analysis of possible systematic differences between the pharmaco- economic studies reporting their results in QALYs and those that are reporting in life years; 5. an empirical comparison of different methods for the estimation of the welfare gains of new pharmaceuticals. Ape
1.1 Background
In 2006, approximately € 5.7 billion was spent on pharmaceuticals in the Netherlands. Like all health care expenditures, expenditures on pharma-ceuticals are rising rapidly. Pharmaceutical expenditures are also fre-quently exceeding the pharmaceutical budget of the Ministry of Health. This leads to a continuous debate about the costs of (innovative) pharma-ceuticals. In this debate, little or no attention is being paid to the benefits of innovative pharmaceuticals such as health gains, improved quality of life, and reduction of costs in health care sectors or in social security. For a truly evidence-based health care policy, the focus of attention should not be health care costs per se, but rather the cost-benefit ratio of health care expenditures. Thus, it is relevant to compare the costs of new drugs with the value of health benefits and the cost reductions for society they gener-ate. In the Netherlands, Goudriaan (2004) has identified the pharmaceutical benefits at the macro level in the form of health care cost reduction. Fur-ther insight in the benefits of innovative pharmaceuticals is crucial in order to upturn the ‘institutionalized distrust' facing the pharmaceutical industry. That requires an analysis of the welfare effects of innovative pharmaceuti-cals. The full benefits of pharmaceuticals are frequently not obtained because of boundaries in public sectors' financing. In addition, there are forms of gov-ernment regulations that impede an efficient and effective use of innova-tive pharmaceuticals. Therefore, there are deadweight losses that make that the welfare gains of new drugs are less than what is potentially ob-tainable. Knowledge and insight in the institutional factors affecting an ef-ficient and effective use of innovative drugs is of great importance in order to evaluate their overall welfare gains. Furthermore, obtaining the largest possible welfare gains of pharmaceuticals not only benefits patients but may also provide incentives to stimulate innovation. Despite the market-based reforms in the Dutch health care system, there are still institutional barriers that hamper the optimal use of innovative pharmaceuticals. At the same time, justification of the pharmacotherapy's Ape
value is increasingly important to public assessors and health care insur-ers. This issue also requires better insight in the added value of innovative pharmaceuticals. The Pharmaceutical Committee of the American Chamber of Commerce in The Netherlands has asked APE and the University of Maastricht to exam-ine the welfare gains of innovative pharmaceuticals including the institu-tional factors that impede their realization. The outcome of this research is presented at the 8th Clingendael European Health Forum. 1.2 Research aim and objectives
The aim of this study is to examine the welfare gains of innovative phar-maceuticals and to identify the institutional factors that impede their reali-zation. In order to achieve this aim, the following research questions are formulated: 1. How do the parameters incorporated in the estimation affect the wel- fare gains of innovative pharmaceuticals? 2. Are there systematic differences in those parameters between coun- 3. What are the welfare gains of innovative pharmaceuticals in the Neth- 4. Which institutional barriers obstruct an efficient and effective use of innovative pharmaceuticals? 1.3 Approach
This study analyses the welfare gains of new drugs by reviewing economic evaluations of innovative pharmaceuticals and by analyzing their findings. The magnitudes of the estimated welfare gains are interpreted relative to the time horizons used in the economic evaluations. Studies included are reports that governmental bodies or quasi-autonomous non-governmental organizations (quangos) take into account in order to develop and imple-ment pharmaceutical policies. Parameters such as the costs included in those studies as well as the rates used to discount the costs and benefits are identified, and their impacts on the estimated welfare gains are ana-lyzed. The parameters are also compared to those of other countries in order to investigate whether systematic differences exist. Ape
In addition, the institutional barriers to an efficient and effective use of in-novative pharmaceuticals are identified by examining the Dutch regula-tions concerning innovative drugs and by reviewing the recent Dutch lit-erature on pharmaceutical policy. 1.4 Overview of the paper
The remainder of the study is as follows. Chapter 2 presents an overview of international guidelines that influence the cost-effectiveness of pharma-ceuticals as well as institutional controls that affect the dispensing of inno-vative drugs. Empirical evidence about the welfare impact of (new) phar-maceuticals and issues of concern about the appropriateness of cost-effectiveness in decision-making are discussed in Chapter 3. Chapter 4 discusses the study design and the research preparation. The international differences in input parameters and the results of the welfare gains are presented in Chapter 5, while the institutional barriers existing in the Netherlands for innovative drugs are analyzed in Chapter 6. Finally, Chap-ter 7 discusses the results and presents the conclusions as well as sug-gests issues for further research. A list of relevant definitions can be found in Appendix A. Ape
INTERNATIONAL POLICIES TOWARDS NEW DRUGS
2.1 Guidelines and thresholds
2.1.1 United Kingdom The National Institute of Clinical Excellence (NICE) in UK was established in 1999 in order to provide guidelines for policy makers within the NHS on new medical technologies. The NICE guidelines target to the methods used in the technology appraisal of new drugs. Although in the guidelines it is explicitly recommended to incorporate costs related to NHS, the guidelines for costs outside this system (e.g., costs related to productivity and social services) are still blurry. The discount rates of both costs and benefits are recommended to be 3.5% in the standard appraisal cases (NICE, 2004). Moreover, there are different values of a maximum acceptable cost per QALY reported in the British literature. However, it seems that the value adopted by NICE is close to £ 30,000 (€ 43,200), although in practice it might be higher (Rawlins & Culyer, 2004). This value is used as an accept-able upper bound of technology appraisals for new pharmaceuticals as well. Despite that threshold, exceptions are possible depending on the ne-cessity of the pharmaceutical for society, and pharmaceuticals with a higher cost-utility ratio may be publicly funded. However, there is still an ongoing discussion between academics and public administrators on the maximum acceptable cost per QALY (Devlin & Parkin, 2004). 2.1.2 Canada In 1989, the Canadian Coordinating Office for Health Technology Assess-ment (CCOHTA) was established. Nowadays this institution is called the Canadian Agency for Drugs and Technologies in Health (CADTH). The aim of this independent organization is to facilitate the appropriate and effec-tive utilization of health technologies within the Canadian health care sys-tem. The guidelines published by CADTH explicitly mention that the costs included in the economic evaluations of pharmaceuticals should be taken from an economic perspective within and outside the health care system Ape
(including productivity costs) (CADTH, 2006). The discount rate suggested for both costs and health outcomes is 5%. Studies in Canada have estimated values of a QALY that range from $ 20,000 to $ 100,000, but there is not an explicitly defined threshold (Laupacis, Feeny, Detsky, & Tugwell, 1992). However, there is an impres-sion that pharmaceutical companies sometimes adjust their model pa-rameters to come just below $ 50,000 per QALY (Harris, Buxton, O'Brien, Rutten, & Drummond, 2001). 2.1.3 New Zealand In New Zealand, the Pharmaceutical Management Agency (PHARMAC) is the institution responsible for the decision over the reimbursement of pharmaceuticals. Between 1993 and 2000, PHARMAC was owned by the Health Funding Authority and the Regional Health Authorities. Since 2000, PHARMAC's is a legally established stand-alone entity. In the guidelines issued by PHARMAC in 2007, the direct and indirect costs are analytically addressed for inclusion in the economic evaluations submitted for consid-eration. However, in the guidelines it is explicitly mentioned not to include the indirect patient costs such as productivity losses and costs related to patient suffering. The discount rates for both effects and costs are recom-mended to be 3.5% (PHARMAC, 2007). There is no threshold defined below which drugs are considered as cost-effective. However, it is implied in the literature that PHARMAC remuner-ates drugs with cost-effectiveness up to $NZ (Rosevear, 2006). 2.1.4 Australia In 1993, the Australian Pharmaceutical Benefits Advisory Committee (PBAC) established an economic sub-committee in order to review and in-terpret economic evaluations of drugs submitted to PBAC. The first guide-lines were issued in 1995 and since then, Australia has been one of the pioneer countries in the implementation and development of economic evaluations in the decision making for pharmaceutical reimbursement (Mitchell, 2002). There was an update of the guidelines in 2002 according to which, costs and benefits should be discounted at a 5% rate (PBAC, 2002). The guidelines address explicitly the costs to be included in the Ape
analyses. It is recommended to include direct and indirect medical costs as well as indirect costs (such as the value of work time gained or lost). In addition, it is required in the financial implications for government budgets to report the medical costs savings (e.g. hospital days, doctor visits, op-eration substitution) and the potential reduction of the burden of illness. Although there is not an officially determined threshold for reimbursing pharmaceuticals, an Australian study reviewed 355 submissions for reim-bursement to PBAC and concluded that the implied value used by PHAR-MAC is about $AU 76,000 (about € 46,000) for a QALY and about $AU 42,000 (€ 25,000) for a life year gained (George, Harris, & Mitchell, 2001). Accordingly, studies in the literature appear to have used this benchmark in economic evaluations (Mauskopf, Cates, Griffin, Neighbors, Lamb, & Rutherford, 2000). 2.1.5 Sweden On 1 October 2002, the Pharmaceutical Benefits Board (the abbreviation in Swedish is LFN) was appointed by the Swedish Government to decide on the reimbursement of drugs. LFN receives the economic evaluations sub-mitted by the pharmaceutical companies, reviews them and within 180 days decides on reimbursement. In April 2003, LFN issued guidelines for the economic evaluations undertaken by the pharmaceutical companies. In these guidelines, the discount rates for benefits and costs are set at 3%. Furthermore, the Swedish guidelines describe the costs that should be in-cluded in the cost-effectiveness studies submitted for consideration. Among others, the indirect costs outside health care (such as productivity costs) are mentioned explicitly. Similar to the abovementioned countries, there is no specifically defined threshold that decision makers use for the reimbursement of a drug. How-ever, studies in Sweden have concluded that the range of the QALY value is between € 25,000 and € 66,000 (Persson, 2007). The average value of this range is € 44,500. 2.1.6 The Netherlands In 1999, the College voor Zorgverzekeringen (CVZ, Health Care Insurance Board) was established in the Netherlands in order to provide independent guidelines for health insurance policies. CVZ has published guidelines con- Ape
cerning the methodology of the economic evaluations of new pharmaceuti-cals. In those guidelines it is explicitly mentioned that costs should include both direct and indirect costs within and outside health care (including productivity costs) (CVZ, 2006). The suggested discount rate for future costs is 4% and for effects in the future, it is 1.5%. Similar to other European countries, the role of health technology assess-ment in health policy setting has been discussed extensively in the Nether-lands over the last years. Moreover, although cost-effectiveness is in the primal set of criteria for the reimbursement of pharmaceuticals, there is no formal threshold considered. Studies have shown that the acceptable cost-effectiveness ratio depends on the therapeutic category of pharmaceuticals (Niessen & Rutten, 2007). In practice, the range of acceptable cost-effec-tiveness ratios varies between € 18,000 and € 20,000 (Pomp, Brouwer & Rutten, 2007). However, there are cases where even though the cost-effectiveness of a medicine was proven (the cost per QALY was lower than that of comparable and accepted pharmaceuticals), the Ministry decided not to include them in the basic package (Berg, van der Grinten, & Klaz-inga, 2004). The current (implicit) value of the threshold is criticized by Dutch experts including a government advisory board (RVZ 2006). RVZ (2006) and Pomp, Brouwer & Rutten (2007) do not propose one constant value of the threshold, but argue that the value of the threshold should be dependent on the disease burden. There is increasing awareness that valuations of QALYs may differ when the QALYs accrue to different patients (Stolk, van Donselaar, Brouwer & Busschbach, 2004). 2.1.7 Summary of guidelines and implied thresholds As can be seen from Table 2.1, Canada and Australia have the highest rec-ommended discount rates for both effects and costs included in the eco-nomic evaluations. Considering the discounting of effects, the Netherlands appears to have the lowest rate among the countries reviewed here. At this point, it should be mentioned that NICE was the first institute that made a distinction between the rate at which effects are discounted and the discount rate for the costs (1.5% and 6% respectively). This distinc-tion corresponds to the current trends in health economic literature (Brouwer, Niessen, Postma, & Rutten, 2006). However, in the latest NICE guidelines, this distinction has disappeared and the suggested discount rate by NICE is 3.5% for both effects and costs. Ape
Regarding the upper limit for the cost-utility ratio in the drug reimburse-ment decision-making, no organization appears to have an explicitly de-fined threshold. However, literature has identified implied thresholds that are likely to be used by decision makers. As Table 2.1 shows, NICE, PBAC, and LFN appear to have the highest implied thresholds, of about € 45,000 for a QALY gained. On the other hand, the implied thresholds used by PHARMAC and CVZ appear to be the lowest among the countries included in our study: about € 20,000 per QALY gained. Finally, the implied thresh-old used by CADTH seems to be somewhere in between (€ 35,000). In general, there does not seem to be a regular adjustment of the threshold to account for inflation. It should be mentioned that the thresholds reported here are taken from the literature and refer to the average recommended benchmarks used by the institutes. However, in every country, exceptions are made and higher thresholds are frequently used. Table 2.1 Summary table of guidelines and thresholds Establishment Guidelines Indirect Discount Discount Implied QALY Country Institute year costs* Effects (%) Costs (%) threshold 5 (€ 35,000) ** 2007 Explicit 3.5 5 (€ 46,000) ** 2006 Explicit 1.5 Outside health care. ** In approximation (exchange rates December 2007). 2.2 Institutional controls and regulations
2.2.1 United Kingdom Table 2.2 provides an overview of the control mechanisms for pharmaceu-tical products in the UK. At the supply side there are controls related to the pricing of pharmaceuticals. In particular, the profit control mechanism refers to the controls on the return on capital or return on sales of phar- Ape
maceuticals.tage of price decrease applied to all products whereas the free price modulation implies that there is no price-comparison with other countries. There are also controls at the sup-ply side targeting to the reimbursement of pharmaceuticals. The UK has an inclusive system where drugs that have received marketing approval are, by default, reimbursed (of course that does not necessarily guarantee that they will be prescribed). Those drugs that are excluded from reim-bursement are said to be on the ‘negative list'. Controls also exist at the demand side of the pharmaceutical market. Fur-thermore, there are guidelines that are primarily based on therapeutic considerations in order to influence the type of medications that doctors prescribe. There are also budgets that provide incentives to doctors to consider costs when selecting between alternative treatments. Generic drugs are promoted and are seen as a tool to contain the rising pharma-ceutical cost and incentives to professionals, such as rewards, are provided to make professional-budgets effective. Finally, there is a co-payment with a flat rate £ 6.85 (about € 9.00) per item with some exemptions (e.g., for pregnant women, children, elderly, and patients suffering from defined conditions). 2.2.2 Canada The Canadian system is fragmented since each provincial jurisdiction (10 of them) has their own philosophy and methods to manage drug pricing and consumption. Some are relatively liberal with more drugs on the list without reimbursement criteria, legislation that stipulates that the province will not contemplate reference-based pricing schemes and a policy that extends the patent life of medicines beyond what the federal authority grants (Quebec). At the complete opposite, is British Columbia that has instituted reference-based pricing on PPIs and anti-hypertensives. How-ever, these provincial differences are going beyond the scope of this pa-per. Thus, there are only general highlights mentioned here. Table 2.2 also summarizes the institutional controls for pharmaceutical products in Canada. At the supply side, there are controls that relate to the pricing of pharmaceuticals. Molecule/class reference pricing implies that drugs with the same active ingredient or therapeutic benefit have the same pricing/reimbursement. Moreover, Canada's Patented Medicine Prices Review board uses cross-country reference pricing to set the gov- 1 Permitted rate of return around 17-21% for new pharmaceutical products. Ape
ernment-approved price for a drug as the median of the prices in other countries. Moreover, in Canada pharmaceutical companies have to apply for reimbursement status for their drugs. If granted, the drugs are placed on the ‘positive list'. At the demand side, physicians are notified of generic alternatives in order to influence prescribing behavior towards less costly medicines. Similar, formularies are lists with preferred drugs. These are used to steer pre-scribing behavior among reimbursed drugs. Finally, a co-payment, in form of fixed fee and deductibles, has been implemented as a cost containment method for pharmaceuticals. 2.2.3 New Zealand The institutional controls for pharmaceutical products in New Zealand are summarized in Table 2.2. Supply side controls include pricing control, i.e. a set of maximum prices, which are periodically recalculated based on the prices in the reference countries. Concerning reimbursement controls, there is a positive list for pharmaceuticals reimbursed. In addition, eco-nomic evaluations of pharmaceuticals are incorporated in the decision-making process of medicine reimbursement. Controls exist also at the demand side of the pharmaceutical market. There are guidelines that are primarily based on therapeutic considerations in order to influence the type of medications that doctors prescribe. Fur-thermore, there is a volume limitation for drug prescriptions per month for each physician. Generic drugs are promoted and are seen as a tool to con-tain rising pharmaceutical costs. Incentives to professionals, such as re-wards, are provided to make professional-budgets more effective. Finally, there is a co-payment with a flat rate per item. 2.2.4 Australia In Australia, there is a set of maximum prices, which are periodically re-calculated based on the prices in the reference countries. Moreover, the reimbursement system for pharmaceuticals operates with a positive list. Controls exist also at the demand side of the pharmaceutical market. Fur-thermore, there are guidelines that are primarily based on therapeutic considerations in order to influence the type of medications that doctors Ape
prescribe. Yet, there is a volume limitation for drug prescriptions per month for each physician. Generic drugs are promoted and are seen as a tool to contain rising pharmaceutical cost. Finally, there is a co-payment with a flat rate per item. Table 2.2 summarizes the institutional controls and regulations concerning the pharmaceutical market in Australia. 2.2.5 Sweden The institutional controls for pharmaceutical products in Sweden are sum-marized in Table 2.2. The pricing part of the supply side controls includes a partial reference pricing control. The set of maximum prices is partly based on other countries and partly based on direct negotiations with pharmaceutical companies. Moreover, the drugs are placed on the ‘positive list'. Further, pharmaceutical economic evaluations are incorporated in the decision-making process as an additional tool to the pharmaceutical reim-bursement process. Controls exist also at the demand side of the pharmaceutical market. Fur-thermore, there are guidelines that are primarily based on therapeutic considerations in order to influence the type of medications that doctors prescribe. There are also budgets that provide incentives to doctors to consider costs when selecting between alternative treatments. Generic drugs are promoted and are seen as a tool to contain rising pharmaceuti-cal cost. Incentives to professionals, such as rewards, are provided to make budgets effective. Finally, there is a co-payment with a flat rate per item. 2.2.6 The Netherlands Table 2.2 presents the institutional controls for pharmaceutical products in the Netherlands. Supply side controls include pricing control, i.e. a set of maximum prices, which are periodically recalculated based on the prices in the four reference countries (Germany, France, Belgium, and UK). How-ever, it is argued that in practice this control affects also the reimburse-ment of the pharmaceutical product. Moreover, in the Netherlands, an ex-clusive system operates, where pharmaceutical companies have to apply for reimbursement status for their drugs. If granted, the drugs are placed on the ‘positive list'. Furthermore, pharmaceutical economic evaluations 2 Exceptions exist for instance for multiple users. Ape
are incorporated in the decision making process as an additional tool to determine the reimbursement price. Table 2.2 Summary table of institutional controls Country Supply side Demand side Pricing Reimbursement Budget to professionals Price cut (4.5%) Generic prescribing Free price modulation Incentives to professionals Molecule/class reference pricing Positive list Generic substitution incentives Formularies Co-payment (flat Reference pricing Practice guidelines Cost-effectiveness Volume limitation to profes- Incentives to professionals Co-payment (flat fee) Maximum price fixing Practice guidelines Volume limitation to profes- Co-payment (flat fee) Pricing due to negotiations Budget to professionals Cost-effectiveness Generic prescribing Incentives to professionals Co-payment (flat fee) Reference pricing Generic prescribing Cost-effectiveness Therapeutic reference pricing Practice guidelines Promotion of parallel imports Incentives to professionals Formularies Co-payment (flat Source: (Brouwers, Silverstein, & Wolff, 2004; Espín & Rovira, 2007; Jacobzone, At the demand side, generics are promoted as a tool to contain rising pharmaceutical costs. Furthermore, there are guidelines that are primarily based on therapeutic considerations in order to influence the type of medi-cations that doctors prescribe. Similarly, in order to favor particular medi-cines, pharmacists receive incentives from authorities, pharmaceutical Ape
companies or wholesalers. In addition, parallel imports are encouraged (i.e., trade of identical medicines sold in another market), and are pro-vided as a way to promote price competition and hence efficiency. There is a limited co-payment policy. Similar, formularies are lists with preferred drugs. Pharmaceuticals are included in the compulsory deductible of € 150 of health care costs that individuals have to pay out of their own pocket and patients have to pay the difference between the actual price of medi-cine and the price limit (reference price). The last control on the demand side is substitution. Generic substitution is normally allowed except when doctors unmark their consent (tick-out). The pharmacist then decides which drug to dispense. This choice is based on strong financial incentives for the pharmacist (bonuses and discounts). 2.2.7 Summary of institutional controls and regulations Institutional controls and regulations exist in each country included in our review. It is remarkable that only half of the countries presented here use explicitly economic evaluations of pharmaceuticals in the decision-making process of medicine reimbursement. Economic evaluations are explicitly used in the reimbursement decision in New Zealand, Sweden, and the Netherlands. In the other countries, economic evaluations are used as a supplementary tool in the decision making process. However, it is not sur-prising that, even if there are some institutional control differences in these counties, a common ‘regulation' trend is followed and a kind of com-munication vessel of knowledge and experience exists among them (van Oostenbruggen, Jansen, Mur, & Kooijman, 2005). Ape
WELFARE EFFECTS: EMPIRICAL EVIDENCE AND THEO-
RETICAL CONCERNS
3.1 The welfare impact of pharmaceuticals
There is substantial empirical evidence in the literature that the innovation of medical technology is worth the increased costs of care (Cutler & McClellan, 2001) and that the return on investments in health is signifi-cantly positive (Bunker, 2001; Bunker, Frazier, & Mosteller, 1994; Luce, Mauskopf, Sloan, Ostermann, & Paramore, 2006). In line with this, Cré-mieux et al. (2007) found that increases in drug spending could be more than offset by decreases in other health care spending without affecting the health of the population. In particular they concluded that increasing drug spending by $Can 1.00 results to a decrease of hospital and physician spending by $Can 1.48 for males and $Can 1.05 for females. Similarly, evidence reported in the international literature indicates that pharmaceuticals provide substantial welfare gains in the form of longer life expectancy and higher quality of life. For instance, Crémieux et al. (2005) concluded that higher spending in drugs in Canada over the period 1981-1998 was associated with significant improvements in infant mortality rates and life expectancies. Moreover, Cutler et al. (2007) have estimated that in the absence of anti-hypertensive drugs the average blood pressure would be 10-13% higher for the American population aged 40 years and older. As a result, the av-erage life expectancy for males and females would be 0.5 and 0.4 years shorter, respectively. The estimated benefit-to-cost ratio of antihyperten-sive drugs is 6:1 (women) to 10:1 (men) and they are potentially even higher, if the impact of those drugs on the quality of life and on labor pro-ductivity is included. Interestingly, Cutler et al. conclude that "private in-surance plans are unlikely to bear the future medical costs of the underuse of effective medicines today" (Cutler, Long, Berndt, Royer, Fournier, Sasser, & Cremieux, 2007). Groot and Maassen van den Brink (2004) have estimated the welfare gains of medicines for severe headache and migraine. They found that the com-pensating income variation of this disease amounts to € 865-1,227 per month. Based on American cost calculations for its treatment ($ 7.80-300 Ape
per month) and considering that in 60-80% of the cases these medicines reduced severe headache migraines effects by 50%, they concluded that the welfare gains of those medicines are higher than their costs (Groot & Maasen van den Brink, 2004). Lichtenberg has examined the effects of new drugs in a series of studies. He found that innovative drugs, although generally more expensive than older ones, lead to a substantial net reduction in the total costs of treating a disease. On average, new drugs resulted in a reduction of $ 71.09 in non-drug spending and a $ 18 increase in prescription costs (Lichtenberg, 2001). In absolute numbers, a one year reduction in the age of drugs util-ized reduces non-drug expenditure 7.2 times as much as the increase in drug expenditures (Lichtenberg, 2002). In addition, he concluded that pa-tients consuming new drugs have lower mortality and morbidity rates than patients who consume older drugs. Moreover, in a large study examining the utilization of 43.000 pharmaceutical products in 12 US States, Lichten-berg estimated the effects of new drugs on the reduction of hospital and nursing home expenditures. He concluded that "even if the upper-bound estimate of the increase in 2003 drug expenditure attributable to the 1997-2003 increase in drug vintage is lower than the sum of the lowest estimates of the 2003 hospital and nursing home expenditure reductions attributable to the 1997-2003 increase in drug vintage" (Lichtenberg, 2006). Other studies have focused on the effects of innovative drugs on produc-tivity within health care. Through the development and availability of new medicines, there is in some cases less need for medical operations, a shorter duration of operations in hospitals, and a substitution of expensive medical interventions by relatively cheaper medicines. Berndt et al. (2000) have estimated that the average costs of depression treatment were 20% lower by shifting from therapeutic to pharmaceutical treatment. An inter-esting side effect of such a reduction is that less personnel is needed in health care when using innovative pharmaceuticals. This might contribute to the solution of (future) labor market problems in health care (Berndt, Bir, Busch, Frank, & Normand, 2000). This conclusion is very relevant for the Dutch situation (e.g., Goudriaan et al. 2005). Summarizing all the abovementioned evidence in highlights, (innovative) pharmaceuticals: ƒ Increase life expectancy and improve quality of life. ƒ Decrease other health care spending such as hospital and physician Ape
ƒ Are likely to improve welfare. ƒ Decrease the need for medical operations and lower the demand for medical personnel. 3.2 Issues of concern
There are several issues currently on the health economics agenda con-cerning the appropriateness of the economic evaluations of pharmaceuti-cals (Hirth, Chernew, Miller, Fendrick, & Weissert, 2000). Although these issues are beyond the scope of this study, they are briefly addressed in order to indicate the relevance of the estimated welfare effects reported in our results. Moreover, the setting of national priorities is a rather difficult process in which health technology assessment is not yet fully developed in a rational manner (Berg, van der Grinten, & Klazinga, 2004). The role of economic evaluations in this process, is also often interfered by political influence and arguments (Harris, Buxton, O'Brien, Rutten, & Drummond, 2001). From a scientific point of view, the consistency and the validity of cost-effectiveness evaluation has been broadly criticized. The context and in-terpretation of cost-effectiveness outcomes highly depends on the value of life used in the analysis. In a meta-analysis, Mrozek (2002) found that there is a substantial range (from $ 1.5-$ 2.5 million) in the ‘value of a statistical life'-values used in economic evaluations. As a result, there is little consistency among the results. An issue that is on the top of the health economics and reimbursement policy agenda is the threshold chosen in order to define cost-effective technologies. The implied thresholds used appears to be arbitrary and in-consistent (Hirth, Chernew, Miller, Fendrick, & Weissert, 2000). It is likely that the threshold is influenced more by budget availability than by socie-tal consensus (McGregor & Caro, 2006; Owens, 1998). The concept of a QALY is sometimes criticized because it is based on a utilitarian principle and it does not reflect important values such as equity and societal preferences. Additionally, the QALY estimates depend on the viewpoint by which they are approached. Different viewpoints such as from the side of patient groups, taxpayers, average citizens, or even epi-demiologists and health economists frequently result in different QALY es-timates (McGregor & Caro, 2006). Ape
Furthermore, the utility estimation methods (e.g., Time-Trade-Off and Standard Gamble) used in economic evaluations are criticized extensively in the literature on their validity and reliability (Laupacis, Feeny, Detsky, & Tugwell, 1992; Oliver, 2003). Another issue of concern about economic evaluations is the discounting of effects and costs. The discount rates applied substantially determine the outcome of the cost-effectiveness study. Differences in discount rates used between studies severely restrict the comparability of outcomes among interventions (Brouwer, Niessen, Postma, & Rutten, 2006). Similar, cost-effectiveness results can favor interventions simply by choosing the ‘ap-propriate' time-horizon and comparator (i.e., alternative intervention) (Hirth, Chernew, Miller, Fendrick, & Weissert, 2000). It is also important to note that there are large differences in cost account-ing methods between countries. These differences make a universal eco-nomic evaluation framework difficult to develop even within the European Union (Mitchell, 2002; van Oostenbruggen, Jansen, Mur, & Kooijman, 2005). In addition, there is rarely any evidence in the pharmaceutical eco-nomic evaluations about the non-health related costs. This implies that the outcomes of these studies have only limited relevance for reimbursement decisions (Harris, Buxton, O'Brien, Rutten, & Drummond, 2001). Summarizing, there are issues of concern about the appropriateness of economic evaluations related to: ƒ the setting of national priorities and the role of economic evaluations in ƒ inconsistencies in the value of life; ƒ the concept of a QALY; ƒ the utility estimation methods; ƒ the arbitrariness of thresholds used in determining cost-effective inter- ƒ the discounting of effects and costs; ƒ the time-horizon used in cost-effectiveness studies; ƒ the drug used as a comparator in cost-effectiveness studies (usually a ƒ the cost calculation method used. Although there are many suggestions in the literature to overcome, the limitations mentioned above for the development of a concrete, reliable, and universal cost-effectiveness framework, still a lot of work needs to be done to achieve it. However, some attempts towards that aim have been Ape
reported in the literature (Cutler, 2007; Jönsson, 2006; Manca & Willan, 2006; Niessen, Grijseels, Koopmanschap, & Rutten, 2007). Our current study, although it identifies all these issues, does not aim to resolve them or to correct for them. This study is based on the existing decision-making process of pharmaceutical reimbursement. Consequently, we take the outcomes of economic evaluations on pharmaceuticals as given – and ignore their limitations and potential biases – as these are the figures that are used explicitly or implicitly in decisions about the reim-bursement of new pharmaceuticals. Our selection of pharmaco-economic studies does not imply any quality judgement over the methodology and the results of the studies concerned. Ape
STUDY DESIGN AND RESEARCH PREPARATION
4.1 Research steps
In order to estimate the welfare gains of innovative pharmaceuticals, the following research steps are undertaken: 1. the selection of pharmaco-economic studies and data collection; 2. data analysis; 3. scenario analysis. These steps are described in the next three sections of this chapter. 4.2 Selection of studies and data collection
4.2.1 Criteria The selection of the studies to be included in the analysis is based on crite-ria that serve the purpose of this study: ƒ policy orientation; ƒ generalizability and international comparability; ƒ drug age; ƒ disease variety; ƒ producers variability; ƒ economic importance. These criteria and the respective selection process are described in the remainder of this section. 4.2.2 Policy orientation The primary object of this study is to examine the outcomes of pharmaco-economic studies used in the formulation of health care policies. For this purpose, the sample is not based directly on studies conducted by phar-maceutical companies but is created by taking into account the perspec-tive of policy makers. From a policy perspective, CVZ is the first source of data extraction in the Netherlands. However, the number of pharmaco- Ape
economic studies publicly available through CVZ is limited. This is due to two factors. The first is that only very recently cost-effectiveness studies for pharmaceuticals are required for the inclusion of a drug in the basic health insurance package. The second factor is that CVZ has only pub-lished studies for drugs that belong to the Annex 1B (Bijlage 1B). This list includes only innovative medicines for which the reimbursement level is unlimited and the maximum price is the average price of four benchmark countries (see Cohen, Faden, Predaris, & Young, 2007; Niezen, Bont, Stolk, Eyck, Niessen, & Stoevelaar, 2007; van Oostenbruggen, Jansen, Mur, & Kooijman, 2005). For the collection of Dutch pharmaco-economic studies, the reports avail-able on the CVZ's website were used. It was confirmed by a responsible person at the CVZ that those studies are the only ones that are publicly available. Pharmaco-economic dossiers for other drugs are confidential and therefore, not accessible. From the 23 studies publicly available at CVZ, only 16 studies report the effects of pharmaceuticals in life years gained (LYG) or quality adjusted life years (QALY's). For the purposes of this study, only these 16 studies are included in the analysis. 4.2.3 Generalizability and international comparability The inclusion of these 16 studies may lead to biased outcomes because they are not a random sample of all innovative pharmaceuticals available to the Dutch population. Therefore, the sample is enriched by pharmaco-economic studies that are more representative for the use of innovative pharmaceuticals by the Dutch population. Such studies include medicines that are widely used and have a high volume of consumption in the Neth-erlands. In addition, we attempt to collect studies for pharmaceuticals available simultaneously in the Netherlands and in other countries. This is because this study attempts also to investigate whether there are system-atic differences in the cost-effectiveness estimations reported in these countries. Therefore, other institutes responsible for the evaluation of pharmaceutical products and the development of relevant policies were also contacted. The websites of NICE and CADTH were visited to collect studies available through those institutes. Again, responsible persons were contacted in these institutes in order to confirm that we include all publicly available studies. From NICE there are 21 studies included in the sample for analy- Ape
sis and from CADTH 5 studies (the numbers do not sum up to the total number of studies due to the common studies in two institutes). One of the top health economists of the Pharmaceutical Benefits Board (LFN) in Sweden was contacted in order to collect studies available to that institute. However, it was explicitly mentioned that there are no independ-ent evaluations conducted by LFN due to the limited time available from the application date for a drug reimbursement until the date for the final decision (180 days by law). The pharmaceutical companies' reports sub-mitted are reviewed by the LFN. Based upon that, the final decision for re-imbursement is being taken. Considering that these reviews and additional comments from LFN are confidential, there can be no studies collected from this institute. Additionally, the Pharmaceutical Management Agency (PHARMAC) in New Zealand and the Pharmaceutical Benefits Advisory Committee (PBAC) in Australia were contacted for relevant studies. However, these studies proved to be inaccessible. Therefore, no studies were collected from them. Furthermore, to enrich the sample with studies concerning medicines with a high turnover in the Netherlands, scientific papers published in interna-tional journals with a high impact are reviewed as well. Therefore, a litera-ture review was conducted in order to collect studies suitable for the analysis. PubMed/Medline, Cochrane Library, Google Scholar, International Journal of Technology Assessment in Health Care, Health Technology As-sessment, Health Affairs, and Pharmacoeconomics were searched for sci-entific articles about cost-effectiveness studies for innovative pharmaceuti-cals. There were three major criteria for the inclusion of studies in the analysis: 1. the studies need to refer to drugs registered in Europe after 1997; 2. the drugs have a relatively high economic impact (turnover) in the 3. the authors must have explicitly mentioned that there are no conflicts of interest related to the content of the articles. Additionally, meta-analysis studies were preferred as an additional crite-rion to keep the selected studies as ‘objective' as possible. As a result, 13 studies are selected for inclusion in the analysis. Appendix B presents these studies. Ape
4.2.4 Age The registration date of a drug was also used to select economic evalua-tion studies for innovative pharmaceuticals. The time-window used for the purposes of this study is 10 years. Thus, studies selected concern pharma-ceuticals that were registered in Europe after 1997. The databases used to obtain data about the registration date of each pharmaceutical are the European Medicines Agency (EMEA) and the Dutch Medicines Evaluation Board (the Dutch abbreviation is CBG). 4.2.5 Disease variety Another criterion in the selection of studies is the disease category of the pharmaceutical evaluated in the pharmaco-economic report. Pharmaceuti-cals for different diseases are selected based on the Anatomical Therapeu-tic Chemical (ATC) classification system. More specifically, the ATC codes of the pharmaceuticals included are screened on the second level in order to capture as many disease categories as possible. In this manner, a vari-ety of pharmaceuticals is included in the sample and their impact on a wide range of population is captured. The ATC code of each pharmaceuti-cal is also found at EMEA and the Dutch CBG. 4.2.6 Producers variability Another aspect taken into account in the selection process is the variation in drug registration holders. The underlying reason is to avoid that the sample of studies only includes products of a limited number of pharma-ceutical companies. Therefore, drugs from a variety of pharmaceutical producers worldwide are included. However, the number of different pro-ducers taken into account is limited by the availability of pharmaco-economic studies. The registration holder of each pharmaceutical product is obtained from EMEA and CBG. 4.2.7 Economic importance The economic and financial importance of the pharmaceuticals is also con-sidered in the selection process. It is reported in the literature that drugs with high sale volumes are more often evaluated (and thus, more impor-tant) than drugs that generate fewer revenues (Lundkvist, Jönsson, & Rehnberg, 2005). Accordingly, pharmaceuticals with high total costs and a Ape
large consumption are included in the study. However, in order to avoid selection bias, this study also investigates the welfare effects of medicines targeted to small populations (i.e., orphan drugs). These drugs have by definition low total costs and low volume. In order to obtain more insight into the economic impact of the selected pharmaceuticals, data for their total costs along with their volume have been collected for the period 2002 - 2006. These data are available from the online database of CVZ for medicines, namely the GIPdatabank. How-ever, this database only includes drugs dispensed outside hospitals. The expenses for the in-hospital medicines are included in the hospital budg-ets, and they are not recorded by CVZ. As a result, we have outpatient data for the total costs of 41 out of the 52 drugs selected. In order to cal-culate the outpatient total costs of all innovative drugs, we used a list from EMEA with all drugs with new chemical substance after 1997. The total outpatient costs for the drugs included in this list were found by the GIP-databank. In particular, the total outpatient costs of the selected drugs capture 13% of the total outpatient costs of pharmaceuticals (both innovative and older) in 2006 in the Netherlands.5% of the outpatient costs of all drugs with new chemical entities registered by the European Medi-cines Agency (EMEA) after 1997. Although, there is no complete data available for the inpatient total costs of the selected drugs, there is strong evidence that these costs are also relatively high. It should be kept in mind that the medicines included in our sample are not generics and thus, the amount of total users of those drugs is not extremely high. It should be also mentioned that the selected drugs are products with off-label us-age that might influence their dispense volume. To give an indication, some of the selected drugs are included in the list with the most expensive medicines in the Netherlands (SFK, 2007). Finally, it should be noted that the calculation of the welfare effects in the next chapter include outpatient and inpatient costs of pharmaceuticals. Considering all the above-mentioned aspects, (52) studies are included. The sample consists of 16 studies from CVZ, 26 from NICE/CADTH and 13 from the literature review (3 studies are common in the CVZ and NICE/CADTH samples). A description of those studies is given in Table 4.1. The first column of the table presents the organizations through which the pharmaco-economic studies are available. The drug entities, their brands, 3 The total outpatient costs in 2006 of 41 out of 52 selected drugs are approxi- mately € 606 million. Ape
their producer company, and their ATC codes are also presented. The se-lected sample concerns products of 24 different pharmaceutical compa-nies. Table 4.1 Drugs included in the study Source No S01EC03 Dorzolamide Merck Sharp & Dohme J07BM01 Human papillomavirus vaccine Gardasil Merck Sharp & Dohme C01EB17 Ivabradine Les Laboratoires Servier omega-3-vetzuren immunotherapy(ASI) 13 A10BH01 Sitagliptine Merck Sharp & Dohme Fludarabine Fludara Temozolomide Temodal 37 N06DX01 Memantine CVZ & CADTH 39 J05AH02 Oseltamivir 46 C10AX09 Ezetimibe Merck Sharp & Dohme 47 N02CC04 Rizatriptan Merck Sharp & Dohme 52 J01XX08 Linezolid Ape
4.3 Data analysis
4.3.1 Adjustment of cost-effectiveness data The first issue in the data analysis is to make the selected international studies as representative to the Dutch socio-economic setting as possible. An adjustment is needed by taking into account specific aspects of the Dutch health care system as well as of the Dutch society. There might be a long list of issues to be adjusted when international studies are attempted to be used within a specific health care setting. For instance, the cost cal-culation methods used in each country vary (see Chapter 2 for details) and thus, the use of the outcomes of non-Dutch cost-effectiveness studies might be inappropriate. However, it is very difficult to adjust for such is-sues in the context of this study because of lack of relevant data. The adjustment that is made in our analysis concerns the domestic-value and time-value of the international cost-effectiveness studies. In particu-lar, Purchasing Power Parities (PPPs) are used in order to convert the in-ternational cost-effectiveness results to the Netherlands and the deflator of GDP is used in order to convert all cost-effectiveness ratios to 2006 val-ues. That means that the initial cost-effectiveness of the innovative drugs has been increased due to the time-value. Although health care inflation is higher than GDP inflation, it is not used because costs included in these studies are also realized outside the health care sector (e.g. productivity costs). For the purpose of clarity, it should be explicitly mentioned at this point that there is no further processing of the cost-effectiveness data. The wel-fare effects are estimated based on the data available in the data sources. In our calculations, all pharmaceuticals are treated identically based on the reports that are publicly available. 4.3.2 The formula used in the estimation of welfare gains Once the data are adjusted, the welfare gains of innovative pharmaceuti-cals can be estimated. In this process, the cost-effectiveness value of each drug is subtracted from the value of a QALY and the result is multiplied with the QALY-effect of the respective drug divided by the time-horizon. Finally, the result is multiplied by the number of patients that have used Ape
that drug in 2006-2007. This is illustrated in equation 4.1. In this equa-tion, the welfare gains of pharmaceuticals are represented by the left hand side of the equation. The time horizon is the period during which health effects are expected. The QALYvalue is the value that society is willing to pay for an additional QALY. This value is implied by the thresholds adopted by policy-makers for the acceptance of cost-effective pharmaceuticals. Further, the cost of an additional QALY is subtracted from its value in order to get its net revenue (or loss). The incremental QALY effect (benefit) per person generated over certain years (time-horizon) is divided by that pe-riod in order to estimate the QALY effect per year. Then, the yearly QALY effect is multiplied by the number of persons receiving these benefits in order to estimate the volume of QALYs in a year. )× ( The intuition behind this formula is to estimate the net benefit of pharma-ceuticals (value of a QALY minus costs for a QALY) and calculate their magnitude by multiplying it by the volume of QALYs (incremental QALY effect divided by the time horizon and multiplied by the actual drug users) generated in the society within a calendar year. In this manner, we are able to estimate the welfare gains (net benefit multiplied by their magni-tude). Box 1 presents an example of the welfare gains calculation based on for-mula 4.1. In the same manner, the welfare gains of the total sample are estimated. Box 1 Example of welfare gain estimation for Faslodex )× ( Incr.QALY effect = (50,000-27,050)×(0.03/20)×4,832= 166,341.6 Time-horizon: 20 Actual users. * Adjusted to 2006 values. A potential drawback of our formula might be that the incremental QALY effects should be divided by the duration of the treatment of each drug instead of the whole time-horizon used in the cost-effectiveness analyses. However, the treatment duration of each drug was difficult (or even im-possible) to obtain from the data sources available to us. For that reason, 4 See Appendix A for a list of defintions. Ape
as an alternative, the time-horizon was used. As a result, the welfare gains estimated in our study might be a conservative estimate since the time-horizon is equal or longer than the treatment duration. Nevertheless, this issue exists only in 9 (out of 52) of the selected drugs since the rest of the sample consists of drugs for chronic diseases and/or have 1-year time-horizon. 4.3.3 Value of a QALY used in our benchmark scenario For the value of a QALY, we tentatively use a value of € 50,000 in our benchmark scenario. A value per QALY of € 50,000 is approximately the average value that cost-utility studies have used in the United States (Bell, Urbach, Ray, Bayoumi, Rosen, Greenberg, & Neumann, 2006). It is also the median value of the implicitly used threshold in the Netherlands (i.e., € 20,000) and what is suggested by RVZ (€ 80,000). In addition, it is slightly higher than what is currently used in the U.K. (Rawlins & Culyer, 2004), Australia (George, Harris, & Mitchell, 2001), and Sweden (Persson, 2007) as the threshold used in policy decisions (see for details Chapter 2). Alternatively, when cost per QALY data is not available then, cost per Life Year Gained (LYG) estimates are used. In 16 cases, both costs per LYG and QALY were reported. In those studies, the average of the costs per LYG reported is € 32,564 while for the respective cases the average cost per QALY is € 12,169. These differences have also been mentioned before in the literature (George, Harris, & Mitchell, 2001). Taking into considera-tion that there is no explicit difference in thresholds for LYG and QALY, it can be argued that having the same threshold for LYG and QALY makes it more likely that a study in which results are reported in cost per LYG will be rejected than a study reported in cost per QALY. Consequently, in our analysis there is no distinction between studies that report in cost/LYG and cost/QALY. It has been argued that different thresholds should be used for different care categories. For example, the threshold for the acceptable cost per QALY for a life saving intervention should be higher than the cost per QALY of a life style moderating interventions. It has also been argued that the threshold should increase with the disease burden (e.g. RVZ, 2006). Ap-plied to pharmaceuticals, different QALY values may be used for preven-tive innovative drugs and therapeutic innovative drugs in policy decisions about reimbursement. However, Niessen and Rutten (2007) found no sub-stantial differences in acceptable cost-effectiveness ratios among those Ape
categories. Furthermore, it is conceptually difficult to assume that the value of life may differ, depending on the health hazard involved: the value of one's life does not change if one is exposed to a different health or mortality risk. As we use a policy perspective in our study and policy makers do not appear to use different thresholds in acceptable costs per QALY for different therapeutic categories of pharmaceuticals, we ignore this issue. 4.3.4 Other parameters taken into account In order to estimate the volume of QALYs generated by the selected drugs in 2006, the number of actual users of each medicine in 2006 is used. The total number of actual users is estimated by combining data from IMS and GipDatabank. In particular, the total number of DDDs (i.e. inpatient and outpatient) of each drug in 2006 (found by IMS) is divided by the ratio DDDs/user over a year (i.e. in 2006) found in GipDatabank. However, in cases where the medicines were only recently introduced in the market and there is no data available, we use the estimates of potential users re-ported in the cost-consequences reports (by CVZ and SFK). In order to examine the existence of systematic differences in costs per QALY estimations in the Netherlands with those in U.K. and Canada, the differences between the discount rates used, the time horizon of the evaluation, the medicine used for comparison (if not placebo), and the type of costs incorporated are considered. The costs are analyzed qualita-tively since there are no quantitative data available. Therefore, the costs reported are grouped in direct and indirect costs inside and outside health care. Finally, the price of the pharmaceuticals at their introduction date to the market and their current price are compared. This is done in order to ex-amine whether the prices of innovative pharmaceuticals are reduced over time and hence, these pharmaceuticals are becoming more cost-effective. The time lag between the registration date and the introduction date of a drug to the market is examined in order to give an indication about the foregone welfare effects of innovative drugs due to a delay in the introduc-tion of new pharmaceuticals. Ape
4.4 Alternative scenarios
In this study, two alternative scenarios are considered in order to provide further insight in the welfare effects of innovative pharmaceuticals. These scenarios are based on alternative valuations of a QALY. While in the benchmark scenario, the value of a QALY is set to be € 50,000, in the al-ternative scenarios this value is set lower and higher corresponding to lit-erature suggestions and international trends. The current upper bound to the cost per QALY used in reimbursement de-cisions in the Netherlands appears to be € 20,000 (with some exceptions), which is lower than the respective upper bounds in U.K., Canada, Sweden, and Australia. Thus, this is the value selected to be the low scenario in our sensitivity analysis. Accordingly, the high alternative scenario is based on the international lit-erature review. The value per QALY of € 50,000 used in this study is a conservative estimate. For example, the RVZ (2006) recently proposed using a value per QALY of € 80,000 as an upper bound to reimburse the costs of medical interventions, while Laupacis et al. (1992) and Cutler et al. (2007) proposed $ 100,000 as a reasonable estimate for a QALY. Most studies on the value of a statistical life year find values above $ 100,000 (see for example, Viscusi & Aldy, 2003). As a result, the upper value of a QALY is set to € 80,000 in the high scenario. These values (€ 20,000 and € 80,000) are also set in order to make the value in the benchmark scenario the median value of the lower and upper scenarios in the sensitivity analysis. Ape
INTERNATIONAL COMPARISONS AND ESTIMATES OF
THE WELFARE GAINS
5.1 International comparison of input parameters
5.1.1 Discount rates In order to investigate the international differences in discount rates used by different organizations, the frequencies of the discount rates reported in each sample are calculated. In Figure 5.1 the frequencies of the discount rates of the effects are presented for each organization as a percentage of the overall organization (or literature) sample. About 76% of the NICE studies have used a 1.5% discount rate for the effects, while less than 45% of the studies in the CVZ sample have used a 1.5% discount rate. About 80% of the CADTH sample used a 5% discount rate and almost 55% of the studies from the literature survey used 1.5% as the discount rate for the effects. Figure 5.1 Frequencies of discount rates of effects The frequency distribution of the discount rates of the costs is presented in Figure 5.2. As it can been seen in this figure, more than 80% of the CVZ studies have used 4% as the discount rate of costs while, almost the 80% of the NICE studies used a 6% discount rate. In the CADTH sample, 80% Ape
of the studies used 5% discount rate for costs and in the literature sample, almost 55% used a 6% discount rate. Detailed information about the fre-quencies as well as descriptive statistics of the discounting rates of effects and costs are presented in Appendix C. Figure 5.2 Frequencies of discounting rates of costs 5.1.2 Time horizons Another interesting perspective in the international comparison of our re-sults are the differences in time-horizons used in the studies. Table 5.1 summarizes the descriptive statistics of the time horizons used by each organization and the studies found in the literature. Table 5.1 Descriptive statistics of time-horizons CVZ * 3 of the 55 studies concern same drugs. The total number of drugs in the sample The average time-horizon used in each sample is close to 15 years except for the CADTH sample. However, the median horizon in years is approxi-mately 5 years for the CVZ and CADTH samples, and amounts to 10 years for the NICE and the literature survey sample. For reasons of clarity, the frequency distribution of the time-horizons is presented in Figure 5.3. The figure presents the frequency in percentages Ape
per sample for each time-horizon. About 65% of the studies in the CVZ, NICE, and literature samples used a time-horizon up to 10 years. In the CADTH sample this is 80%. More details about the descriptive statistics of the time-horizons are presented in Appendix D. Figure 5.3 Frequencies of time-horizon 5.1.3 Costs included Costs included in the cost-effectiveness studies in our analysis provide an-other valuable perspective in the international comparison. All of the stud-ies include direct costs of health care. However, the definition of health care differs between countries. In addition, there are differences observed in the direct and indirect costs outside the health care sector. Table 5.2 presents four categories of direct and indirect costs outside health care, where differences are observed. It appears that the traveling costs of patients are included in only 4 of the CVZ studies, while no other institute has reported these costs. Similar findings pertain to the costs related to the time of patients and to medical research. These are reported in only 2 CVZ studies. In addition, only 1 of the CVZ studies has included nursing home costs related to Alzheimer's disease. The number of studies that have included productivity costs is higher. In the CVZ sample 7 studies have taken these costs into account compared to only one study in the NICE sample and 2 in the literature sample. The differences in the cost-accounting studies reflect the differences of the health care systems and the health care cost accounting. For example, Ape
nursing home costs are as a rule included in the Dutch health care system costs, while in England they are incorporated outside the NHS. It should also be mentioned that that costs of productivity losses and nursing homes have an income effect and lower the cost-effectiveness estimates of phar-maceuticals. In other words, when for instance productivity costs are in-cluded, the cost-effectiveness is lower (than if they are not included) due to the cost savings outside the health care sector. Table 5.2 Frequencies of indirect costs outside health care Indirect costs outside health care Freq. % Freq. % Freq. % Freq. % travelling time costs productivity nursing home 5.1.4 Comparative medicine The last issue we address in the international comparison of the studies included in our analysis is the comparable medicine or intervention used in the cost-effectiveness studies. By definition, studies that include placebo as the comparable alternative have higher incremental costs but, also higher incremental effects than studies that use another medicine (usually the current broadly used) for comparison. Table 5.3 presents the frequency distribution of comparable alternatives used in each sample. The comparable alternatives are grouped into a) pla-cebo or medical intervention, and b) current drug in use. As it appears, 47% of the CVZ studies used the current drug in use to compare, while in about 40% of the NICE and CADTH studies use the existing medication. Ape
This percentage appears to be higher in the literature studies, namely 54%. Table 5.3 Frequencies of comparators CVZ Freq. % Freq. % Freq. % Freq. % 16 100 21 100 5 100 13 100 5.2 Welfare gains in the Netherlands
The total welfare gains from the different samples are presented in Table 5.4. It should be mentioned that in this table, the total welfare gains of the NICE/CADTH sample are reduced by the welfare effects of the drugs that are also included in the CVZ-sample. Thus, we do not count them twice and we give priority to the results derived from the CVZ-sample. The total welfare gains of the three samples in the benchmark scenario are about € 1.7 billion in 2006. This amounts to about 0.3% of the GDP in the Netherlands in 2006 and about 30% of the Dutch pharmaceutical spending in the same year. In the lower alternative scenario the total welfare ef-fects are about € -230 million, whereas in the upper alternative scenario the welfare effects of the overall sample are almost € 3.6 billion. Table 5.4 Results from benchmark and the lower & upper scenarios from the overall sample (values in thousands of euros) Lower scenario Benchmark scenario Upper scenario QALY=€ 50,000 QALY=€ 80,000 NICE/CADTH* -215,211 Literature -69,722 -232,338 1,659,116 3,550,570 The drugs overlapping with the CVZ sample are excluded. ** Dutch GDP in 2006 is about € 534 billion. As can be seen from Table 5.4, the aggregated welfare effects of the CVZ sample are in all cases positive. However, in the cases of the two other 5 The Dutch pharmaceutical spending in 2006 is approximately € 5.7 billion. Ape
samples (i.e. NICE/CADTH and literature) there are negative aggregated welfare gains in the lower benchmark scenario. Therefore, it should be clearly mentioned here what are the likely causes for these negative wel-fare effects. To begin with, a value of a statistical life year of € 20,000 is very low and at odds with all empirical findings in this area. Furthermore, the average cost-effectiveness estimates in the NICE/CADTH and literature samples are apparently higher than € 20,000, which is the value of a QALY taken in the lower alternative scenario. By definition this produces negative results. A possible explanation for that is that there are different costing methods used in the NICE/CADTH and literature cost-effectiveness studies. As was already mentioned in Table 5.2, CVZ's studies take into account the impact of innovative medicines in other (non)health care sectors in the form of costs saved, which results in a lower cost-effectiveness estimation. This is not the case with the cost computations in the NICE/CADTH and literature samples. Thus, one can argue that studies derived from the NICE/CADTH and literature sample need to be included with a value of a QALY higher than € 20,000. In addition, as it appears in Table 5.4, the NICE/CADTH sample is an outlier that determines almost entirely the overall welfare ef-fects. Therefore, we should be very careful in interpreting the negative overall results in the lower alternative scenario (i.e., when the QALYvalue is € 20,000). There are also substantial differences in the welfare effects estimates con-cerning the 3 drugs that are available in both the CVZ and NICE/CADTH samples. These differences are presented in Appendix F. In order to ex-plain the numerical differences presented in Table F.1, the input parame-ters used by the institutions are examined. These differences are pre-sented in Table F.2. Apparently, differences in inputs lead to substantially differences in results. That should be kept in mind when we interpret these results. These differences in inputs limit comparisons between the welfare effects for individual drugs. There is evidence that the prices of innovative pharmaceuticals are re-duced within the patent period (for details see Appendix G.1). That means that the welfare effects of innovative drugs might be higher if their future price reduction is taken into account. Finally, there are delays in the introduction of a medicine in the market after its registration (see Appendix G.2). In our sample, the average intro-duction delay is 7 months. This delay results approximately in € 970 mil- Ape
lion (=1,659million×7÷12) foregone welfare gains based on the bench-mark scenario of the total sample. Overall, delays in the introduction of innovative pharmaceuticals in the market have a negative impact on soci-ety. Ape
INSTITUTIONAL BARRIERS
6.1 Introduction
In the preceding chapter, we have tentatively estimated the welfare gains of innovative drugs. The estimated welfare gains are quite substantial, and could be even higher in the absence of institutional barriers that hamper an efficient and effective use of innovative pharmaceuticals. These barriers cause deadweight losses that make that the welfare gains of new drugs are less than what is potentially achievable. As shown in the previous chapter, delays in the introduction of the 52 new medicines lead to a deadweight loss of € 970 million. A clear understanding of the institutional factors affecting an efficient and effective use of innovative drugs is impor-tant in order to realize the potential welfare gains and to strengthen the incentives for innovation. Pharmaceutical policy-making is highly politicized. Governments try to regulate few markets as much as they do the pharmaceutical market. Mostly government controls are concerned with issues such as public health, access to safe and effective medicines, and quality of health care. However, some regulations primarily focus on cost containment with re-spect to pharmaceuticals, and affect authorization and reimbursement of innovative medicines. These regulations could lead to deadweight losses, because they often do not account for efficiency effects, such as the im-pact on the rest of the health care sector and other government programs. Usually the side effects of regulations have to do with boundaries in public sector's financing. In addition, there exist ‘perverse' incentives in the fi-nancing of health insurers, which hamper the substitution of inpatient care by cheaper outpatient care and pharmaceuticals. In this chapter, an at-tempt is made to identify the institutional barriers that impede the realiza-tion of the potential welfare gains of innovative pharmaceuticals. The analysis in this chapter focuses on the Dutch situation, but may also be relevant for other countries. 6.2 Boundaries in the financing of the different sectors
The benefits of innovative pharmaceuticals are often substantial. The utili-zation of innovative pharmaceuticals results in health gains, better quality Ape
of life, lower worker absenteeism, cost savings in social security and cost savings in other health care sectors. Policy discussions on innovative pharmaceuticals are generally confined to the costs of these pharmaceuti-cals. Often, cost savings in other parts of health care or in other sectors caused by the use of innovative pharmaceuticals are not or only partly taken into account by policy makers. For instance, innovative pharmaceu-ticals contribute to preventing hospital admissions and reducing average length of stay in hospitals. However, boundaries in the financing of outpa-tient pharmaceutical care and hospital care hamper a cost effective shift of financial resources from hospital care to pharmaceutical care. This situa-tion is made worse by the incentives in the financing of health care insur-ers (see Section 6.4). It is obvious that the current boundaries in the fi-nancing of health care impede substitution of hospital care by outpatient care and reduce the welfare gains of innovative drugs. Likewise, current boundaries in the financing of health care and social security weaken the potential impact of innovative medicines on reducing worker absenteeism and disability. Again, this might lead to deadweight losses. More compre-hensive financing schemes are needed to prevent deadweight losses. 6.3 Authorization and reimbursement
The process of market authorization and reimbursement of new pharma-ceuticals is a crucial factor in the realization of welfare gains. Delays in the process of authorization prevent the realization of welfare gains. Of course, the authorization of new pharmaceuticals needs to be based on a careful process of decision-making, and therefore it is time consuming. However, unnecessary delays in the process of market authorization and decision-making on reimbursement have to be avoided. In general, new pharmaceuticals that are therapeutically equivalent to ex-isting pharmaceuticals can be easily included in the current Dutch refer-ence price system (GVS), although the concept of therapeutic equivalence is subject to discussion. The reimbursement limit is determined by the first introduced (cheapest) drug in the therapeutic cluster, which is usually an off-patent drug. Consumers are confronted with co-payments when the actual price exceeds the reimbursement limit. Older off-patent drugs serve as a benchmark for reimbursement decisions. This mechanism stimulates cost containment, but does not stimulate innovation, price competition or efficiency. Consumers or prescribers are directed towards non-innovative drugs, which lead to deadweight losses when the concept of therapeutic equivalence is inappropriately used. Furthermore, quality of life of patients Ape
may be negatively affected as new pharmaceuticals frequently have fewer side effects than existing (off-patent) pharmaceuticals. When there are no therapeutically equivalent substitutes for the new inno-vative drug, an extensive assessment procedure is applied. As of 2005 a mandatory pharmaco-economic study is part of the assessment procedure. The mandatory study is subject to guidelines of CVZ. It has been shown elsewhere in our report that various countries use different guidelines for pharmaco-economic studies, which may affect reimbursement decisions. Our impression is that the Dutch guidelines are relatively prudent com-pared to international standards. Recently, the Dutch ministry of Health has acknowledged that the current GVS, based on the concept of government regulation, does not lead to op-timal results. Instead, competitive forces have to be strengthened in order to achieve more socially acceptable outcomes. The current reference price system will phased out gradually. In the long run more incentives for inno-vation will be in force. 6.4 Imperfect incentives for health insurers
The basic package of the Dutch health insurance system (‘Zorgverzeker-ingswet') is financed by income-related premiums and nominal premiums. The income-related premiums are collected by the tax administration and are subsequently allocated by CVZ to the health insurers. CVZ uses a risk adjustment formula to compute the prospective budgets of the health in-surers (see APE/ministerie van VWS 2007 for a description). Health insurers bear the full financial risks of all non-hospital costs (includ-ing all outpatient pharmaceutical costs). If the prospective budget for the non-hospital costs for some or all health insurers turns out to be insuffi-cient afterwards, the financial risk has to be borne by the health insurers. This is a strong incentive for health insurers to contain non-hospital costs, such as pharmaceutical costs. On the other hand, the prospective budget that health insurers receive for hospital care is supplemented by a generous ex-post compensation sys-tem. Consequently, health insurers are only partly financial responsible for hospital costs. For them additional hospital costs are partly a free lunch compared to pharmaceutical cost and other outpatient costs. This ‘per-verse' incentive discourages substitution of inpatient care by outpa- Ape
tient/pharmacotherapeutic care. Health insurers have little incentives for cost containment in the hospital sector in favor of outpatient care and in-novative pharmaceutical care even when the utilization of innovative pharmaceuticals is more cost effective. In short, the main drawbacks of the unequal risk regimes of hospital care and non-hospital care in the risk adjustment system are: ƒ It is not in line with the politically and socially desired shift towards outpatient health care. ƒ It lowers the efficiency of the Dutch health care system. ƒ It hampers the realization of welfare gains attributed to the utilization of innovative pharmaceuticals. ƒ It may reduce the potential gains in quality of life of patients, as they do not receive the best available pharmaceuticals. As of January 2008, the ex-post compensation system for hospital care has become somewhat less generous, but is still in force. In addition, it should be noted that in December 2007 the Dutch parliament has passed a motion (motie Omtzigt c.s.) in which the minister of Health has been asked to remove the present perverse incentives in the risk adjustment system. 6.5 Conclusions
Institutional barriers hamper an efficient and effective use of innovative pharmaceuticals. As a result, it is to be expected that the welfare gains of new pharmaceuticals are less than what is potentially achievable. The fol-lowing institutional barriers are currently causing deadweight losses: ƒ boundaries in the financing of outpatient pharmaceutical care versus inpatient care and social security; ƒ the process of market authorization and reimbursement of new phar- ƒ imperfect incentives for health insurers. Ape
DISCUSSION AND CONCLUSIONS
7.1 Discussion
Analyzing the results of our study, it appears that small changes in inputs may lead to substantial differences in outputs. In particular, the variance in the costs-effectiveness of pharmaceuticals as a result of the time-horizons used, costs incorporated, discount rates adopted and, alternative drugs compared, lead to large differences in cost-effectiveness estimates and consequently in the estimates of the welfare gains. For instance, long time-horizons tend to favor the cost-effectiveness of a pharmaceutical, while including costs incurred only within health care makes a drug less cost-effective. Considering that, there is a need for more specific arrangements about the way cost-effectiveness studies are conducted for policy-making. It is rather arbitrary to use cost-effectiveness analyses in pharmaceutical reim-bursement if their consistency is doubtful and their scope is narrowly de-fined. Such an approach could lead to an inefficient resource allocation within society. Therefore, a well-defined and thoroughly developed frame-work could facilitate rational decision-making. This would contribute to a more efficient allocation of the scarce health care resources. Similar, at the international level, there is no international framework by which comparisons can be made. Differences in approaches, cost defini-tions, discount rates and time-horizons limit a valuable exchange of inter-national findings and knowledge. An international framework for economic evaluation of pharmaceuticals could benefit decision-making worldwide by providing guidelines based on international experiences. However, it should be mentioned that in order to develop an international framework, many parameters within each health care system have to be aligned. A representative example are the costs incorporated in the cost-effectiveness analyses in different countries. In many countries, the costs outside health care such as productivity loss and social security are not included, while in other countries, for instance in the Netherlands, they are taken into account. For that reason, differences in cost accounting among countries lead to different estimations of welfare effects worldwide. Ape
It should be also mentioned that in our research we have not examined the research and development (R&D) spending and incentives. The Neth-erlands is a country with relatively small pharmaceutical R&D programs. Therefore, our approach is applicable only to countries with small R&D spending and incentives. If a study examines the welfare effects of innova-tive pharmaceuticals in a country with large R&D programs, for instance in the United States, then the R&D costs and benefits should be included in the analysis. 7.2 Conclusions
7.2.1 International differences It follows from our analysis that there are substantial international differ-ences in the input parameters for the cost-effectiveness analysis. Differ-ences in cost-accounting methods, time-horizons, comparator interven-tion, and discount rates exist and affect the cost-effectiveness results on which policy decisions are based. In particular, close to 80% of the NICE sample used 1.5% discount rate for costs while only the 45% of the CVZ used that rate. On the other hand, 80% of the NICE sample used 6% discount rate for costs, whereas the same percentage of the CVZ sample used 4% as discount rate for costs. With respect to the time-horizons there are no significant differences among the sample since, 65% of the CVZ, NICE, and literature samples used a time-horizon up to 10 years. There are substantial differences found in the cost accounting methods since 44% of the CVZ sample included productivity costs, while only the 5 % of the NICE sample and 15% of the literature sample included these costs. Other indirect costs such as traveling (25%) and time (12%) costs of patients as well as costs of nursing homes (6%) are included only in parts of the CVZ sample. In addition, 50% of the CVZ sample used placebo as comparative inter-vention, while this percentage is about 60 in the NICE and CADTH samples and 46 at the literature sample. Finally, in all countries QALY thresholds are not regularly updated to account for inflation. Ape
7.2.2 Welfare effects in the Netherlands In the benchmark scenario, we have used € 50,000 as the valuation of a QALY. In that case, the estimated welfare gains of the 52 selected innova-tive pharmaceuticals amount to about € 1.7 billion, which is 0.3% of the Dutch GDP in 2006 and 30% of the total pharmaceutical spending in the Netherlands in the same year. In the lower alternative scenario where the value of a QALY is set at a lower bound of € 20,000, the resulted estimates of welfare gains are approximately € -230 million. In the higher alternative scenario where the QALYvalue is € 80,000, the overall welfare gains are € 3.5 billion in 2006. It should be kept in mind that the welfare effects from the NICE/CADTH sample are an outlier and substantially determine the welfare effects of the overall sample. Therefore, the differences in the input parameters and international inconsistencies in thresholds are reflected in the overall re-sults. Moreover, by taking the results of the benchmark scenario into account, it is estimated that there are approximately € 970 million welfare gains that are not realized due to delays in the introduction (7 months in this study) of innovative pharmaceuticals in the market. In addition, it should be noted that the welfare gains are likely to be higher when the price reduc-tion of existing innovative medicines over time is taken into account. 7.2.3 Institutional barriers Institutional barriers hamper an efficient and effective use of innovative pharmaceuticals. As a result, it is to be expected that the welfare gains of innovative pharmaceuticals are less than what is potentially achievable. The following institutional barriers are currently causing deadweight losses: ƒ boundaries in the financing of outpatient pharmaceutical care versus inpatient care and social security; ƒ the process of market authorization and reimbursement of new phar- ƒ ‘perverse' incentives for health insurers. Ape
7.3 Limitations of the study
As a pilot investigation of the pharmaceutical welfare effects and consider-ing the international sample used in our analysis, the limitations of this study derive from inconsistencies in the economic evaluations, interna-tional differences, limited data availability, and lack of consensus about the appropriate valuation of a QALY. The sample used in the study is limited since it includes only 52 innovative drugs and biased because it mainly consists of drugs directed to specific populations. Another limitation is that we have used heterogeneous eco-nomic evaluations from different sources. Thus, the inconsistencies exist-ing in the economic evaluations among different institutions are also incor-porated in our sample. As a methodological limitation of this study, it should be mentioned that the adjustment for the time-horizons used in our analysis could be consid-ered as rough. It should be acknowledged that the incremental QALY ef-fects possibly are not normally distributed throughout the time-horizon. Such an adjustment is, however, time-consuming and goes beyond the scope of this study. 7.4 Suggestions for future research
In this pioneer study of the welfare effects of innovative pharmaceuticals there are many issues identified that could be part of further research. In particular, future research could include a more generalized sample with regard to the number of pharmaceuticals included as well as to the popu-lation that uses these pharmaceuticals. Thus, a more representative calcu-lation of the welfare effects of innovative pharmaceuticals can be made. Further research might also focus on the development of a framework that would facilitate the comparability of internationals studies. In this frame-work, differences in costs incorporated, discount rates applied, and evalua-tion methods of drug-effects implemented in cost-effectiveness studies might be overcome. In addition, further research might develop a more sophisticated model that adjusts the welfare gains for the time-horizons used in economic evaluations of pharmaceuticals. The use of treatment duration might be a plausible yet, difficult problem. In this manner, the comparison among alternative pharmaceuticals might be enhanced and Ape
thus, decision-making for resource allocation will incorporate all options available. Future research might also separate the sample of pharmaceuticals in therapeutic groups or according to disease burden. Then we can examine differences among them and implement different values of a QALY in each category. In this manner, current decision-making will be reflected ade-quately to the estimated welfare gains. Moreover, further research might examine whether there are systematic differences between cost-effectiveness studies that report in LYGs and QA-LYs and use if necessary, different thresholds for these two categories. In this way, potential differences in LYGs and QALYs will be reflected in the welfare gains estimations. Finally, an empirical comparison of different methods for the estimation of welfare gains of new pharmaceuticals is desirable (see Section 3.1). Ape
LITERATURE
APE/ministerie van VWS (2007). Beschrijving van het risicoverevenings- systeem van de Zorgverzekeringswet. Den Haag: ministerie van VWS. Bell, C.M., Urbach, D.R., Ray, J.G., Bayoumi, A., Rosen, A.B., Greenberg, D., et al. (2006). Bias in published cost effectiveness studies: sys-tematic review. BMJ, 332, 699-703. Berg, M., van der Grinten, T., & Klazinga, N. (2004). Technology assess- ment, priority setting, and appropriate care in Dutch health care. International Journal of Technology Assessment in Health Care, 20(1), 35-43. Berndt, E.R., Bir, A., Busch, S.H., Frank, R.G., & Normand, S.-L.T. (2000). The medical treatment of depression, 1991-1996: productive effi-ciency, expected outcome variations, and price indexes. Cam-bridge: National Bureau of Economic Research. Brouwer, W.B.F., Niessen, L.W., Postma, M.J., & Rutten, F.F.H. (2006). Need for differential discounting of costs and health effects in cost effectiveness analyses. BMJ, 331, 446-448. Brouwers, C.A., Silverstein, M.B., & Wolff, T. (2004). Adverse Conse- quences of OECD government interventions in pharmaceutical mar-kets on the U.S. economy and consumer: The Boston Consulting Group. Bunker, J.P. (2001). The role of medical care in contributing to health im- provements whithin societies. International Journal of Epidemiol-ogy, 30, 1260-1263. Bunker, J.P., Frazier, H.S., & Mosteller, F. (1994). Improving health: measuring effects of medical care. The Milbank Quarterly, 72(2), 1994. CADTH. (2006). Guidelines for economic evaluation of pharmaceuticals: Canada. Ottawa: Canadian Agency for Drugs and Technologies in Health. Clegg, A., Bryant, J., Nicholson, T., McIntyre, L., De Broe, S., Gerard, K., et al. (2002). Clinical and cost-effectiveness of donepezil, rivastig-mine, and galantamine for Alzheimer's disease. International Jour-nal of Technology Assessment in Health Care, 18(3), 497-507. Cohen, J., Faden, l., Predaris, S., & Young, B. (2007). Patient access to pharmaceuticals: an international comparison. Eur J Health Econ, 8, 253-266. Ape
Cook, J.R., Yin, D., Alemao, E., Davies, G., Krobot, K.J., Veltri, E., et al. (2004). Cost-effectiveness of ezetimibe coadministration in statin-treated patients not at cholesterol goal: application to Germany, Spain and Norway. Pharmacoeconomics, 22(3), 49-61. Crémieux, P.-Y., Meilleur, M.-C., Ouellette, P., Petit, P., Zelder, M., & Pot- vin, K. (2005). Public and private pharmaceutical spending as de-terminants of health outcomes in Canada. Health Economics, 14, 107-116. Crémieux, P.-Y., Ouellette, P., & Petit, P. (2007). Do drugs reduce utilisa- tion of other health care resources? Pharmacoeconomics, 25(3), 209-221. Cutler, D. M. (2007). The lifetime costs and benefits of medical technol- ogy. Journal of Health Economics, 26, 1081-1100. Cutler, D. M., Long, G., Berndt, E. R., royer, J., Fournier, A.-A., Sasser, A., et al. (2007). The value of hypertensive drugs: a perspective on medical innovation. Health Affairs, 26(1), 97-110. Cutler, D. M., & McClellan, M. (2001). Is technological change in medicine worth it? Health Affairs, 20(5), 11-29. CVZ (2006). Richtlijnen voor farmaco-economisch onderzoek, geactuali- seerde versie. Diemen: College voor Zorgverzekeringen. Czoski-Murray, C., Warren, E., Chilcott, J., Beverly, C., Psyllaki, M. A., & Cowan, J. (2004). Clinical effectiveness and cost-effectiveness of pioglitazone and rosiglitazone in the treatment of type 2 diabetes: a systematic review and economic evaluation. Health Technology As-sessment, 8(13). De Cock, E., Hutton, J., Canney, P., Body, J. J., Barrett-Lee, P., Neary, M. P., et al. (2005). Cost-effectiveness of oral ibandronate compared with intravenous (i.v.) zoledronic acid or i.v. generic pamidronate in breast cancer patients with metastatic bone disease undergoing i.v. chemotherapy. Support Care Cancer, 13(975-986). Devlin, N., & Parkin, D. (2004). Does NICE have a cost-effectiveness threshold and what other factors influence its decisions? A binary choice analysis. Health Economics, 13(5), 437-452. Espín, J., & Rovira, J. (2007). Analysis of differences and commonalities in pricing and reimbursement systems in Europe: European Commis-sion. George, B., Harris, A., & Mitchell, A. (2001). Cost-effectiveness analysis and the consistency of decision making: Evidence from pharmaceu-tical reimbursement in Australia (1991 to 1996). Pharmacoecono-mics, 19(11), 103-1109. Goudriaan, R. (2004). Recepten voor de extramuralisering van de zorg, Economisch Statistische Berichten, 89, 571-573. Ape
Goudriaan, R., & Vemer, P. (2003). Als het geld de patiënten volgt: Een analyse van geneesmiddelen en extramuralisering op macroniveau. Den Haag: APE bv. Goudriaan, R., van den Hauten, M., Bartelings, H., & Snijders, R.W.D.J. (2005). Arbeidsmarkt, arbeidsproductiviteit en vergrijzing: Interna-tionale ervaringen met oplossingen voor personeelstekorten in de zorgsector. Den Haag APE bv. Groot, W., & Maasen van den Brink, H. (2004). A direct method for esti- mating the compensating income variation for severe headache and migraine. Social Science & Medicines, 58, 305-314. Harris, A., Buxton, M., O'Brien, B., Rutten, F., & Drummond, M. (2001). Using economic evidence in reimbursement decisions for health technologies: experience of 4 countries. Paper presented at the iHEA conference, York. Hirth, R. A., Chernew, M. E., Miller, E., Fendrick, A. M., & Weissert, W. G. (2000). Willingness to pay for a quality-adjusted life year: in search of a standard. Medical Decision Making, 20(3), 232-242. Iskedjian, M., Walker, J. H., Gray, T., Vicente, C., Einarson, T. R., & Gehshan, A. (2005). Economic evaluation of Avonex (interferon beta-1a) in patients following a single demyelinating event. Multiple Sclerosis, 11, 542-551. Jacobzone, S. (2000). Pharmaceutical policies: reconciling social and in- dustrial goals (No. 40): OECD. Jönsson, B. (2006). Time for a common standard for cost-effectiveness in Europe?. Eur J Health Econ, 7, 223-224 Laupacis, A., Feeny, D., Detsky, A. S., & Tugwell, P. X. (1992). How at- tractive does a new technology have to be to warrant adoption and utilization? Tentative guidelines for using clinical and economic evaluations. CAN MED ASSOC, 146(4), 473-481. Lichtenberg, F. (2002). Benefits and costs of newer drugs: an update. Cambridge: National Bureau of Economic Research. Lichtenberg, F. (2001). Are the benefits of newer drugs worth their costs? Evidence from the 1996 MEPS. Health Affairs, 20(5), 241-251. Lichtenberg, F. R. (2006). The effect of using newer drugs on the admis- sions of elderly Americans to hospitals and nursing homes: state-level evidence from 1997 to 2003. Pharmacoeconomics, 24(3), 5-25. Luce, B. R., Mauskopf, J., Sloan, F. A., Ostermann, J., & Paramore, L. C. (2006). The return on investment in health care: from 1980 to 2000. Value In Health, 9(3), 146-156. Lundkvist, J., Jönsson, B., & Rehnberg, C. (2005). Cost-effectiveness of new drugs: a systematic review of published evidence for new Ape
chemical entity drugs introduced on the Swedish market 1987-2000. International Journal of Technology Assessment in Health Care, 21(2), 187-193. Manca, A., & Willan, A. R. (2006). 'Lost in translation': accounting for be- tween-country differences in the analysis of multinational cost-effectiveness data. Pharmacoeconomics, 24(11), 1101-1119. Mauskopf, J. A., Cates, S. C., Griffin, A. D., Neighbors, D. M., Lamb, S. C., & Rutherford, C. (2000). Cost effectiveness of Zanamivir for the treatment of influenza in a high risk population in Australia. Phar-macoeconomics, 17(6), 611-620. McCormack, P. L., & Foster, R. H. (2006). Rizatriptan: a pharma- coeconomic review of its use in the acute treatment migraine. Pharmacoeconomics, 23(12), 1283-1298. McEwan, P., Dixon, S., Babbolal, K., Conway, P., & Currie, C. J. (2006). Evaluation of the cost effectiveness of sirolimus versus tacrolimus for immunosuppression following renal transplantation in the UK. Pharmacoeconomics, 24(1), 67-79. McGregor, M., & Caro, J. J. (2006). QALYs: are they helpful to decision makers? Pharmacoeconomics, 24(10), 947-952. McIntosh, E. D. G., Conway, P., Willingham, J., & Lloyd, A. (2003). The cost-burden of paedriatic pneumococcal disease in the UK and the potential cost-effectiveness of prevention 7-valent pneumococcal conjugate vaccine. Vaccine, 21, 2564-2572. Mitchell, A. S. (2002). Antipodean assessment: activities, actions, and achievements. International Journal of Technology Assessment in Health Care, 18(2), 203-212. Moore, D., Meads, C., Roberts, L., & Song, F. (2002). The effectiveness and cost-effectiveness of somatostatin analogues in the treatment of acromegaly: ARIF, University of Birmingham. Mrozek, J. R., & Taylor, L. O. (2002). What determines the value of life? A meta-analysis. Journal of Policy Analysis and Management, 21(2), 253-270. NICE, (2004). Guide to the methods of technology appraisal. London: Na- tional Institute of Clinical Excellence. Niessen, L., & Rutten, F. (2007). Economics-based clinical practice guide- lines. Paper presented at the Medisch Technologie, Kosteneffec-tiviteit en Richtlijnen, Ministry of Health, Welfare, and Sport. Niessen, L. W., Grijseels, E., Koopmanschap, M., & Rutten, F. (2007). Eco- nomic analysis for clinical practice- the case of 31 national consen-sus guidelines in the Netherlands. Journal of Evaluation in Clinical Practice, 13, 68-78. Ape
Niezen, M., Bont, A. d., Stolk, E., Eyck, A., Niessen, L., & Stoevelaar, H. (2007). Conditional reimbursement within the Dutch drug policy. Health Policy, 84, 39-50. Oliver, A. (2003). For debate: putting the quality into quality-adjusted life years. Journal of Public Health Medicine, 25(1), 8-12. Owens, D. K. (1998). Interpretation of cost-effectiveness analyses. J Gem Intern Med., 13(10), 716-717. PBAC. (2002). Guidelines for the pharmaceutical industry on preparation of submissions to Pharmaceutical Benefits Advisory Committee. Australia: Commonwealth Department of Health and Ageing. Persson, U. (2007, 8 October). Using cost-effectiveness in practice- Swedish experience. Paper presented at the Medische Technologie, Kosteneffectivieit ei Richtlijnen, Dutch Ministry of Health, Welfare, and Sports. PHARMAC. (2007). Prescription for pharmacoeconomic analysis: Methods for cost-utility anlaysis. New Zealand: Pharmaceutical Management Agency. Plosker, G. L., & Figgitt, D. P. (2004). Repaglinide:a pharmacoeconomic review of its use in type 2 diabetes mellitus. Pharmacoeconomics, 22(6), 389-411. Plosker, G. L., & Figgitt, D. P. (2006). Linezolid: a pharmacoeconomic re- view of its use in serious gram-positive infections. Pharmacoeco-nomics, 23(9), 945-964. Pomp, M., Brouwer, W., & Rutten, F. (2007). QALY-tijd, nieuwe medische technologie, kosteneffectiviteit en richtlijnen (No. 152): Centraal Planbureau. Rawlins, M. D., & Culyer, A. J. (2004). National Institute for Clinical Excel- lence and its value judgments. BMJ, 329, 224-227. Rosevear, M. (2006). PHARMAC and Herceptin for early-stage breast can- cer in New Zealand: Herceptin or deception? Journal of the New Zealand Medical Association, 119, 1235. SFK. (2007). Monitor Dure Geneesmiddelen 2004, 2005 en 2006: Stichting Farmaceutische Kengetallen. Stolk, E.A., van Donselaar, G., Brouwer W.B.F., & Busschbach, J.J.V. (2004). Reconciliation of economic concerns and health policy: Il-lustration of an equity adjustment procedure using proportional shortfall. Pharmacoeconomics, 22(17), 1097-1107. van Oostenbruggen, M. F., Jansen, R. B., Mur, K., & kooijman, H. (2005). Penny and pound wise. Pharmacoeconomics, 23(3), 219-226. Veenstra, D. L., Sullivan, S. D., Dusheiko, G. M., Jacobs, M., Aledort, J. E., Lewis, G., et al. (2007). Cost-effectiveness of peginteferon a-2a compared with lamivudine treatment in patients with HBe-antigen- Ape
positive chronic hepatitis B in the United Kingdom. European Jour-nal of Gastrenterology & Hepatology, 19, 631-638. Viscusi, W. K., & Aldy, J. E. (2003). The value of a statistical life: a critical review of market estimate throughout the world. The Journal of Risk and Uncertainty, 27(1), 5-76. Wlodarzczyk, J. H., Cleland, L. G., Keogh, A. M., McNeil, K. D., Perl, K., Weintraub, R. G., et al. (2006). Public funding of bosentan for the treatment of pulmonary artery hypertension in Australia. Pharma-coeconomics, 24(9), 903-915. Ape
APPENDIX A: LIST OF DEFINITIONS

Comparator
The standard intervention against which the intervention under appraisal is
compared. The comparator can be no intervention, for example, best sup-
portive care.
Cost-effectiveness analysis
An economic study design in which consequences of different interventions
are measured using a single outcome, usually in ‘natural' units (for exam-
ple, life-years gained, deaths avoided, heart attacks avoided, or cases de-
tected). Alternative interventions are then compared in terms of cost per
unit of effectiveness.

Deadweight loss
In economics, a deadweight loss (also known as excess burden) is a loss
of economic efficiency that can occur when the equilibrium for a good or
service is not Pareto optimal. In other words, either people who would
have more marginal benefit than marginal cost are not buying the good or
service or people who would have more marginal cost than marginal bene-
fit are buying the product.
Discounting
Costs and benefits incurred today are usually valued more highly than
costs and benefits occurring in the future. Discounting health benefits re-
flects individual preference for benefits to be experienced in the present
rather than in the future. Discounting costs reflects individual preference
for costs to be experienced in the future rather than in the present.
Incremental cost-effectiveness ratio (ICER)
The ratio of the difference in the mean costs of a technology compared
with the next best alternative to the differences in the mean outcomes.
Life-years gained (LYG)
Average years of life gained per person because of the intervention.
Meta-analysis
A statistical technique for combining (pooling) the results of a number of
studies that address the same question and report on the same outcomes
Ape
to produce a summary result. The aim is to derive more precise and clear
information from a large data pool.
Quality-adjusted life year (QALY)
An index of survival that is adjusted to account for the patient's quality of
life during this time. QALYs have the advantage of incorporating changes
in both quantity (longevity/mortality) and quality (morbidity, psychologi-
cal, functional, social, and other factors) of life. Used to measure benefits
in cost–utility analysis.
Systematic review Research that summarizes the evidence on a clearly
formulated question according to a predefined protocol. Systematic and
explicit methods to identify, select and appraise relevant studies, and to
extract, collate and report their findings are used. Statistical meta-analysis
may or may not be used.

Threshold value
A value often set by policy makers, who may decide that only interven-
tions with an ICER below the threshold are cost effective (and therefore
should be reimbursed).
Time horizon
The time span used in health technology assessment that reflects the pe-
riod over which the main differences in health effects and use of health
care resources between interventions are expected to be experienced.
Time trade-off
A method used to measure utility (for example, health states). The utility
value is measured by finding the point at which the respondent cannot
choose between two scenarios. For chronic illness, the choice is between
the illness for a period of time and perfect health for a shorter time, both
followed by death. For short-term illness, the choice is between the illness
for a period of time and a worse health state for a shorter time, both fol-
lowed by the same specified outcome.
Utility
A measure of the strength of a person's preference for a specific health
state in relation to alternative health states. The utility scale assigns nu-
merical values on a scale from 0 (death) to 1 (optimal or ‘perfect' health).
Health states can be considered worse than death and thus have a nega-
tive value.
Ape
APPENDIX B: RESULTS FROM LITERATURE REVIEW
Authors & Date (Wlodarzczyk, Cleland, Keogh, McNeil, Perl, Weintraub, & Pharmacoeconomics Monary Artery Hypertension in Australia Cost Effectiveness and Risk Sharing Wlliams, 2006) (Iskedjian, Walker, Gray, Economic evaluation of Avonex (interferon beta-1a) in patients following a single de- Vicente, Einarson, & Gehshan, Multiple Sclerosis myelinating event 2005) (Moore, Meads, Roberts, & ARIF, University of Birming- The Effectiveness and Cost-effectiveness of Somatostatin Analogues in the Treatment Somatuline Song, 2002) (Clegg, Bryant, Nicholson, International Journal of Tech- Clinical and cost-effectiveness of Donepezil, Rivastigmine, and Galantamine for Alz- McIntyre, De Broe, Gerard, & nology Assessment in health heimer's disease: a systematic review (Veenstra, Sullivan, Dusheiko, European Journal of Gas- Cost-effectiveness of peginterferon a-2a compared with lamivudine treatment in pa- Jacobs, Aledort, Lewis, & trenterology & Hepatology tients with HBe-antigen-positive chronic hepatitis B in the United Kingdom Patel, 2007) (Czoski-Murray, Warren, Chil- Health Technology Assess- Clinical effectiveness and cost-effectiveness of pioglitazone and rosiglitazone in the cott, Beverly, Psyllaki, & treatment of type 2 diabetes: a systematic review and economic evaluation Cowan, 2004) (Cook, Yin, Alemao, Davies, Cost-Effectiveness of Ezetimibe Coadministration in Statin-Treated Patients not at Krobot, Veltri, Lipka, & Badia, Pharmacoeconomics Cholesterol Goal Application to Germany, Spain and Norway (McCormack & Foster, 2006) Pharmacoeconomics Rizatriptan A Pharmacoeconomic Review of its Use in the Acute Treatment of Migraine (Plosker & Figgitt, 2004) Pharmacoeconomics Repaglinide A Pharmacoeconomic Review of its Use in Type 2 Diabetes Mellitus (McIntosh, Conway, Willing- The cost-burden of paediatric pneumococcal disease in the UK and the potential cost- ham, & Lloyd, 2003) effectiveness of prevention using 7-valent pneumococcal conjugate vaccine (McEwan, Dixon, Babbolal, Evaluation of the Cost Effectiveness of Sirolimus versus Tacrolimus for Immunosup- Pharmacoeconomics Conway, & Currie, 2006) pression Following Renal Transplantation in the UK (De Cock, Hutton, Canney, Cost-effectiveness of oral ibandronate compared with intravenous (i.v.) zoledronic acid Body, Barrett-Lee, Neary, & Support Care Cancer or i.v. generic pamidronate in breast cancer patients with metastatic bone disease undergoing i.v. chemotherapy (Plosker & Figgitt, 2006) Pharmacoeconomics Linezolid A Pharmacoeconomic Review of its Use in Serious Gram-Positive Infections Ape
APPENDIX C: STATISTICS OF DISCOUNTING RATES FOR EFFCTS AND COSTS
Table C.1 Frequencies of discounting rates for effects and costs Frequency Percent Frequency Percent Frequency Percent Frequency Percent Frequency Percent discount rate Total 16 * 3 of the 55 studies concern same drugs. The total number of drugs in the sample amounts to 52. Ape
Table C.2 Descriptive statistics of discounting rates for effects and costs CVZ * 3 of the 55 studies concern same drugs. The total number of drugs in the sample amounts to 52. Ape
APPENDIX D: STATISTICS OF TIME-HORIZONS
Table D1 Frequencies of time-horizons Time-horizon Frequency Percent Frequency Percent Frequency Percent Frequency Percent Frequency Percent 16 100 21 100 5 100 13 100 55* 100 * 3 of the 55 studies concern same drugs. The total number of drugs in the sample amounts to 52. Ape
APPENDIX E: WELFARE GAINS RESULTS PER SAMPLE
For the sake of transparency, we have chosen to apply a uniform method and the same QALY value for all pharmaceuticals. It is possible that this uniform approach leads to an under- or overestimation of the welfare ef-fects of some individual pharmaceuticals. Results from the CVZ sample Table E.1 presents the results of the benchmark scenario using the sample collected from CVZ. The results are based on the cost-effectiveness re-sults, the incremental effects per patient, and the time-horizon reported in the CVZ-studies as well as they incorporate the user-population. Those inputs are analyzed by equation 4.1. As it appears in column concerning the benchmark scenario, the welfare gains of the selected innovative drugs are substantially positive. The welfare gains reported in QALYs ranges from approximately € -2 million to € 90 million. The total welfare gains are approximately € 213 million for the QALY estimates for the year 2006. As already explained in Section 5.2, the negative welfare effects that ap-pear in some cases imply that the respective values of a QALY used in these calculations were lower than the cost-effectiveness of the drugs. It should be also mentioned that comparisons among welfare gains of indi-vidual drugs is not possible since, there are differences in the inputs such as costs incorporated, comparator etc. It can be seen in the table that some medicines have higher welfare im-pact than others. The reason of such difference is due to the differences in cost-effectiveness, in incremental effects per user, the time-horizon used in the studies, and/or the drug-user population. For example, in the case of Tamiflu® although the time horizon is just 1 year, cost-effectiveness es-timation is low and the population uses this vaccination is substantially high. Thus, this drug has also large welfare impact. Moreover, in the lower scenario the welfare effects of the CVZ sample ap-pear to be € 55 million. That means that even reducing the value of a QALY there is still positive welfare impact of the drugs included in our sample. This is because the costs per QALY generated by these drugs are lower than the value that society is willing to pay for a QALY. Ape
Table E.1 Results from the benchmark and the lower & upper scenarios from the CVZ sample (values in thousands of euros) Lower scenario Benchmark scenario Upper scenario QALY=€ 50,000 QALY=€ 80,000 C01EB17 Procoralan J07BM01 Gardasil N.A. OncoVAX -4,150 L02BA03 Faslodex * No ATC assigned.
On the other hand, the results in the upper alternative scenario shows that
the generated welfare gains of the CVZ sample are about € 371 million.
That means that by increasing the value of a QALY from € 50,000 to
€ 80,000 the welfare effects, according to this sample, are increased sub-
stantially.
Results from the NICE/CADTH sample Similarly, Table E.2 presents the welfare effects of the innovative drugs included in the samples of NICE and CADTH. The variation in the results in the benchmark scenario is approximately from € -110 million to € 350 mil-lion. The overall welfare effects of this sample are in the benchmark sce-nario about € 1 billion. This amount is much larger than the welfare effects of the CVZ sample in the benchmark scenario. It should be taken into ac-count that the CVZ sample includes 16 drugs and the NICE/CADTH sample includes 21 drugs. Ape
When the value of a QALY is decreased to € 20,000 (lower alternative sce-nario), the welfare effects of this sample is approximately € -260 million while, when the QALY value is € 80,000 (upper alternative scenario) the total welfare effects becomes about € 2.3 billion in a year. In this sample, there are also pharmaceuticals with higher impact on the total welfare effects. These drugs are Xigris®, Lipitor®, and Zyban®, which are used in arthritis, heart diseases, and smoking addiction respectively. These drugs have low costs per QALY, large time-horizons, and relatively large user population. Ape
Table E.2 Results from the benchmark and the lower & upper scenarios from the NICE/CADTH sample (values in thousands of euros) Lower scenario Benchmark scenario Upper scenario Source ATC code Brand QALY=€ 50,000 QALY=€ 80,000 -243,057 -110,999 CADTH J05AH02 Tamiflu* CADTH L04AA11 Enbrel CADTH LO4AA12 Remicade CADTH M01AH01 Celebrex CADTH N05AX08 Risperdal * Common with the CVZ sample. Ape
Results from the literature sample The welfare effects calculated from the literature sample are presented in Table E.3. The range of the welfare effects in the benchmark scenario of this sample ranges approximately from € -3 million to € 338 million. The total welfare gains calculated from this sample are about € 409 million in year 2006. In the lower alternative scenario the welfare effects of the lit-erature sample are close to € -70 million while, in the upper alternative scenario the welfare effects become approximately € 888 million. Maxalt® has an enormous impact on the results of this sample. This drug is used in migraine treatment and has a relatively high amount of users. It also has high incremental QALY effect per person and these are the under-lying reasons for its substantial welfare impact comparing to the rest lit-erature sample. Apparently, Maxalt® is such an outlier in that sample that it determines the aggregate welfare effects of the whole sample. Without that drug the welfare effects would have substantially lowered. Table E.3 Results from the benchmark and the lower & upper scenarios from the Literature sample (values in thousands) Lower scenario Benchmark scenario Upper scenario QALY=€ 50,000 QALY=€ 80,000 C02KX01 Tracleer H01CB03 Somatuline J07AL02 Prevenar L04AA10 Rapamune Ape
APPENDIX F: DIFFERENCES CONCERNING COMMON DRUGS
Table F.1 Welfare gain differences between samples (values in thousands) Benchmark Benchmark Difference (CVZ - NICE/CADTH) Benchmark Table F.2 Input differences among institutes Incr. cost- Comparative incr. QALY time ho- discount discount effective- pharmaceutical € 8,876 acetylsalicylzuur € 9,211 aspirin -€ 7,870 Paracetamol symptomatic relief 0.009 1 N.A. N.A. -€ 14,410 placebo € 69,242 placebo * Values converted to 2006 Dutch Euro. Ape
APPENDIX G: PRICE REDUCTIONS AND INTRODUCTION DE-
LAYS
The evolution of prices of pharmaceuticals The data provided by the pharmaceutical companies, which are members of the American Chamber of Commerce, give some indication of the price change of innovative drugs over time. The total sample for this estimation consists of 18 innovative drugs. The drug-ages (years of existence in the market) are grouped in clusters of less or equal than/to 3 years in the market, four to six years and equal or more than 7 years. These clusters are chosen by considering a comparable number of drugs for each cluster. As is illustrated in Figure G.1 the average price change for the cluster ≤ 3 years is -0.07%, for the 4-6 years cluster is -3.58%, and for the cluster > 7 years the change is -3.81%. It should be mentioned that inflation is not taken into account in these changes. Thus, although the nominal price changes in the cluster of ≤ 3 years are minor, the real price reductions are considerably to be higher due to inflation. Another issue for consideration is that the price changes are, among others, influenced by the reference pricing system existing in the Netherlands (for further details on that issue see Chapters 2 and 6). Figure G.1 Average price changes of innovative drugs Price change s
c
ge -3,00%
Ape
Delays in the introduction of new pharmaceuticals Based on a sample of 13 innovative drugs included in our study, an esti-mation of the delay in the introduction of a drug to the market after the registration date can be made. These figures are presented in Table G.1. Table G.1 Delay in the introduction of new pharmaceuticals in months N: Standard deviation: As it appears, the average length of delay is 7 months relative to the reg-istration date. This value appears to be reported also in the literature as the time delay from the marketing approval until the actual reimburse-ment. A cross-national comparison shows that these delays are higher than those in the UK are, but are lower than in France (see Cohen et al., 2007). In our results the minimum delay is 1 month while the maximum delay appears to be 24 months. That means by taking the average intro-duction delay – i.e., 7 months - the foregone welfare effects based on the benchmark scenario of the total sample are approximately € 970 million (=1,664 million×7÷12). Ape

Source: http://amchampc.org/files/2012/04/Welfare-effects-of-innovative-pharmaceuticals-APE-20081.pdf