Value in Health Regional Issues Vol. 5 (focusing on Latin America)
Transcripción
Value in Health Regional Issues Vol. 5 (focusing on Latin America)
VALUE IN HEALTH REGIONAL ISSUES 5C (2014) 73–74 Available online at www.sciencedirect.com journal homepage: www.elsevier.com/locate/vhri EDITORIAL One Small Step for Latin America, One More Leap for Value in Health Regional Issues It is our pleasure to introduce this issue of Value in Health Regional Issues (ViHRI), in this case focusing on Latin America. It is now the eighth issue in our 3 years of existence, and the third looking at Latin America. Our journal is steadily growing, with articles of increasing diversity and rigor [1]. Our mission is challenging: to encourage and enhance the science of health economic and outcomes research and its use in supporting health care decisions in Asia, Latin America, Central and Eastern Europe, and Africa. To better reach our goals, we are applying to MEDLINE and are optimistic about being included in 2015. If we succeed, the journal will be indexed from its birth in 2012, and our visibility and impact will be “boosted.” Starting in 2015, the Value in Health Regional Issues Excellent Article Award will be announced and given for the best article published each year. While we await MEDLINE indexing and the eventual impact factor as a widespread, though imperfect metric for a scientific journal [2], other metrics reflect our journal’s increasing dissemination. Total ViHRI article downloads from the ISPOR Web site have steadily increased from 1,280 in 2012 to 24,113 through the first 6 months of 2014. In the case of ViHRI Latin America, we had 1,065 in 2012, 20,382 in 2013, and 5,791 through the first 6 months of 2014. From 2012 to date, we have received more than a hundred article submissions to ViHRI Latin America. Colombia (n ¼ 23), Brazil (n ¼ 21), Mexico (n ¼ 16), Chile (n ¼ 10), and Argentina (n ¼ 9) are the top countries sending their scientific output, as well as United States (n ¼ 12) though all the articles had at least one author from Latin America and the study has to be about a regional topic or population. Although two thirds of the submissions are in English, we encourage Spanish and Portuguese submissions to reduce barriers to the publication of local good quality research. Almost half the submissions are economic evaluations; other frequent study topics are patient-reported outcomes, health policy analyses, conceptual articles, and comparative effectiveness. As of October 2014, ViHRI Latin America has 7 of the 25 most cited ViHRI articles overall [3]. Our present issue builds on the previous ones [4], and is an adequate reflection of our scope, in terms of both study and research topics and geographic diversity, and it also mirrors current priority issues in health and disease burden [5]. There are several economic evaluations focusing on older and newer health technologies for noncommunicable diseases. These cover selected statins in Brazil and Colombia [6], biologics for psoriasis in Brazil [7] and Uruguay [8], as well as studies on diabetes, asthma, osteoarthritis, drug administration systems in intensive care, and costs of pesticide intoxications. Interesting and timely health policy and conceptual articles describe the process of updating the Mexican Health Care Formulary and Supply Catalog Cuadro Básico [9]; reflect on the relation of health, socioeconomic status, and migration [10]; provide guidance on adapting global pharmacoeconomic models [11]; and debate a controversial issue about doctor supply [12]. We hope you enjoy the issue. It should facilitate and contribute to evidence-based health policy and decision making in our region. Though significant heterogeneity still remains–as shown in a recent publication about national health benefit plans [13]–Latin America is increasingly using evidence/based policy to improve population health in an efficient and equitable way. Acknowledgments We thank Maria Cook and Jerusha Harvey from ISPOR for their support and review of the editorial and Terry Materese from Elsevier for providing insight about journal metrics. Federico Augustovski, MD, MSc, PhD Faculty of Medicine, School of Public Health, University of Buenos Aires, Buenos Aires, Argentina Marcos Bossi Ferraz, MD, PhD São Paulo Center for Health Economics (GRIDES), São Paulo, Brazil Department of Medicine, Sao Paulo School of Medicine, Federal University of São Paulo, São Paulo, Brazil J. Jaime Caro, MDCM, FRCPC, FACP Evidera, Lexington, MA, USA Faculty of Medicine, McGill University, Montreal, QC, Canada Victor Zárate, MD, MSc Faculty of Medicine, Los Andes University, Santiago, Chile 2212-1099/$36.00 – see front matter Copyright & 2014, International Society for Pharmacoeconomics and Outcomes Research (ISPOR). Published by Elsevier Inc. http://dx.doi.org/10.1016/j.vhri.2014.10.002 Conflict of interest: The authors have indicated that they have no conflicts of interest with regard to the content of this article. Source of financial support: The authors have no other financial relationships to disclose. 74 VALUE IN HEALTH REGIONAL ISSUES 5C (2014) 73–74 R EF E R EN CE S [1] Yang BM, Lee K, Guo J. A year of significant progress for Value in Health Regional Issues. Value Health Regional 2014;3C:172–3. [2] Bornmann L, Marx W, Gasparyan AY, Kitas GD. Diversity, value and limitations of the journal impact factor and alternative metrics. Rheumatol Int 2012;32:1861–7. [3] Most cited Value in Health Regional Issues articles. Available from: http:// www.journals.elsevier.com/value-in-health-regional-issues/ most-cited-articles/. [Accessed October 4, 2014]. [4] Augustovski F, Caro JJ, Zárate Barahona V, Bosi Ferraz M. Introducing the new Value in Health Regional Issues journal: the Latin America region edition. Available from: http://www.ispor.org/ValueInHealth/ShowValueInHealth.aspx?issue=722CCE35-54A6-4927-84A2-F538E53129D3. [Accessed October 6, 2014]. [5] Murray CJ, Vos T, Lozano R, et al. Disability-adjusted life years (DALYs) for 291 diseases and injuries in 21 regions, 1990–2010: a systematic analysis for the Global Burden of Disease Study 2010. Lancet 2012;380:2197–223. [6] Mould-Quevedo JF, Gutierrez Ardilia MV, Ordoñez JE, Vargas Zea N. Cost-effectiveness analysis of atorvastatin versus rosuvastatin in primary and secondary cardiovascular prevention populations in Brazil and Colombia. Value Health Regional 2014;5C:48–57. [7] Salgado Riveros B, Klarmann Zeigerlmann P, Correr CJ. Costeffectiveness of biologic agents in the treatment of moderate to severe psoriasis: a Brazilian Public Health Service (SUS) perspective. Value Health Regional 2014;5C:65–72. [8] Gonzalez-Vacarezza N, Gehling Bertoldi E, Deminco A, Martinez Asuaga M. Análisis costo-utilidad del uso de adalimumab, etanercept e infliximab para el tratamiento de la artritis psoriásica en Uruguay [Spanish]. Value Health Regional 2014;5C:58–64. [9] Rizo Rios P, Gonzalez Rivera A, Rivas Oropeza I, Campos Ramirez O. The update of the health care formulary and supply catalog in the context of the health technology assessment. Value Health Regional 2014;5C:29–34. [10] Cabieses B. La compleja relación entre posición socioeconómica, estatus migratorio y resultados de salud en Chile: discusión crítica de la evidencia científica [Spanish]. Value Health Regional 2014;5C:1–6. [11] Mullins D, Onwudiwe NC, Tannus G, et al. Guidance document: global pharmacoeconomic model adaption strategies. Value Health Regional 2014;5C:7–13. [12] Westphalen AC, Stefani SD. More doctors: thoughts about a controversial healthcare policy. Value Health Regional 2014;5C:73–5. [13] Health benefit plans in Latin America: a regional comparison. In: Giedion U, Bitrán R, Tristao I,eds. Inter-American Development Bank Social Protection and Health Division. Washington, DC: Inter American Development Bank, 2014. VALUE IN HEALTH REGIONAL ISSUES 5C (2014) 75–77 Available online at www.sciencedirect.com journal homepage: www.elsevier.com/locate/vhri EDITORIAL More Doctors: Thoughts about a Controversial Health Care Policy The recent economic downfall has affected health care systems worldwide, sparking renewed discussions about the need for major reforms in many countries. In the wake of social protests, the Brazilian government swiftly proposed reforms addressing the poor distribution and shortage of medical personnel in underserved communities. The topic has gained substantial laymedia attention, but most debates have been driven by political motivations, rather than consensus of public health experts. We use the case of Brazil to illustrate potential problems of hasty efforts to solve systemic issues without addressing the underlying causes that have led to insufficient training, scarce infrastructure, and lack of public engagement on decision making. Brazil, a country of 198,000,000 people, has a decentralized universal health care system in which states and municipalities manage health care services [1,2]. The “Unified Health System” is considered by many an exemplary model [3]. The country has a broad free vaccination program, and nearly all children receive the full schedule [3,4]. It hosts an internationally renowned program that provides universal access to HIV prevention and treatment services [3,5]. And recent substantial investments have been effective in decreasing cases of tuberculosis [6]. Maintenance of this large enterprise, however, is costly, and federal health care spending is minimal [3]. Despite an increase in the per capita expenditure on health from US $107 to US $466, it is still below the world average of US $599 [4]. In Latin America, Brazil is behind Argentina (US $489), Costa Rica (US $553), and Uruguay (US $619) [4]. Compared with developed countries, the gap is abysmal; for instance, The Netherlands spends US $4820 per person [4]. According to the World Health Organization, health care spending in Brazil represents approximately 9.0% of its gross domestic product [4], yet half of the expenditures predominantly benefit 20% of the population with access to private health plans [7]. One of the indexes the World Health Organization uses to determine whether the health workforce is sufficient to address a country’s needs is the physician density, expressed by the number of physicians per 10,000 people. In Brazil, there are 17.6 physicians per 10,000 people, a figure slightly above the world average [4]. Although this number could be larger, only two countries in South America have a greater physician density, Uruguay (37.4/10,000) and Argentina (31.6/10,000) [4,8]. The physician density in the United States is 24.2/10,000, only a few positions above Brazil in the world ranking [4]. Yet most of the Brazilian physicians are located in or around major centers, especially in southern regions [9]. And 10 of the 26 states have less than 10 physicians per 10,000 habitants [9]. This phenomenon, however, is not exclusive to Brazil; about 90% of American physicians are located in regions with a high population density and economic importance [10]. The distribution of health care providers was studied by professor Malcolm Brown, from the University of Calgary, in Canada [11]. He confirmed what was already suspected, the lowest practitioner to population ratio is found in rural areas and the highest in large urban areas. More importantly, though, he suggests that these distributions are highly stable and difficult to change through normal types of public policy interventions. Similarly, Gravelle and Sutton [12] published in 1998 a short report in which they discuss the inequality in general practitioners provision in England and the importance of adequate metrics of physician distribution. On June 21, 2013, after days of protests, Brazilian President Dilma Rousseff spoke to the country [13]. Among various reforms, she announced that thousands of foreign doctors would be recruited to expand the public health system to underserved communities. Brazilian medical institutions immediately positioned themselves against such measure, arguing that the problem is not a shortage of physicians but a distribution inequity that results from poor working conditions in smaller towns. Soon an alternate plan, “More Doctors,” was implemented by the Ministry of Health to recruit 15,460 physicians for a period of 3 to 6 years [14]. This represents a 4% increase in the total number of physicians, and all should be allocated to areas where the greatest impact in public health is expected [4,9,14]. Although this new program gives preference to national graduates, foreign graduates are allowed to work in the country without medical training recertification [15]. Most of these physicians come from Cuba, through an agreement signed by the government of both countries [16]. Similar projects have been implemented in Venezuela, “Mission Inside the Neighborhood,” and other Latin American and Caribbean countries, “Operation Miracle” [17,18]. Heated debates have ensued following the announcement of More Doctors. Local medical groups accused the government of being driven by politics and ideology, rather than by an interest in the population’s welfare. Most of the Brazilian physicians agree that many areas of the country are underserved, but the general consensus is that the proposed short-term solution is prone to failure. The suggested alternative solution is the development of a long-term plan of substantial investments in basic infrastructure and personnel, along with the development of a career plan for health professionals employed by the government [19]. However, the government denies playing politics because the need for physicians is real, including in the outskirts of metropolitan areas. Furthermore, the authorities refute other criticisms on the basis of the fact that more than 60% of the physicians enrolled in the program are Brazilian graduates and that foreign physicians are required to attend a 2-week preparatory course and pass a medical knowledge examination. Finally, these Conflicts of interest: The authors have indicated that they have no conflicts of interest with regard to the content of this article. Source of financial support: The authors have no other financial relationships to disclose. 76 VALUE IN HEALTH REGIONAL ISSUES 5C (2014) 75–77 physicians are not allowed to practice outside the allocated area and, after 3 years, they must validate their medical license to stay in Brazil. Despite these measures, questions remain. More than 90% of foreign graduates passed the examination, which has approval and conditioning thresholds of 50% and 30%, respectively. This pass rate is significantly higher than those seen in other similar tests. In 2012, for example, 55% of the applicants failed an examination with a 60% approval threshold applied by the Medical Board of São Paulo [20]. Although the test was mandatory for obtaining the state medical license, its results had no influence on certification. Brazilian graduates are not required to undergo examinations such as the United States Medical Licensing Examination, and this has undermined the arguments of medical groups and reinforced the government’s actions. Also of much significance are legal matters because today it is unclear how much liability international graduates, in particular foreign citizens, have in case of misconduct and malpractice [21]. Finally, it is possible that many of these physicians, if unable to obtain recertification, might resort to illegal practice or legal motions to continue working in Brazil after leaving the program. Reviving the Brazilian health care system will require more than a temporary increase in the number of physicians in the medically deprived rural areas and slums. Makaroff et al. [22] recently discussed the limitations of simply increasing the number of physicians in the United States as a solution for the poor distribution of resources. The problem is multifaceted and includes poor management, inadequate distribution of funds and human resources, and lack of infrastructure. Although the solution to these various problems is not simple, the first step toward this goal is better understanding of their causes. Data collection efforts should be planned and processed through the use of internationally standardized classifications (to enhance crossnational comparability) and should be a collaborative action of all the health system stakeholders, including providers, government, nongovernmental organizations, and international agencies [23]. The number of possible indicators needed to describe the profile and monitor the progress of any program is extensive. For example, Buchan [24] recently published a study describing the benefits and limitations of measuring personnel “turnover” and “stability,” both correlated with cost reduction, productivity improvement, and better care outcomes. And Russell et al. [25] demonstrated that population size and geographical locations are strongly associated with the retention of family physicians in underserved areas of Australia. These same types of data can and must be used in the case of Brazil, or of any other country facing similar problems. These programs, however, are not static processes, but dynamic and with multiple parts that may or may not achieve their intended objectives. Accordingly, in addition to wisely choosing the variables that will measure the degree of success of any implemented policies, models such as the plando-study-act cycles should be used to assess, change, and reassess any proposed efforts [26]. The utilization of international personnel is adequate and comprehensible in emergency situations. Foreign physicians have had a significant impact in the aftermath of Haitian earthquakes in 2010 [27]. And Doctors Without Borders and other relief groups constantly deploy for temporary missions [28]. It is beyond the scope of this reflection, however, to determine whether the Brazilian situation represents an emergency. The fact is that rather than engaging all stakeholders, the government imposed a program through an executive order [15]. Now it is time for both sides to come together and plan for the future, as the nation faces a lack of well-defined long-term projects tied to More Doctors. This episode should serve as a reminder of the importance of continuous discussions among national and international representatives of the health care community. In a world that is rapidly changing, new and updated principles, guidelines, and regulations for the delivery of adequate services are crucial. Antonio C. Westphalen, MD, PhD Department of Radiology and Biomedical Imaging, University of California, California, San Francisco Stephen D. Stefani, MD Mãe de Deus Cancer Institute, Unimed Foundation, Porto Alegre, RS, Brazil 2212-1099/$36.00 – see front matter Copyright & 2014, International Society for Pharmacoeconomics and Outcomes Research (ISPOR). Published by Elsevier Inc. http://dx.doi.org/10.1016/j.vhri.2014.10.001 R EF E R EN C ES [1] 2010 Census [in Portuguese]. Rio de Janeiro, Brazil: Brazilian Institute of Geography and Statistics, 2010. [2] Unified Health System (SUS). Portal Brasil. Brasilia, Brazil: Presidential Office of Social Communications, 2010. [3] Kepp M. Cracks appear in Brazil’s primary health-care programme. Lancet 2008;372:877. [4] World Health Statistics 2013. Geneva, Switzerland: World Health Organization, 2013. [5] AIDSInfo Brazil. Geneva, Switzerland: Joint United Nations Programme on HIV and AIDS, 2011. [6] Global Tuberculosis Report 2012 - Annex 2 - Country Profiles. Geneva, Switzerland: World Health Organization, 2012. [7] Abdala V. Public sector represents only 42% of expenses with healthcare in the country [in Portuguese]. Rio de Janeiro, Brazil: Agência Brasil, 2013. [8] Argentina. The World Fact Book. Washington: Central Intelligence Agency, 2013. [9] Scheffer M, Cassenote A, Biancarelli A. Medical demographics in Brazil [in Portuguese]. São Paulo, Brazil: Cenários e Indicadores de Distribuição, Brazilian Federal Board of Medicine/Medical Board of the State of São Paulo, 2013. [10] Boukus ER, Cassil A, O’Malley AS. A snapshot of U.S. physicians: key findings from the 2008 Health Tracking Physician Survey. Data Bull (Cent Stud Health Syst Change) 2009;35:1–11. [11] Brown MC. Using Gini-style indices to evaluate the spatial patterns of health practitioners: theoretical considerations and an application based on Alberta data. Soc Sci Med 1994;38:1243–56. [12] Gravelle H, Sutton M. Trends in geographical inequalities in provision of general practitioners in England and Wales. Lancet 1998;1910:352. [13] Brazilian President’s Dilma Rousseff Address to the Nation [in Portuguese]. Portal do Planalto. Brasília, Brazil: Office of Social Communications, 2013. [14] Health Portal News [in Portuguese]. Portal da Saúde - SUS. Brasília Brazil: Communications Office, Ministry of Health, 2013. [15] Executive Order 621 [in Portuguese]. Brasília, Brazil: Legal Affairs, Presidential Executive Office, 2013. [16] Palma N. (Ministry of) Health signs agreement with PAHO to attract more physicians [in Portuguese]. Portal da Saúde - SUS. Brasília, Brazil: Communications Office, Ministry of Health, 2013. [17] Westhoff WW, Rodriguez R, Cousins C, et al. Cuban healthcare providers in Venezuela: a case study. Public Health 2010;124: 519–24. [18] Zakrison TL, Armada F, Rai N, et al. The politics of avoidable blindness in Latin America–surgery, solidarity, and solutions: the case of Misión Milagro. Int J Health Serv 2012;42:425–37. [19] Federal Board of Medicine. Medical Boards Campaign for the Development of a Career Plan [in Portuguese]. Brasília, Brazil: Brazilian Federal Board of Medicine, 2013. [20] São Paulo Board of Medicine. Results of CREMESP Examination [in Portuguese]. São Paulo, Brazil: Medical Board of São Paulo, 2012. VALUE IN HEALTH REGIONAL ISSUES 5C (2014) 75–77 [21] Supreme Court of Brazil. “More Doctors” is Challenged in the Supreme Court [in Portuguese]. Notícias STF, Supreme Federal Court of Brazil, 2013. [22] Makaroff LA, Green LA, Petterson SM, et al. Trends in physician supply and population growth. Am Fam Physician 2013;87. [23] Diallo K, Zurn P, Gupta N, et al. Monitoring and evaluation of human resources for health: an international perspective. Hum Resour Health 2003;1:3. [24] Buchan J. Reviewing the benefits of health workforce stability. Hum Resour Health 2010;8:29. 77 [25] Russell DJ, Humphreys JS, McGrail MR, et al. The value of survival analyses for evidence-based rural medical workforce planning. Hum Resour Health 2013;11:65. [26] Gillam S, Siriwardena AN. Frameworks for improvement: clinical audit, the plan-do-study-act cycle and significant event audit. Qual Prim Care 2013;21:123–30. [27] Burnett J. Cuban doctors unsung heroes of Haitian earthquake. Weekend Edition NPR News 2010. [28] Médecins Sans Frontières Activity Report 2011. Doctors Without Borders, 2011. VALUE IN HEALTH REGIONAL ISSUES 5C (2014) 29–34 Available online at www.sciencedirect.com journal homepage: www.elsevier.com/locate/vhri The Update of the Mexican Health Care Formulary and Supply Catalog in the Context of the Health Technology Assessment Pedro Rizo Ríos, MD, MHSM1,*, Aurora González Rivera, MD2, Itzel Rivas Oropeza, MScHE1, Odette Campos Ramírez, MD1 1 Consejo de Salubridad General, Ciudad de México, México; 2Instituto Nacional de Pediatría, Ciudad de México, México AB STR A CT One of the instruments Mexico has available for the optimization of resources specifically allocated to health technologies is the Health Care Formulary and Supply Catalog (Cuadro Básico y Catálogo de Insumos del Sector Salud [CBCISS]). The aim of the CBCISS is to collaborate in the optimization of public resources through the use of technologies (supplies) that have proven their safety, therapeutic efficacy, and efficiency. The importance of the CBCISS lies in the fact that all public institutions within the National Health System must use only the established technologies it contains. The implementation of strategies that strengthen the CBCISS update process allows it to be thought of as an essential regulatory tool for the introduction of health technologies, with relevant contributions to the proper selection of cost-effective interventions. It ensures that each supply included on the list meets the criteria sufficient and necessary to ensure efficacy, safety, effectiveness, and, of course, efficiency, as evidence supporting the selection of suitable technologies. The General Health Council (Consejo de Salubridad General [CSG]) is a collegial body of constitutional origin that—in accordance with its authority— prepares, updates, publishes, and distributes the CBCISS. To perform these activities, the CSG has the CBCISS Inter-institutional Commission. The CBCISS update is performed through the processes of inclusion, modification, and exclusion of supplies approved by the Interior Commission. The CBCISS update process consists of three stages: the first stage involves a test that leads to the acceptance or inadmissibility of the requests, and the other two focus on an indepth evaluation for the ruling. This article describes the experience of health technology assessment in Mexico, presents the achievements and outlines the improvements in the process of submission of new health technologies, and presents a preliminary analysis of the submissions evaluated until December 2012. During the analysis period, 394 submissions were received. After confirming compliance with the requirements, 59.9% of the submissions passed to the next stage of the process, technology assessment. In the third stage, the committee approved 44.9% of the submissions evaluated. The improvements established in the country in terms of health technology assessment allowed choosing the technologies that give more value for money in a context of public health institutions. Keywords: efficiency, health care, health technologies, health technology assessment. Introduction One of the instruments Mexico has available for the optimization of resources for HTs is the Health Care Formulary and Supply Catalog (Cuadro Básico y Catálogo de Insumos del Sector Salud [CBCISS]) [1], and as such, the strategies developed to strengthen it must be favorable in terms of meeting the growing demands for technology and health care services in a difficult economic and social environment. The search for equity, quality of care, and efficiency has been one of the challenges facing public health systems (PHSs) worldwide. In that context, health technology assessment plays a significant role in gaining greater benefits for the health care field and planning of the rational use of resources. As a result, recent decades have been characterized by significant progress in the supply and availability of new health technologies (HTs). Many of these technologies have budgetary implications owing to the high costs of their incorporation. Despite the benefits offered by these new technologies as compared with existing ones, their incorporation does not always solve the population’s major health issues and they are not effectively accessed by all public sector institutions. Copyright & 2014, International Society for Pharmacoeconomics and Outcomes Research (ISPOR). Published by Elsevier Inc. The Health Care Formulary and Supply Catalog The CBCISS is a document that includes, defines, and encodes all medical supplies used by public institutions to provide health services to the population. Conflict of interest: The authors have indicated that they have no conflicts of interest with regard to the content of this article. * Address correspondence to: Pedro Rizo Ríos, Dirección de Priorización en el Consejo de Salubridad General, Jesús Lecuona Manzana 110 Lote 2-A, Colonia Ampliación Miguel Hidalgo, Delegación Tlalpan, C.P. 14250, México, D.F., Mexico. E-mail: [email protected]. 2212-1099$36.00 – see front matter Copyright & 2014, International Society for Pharmacoeconomics and Outcomes Research (ISPOR). Published by Elsevier Inc. http://dx.doi.org/10.1016/j.vhri.2014.08.002 2 VALUE IN HEALTH REGIONAL ISSUES 5C (2014) 29–34 Its origins date to 1846, the year the first Mexican Pharmacopeia [2] was published, and its development over time has set the standards for PHSs. In 1975, a presidential decree [3] was published that established that all public health institutions must have a drug formulary (Cuadro Básico de Medicamentos [CBM]). That same year, the Public Sector Drug Formulary Commission was formed, responsible for the development of the first sectoraltype CBM in 1977 [4]. Later, in 1983, biological products and laboratory reagents, instruments and medical equipment, wound dressing materials, and prosthetics were added to the CBM [5]. In 1996, for the purpose of facilitating information about the supplies used in the support of diagnosis and treatment, the CBM was divided into two parts, on the basis of the type of medical unit: one referred to as formulary for primary care and the other called catalog for secondary and tertiary care [6], just as we know it today. The aim of the CBCISS is to ensure the optimization of public resources through the use of technologies (supplies) that have proven their safety, therapeutic efficacy, and efficiency. The importance of the CBCISS lies in the fact that all public institutions within the National Health System must use only the established technologies it contains. To reinforce its compulsory nature, the bylaws of the CBCISS Inter-institutional Commission were published in the Official Gazette of the federation in February 2003, which specify the powers of the Inter-institutional Commission and Specific Technical Committees (Comité Técnico Específico [CTEs]) as well as institutional representatives. In addition, it describes the CBCISS update procedure [7]. efficiency. On the one hand, these strategies help to overcome the inefficient use of limited resources, as one of the main obstacles identified in the health world, and on the other hand, they align the efforts of the institutions and stakeholders involved in the process, avoiding the duplication of effort in the initial evaluation. The General Health Council (Consejo de Salubridad General [CSG]) is a collegial body of constitutional origin that—in accordance with its authority—prepares, updates, publishes, and distributes the Formulary for the primary health care level and the Supply Catalog for the secondary and tertiary levels. To perform these activities, the CSG has the CBCISS Interinstitutional Commission, which is composed of representatives of the Ministry of Health, the Mexican Social Security Institute, the Government Employees’ Social Security and Services Institute, the National System for Integral Family Development, the Ministry of National Defense, the Secretariat of the Navy, Petróleos Mexicanos, and the Federal District’s Ministry of Health. For the purposes of updating the CBCISS, the commission has a Technical Secretariat under the Directorate-General for Prioritization and Specific Technical Committees (CTE) for medicines, wound dressing materials, diagnostic aids, and instruments and medical equipment, in addition to the Technical Committees for herbal remedies, homeopathic medicines, and acupuncture supplies, which were recently formed as part of the strategy to strengthen the upgrade process. Each CTE consists of representatives of the commission’s full members and is coordinated by its technical secretary. Strategies for Strengthening the Update Process Update Process We know that as a country becomes more developed, the burden and weight of disease increase, especially for noncommunicable diseases that accompany the epidemiological transition. This translates into new health care needs and causes the development of new technologies that revolutionize medical practice and contribute to the challenge facing the PHSs, which is to improve the quality of care and supplies, making them safe, effective, efficacious, and efficient services and supplies. Against this backdrop, the CBCISS Inter-institutional Commission of the General Health Council, through the Directorate for Prioritization, was given the task of strengthening the CBCISS update process, implementing strategies that contribute to the effective and efficient use of technology in PHSs. Some of these strategies were as follows: The CBCISS update is performed through the processes of inclusion, modification, and exclusion of supplies approved by the Interior Commission. Update requests are received at the CSG during the first five working days of each period (January–April, May–August, September–December), in accordance with the bylaws, and they are passed to the Directorate General for Prioritization, which, in turn, acts as the Technical Secretariat of the Commission to initiate the review and assessment process. According to what was instituted by the new bylaws, CBCISS update requests related to supplies may be made by public institutions that serve as health care providers; scientific organizations; academies and specialty boards; government institutions; members of the commission; the secretary and the chairman of the General Health Council; and providers or manufactures of technology, who, to date, make the majority of requests. The organization of a structure dedicated to the assessment and evaluation of the information contained in the requests, primarily those that include HTs. The amendment and approval of the bylaws of the CBCISS Inter-institutional Commission, published in the Official Gazette in June 2011 [8], to have a legal regulatory framework that supports the process. The bylaws contain seven sections, which provide for the activities and responsibilities of the commission, the requirements that CBCISS update requests must meet, and the timing thereof. The development of the Guidelines for the Evaluation of Supplies (Guía de Evaluación de Insumos [GEI]) [9] with the application of standardized criteria based on scientific methods to evaluate the clinical, epidemiological, and economic evidence from the studies presented in the update proposals and the process to obtain a ruling. The strategy consists of three key, interconnected themes for the update process: transparency, scientific evidence, and Stages of the Upgrade Process The CBCISS update process consists of three stages. The first stage involves a checklist that leads to the acceptance or inadmissibility of the requests, and the other two focus on an in-depth evaluation for the ruling. The first stage, called review and assessment of update requests, leads to verification of compliance by the Technical Secretariat with the requirements instituted in the bylaws (explicitly expressed, for the most part, in Article 28), to ensure integrity, internal consistency, and accuracy: Request indicating generic name, code, and reasons for the update request; Current Health Registration issued by the a decentralized organ of the Department of Health with technical, VALUE IN HEALTH REGIONAL ISSUES 5C (2014) 29–34 administrative, and operational autonomy, whose mission is to protect the population against sanitary risks, through sanitary regulation Federal Commission for the Protection against Sanitary Risk (Comisión Federal para la Protección contra Riesgos Sanitarios, [COFEPRIS]), certified by a public notary, except for supplies that do not require it, as defined by COFEPRIS; Declaration that the manufacturer has a technovigilance or pharmacovigilance system and immediate notification in accordance with official regulations of the country (NOM220-SSA1-2004) [10]; Complete economic evaluation (EE) study as defined in the Guidelines for Conducting Economic Evaluation Studies for the Update of the Health Care Formulary (Guía para la conducción de estudios de evaluación económica para la actualización del Cuadro Básico de Insumos del Sector Salud en México [GCEEE]); Scientific studies that support the benefits of the proposed technology over other existing technologies with the same indications in the CBCISS and/or Clinical Practice Guidelines, as well as the values used to populate the economic models; Document with the generic proposal for the technology requested for inclusion, in the correct format; A sworn statement that the supply item does not infringe on patents nor is it under dispute and that the information is accurate; Budget impact analysis for cases of low-incidence diseases with social consequences (Article 29). When it is determined that a request meets the requirements, the applicant is notified in writing of the decision to continue the process, and the approved requests, the accompanying documentation, and the review and assessment form for the requests are sent to the appropriate committee. In the second stage, known as assessment of the evidence, the CTEs and the secretariat begin to get involved. In this stage, a critical analysis of the clinical epidemiological studies is performed, using instruments and forms developed by organizations specializing in the topic (as established in the GEI), to evaluate the methodological aspects that determine the internal and external validity, as well as the risks and benefits associated with its use for health care in the target population, with the corresponding direct impact on public resources. With respect to the critical reading of EE studies, the following aspects are considered: Is this economic evaluation valid? How are costs evaluated and compared with the effects or consequences? Will the results help optimize the use of public resources available for the funding of health interventions? These questions allow the analysis to focus considerably on the robustness of the structure of the economic models and their uncertainty, the plausibility of the data used to feed into the models, the probability of making erroneous decisions and their consequences, and the applicability of the results to the target population. The EE allows for a comparison of supplies, in terms of not only the costs they generate but also the health effects they produce. In this sense, the use of EE evidence seeks to improve efficiency in the allocation of limited resources, systematically selecting the safest, most effective, and least expensive option among alternatives that compete for the same resources. This evidence, however, presents limitations in terms of fully evaluating the allocation of resources and may be seriously affected by the models used or the necessary assumptions made when building them. When the time comes to take decisions, the members of the committee must be aware of these limitations and the uncertainty associated with results of this type of evaluation. 3 The EE study as a requirement for updating the CBCISS was established by the Inter-institutional Commission in May 2003, prompting the development and publication of the Guidelines for Conducting Economic Evaluation Studies for the Update of the Health Care Formulary (GCEEE) in 2008 [11]. Since then, these guidelines have made it possible to carry out a methodological review consistent with standardized criteria from an economic perspective. For the third stage of the ruling, the CTEs convene on a regular basis (monthly), seeking consensus to issue a unanimous decision, considering the opinions and recommendations of each institution’s evaluator and, where appropriate, expert opinions. Essentially, the conclusions and recommendations are backed by prior review and analysis of the quality of clinical and economic evidence; the balance between the risks, benefits, and costs that the new technology represents in comparison to existing alternatives; the scientific basis of the Clinical Practice Guidelines in Mexico and/or other countries where its use is recommended; the consideration of institutional resources available for the intervention; and any other situations relevant to the social and ethical context. If any of the requests do not meet the established criteria, the technology will not be considered for the CBCISS update. Through its CTEs, however, the commission has the authority to make exceptions when there are reasonable grounds to do so. These grounds include the needs of patients with the disease that the supply will target; national health programs aligned to priority public health needs that need technology to operate and must be addressed in a timely manner; the effect that the new supply will have on the organization of health care services at the institutions; and the potential long-term benefit that the addition of the supply may have. The evaluation process culminates during this stage with the deliberation and ruling that occurs during the committee meeting, taking into consideration the recommendations reached through consensus, taking 90 calendar days from the date of submission, to give a resolution. The Technical Secretariat prepares the supply item’s descriptive document for the applicable requests, and once approved by the CTE, it is included in the CBCISS update project that is made available for 10 calendar days on the CSG Web site for comments from stakeholders. After this time, the comments received are addressed and the update is then published in the Official Gazette of the federation. The update takes effect the day after its publication, and the process is complete. Preliminary Results To describe the experience gained, evaluate the achievements made, and improve the process, a preliminary analysis of the results obtained of four periods from September 2011 has been performed. From September 2011 to December 2012, 394 update proposals were received over four periods (September–December 2011, January–April, May–August, and September–December 2012). All the requests received were reviewed and evaluated by the commission’s Technical Secretariat within the prescribed period of 15 days and using the criteria established in the bylaws and the GEI. Of the requests received, 62% were for medicines, 22% for wound dressing materials, 8.4% for instruments and medical equipment, and 7.1% for diagnostic aids (Fig. 1). Within each period for receiving requests, the constant number of requests received for medicines (mean ¼ 61) and diagnostic aids (mean ¼ 7) stands out. In contrast, there was an increasing trend in terms of the number of requests received for 4 VALUE IN HEALTH REGIONAL ISSUES 5C (2014) 29–34 33 (8%) CBCISS update requests received 28 (7%) Medicines Wound dressing materials 89 (23%) Diagnostic aids 244 (62%) Instruments and medical equipment Fig. 1 – CBCISS update requests received from September 2011 to December 2012. CBCISS, Health Care Formulary and Supply Catalog. Source: Database of the Commission’s Technical Secretariat [12]. methodology based on critical analysis of the scientific evidence present in the clinical, epidemiological, and economic studies that accompanied each update proposal, applying at all times transparency, consistency, and legitimacy backed by a legal framework, as established by an amendment to the commission’s bylaws. At the same time, it defines the rights and responsibilities of the shareholders involved, as well as the guidelines and basic principles that direct them, with the opportunity to take decisions in a reasonable time frame. The implementation of strategies that strengthen the CBCISS update process allows it to be thought of as an essential regulatory tool for the introduction of health technologies, with relevant contributions to the proper selection of cost-effective interventions. It ensures that each supply included on the list meets the criteria sufficient and necessary to ensure efficacy, safety, effectiveness, and, of course, efficiency, as evidence supporting the selection of suitable technologies. Besides stimulating the optimization of public resources allocated to the country’s health issues, the CBCISS, as a reference tool, offers the opportunity to have a single system of classification and coding of supplies at the Federal Public Administration, thus contributing to the standardization of purchasing policies for the public institutions within the National Health System. wound dressing materials (10–33) and instruments and medical equipment (7–19). Furthermore, a global average of 98 requests was determined for the four basic lists (Fig. 2). After confirming compliance with established requirements in due time and manner, 59.9% of the requests reviewed passed on to the next stage of the process, the evaluation, which was held jointly by the CTEs and the Technical Secretariat. At this stage, an increasing trend in the number of requests evaluated was noted. For the list of medicines, the increase from the first period to the fourth was from 41% to 75%; for wound dressing materials, the increase was from 30% to 76%; for diagnostic aids, it was from 25% to 50%; and for instruments and medical equipment, it was from 42% to 81% (Fig. 3). In the third stage, the committee approved 44.9% of the requests evaluated over the four update periods (Fig. 4), which corresponded to 28% for medicines, 25% for wound dressing materials, 21% for diagnostic aids, and 27% for instruments and medical equipment. Discussion The strengthening of the selection process for technologies to be included in the CBCISS was performed under a rigorous CBCISS update requests received 80 70 70 60 56 60 58 50 40 33 26 30 20 10 20 10 8 7 19 8 3 6 6 4 0 Sept.-Dec. (2011) Jan.-Apr. May.-Aug. Sep.-Dec. Medicines Wound dressing materials Diagnostic aids Instruments and medical equipment Fig. 2 – CBCISS update requests received by the Technical Secretariat. CBCISS, Health Care Formulary and Supply Catalog. Source: Database of the Commission’s Technical Secretariat [12]. 5 CBCISS upgrade requests evaluated by Specific Technical Committee VALUE IN HEALTH REGIONAL ISSUES 5C (2014) 29–34 50 45 40 35 30 25 20 15 10 5 0 44 41 38 29 25 19 11 3 2 6 3 Sept.-Dec. (2011) 5 3 Jan.-Apr. 3 3 1 May.-Aug. Sep.-Dec. Medicines Wound dressing materials Diagnostic aids Instruments and medical equipment Fig. 3 – CBCISS upgrade requests evaluated by the Specific Technical Committee. CBCISS, Health Care Formulary and Supply Catalog. Source: Database of the Commission’s Technical Secretariat [12]. CBCISS update requests approved Furthermore, the development and implementation of the GEI, in addition to the GCEEE, with strict adherence to the bylaws, specifically Section VI “On the Procedure for Updating Supplies,” allows for the incorporation of scientific evidence of technological advances in medical care, ensuring its quality. It also allows the efforts of different institutions involved in updating the CBCISS to converge, align, and complement each other, avoiding duplication of effort. It uses standardized criteria based on the health needs of the Mexican population, and its use also favors transparent processes. With the results obtained, we note that the highest percentage of update requests are for medicines and that while requests for diagnostic aids, instruments and medical equipment, and wound dressing materials are minimal, the trend is growing, especially for wound dressing materials. In addition, it was noted that of the 394 total requests received during the four periods, only 59.9% were accepted and allowed to continue with the upgrade process, and so, the question arises as to what were or are the possible reasons for not meeting the explicit requirements set forth in the bylaws and GEI. In analyzing this very event, by observing what occurred period by period, it seems that the trend of requests accepted 20 and sent to the second stage (evaluation) is on the rise, especially for medicines and wound dressing materials. It should be mentioned, however, that after observing the results from the first period (September–December 2011), after the publication of the new bylaws as “operational policy,” the applicant was allowed to respond to the observations of the Technical Secretariat by resubmitting the request with the appropriate changes, within a maximum period of 72 hours, beginning upon receipt of the initial response. It is also important to detail that after receiving the response (amended request), the Technical Secretariat reviews and assesses it and then issues a definitive answer. If a reply is not received from the applicant within the stipulated time frame, the Technical Secretariat’s first response is considered final. With this “policy,” the increase in applications that advance to the evaluation stage has been notable, but of these, only 44.9% are ruled as applicable to the CBCISS update, a fact that will be discussed in a later analysis. There is no doubt that with the strategies developed, there has been progress in the implementation of the project to strengthen the update process with a regulatory framework and rational use of HTs, promoting transparency, quality of evidence, and efficiency of the national regulatory authority. 21 19 16 15 15 13 10 7 4 5 2 0 1 3 2 0 1 1 1 0 Sept.-Dec. (2011) Jan.-Apr. May.-Aug. Sep.-Dec. Medicines Wound dressing materials Diagnostic aids Instruments and medical equipment Fig. 4 – CBCISS update requests approved. CBCISS, Health Care Formulary and Supply Catalog. Source: Database of the Commission’s Technical Secretariat [12]. 6 VALUE IN HEALTH REGIONAL ISSUES 5C (2014) 29–34 While the health technology assessment was considered a fundamental component of the CBCISS update process in terms of decision making and the ruling, EE studies were taken as part of the informational process and a support tool rather than a decision-making method for selecting the HT, clarifying that the main theme of the process was not “if one had to choose” but rather “how to make the decision to choose” a proposal to be accepted for the CBCISS update. In conclusion, it is important to clarify that we are immersed in a dynamic process in which weaknesses can be observed that give us the opportunity to improve the process on a periodic basis and/or make the necessary changes to establish a balance between the costs and benefits of HTs, where price is determined by the value of health and the value of health is determined by the quality of the evidence and the ability to determine adequate health spending. Although we do not know how much health spending is enough to meet our endless health needs when resources are in short supply, by setting cost-effectiveness thresholds, such as gross domestic product per capita, a balance can be achieved between the sectors involved and the needs of society aimed at real population's health improvement. These and other questions will be tackled as advances are made in the process, the necessary adjustments are carried out, and agreements and/or negotiations with stakeholders are reached to address the health pressures that vex the population and to mitigate the burden of disease with efficacious, effective, and efficient technologies. Source of financial support: The authors have no other financial relationships to disclose. R EF E R EN C ES [1] General Health Council. Drug Formulary: Health Care Formulary and Supply Catalog Inter-institutional Commission (2011 ed.). Available from: http://www.csg.gob.mx/descargas/pdfs/ED2011MEDI.pdf. [Accessed March 24, 2014]. [2] Secretariat of Health. Instituto Nacional de Psiquiatría Ramón de la Fuente Muñiz, Institutional Medication Plan, Clinical Services Division, Mexico, 2012. [3] General Health Council. Public Sector Drug Formulary. Official Gazette April 9, 1975. [4] General Health Council. Public Sector Drug Formulary. Official Gazette December 2, 1977:1–3. [5] Secretariat of Health. Agreement establishing the Health Care Supply Formulary. Official Gazette June 9, 1983:22–4. [6] General Health Council. Health Care Formulary and Biological Products and Reagents Catalog. Official Gazette September 29, 1997:12–3. [7] Bylaws of the Health Care Formulary Inter-institutional Commission, Official Gazette of the Federation, February 19, 2003. [8] General Health Council. Bylaws of the Health Care Formulary Interinstitutional Commission, Official Gazette of the Federation, June 22, 2011. [9] General Health Council. Guidelines for the evaluation of supplies. 2011. Available from: http://www.csg.salud.gob.mx/descargas/pdfs/cuadro_ basico/guia_eval_insumos11052011.pdf. [Accessed March 24, 2014]. [10] Mexican Official Standard NOM-220-SSA1-2012, Installation and operation of pharmacovigilance. Official Gazette of the Federation: January 7, 2013. Available from: http://www.dof.gob.mx/nota_detalle. php?codigo=5284236&fecha=07/01/2013. [Accessed March 24, 2014]. [11] General Health Council, Instituto Nacional de Salud Pública. Guidelines for conducting economic evaluation studies for the update of the health care formulary. 2008. Available from: http://www.csg.salud.gob.mx/descargas/pdfs/cuadro_basico/ GUxA_EVAL_ECON25082008_2_ech.pdf. [Accessed March 24, 2014]. [12] General Health Council. Database of the Inter-institutional Commission’s Technical Secretariat of the Health Care Formulary and Supply Catalog, prepared with requests received from September 2011 to December 2012. México D.F: General Health Council, 2012. VALUE IN HEALTH REGIONAL ISSUES 5C (2014) 14–19 Available online at www.sciencedirect.com journal homepage: www.elsevier.com/locate/vhri A Comparison of 1-Year Treatment Costs in Patients with Type 2 Diabetes Following Initiation of Insulin Glargine or Insulin Detemir in Argentina Joaquín E. Caporale, BEcon, MSc1,*, Andrés Pichón-Riviere, MD, MSc, PhD1, Andrea G. Beratarrechea, MD, MSc1, Cristian Von Schulz-Hausmann, MD2, Federico Augustovski, MD, MSc, PhD1 1 Institute for Clinical Effectiveness and Health Policy (IECS), Buenos Aires, Argentina; 2Sanofi Aventis, Buenos Aires, Argentina AB STR A CT Objective: To estimate and compare type 2 diabetes mellitus treatment costs in insulin-naive patients following initiation of therapy with either insulin glargine (IG) or insulin detemir (ID) over 1-year time horizon from a payers’ perspective in Argentina. Methods: We used a pharmacoeconomic model based on a randomized trial comparing IG and ID (Rosenstock J, Davies M, Home PD, et al. A randomised, 52-week, treat-to-target trial comparing insulin detemir with insulin glargine when administered as add-on to glucoselowering drugs in insulin-naive people with type 2 diabetes. Diabetologia 2008;51:408–16) and Argentinean sources. Clinical, resource use, and cost data were combined to estimate direct medical costs (insulin, test strips, and needles) during the first year. Price per international unit of insulin is similar for IG and ID in the local market. Deterministic analysis was performed on insulin unit cost and probabilistic sensitivity analyses on clinical, resource use, and unit costs to evaluate contribution to variance on the difference in total annual treatment cost. Results: Annual mean treatment cost (Argentinean pesos 2013) was AR $6229 for IG and AR $9257 for ID, showing 33% total cost reduction with IG (AR $3028; exchange rate US $1.00 ¼ AR $5.30). Probabilistic sensitivity analysis showed that IG was cost saving in 88% of the simulations. The most influential parameter was the difference in insulin dose requirements. Threshold analysis showed that if the unit price of ID is reduced by 43%, ceteris paribus, the total annual costs per person for both insulin regimens would be the same. Conclusions: From a payer’s perspective in Argentina, cost savings related to the use of IG represented one third of total treatment costs. Sensitivity analyses confirmed the robustness of these results. Keywords: cost comparison, insulin detemir, insulin glargine, type 2 diabetes. Introduction landmark randomized controlled trials and meta-analysis of randomized controlled trials (RCTs) established that intensive glucose-lowering treatment reduces microvascular complications, and follow-up data from these studies suggest that intensive treatment also lowers macrovascular risk in T2DM [5,7–11]. When considering effectiveness, tolerability, and cost of the various diabetes treatments, insulin is not only the most potent but also the most cost-effective intervention [12,13]. In spite of the existing evidence, there has been a stepwise introduction of glucose-lowering interventions, with the final step of insulin therapy being administered 10 to 15 years after diagnosis [14]. Both patients and physicians are often reluctant to start insulin because of fears of painful injections, hypoglycemia, and weight gain [15–17]. In recent years, long-acting insulin analogues, insulin glargine (IG) and insulin detemir (ID), were introduced and proposed as a therapeutic alternative with the potential to overcome some of these barriers as data from trials Type 2 diabetes mellitus (T2DM) is a serious public health problem due to its high prevalence and the development of chronic complications (retinopathy, nephropathy, peripheral vascular disease, ulcers, diabetic foot and amputations, cardiovascular disease, and stroke), which increases resource use and socioeconomic costs, especially in developed countries [1]. The economic burden raised by diabetes is challenging health care systems. According to the World Health Organization, direct health care cost of diabetes-related illnesses ranged from 5% to 13% of a country’s annual health care budget, depending on local prevalence and treatment costs [2]. In Argentina, diabetes affects 11.9% of the population [3] and is estimated to represent a high proportion of total health expenditure [2]. It has been clearly established that the development and progression of complications can be effectively prevented or delayed through tight glycemic control [4–6]. A number of Copyright & 2014, International Society for Pharmacoeconomics and Outcomes Research (ISPOR). Published by Elsevier Inc. Conflict of interest: Funding for this study was provided by Sanofi Aventis Argentina. C. Schulz-Hausmann is an employee of Sanofi Aventis Argentina. All the other authors are employees of IECS, an independent research organization, and maintained independent scientific control over the study, including data analysis and interpretation of final results. * Address correspondence to: Joaquín E. Caporale, Institute for Clinical Effectiveness and Health Policy (IECS), Dr Emilio Ravignani 2024, C1414CPV - Buenos Aires, Argentina. E-mail: [email protected]. 2212-1099$36.00 – see front matter Copyright & 2014, International Society for Pharmacoeconomics and Outcomes Research (ISPOR). Published by Elsevier Inc. http://dx.doi.org/10.1016/j.vhri.2014.02.007 15 VALUE IN HEALTH REGIONAL ISSUES 5C (2014) 14–19 and meta-analysis showed a lower rate of symptomatic, overall, and nocturnal hypoglycemia in patients treated with either IG or ID compared with neutral protamine hagedorn (NPH) insulin [18]. According to the American Diabetes Association and the European Association for the Study of Diabetes guidelines for the management of T2DM, insulin could be initiated with either oncedaily NPH insulin or long-acting insulin analogues [19]. Regimes involving long-acting insulin analogues can achieve clinically important improvements in glycemic control similar to those achieved with NPH, but with less risk of hypoglycemia [20,21]. Studies that compared IG and ID in patients with T2DM showed that both analogues did not differ in efficacy and safety profiles [22–25]. The economic impact of the use of these insulins was estimated in Spain by Guisasola et al. [26] on the basis of the only 52-week randomized trial to date (Rosenstock et al.) [22], which compared clinical outcomes related to the addition of basal insulin analogues ID or IG in a sample of 582 insulin-naive patients with T2DM who were inadequately controlled with oral glucose-lowering drugs. In this study, it was found that the use of IG instead of ID would result in annual saving on treatment costs of 34% or 534.96 (€ 2006) for a patient with T2DM. Pscherer et al. [27] compared treatment costs of IG with those of ID, both combined with bolus insulin as part in patients with T2DM in Germany. The authors concluded that IG may represent a cost-saving option for patients with T2DM in this country, with potential annual cost savings of €684 (19%) per patient compared with ID at 2008 prices. In contrast, a retrospective cohort analysis of health care claims data in a large US managed care organization (since May to December 2006) found that patients receiving ID incurred lower diabetes-related medical costs ($707 vs. $1510; P ¼ 0.03) and total health care costs ($2261 vs. $3408; P ¼ 0.03) than did those using IG [28]. We found many other similar cost comparison studies between these insulins for many countries [26,27,36], but none of them was for any Latin American country. The Latin-American Diabetes Association guidelines recommend the use of insulin analogues when hypoglycemia is limiting glycemic control [29]. Up to date, no studies in Latin America have compared the economic impact of the use of IG versus ID. This study attempts to estimate and compare the economic implications of IG and ID therapy initiation in insulin-naive patients with T2DM with 1-year time horizon, from a payer’s perspective in Argentina incorporating a probabilistic sensitivity analysis (PSA). Table 1 – Clinical parameters. Parameter Initial mean body weight (kg) Final weight (kg)* Initial doses (IU) Final doses (IU/kg) Average dose (IU) Insulin glargine Insulin detemir Once daily injection Twice daily injection 87.4 87.4 87.4 91.3 12 0.44 26.09 89.7 12 0.52 29.32 91.1 12 1 51.55 Note. Estimated and adapted from Rosenstock et al. [22]. * This indicator was calculated adding to the initial mean body weight the mean change registered at the end of the trial for each insulin scheme. This trial informed the initial and final doses of each insulin regimen, so an average total dose per each insulin regimen was estimated on the basis of the initial dose (12 international unit [IU] for all patients) and the final dose per insulin regimen reported in Rosenstock et al. [22] considering a linear titration over the 52 weeks. This is a conservative assumption, given that 80% of the patients requiring ID twice daily (n ¼ 103) were transferred to this scheme during the first 12 weeks of treatment. Average total dose Final dose per kgInitial dose per kg ¼ Initial dose per kg þ 2 Final body weight As the equation shows, mean dose per each insulin regimen was calculated using patient’s final weight. As in Rosenstock et al. [22], only the mean initial and incremental body weight from baseline at 52th week were reported and the final body weight was estimated as the initial plus the incremental one. As in the case of the average total dose, this estimation was also a conservative assumption for IG because the final body weight estimated was higher for IG than for ID per each insulin regime. Because the difference in hypoglycemic events was neither clinical nor statistically significant among both regimens (0.04 episodes per patient-year) [22], it was not considered in the model. Costs associated with the change in body weight of each insulin regimen are a relatively new issue and usually not included in the literature, and they were not considered in ours because of the difficulty in identifying an unbiased cost estimate for the Argentinean context. Methods We used a pharmacoeconomic exercise based on Guisasola et al. [26], and Pscherer et al. [27] constructed on MS Excel based on the results of Rosenstock et al. [22]. Although other trials comparing the efficacy and safety of both insulin have been published [23– 25], the study by Rosenstock et al. [22] is the only trial to date that compared IG and ID in an annual duration of treatment in insulin-naive patients with T2DM. Clinical, resource use, and cost data were combined in the model to estimate annual direct medical costs associated with the use of insulin, test strips, and needles required during the first year of insulin treatment in T2DM. Clinical Parameters Table 1 lists the clinical parameters for each insulin regime. At the end of follow-up, 55% of the patients treated with ID required twice-daily application. All patients treated with IG required once-daily injections. Utilization of Resources and Cost Parameters The commercial forms considered for insulins were Lantus Solostar and Levemir FlexPen for IG and ID, respectively. This decision is based on the fact that both presentations are the only ones available in the Argentinean market that contain the same quantity of insulin (five prefilled pens of 3 mL with 100 IU/mL). In relation to the use of needles, we assumed a utilization rate of one per each insulin application. Finally, regarding the use of test strips, a consumption rate of three and six units per week for once-daily and twice-daily injection scheme, respectively, was assumed. Both assumptions were based on expert opinion of diabetes specialists. It is recognized that a lower number of applications may have advantages in terms of quality of life, but this issue will not be considered in monetary terms in this cost comparison exercise because it is out of scope of this article and because of the absence of local estimates. Monetary values for insulins were obtained from the Argentinean market. Unit prices for the commercial forms considered 16 VALUE IN HEALTH REGIONAL ISSUES 5C (2014) 14–19 were AR $1126.88 and AR $1066.54 for IG and ID, respectively, both valid since April 30, 2013. To customize the situation to our perspective, a 35% discount over market prices was used, assuming the presence of purchase power from large payers. Unit costs for test strips and needles were obtained from the Argentinean market; we considered Accu-chek active glucose per 50 strips from ROCHE (AR $6.4656 per strip) and Novofine 30 G 100 and 70 (AR $1.5574 per needle). All unit prices and costs include valueadded tax and are expressed in local currency, year 2013 (mean exchange rate US $1.00 ¼ AR $5.30). Annual costs were calculated for insulins, needles, and test strips. For each insulin, the cost per unit of insulin was multiplied by the corresponding average total dose, estimated by the equation, and by 365 days of treatment. Regarding needles and test strips, total annual costs for IG were calculated by multiplying the unit price per the utilization rate and 365; for ID, these were calculated in the same way as in the IG regimen up to the 12th week, and for the remaining 40 weeks, they were calculated in a similar way but ponderating by the proportion of people with oncedaily and twice-daily dose, as expressed in the following equation: ð12 7 Unit Cost Use1 Þ þ ð40 7Þ ½ð%1 Unit Cost Use1 Þ þ ð%2 Unit Cost Use2 Þ where Use1 and Use2 refer to the utilization rate for those on a once-daily and twice-daily dose scheme with ID, respectively, and %1 and %2 refer to the proportion of people on a once-daily and twice-daily dose scheme with ID. Sensitivity Analysis Base-case values with their corresponding distributions, associated parameters, and commentaries are expressed in Table 2. Deterministic analysis was performed to evaluate the change in the unit price of ID so as to get a null difference in the total annual costs per person for each insulin regimen. Then, PSA was performed using Monte-Carlo technique and considering 1950 iterations at a confidence level of 95% for mean and SD; this number of trials was enough to reach an accurate forecast for our results. Final doses of insulin in units per kilogram (IU/kg) were simulated under a normal distribution considering as base-case values those reported in Rosenstock et al. [22], with an inferior limit of 0.15 IU/kg [19], assuming a common SD of 50% for each regimen, which equals 0.22 IU/kg, 0.26 IU/kg, and 0.50 IU/kg for IG, ID once-daily, and ID twice-daily dose scheme, respectively. Because Rosenstock et al. [22] did not inform measures of dispersion for insulin doses, the proportion for the SD of final doses of insulin was based on Raskin et al. [25]. The proportion of patients treated with ID receiving oncedaily dose was assumed to follow a beta distribution. Unit prices of insulins were not included in the PSA; instead, the discount over both market prices was included, assuming it to be similar for both insulins. The distribution type and parameters for this variable were derived from experts’ opinion. Unit costs for test strips and needles are both assumed to follow a normal distribution considering as base-case values those from the sources already mentioned; the SD in both cases was assumed to be 50%. Results The overall annual treatment costs (insulin, needles, and test strips) per person in the first year were AR $6229 and AR $9257 for IG and ID, respectively (Table 3). The insulin component represented 75% of the total treatment cost for both insulin regimens. Although unit prices of insulins were very similar, treatment of Table 2 – Probabilistic sensitivity analysis: Variables, distributions, and parameters. Variable Base value Distribution Parameters Final dose per kilogram for insulin glargine (IU/kg) 0.44 Normal Mean ¼ 0.44 SD ¼ 0.22 Min ¼ 0.15 Final dose per kilogram for once-daily injection for insulin detemir (IU/kg) Final dose per kilogram for twice-daily injection for insulin detemir (IU/kg) Completers with once-daily injections with insulin detemir (%) 0.52 Normal 1 Normal 45% Beta Discount over unit price of insulins (%) 35% Beta Unit cost of test strips (AR $) 6.4656 Normal Unit cost of needle (AR $) 1.5574 Normal Mean ¼ 0.52 SD ¼ 0.26 Min ¼ 0.15 Mean ¼ 1.00 SD ¼ 0.50 Min ¼ 0.15 Min ¼ 0% Max ¼ 100% α ¼ 105 β ¼ 129 Min ¼ 15% Max ¼ 50% α¼5 β¼4 Min ¼ 0 Max ¼ 50 Mean ¼ 6.46 SD ¼ 3.2328 Min ¼ 0 Max ¼ 50 Mean ¼ 1.55 SD ¼ 0.7787 IU, international unit; max, maximum; min, minimum. Comments Mean extracted from Rosenstock et al. [22]. SD was assumed equal to 50% based on Raskin et al. [25]. Minimum value was assumed according to general local practice. As above As above Parameters extracted from Rosenstock et al. [22]. Completers with twice-daily injections were calculated as 1 minus this variable. Parameters were set to shape local expert’s opinion. Extracted from Argentinean market. SD assumed to reach 50% according to personal communication with experts. Extracted from Argentinean market. SD assumed to reach 50% according to personal communication with experts. 17 VALUE IN HEALTH REGIONAL ISSUES 5C (2014) 14–19 Table 3 – Base case: Total annual costs per treatment. Annual costs ( AR $) Insuline glargine Insulin detemir Difference (insulin detemir – insulin glargine) 4649.43 1011.40 568.45 6229.29 7008.68 1439.66 809.15 9257.48 2359.24 428.25 240.70 3028.19 Insulin Test strips Needles Total patients with T2DM initiating IG was 33% less costly than of those with T2DM initiating ID, mainly due to daily doses requirement. The results of a deterministic threshold analysis indicated that if the unit price of ID net of the discount was reduced by 43% of its original, ceteris paribus, the total annual costs per person for both insulin regimens in the base case would be the same. This price reduction would be enough to offset those savings generated by the IG regime on needles and on self-monitoring blood glucose costs. In the PSA, the estimated mean value for the difference in total annual treatment costs between IG and ID under a lognormal distribution was marginally higher than the estimated value in the base case (AR $3114.6 vs. AR $3028.2). Fig. 1 shows density and cumulative distributions, respectively, for the difference in total annual treatment cost. This difference was positive, meaning that IG was cost saving, in 87% of the simulations. Table 4 summarizes the sensitivity of the results (correlation and contribution to variance) of the assumptions made in the model, that is, the influence of each variable incorporated into PSA over mean difference in total annual treatment costs. In our estimations, final doses per kilogram of both insulin regimens were main determinants for the variance in our result. All these variables had the expected sign and the expected strength over them. As shown, variations in the final doses (IU/kg) of ID had a growing effect over the variance in our result as we moved from once-daily dose to twice-daily dose scenarios. Discussion In our study, mean total annual costs in the first year after initiating insulin treatment among insulin-naive patients with T2DM were lower when treated with IG than with ID, AR $6229 and AR $9257 for IG and ID, respectively (Table 3). But when patients were treated with once-daily dose of ID or IG, such difference narrowed in a significant way (5%). Because the unit prices of commercial forms used for both insulins were similar, this difference in cost was mainly explained by the difference in insulin dose requirement. Costs were also higher in the ID group for strips and needles, mainly as a result of twice-daily dosing in 55% of the patients who initiate insulin therapy with ID. Costs associated with hypoglycemic events and other complications were not accounted for in this analysis because the difference in hypoglycemic events was neither clinical nor statistically significant among both regimens in the trial [22]. Studies suggest that higher doses of insulin are required for patients with T2DM when using ID than when using other basal insulins [30–32]. Higher requirements of insulin doses for ID (0.82 IU/kg vs. 0.59 IU/kg) were observed in a 52-week target-to-treat trial that compared IG versus ID efficacy and safety in a basalbolus regimen with mealtime insulin as part in patients with T2DM [23]. Similar results were found in a 24-week treat-to-target 120 Frequency 80 40 Cost saving 87% 0 Accumulated Probability 100 80 87% 60 40 20 0 -$ 4,266 Cost saving -$ 1,254 $ 1,759 $ 4,772 $ 7,784 $10,495 Fig. 1 – Probabilistic sensitivity analysis: Density and cumulative distributions of difference on total annual costs (insulin detemir – insulin glargine). Density and cumulative distributions are plotted in relation to the absolute and relative frequency of iterations, respectively. 18 VALUE IN HEALTH REGIONAL ISSUES 5C (2014) 14–19 Table 4 – Probabilistic sensitivity analysis: Multivariable correlation to and contribution to variance on difference on total annual costs. Variable Final dose for twice-daily injection for ID (IU/kg) Final dose for IG (IU/kg) Final dose for once-daily injection for ID (IU/kg) Unit cost of test strips (AR $) Unit cost of needle (AR $) Discount over unit price of insulins Completers (%) with oncedaily dose of ID Correlation* Contribution to variance 0.741522 0.550503 0.58027 0.310636 (0.337116) 0.096609 0.094218 0.008888 0.061392 0.04462 0.003773 (0.001993) 0.03341 (0.001117) ID, insulin detemir; IG, insulin glargine; IU, international unit. * Rank correlation estimates between each variable and the difference in total annual costs (ID vs. IG). trial comparing initiation of IG once daily versus ID twice daily [24]. Raskin et al. [25], however, showed no differences in insulin doses (0.70 vs. 0.67) in the subgroup of insulin-naive patients with T2DM when treated with insulin as part in a basal-bolus regimen. A study conducted in Argentina that included 607 insulintreated patients with T2DM followed by diabetes specialists showed that 82% of 51 patients treated with IG received one daily dose and 18% two daily doses. Regarding ID, 66% of 45 patients received two daily doses and 34% only one daily dose. Mean daily doses of insulin reported in this study were 30.7 ⫾ 17.9 UI/d for IG and 36.1 ⫾ 23.8 UI/d for ID; no distinction was made for patients who received one or two daily injections [33]. Although once-daily and twice-daily insulin regimens for IG and ID were similar to those reported in Rosenstock et al. [22], the mean daily dose of ID differed. In Rosenstock et al. [22], this value doubled the dose reported in this study. The reasons for the increased dose requirement for ID are not clear, but it is likely to be related to the pharmacologic properties of the analogue [34,35]. We decided to include in our study the only RCT that evaluated clinical outcomes in insulin-naive patients with T2DM treated with IG and ID as add-on therapy to glucoselowering agents. We excluded trials using less rigorous methods, such as quasi-experimental studies or nonrandomized studies because RCTs provide stronger and more unbiased estimations of the effect of interventions. We also decided not to use the study that described insulin therapy in Argentina because it has an observational design and included a small number of patients with IG and ID treated by diabetes specialists. This restriction clearly limited the eligibility of studies for this exercise to only one RCT with 582 participants. Our results are consistent with those described in cost comparison studies in other countries and with a cohort study in which patients with T2DM who initiated a basal-bolus insulin regimen with IG had reduced costs of treatment than did those treated with ID in UK routine general practice [36]. In this study, the median cost of prescriptions for classifiable antidiabetic therapy was 28.1% lower among those treated with IG than in those treated with ID (£1014 vs. £1410), a difference of £397 per person-year. The largest single contribution to this difference arose from the difference in insulin cost (32% lower), reagent cost (16% lower), pen delivery devices (50% lower), and sharps (16% lower) in the IG group compared with the ID group. These findings may have coverage and policy implications. As studies to date show that ID and IG have similar efficacy [22,24], at similar commercial unit cost, treatment with IG over ID would be preferred in insulin-naive patients with T2DM as part of a basal-bolus regimen because this decision could achieve a substantial cost saving. Our results based on the study by Rosenstock et al. [22] state that per 1000 patients treated with IG, an overall cost saving per year of AR $3,028,190 could be expected. Also, in the case that a broader perspective as the societal one was considered, this difference in cost would probably be higher, due to the extra time and discomfort associated with a more frequent dosing scheme of ID [37]. Further research, however, needs to be undertaken to evaluate real-world utilization, the long-term cost and cost-effectiveness of IG over ID, as well as the consistency of the finding of different dosing requirements between these insulins in our country. Our findings analyze treatment costs in insulin-naive patients with T2DM during the first year of initiating a therapy with either IG or ID. This study estimated the annual difference in treatment cost of ID and IG schemes in patients with T2DM in Argentina. Initiating insulin treatment with IG in patients with T2DM was associated with lower costs compared with ID use. T2DM requires substantially higher doses of ID, which translates into higher insulin costs From a payer’s perspective, cost savings related to the use of IG represented one third of total treatment costs. Sensitivity analyses confirmed the robustness of these results. Acknowledgment We acknowledge the contributions of Juana Patricia Sánchez, MD, who collaborated in the preparation of the manuscript. Source of financial support: This study was financed by an unrestricted grant provided by Sanofi-Aventis, Argentina. The authors are responsible for the design and conduction of the study; data collection, analysis, and interpretation of the data; and preparation of the manuscript. R EF E R EN C ES [1] Ringborg A, Cropet C, Jonsson B, et al. Resource use associated with type 2 diabetes in Asia, Latin America, the Middle East and Africa: results from the International Diabetes Management Practices Study (IDMPS). Int J Clin Pract 2009;63:997–1007. [2] IDF. IDF Diabetes Atlas (4th ed.), 2009. Available from: http://www.idf. org/sites/default/files/da5/IDF%20Diabetes%20Atlas%204th%20edition. pdf. [Accessed April 3, 2014]. [3] Daniel Ferrante, Virgolini M. Encuesta Nacional de Factores de Riesgo 2005: resultados principales. Rev Argent Cardiol 2007;75:20–9. [4] The effect of intensive treatment of diabetes on the development and progression of long-term complications in insulin-dependent diabetes mellitus. The Diabetes Control and Complications Trial Research Group. N Engl J Med 1993;329(14):977–86. [5] Intensive blood-glucose control with sulphonylureas or insulin compared with conventional treatment and risk of complications in patients with type 2 diabetes (UKPDS 33). UK Prospective Diabetes Study (UKPDS) Group. Lancet 1998;352:837–53. [6] Gaede P, Lund-Andersen H, Parving HH, Pedersen O. Effect of a multifactorial intervention on mortality in type 2 diabetes. N Engl J Med 2008;358:580–91. [7] Ohkubo Y, Kishikawa H, Araki E, et al. Intensive insulin therapy prevents the progression of diabetic microvascular complications in Japanese patients with non-insulin-dependent diabetes mellitus: a randomized prospective 6-year study. Diabetes Res Clin Pract 1995;28:103–17. [8] Patel A, MacMahon S, Chalmers J, et al. Intensive blood glucose control and vascular outcomes in patients with type 2 diabetes. N Engl J Med 2008;358:2560–72. VALUE IN HEALTH REGIONAL ISSUES 5C (2014) 14–19 [9] Boussageon R, Bejan-Angoulvant T, Saadatian-Elahi M, et al. Effect of intensive glucose lowering treatment on all cause mortality, cardiovascular death, and microvascular events in type 2 diabetes: meta-analysis of randomised controlled trials. BMJ 2011;343:d4169. [10] Hemmingsen B, Lund SS, Gluud C, et al. Intensive glycaemic control for patients with type 2 diabetes: systematic review with meta-analysis and trial sequential analysis of randomised clinical trials. BMJ 2011;343: d6898. [11] Holman RR, Paul SK, Bethel MA, et al. 10-year follow-up of intensive glucose control in type 2 diabetes. N Engl J Med 2008;359:1577–89. [12] Nathan DM. Clinical practice: initial management of glycemia in type 2 diabetes mellitus. N Engl J Med 2002;347:1342–9. [13] Swinnen SHJ, De Vries J Hans. Insulin therapy for type 2 diabetes. Diabetes Care 2009;32(Suppl. 2):S253–9. [14] Nathan DM. Initial management of glycemia in type 2 diabetes mellitus. N Engl J Med 2002;347:1342–9. [15] Nakar S, Yitzhaki G, Rosenberg R, Vinker S. Transition to insulin in type 2 diabetes: family physicians’ misconception of patients’ fears contributes to existing barriers. J Diabetes Complications 2007;21(4):220–6. [16] Polonsky WH, Fisher L, Guzman S, et al. Psychological insulin resistance in patients with type 2 diabetes: the scope of the problem. Diabetes Care 2005;28:2543–5. [17] Polonsky WH, Jackson RA. What’s so tough about taking insulin? Addressing the problem of psychological insulin resistance in type 2 diabetes. Clin Diabetes 2004;22:147–50. [18] Horvath K, Jeitler K, Berghold A, et al. Long-acting insulin analogues versus NPH insulin (human isophane insulin) for type 2 diabetes mellitus. Cochrane Database Syst Rev 2007: CD005613. [19] Nathan DM, Buse JB, Davidson MB, et al. Medical management of hyperglycemia in type 2 diabetes: a consensus algorithm for the initiation and adjustment of therapy: a consensus statement of the American Diabetes Association and the European Association for the Study of Diabetes. Diabetes Care 2009;32:193–203. [20] Horvath K, Jeitler K, Berghold A, et al. Long-acting insulin analogues versus NPH insulin (human isophane insulin) for type 2 diabetes mellitus. Cochrane Database of Systematic Reviews 2007;(2): CD005613. [21] Duckworth W, Davis SN. Comparison of insulin glargine and NPH insulin in the treatment of type 2 diabetes: a review of clinical studies. J Diabetes Complications 2007;21:196–204. [22] Rosenstock J, Davies M, Home PD, et al. A randomised, 52-week, treatto-target trial comparing insulin detemir with insulin glargine when administered as add-on to glucose-lowering drugs in insulin-naive people with type 2 diabetes. Diabetologia 2008;51:408–16. [23] Hollander P, Cooper J, Bregnhoj J, Pedersen CB. A 52-week, multinational, open-label, parallel-group, noninferiority, treat-to-target trial comparing insulin detemir with insulin glargine in a basal-bolus regimen with mealtime insulin aspart in patients with type 2 diabetes. Clin Ther 2008;30:1976–87. [24] Swinnen SG, Dain MP, Aronson R, et al. A 24-week, randomized, treatto-target trial comparing initiation of insulin glargine once-daily with [25] [26] [27] [28] [29] [30] [31] [32] [33] [34] [35] [36] [37] 19 insulin detemir twice-daily in patients with type 2 diabetes inadequately controlled on oral glucose-lowering drugs. Diabetes Care 2010;33:1176–8. Raskin P, Gylvin T, Weng W, Chaykin L. Comparison of insulin detemir and insulin glargine using a basal-bolus regimen in a randomized, controlled clinical study in patients with type 2 diabetes. Diabetes Metab Res Rev 2009;25:542–8. Guisasola A, Llorente C, Rubio-Terrés C. Cost analysis of type 2 diabetes mellitus treatment with glargine insulin or detemir insulin in Spain. Rev Esp Econ Salud 2007;6:304–10. Pscherer S, Dietrich ES, Dippel FW, Neilson AR. Comparison of one-year costs of type 2 diabetes treatment with insulin glargine or insulin detemir in a basal supported oral therapy (BOT) in Germany. Int J. Clin Pharmacol Ther 2010;48:129–37. Borah BJ, Darkow T, Bouchard J, et al. A comparison of insulin use, glycemic control, and health care costs with insulin detemir and insulin glargine in insulin-naive patients with type 2 diabetes. Clin Ther 2009;31:623–31. ALAD. ALAD guidelines for the diagnosis, control and treatment of type 2 diabetes mellitus. Rev Asoc Lat Diab 2000;8(Suppl.1):101–67. Raslova K, Bogoev M, Raz I, et al. Insulin detemir and insulin aspart: a promising basal-bolus regimen for type 2 diabetes. Diabetes Res Clin Pract 2004;66:193–201. Hermansen K, Davies M, Derezinski T, et al. A 26-week, randomized, parallel, treat-to-target trial comparing insulin detemir with NPH insulin as add-on therapy to oral glucose-lowering drugs in insulinnaive people with type 2 diabetes. Diabetes Care 2006;29:1269–74. Umpierrez GE, Hor T, Smiley D, et al. Comparison of inpatient insulin regimens with detemir plus aspart versus neutral protamine hagedorn plus regular in medical patients with type 2 diabetes. J Clin Endocrinol Metab 2009;94:564–9. Luongo AM, López González E, Garcia AB, et al. Evaluation of the treatment of type 2 diabetes by specialists in Argentina. Revista Alad 2011;1:78–87. Porcellati F, Rossetti P, Ricci NB, et al. Pharmacokinetics and pharmacodynamics of the long-acting insulin analog glargine after 1 week of use compared with its first administration in subjects with type 1 diabetes. Diabetes Care 2007;30:1261–3. Hompesch M, Ocheltree SM, Wondmagegnehu ET, et al. Pharmacokinetics and pharmacodynamics of insulin lispro protamine suspension compared with insulin glargine and insulin detemir in type 2 diabetes. Curr Med Res Opin 2009;25:2679–87. Poole Chris D, Tetlow T, McEwan P, et al. The prescription cost of managing people with type 1 and type 2 diabetes following initiation of treatment with either insulin glargine or insulin detemir in routine general practice in the UK: a retrospective database analysis. Curr Med Res Opin 2007;23(Supp. 1):S41–8. Davis TM, Clifford RM, Davis WA. Effect of insulin therapy on quality of life in type 2 diabetes mellitus: The Fremantle Diabetes Study. Diabetes Res Clin Pract 2001;52:63–71. VALUE IN HEALTH REGIONAL ISSUES 5C (2014) 20–24 Available online at www.sciencedirect.com journal homepage: www.elsevier.com/locate/vhri Economic Evaluation of Four Drug Administration Systems in Intensive Care Units in Colombia Diego Rosselli, MD, MSc1,*, Juan David Rueda, MD1, María Daniela Silva, MD1, Jorge Salcedo, PharmD, MBA2 1 Clinical Epidemiology and Biostatistics Department, Pontificia Universidad Javeriana Medical School, Bogota, Colombia; Pharmaceutical Integrated Service (SEIFAR), Cali, Colombia 2 AB STR A CT Introduction: Intensive care units (ICUs) are the most frequent setting for serious medical errors, which not only have serious health consequences but also an economic impact. In this article, using a theoretical model, we evaluate four medication administration systems: conventional preparation by nursing staff, MINIBAG Plus delivery system, compounding center preparation, and premix drugs. Methods: We designed a decision tree model from a third-party payer perspective, and the time horizon of the acute event. Local costs, in Colombian pesos (US $1 ¼ 1784 COP$), were obtained from tariff manuals, medication costs from Sismed information system, and clinical variables from the published literature, and uncertainty was dealt with by an expert panel. The drug used for the model was dopamine. Results: Average costs for each dopamine dose delivered were $46,995 for premix, $47,625 for compounding center, $101,934 for MINIBAG Plus, and $108,870 for drug prepared in the ICU. The variability of these results is higher for compounding center than for premix, and even higher for MINIBAG Plus and nurse delivery. Conclusions: The use of premix drugs can be a costsaving strategy, which decreases medical errors in drug administration in the ICU, particularly if it is part of an integral error reduction program. Keywords: drug administration schedule, drug delivery systems, economics, intensive care units, medication errors. Introduction Error prevalence at ICUs is uncertain; estimation of the frequency of errors during drug administration ranges between 1.2 and 947 per 1000 patient-days in adults [3]. Such range results from medical error reporting mechanisms; for example, in independently reported trials, the prevalence is lower. An example is the study by Taxis and Barber [6] in which 1328 patients from 113 ICUs in 27 countries were included, with Brazil and Argentina as the only Latin American representatives. Types of errors evaluated were omissions, wrong drug, wrong dose, wrong administration route, and improper dosing time. The estimated prevalence in this study was 74.5 errors per every 100 patient-days; interestingly, 19% of the participating ICUs reported no errors during the study period. The incidence of errors with injectable medications is higher than with other forms of medications [7]. Of the five stages in IV medication administration (prescription, transcription, dispensation, administration, and monitoring), the drug administration phase is most prone to errors [8], which can be further classified as follows: omission, inadequate dosing, inadequate concentration, wrong medication, incorrect technique, incorrect administration route, improper administration rate, incorrect dosing time, and wrong patient. Consequences of these errors are also classified, in increasing severity, with a lettered-scale that ranges from B to I, with B corresponding to a wrong medication that is not administered and I to an error that The aphorism primum non nocere—above all, do no harm— included in the Hippocratic Oath has regained particular relevance 2500 years later. In health care systems worldwide, effectiveness, as the final goal for therapeutic interventions, has been placed in a balance in which safety and cost also play primary roles [1]. Much has been written on medical error (a mistake by any health care team member, not only physicians) since the controversial book To Err Is Human was published [2], which showed how errors could lead to 1 million injuries and about 100,000 deaths per year in the United States alone. Nowhere else is the “error” issue as sensitive as in the intensive care unit (ICU) [3]. The combination of high complexity, interventional diversity, and critically ill patients make the ICU particularly vulnerable to medical errors [4]. A multinational trial that included 205 ICUs showed 38.8 “incidents” per 100 patientdays in five domains: intravenous (IV) lines and accesses, airway management, equipment, alarms, and medications [5]. It is estimated that 1 of every 10 IV infusions at the ICUs are either erroneously prepared or administered [3]. When only medication errors are considered, the estimation is 10.5 incidents per 100 patient-days in the prescription and administration stages [5]. Copyright & 2014, International Society for Pharmacoeconomics and Outcomes Research (ISPOR). Published by Elsevier Inc. Conflict of interest: This study was supported by an unrestricted grant from Baxter Colombia. * Address correspondence to: Diego Rosselli, Faculty of Medicine, Clinical Epidemiology and Biostatistics Department, Pontificia Universidad Javeriana, Carrera 7 No. 40-62, Piso 2 Hospital San Ignacio, Bogotá, Colombia. E-mail: [email protected]. 2212-1099$36.00 – see front matter Copyright & 2014, International Society for Pharmacoeconomics and Outcomes Research (ISPOR). Published by Elsevier Inc. http://dx.doi.org/10.1016/j.vhri.2014.05.001 21 VALUE IN HEALTH REGIONAL ISSUES 5C (2014) 20–24 Table 1 – Classification of medical errors by Calabrese et al. [9], as per error type clustering used by the authors and the corresponding incidence probability (once an error has occurred). Category A B C D E F G H I Definition Quasi error Wrong drug, not administered Medication administered without consequences Monitoring required but no symptoms Symptoms development, management required Hospital stay is required or increased Permanent consequences develop Risk of death (anaphylaxis, cardiac arrest) Death results in the patient’s death (Table 1). The time horizon of medication error is, therefore, quite variable. Most errors lead to either minor harm or no harm at all, while it has been estimated that about 2% of the errors cause significant injuries to patients [9]. Calabrese et al. [9] proposed another error classification considering both drug administration route (e.g., subcutaneous administration, IV bolus, and IV infusion) and drug class (e.g., antibiotics, sedatives, vasopressors, and insulin). Despite the fact that no other potential medication errors (microbiological contamination, compounding errors such as wrong dilutant or drug concentration miscalculation) were considered in this trial, error rate exceeded the one reported in the previously described study as it reached 74.5 errors per 100 patient-days. Overall, an error occurs in about 7% of all parenteral drug administrations [10]. Medication dosing errors (118 of 861 recorded errors) had the most serious consequences as they resulted in permanent harm in three patients and in the death of other three patients. Other studies [3,11] report even higher error rates regarding medication administration at the ICUs, reaching up to 1 daily error per patient, on average [12]. Apart from having severe consequences on the patients’ health, errors have serious financial consequences [13]. In a sample of ICU patients in Switzerland (n ¼ 333), Nuckols et al. [14] showed preventable IV drug administration–related adverse events in 94 patients (28%). Such adverse events were associated with an extended hospital stay (mean 4.8 days) as well as with increased costs (mean US $4500) versus the control group. The objective of this study was to develop an economic evaluation model to estimate the costs and outcome impact of four different IV drug administration systems (assuming the same drug) at the ICU setting in Colombia. Error type Probability (%) Reference 92 [10] Mild harm 6.33 [10] Moderate harm 1.44 [6] Severe harm 0.42 [6] Death 0.21 [6] No harm 1 year. Costs are in Colombian pesos (COP $) as for 2012 (as reference, the exchange rate for July 2012 was US $1 ¼ COP $1784). Model Assumptions The main assumption of the model is that error incidence rates at Colombian ICUs are similar to those published elsewhere. Calculations are based on a “typical” ICU adult patient. Error risks To estimate error risks, a literature review was carried out using MeSH “Medication Errors,” “Drug Administration Schedule,” “Drug Delivery Systems,” and “Intensive Care Units” as search criteria. A total of 272 abstracts were reviewed; 27 articles were selected for full-text review, of which 20 [4,5,9–11,13–27] reported error rates. Among 113 ICUs from 27 countries, the only data from the region came from Brazil [17] and from 6 ICUs (3 Argentinean and 3 Brazilian) included in the study by Valentin et al. [5]. Error probabilities used in the model (based on Nuckols et al. [14]), for each individual drug administration, were as follows: 0.01 for compounding center preparation (by pharmacists), 0.08 for bedside preparation by nursing staff, 0.03 for MINIBAG Plus, and 0.0027 for premix drugs. For our base-case scenario, we selected the error rates of Taxis and Barber [6] and Klopotowska et al. [10] because they were on the conservative side (we preferred to underestimate the risk) and because they include a wide range of error consequences. Only one reference [15] included error rates for compounding centers versus premix medication. For the sensitivity analysis, we arbitrarily assumed a wide range after discussions with our expert panel. In error probabilities, we considered it unpractical to convert rates to probabilities. We used rates as probabilities because of the short time frame. Methods We designed a decision tree–type economic model using TreeAge Pro Healthcare 2009 (Fig. 1). Four alternatives for IV drug delivery were considered: use of premix drugs, compounding center preparation, bedside preparation at the ICU by a nurse (but using a buretrol set), and MINIBAG Plus use (flexible closed system bag with a vial adaptor—a point-of-care activated device). Baseline data included in the model were extracted from international medical journal publications (see Table 1) and subsequently discussed and validated by an expert panel independently selected by the investigators (with no sponsor participation). The panel was composed of an internal medicine specialist, a surgeon, two physicians specialized in pharmacology, and a pharmacist. Dopamine was selected as the drug for the model because it shows a larger cost difference between premix and competitors. We used a third-party payer perspective (Colombian health system), and the time horizon was the length of ICU stay (which is similar to that reported in the literature); no discount rate was applied because the period of analysis was shorter than Fig. 1 – Decision tree outline. Errors include preparationrelated, contamination, and biological risk errors. ICU, intensive care unit. 22 VALUE IN HEALTH REGIONAL ISSUES 5C (2014) 20–24 Costs Cost calculation for compounding center preparation Initially, costs were obtained from a large compounding center that prepares medications and parenteral nutrition solutions for four high-complexity hospitals in Bogota. The microcosting technique was used, in which administrative costs were distributed between oncology drugs, parenteral nutrition, antibiotics, and other medications. Staff at this center includes 99 people, 78 of whom are pharmacists on different working schedules. Payroll is almost COP $ 350 million, of which 30% corresponds to the preparation of antibiotics and 52% to the preparation of other drugs. On a monthly basis, 150,000 units are prepared (antibiotics 50,000, other drugs 95,000, remaining ones corresponding to oncology drugs and parenteral nutrition). These data result in an estimated unit value of COP $17,552 for chemotherapy, COP $14,306 for total parenteral nutrition, COP $2,050 for antibiotics, and COP $1,888 for other drugs, before applying a profit margin; because of economies of scale, this last cost was the one used for the lower limit of our model. For estimation of costs, three additional hospital mixing centers were studied, in which unit preparation costs ranged between COP $3,938 and COP $14,306 through microcosting techniques (we used COP $4,923 and COP $17,883 as margins for the sensitivity analysis, applying a 25% profit margin). This wide range of costs could be accounted for by the proportion of resources allocated to the preparation of oncology drugs and parenteral nutrition products (whose costs and profit margin are both higher) and the volume of prepared medications. In the model, for the base case, we used an average cost of COP $13,468 per drug prepared at the compounding center. Cost calculation for nursing staff preparation of buretrol, and for premix drugs These costs include IV drug preparation time at the ICU. For this estimation, a chronometer was used while preparing and administering a convenience sample of 130 drugs for IV application at a university hospital ICU in Bogota. Prolonged dilution medications and drugs with no premix equivalent were excluded (e.g., piperacillin/tazobactam). On average, hand washing took 35.7 ⫾ 18.5 seconds, mouth cover and gloves use 1.9 ⫾ 4.5 seconds, drug preparation 135.2 ⫾ 250.0 seconds, and drug administration 64.8 ⫾ 35.2 seconds, for 382.7 ⫾ 513.2 seconds in total for each administered drug. This time (slightly over 6 minutes) was the time used for the nursing staff and buretrol base case. The literature review showed similar findings at a cardiology ICU in Houston, Texas (91.8 seconds for drug preparation, 59.8 seconds for drug administration, total time 313 seconds) [27]. The cost was calculated on the basis of a professional nurse salary of $12,495 per hour (market price). For the premix arm, same times for hand washing and mouth cover/gloves use were used. In this case, drug preparation time was shorter (24.1 ⫾ 27.2 seconds) and took into account the time corresponding to cart medication collection and bag labeling. As for administration time, same nurse delivery base-case items were considered, that is, time for reaching the patient’s bed, for positioning of protection elements, and for administering the drug. The total time required in this arm was 136.2 ⫾ 58.0 seconds. Drug costs For the base-case scenario, we selected dopamine for cost calculations because this was the drug with the greatest price difference between the normal preparation and the premixed. We assumed that if there were any cost savings with premixed dopamine, where premixed preparation was the most expensive compared with regular drugs, any potential difference in favor of premix would be even greater with other ICU medications. Dopamine cost was obtained from annual sales (March 2011March 2012) according to Sismed (an official source of medication price and sales volume information). In total, 17,555 units were sold during this 12-month period. Because the unit value and market share of dopamine were both known, we estimated a weighted mean cost of COP $2910. For premixed dopamine, we used the unit cost supplied by the manufacturer (COP $17,930), as well as for MINIBAG Plus delivery system (COP $4,800). Error costs In 2011, Carey and Stefos [12] published in Health Economics an article in which they calculated the frequency of a series of errors and quantified their costs. On the basis of a population from the US veterans database (n ¼ 71.349 patients at risk), the researchers estimated the median costs of a medical care–associated infection (US $42,309–US $60,199), sepsis (US$25,891–US $30,515), and a work accident (pinching or laceration) (US$ 5,059–US $9,448). The cost for the latter was higher than the one reported by Rivard et al. [28] in 2008 (US $3359). Because costs in Colombia are significantly lower [29], following discussion with the experts we assumed that harmless errors (which account for 91.6% of the errors) do not incur costs, minor errors (6.3%) incur an additional cost of COP $50,000 each, intermediate errors (those leading to extended hospital stay by 2.4 days, and accounting for 1.4% of the errors) incur costs of COP $3.1 million, and serious errors (0.6%) incur costs of COP $3.56 million. Despite the literature review, error costs were finally estimated by the expert panel, who believe that these costs are much lower than those reported in the literature. Sensitivity analysis We started with a tornado diagram to find which variables are critical for the model. Then, we performed univariate and bivariate sensitivity analysis. Because of the uncertainty surrounding many of the variables, wide ranges have been used (Table 2). We also performed a probabilistic sensitivity analysis using uniform distributions for all the variables except for error costs and compounding center costs, for which we used a gamma distribution (Table 3). Results According to the model, mean costs of a single IV dopamine dose administration at the ICU, including error costs, would be as follows: COP $46,995 for the premix drug, COP $47,625 for the compounding center preparation, COP $101,934 for the MINIBAG Plus delivery system, and COP $108,870 for drug preparation by an ICU nurse. Variation within these mean costs is higher for compounding center preparation than for premix drugs, due to the number of variables involved in the respective processes. This also applies to MINIBAG and nursing staff preparation, compared with premix drugs. According to the model, the rate of errors that lead to harm (expressed as the number of errors per 10,000 administered medications) is 1 for premix drugs, 2 for compounding center preparations, 4 for MINIBAG Plus, and 17 for nurse drug delivery (which is equivalent to 1 harming error per every 600 administered medications). A similar calculation for deaths attributable to a drug administration error, per every 100,000 IV medications administered, resulted in 6 deaths for premix drugs, 21 for compounding center preparations, 42 for MINIBAG Plus, and 168 for nurse staff preparation delivered with buretrol. 23 VALUE IN HEALTH REGIONAL ISSUES 5C (2014) 20–24 Table 2 – Selected variables used in the model, with wide intervals used for the sensitivity analysis. Variable Compounding center preparation Compounding center application Bedside application by nursing staff MINIBAG Plus application Premix drugs application Professional nurse salary (h) Cost of dopamine Day in ICU Cost of mild error Cost of moderate error Cost of severe error Base case ($) Low ($) High ($) Source 13,468 30,582 100,639 92,894 28,885 12,495 5,392 1,291,523 50,000 3,100,000 3,560,000 4,923 22,752 74,226 68,496 21,614 9,375 2,500 600,000 25,000 1,500,000 1,800,000 22,013 25,240 83,410 76,190 23,183 15,620 15,000 2,000,000 100,000 6,000,000 7,200,000 Author’s estimation Author’s estimation Author’s estimation Author’s estimation Author’s estimation Market prices Weighted mean of market price Costs of local ICU Expert panel estimation Expert panel estimation Expert panel estimation ICU, intensive care unit. Sensitivity Analysis The tornado simulation (Fig. 2) showed which variables could change the decision (using the ranges preestablished by the authors). The variable with the greatest effect on total cost is the nurse salary, affecting all the four alternatives, but not modifying the final decision. Critical variables were the costs of medication (in this case dopamine), and of compounding center preparation, as well as the probabilities of errors that could cause harm, and the rate of error with premix medication. We found in the one-way sensitivity analysis that if each preparation in the compounding center costs more than $13,468, premix would be cost saving. This cost is one of the variables with higher uncertainty in our model, so each institution should be careful when estimating this particular cost. The difference between the premix cost and plain medication is also important. The two-way sensitivity analysis of the cost of premix and plain dopamine shows that up to a difference of $15,000 in prices, the premix would still be cost saving. If we assume ranges from $1,000 to $10,000 for plain dopamine and $18,000 to $20,000 for premix dopamine, the premix will be cost saving 42% of the times for the possible combinations. Another two-way sensitivity analysis for the error probability of compounding center and premix shows that costs favor one or the other in a very similar rate. In an in-hospital compounding center, the probability of error is higher, which would favor the premix alternative. Discussion This report has certain limitations: First, it is mainly based on foreign studies, error probabilities are not adjusted to medication type, and local costs are approximate. Drug delivery times were measured at a single ICU; however, they were not very different from what has been reported in the literature, except for the article by Rivard et al. [28] who reported shorter times; this might be because their study was undertaken at a cardiology ICU in which fewer antibiotics are used; preparation times of these drugs are lengthier. Our per-error costs were lower than those reported in the literature; according to Nuckols et al. [14], adverse events associated with drug errors increased hospitalization costs by 53% (on average, US $6,647 were added to the baseline costs of US $12,529) and were related to a longer stay (4.8 days more). Mortality rate attributable to errors is also probably underestimated in our model. Rothschild et al. [13] in 1490 patient-days found 54 cases of errors, which either led to death (n ¼ 2), imminent danger (n ¼ 5), or “some” consequence (n ¼ 28) or prolonged the length of stay (n ¼ 19). In the study by Thomas et al. [30], among 3691 incidents resulting in harm, 13 (0.35%) led to the patient’s death, 40 (1.1%) to serious harm, 327 (8.9%) to moderate harm, and 857 (23.2%) to minor harm, while 2454 (66.5%) led to no harm at all. With these data in mind, it is likely that our model underestimated the financial impact of medical errors. Further national studies on the subject are required [29]. Sales of dopamine in Colombia, according to official data, accounted for 17,555 vials in a 12-month period (2011–2012). If our assumptions are correct, the Colombian health system could save up to COP $1.1 billion (around half a million US dollars) if premixed dopamine were used. Of the four drug delivery systems, the one showing higher cost variability is the compounding center drug preparation, for two reasons: the first one is economies of scale. A large mixing center can expect lower marginal costs, thereby having significant differences compared with the final costs of a small compounding center. The other one is opportunity cost; oncological Table 3 – Probabilities are per error occurring in IV drug administration in the ICU, with wide intervals used for the sensitivity analysis. Error Error probabilities for bedside preparation by nursing staff Error probabilities for MINIBAG Plus Rate of error compounding center Rate of error premix Probability of “mild” consequences error Probability of “moderate” consequences error Probability of “severe” consequences error ICU, intensive care unit; IV, intravenous. Base case Low High Source 0.08 0.03 0.01 0.0027 0.063 0.014 0.004 0.04 0.005 0.005 0.001 0.010 0.001 0.001 0.12 0.05 0.05 0.05 0.300 0.050 0.020 [14] [14] [15] [15] [10] [6] [6] 24 VALUE IN HEALTH REGIONAL ISSUES 5C (2014) 20–24 Fig. 2 – Tornado diagram showing variables with the greatest effect on the results. Dark bar shows critical variables, those that can change the least costly alternative between compounding center and premix. The x-axis represents the change in mean cost. ICU, intensive care unit. medications and parenteral nutrition products tend to have a larger profit margin because they have a higher wholesale price (and profit is generally a relatively constant proportion of this price). The main lesson learned is that medication administration errors, particularly in the ICU, can be both more common and more costly than generally assumed. Efforts to prevent them should not only consider quality and safety issues but could also be intended to reduce costs. Conclusions The use of premix drugs at the ICU setting, as part of a strategy aimed at reducing errors, might improve the quality of health care, reducing adverse events attributable to drug administration and reducing costs for the health system. Source of financial support: This study was financially supported by Baxter Laboratories S.A. R EF E R EN CE S [1] Moreno RP, Rhodes A, Donchin Y. European Society of Intensive Care. Patient safety in intensive care medicine: the Declaration of Vienna. Intensive Care Med 2009;35:1667–72. [2] Institute of Medicine. To Err Is Human: Building a Safer Health System. Washington, DC: National Academy Press, 1999. [3] Kane-Gill S, Weber RJ. Principles and practices of medication safety in the ICU. Crit Care Clin 2006;22:273–90. [4] Tissot E, Cornette C, Demoly P, et al. Medication errors at the administration stage in an intensive care unit. Intensive Care Med 1999;25:353–9. [5] Valentin A, Capuzzo M, Guidet B, et al. Errors in administration of parenteral drugs in intensive care units: multinational prospective study. BMJ 2009;338:b814. [6] Taxis K, Barber N. Ethnographic study of incidence and severity of intravenous drug errors. BMJ 2003;326:684. [7] Santell JP. Medication errors. 2008. Available from: http://www.ashp. org/DocLibrary/Policy/PatientSafety/IVSafety/ErrorReports.pdf. [Accessed May 12, 2013]. [8] Hicks RW, Cousins DD, Williams RL. Selected medication-error data from USP’s MEDMARX program for 2002. Am J Health Syst Pharm 2004;61:993–1000. [9] Calabrese AD, Erstad BL, Brandl K, et al. Medication administration errors in adult patients in the ICU. Intensive Care Med 2001;27: 1592–8. [10] Klopotowska JE, Kuiper R, van Kan HJ, et al. On-ward participation of a hospital pharmacist in a Dutch intensive care unit reduces prescribing errors and related patient harm: an intervention study. Crit Care 2010;14:R174. [11] Ridley SA, Booth SA, Thompson CM. Intensive Care Society’s Working Group on Adverse Incidents. Prescription errors in UK critical care units. Anaesthesia 2004;59:1193–200. [12] Carey K, Stefos T. Measuring the cost of hospital adverse patient safety events. Health Econ 2011;20:1417–30. [13] Rothschild JM, Landrigan CP, Cronin JW, et al. The Critical Care Safety Study: the incidence and nature of adverse events and serious medical errors in intensive care. Crit Care Med 2005;33:1694–700. [14] Nuckols TK, Paddock SM, Bower AG, et al. Costs of intravenous adverse drug events in academic and nonacademic intensive care units. Med Care 2008;46:17–24. [15] Flynn EA, Pearson RE, Barker KN. Observational study of accuracy in compounding i.v. admixtures at five hospitals. Am J Health Syst Pharm 1997;54:904–12. [16] Benkirane RR, Abouqal R, Haimeur CC, et al. Incidence of adverse drug events and medication errors in intensive care units: a prospective multicenter study. J Patient Saf 2009;5:16–22. [17] Bohomol E, Ramos LH, D’Innocenzo M. Medication errors in an intensive care unit. J Adv Nurs 2009;65:1259–67. [18] Chapuis C, Roustit M, Bal G, et al. Automated drug dispensing system reduces medication errors in an intensive care setting. Crit Care Med 2010;38:2275–81. [19] Colpaert K, Claus B, Somers A, et al. Impact of computerized physician order entry on medication prescription errors in the intensive care unit: a controlled cross-sectional trial. Crit Care 2006;10:R21. [20] DeYoung JL, Vanderkooi ME, Barletta JF. Effect of bar-code-assisted medication administration on medication error rates in an adult medical intensive care unit. Am J Health Syst Pharm 2009;66:1110–5. [21] Fahimi F, Ariapanah P, Faizi M, et al. Errors in preparation and administration of intravenous medications in the intensive care unit of a teaching hospital: an observational study. Aust Crit Care 2008;21:110–6. [22] Ford DG, Seybert AL, Smithburger PL, et al. Impact of simulation-based learning on medication error rates in critically ill patients. Intensive Care Med 2010;36:1526–31. [23] Garrouste-Orgeas M, Soufir L, Tabah A, et al. A multifaceted program for improving quality of care in intensive care units: IATROREF study. Crit Care Med 2012;40:468–76. [24] Garrouste-Orgeas M, Timsit JF, Vesin A, et al. Selected medical errors in the intensive care unit: results of the IATROREF study: parts I and II. Am J Respir Crit Care Med 2010;181:134–42. [25] Kopp BJ, Erstad BL, Allen ME, et al. Medication errors and adverse drug events in an intensive care unit: direct observation approach for detection. Crit Care Med 2006;34:415–25. [26] Manias E. Errors in administration of parenteral medications are a serious safety problem in intensive care units. Aust Crit Care 2009;22:141–3. [27] Rogers AE, Dean GE, Hwang WT, Scott LD. Role of registered nurses in error prevention, discovery and correction. Qual Saf Health Care 2008;17:117–21. [28] Rivard PE, Luther SL, Christiansen CL, et al. Using patient safety indicators to estimate the impact of potential adverse events on outcomes. Med Care Res Rev 2008;65:67–87. [29] Segura O, Maldonado C. Las reacciones adversas a medicamentos: una aproximación desde el punto de vista económico [in Spanish]. Biomédica 2003;23:401–7. [30] Thomas AN, Panchagnula U, Taylor RJ. Review of patient safety incidents submitted from critical care units in England & Wales to the UK National Patient Safety Agency. Anaesthesia 2009;64:1178–85. VALUE IN HEALTH REGIONAL ISSUES 5C (2014) 25–28 Available online at www.sciencedirect.com journal homepage: www.elsevier.com/locate/vhri Asthma Control and Cost in Latin America Laura S. Gold, PhD1,*, Federico Montealegre, PhD2,3, Felicia C. Allen-Ramey, PhD4, Jose Jardim, MD5,6, Raul Sansores, MD7, Sean D. Sullivan, PhD1 1 Pharmaceutical Outcomes Research and Policy Program, School of Pharmacy, University of Washington, Seattle, Washington, USA; Baxter Bioscience, Guaynabo, Puerto Rico; 3Merck & Co., Inc, Carolina, Puerto Rico, USA; 4Merck & Co., Inc, Whitehouse Station, NJ, USA; 5Pulmonary Rehabilitation Center, Federal University of São Paulo (Unifesp), São Paulo, Brazil; 6Associação de Assistência à Criança Deficiente (AACD), São Paulo, Brazil; 7Instituto Nacional de Enfermedades Respiratorias Ismael Cosio Villegas, México, D.F., México 2 AB STR A CT Objective: Few patients with asthma have disease that is wellcontrolled, particularly in Latin American countries. The purpose of this study was to investigate whether partly controlled and uncontrolled asthma are associated with increased costs for asthma-related medications and health care utilization compared with wellcontrolled asthma in five Latin American countries. Methods: Using the Global Initiative for Asthma guidelines, we classified respondents from the Latin American Asthma Insights and Management survey into those with well-controlled, partly controlled, and uncontrolled asthma and compared the utilization of health care services and costs among these groups. Results: Most respondents to our survey (93%) had asthma that was classified as partly controlled or uncontrolled. Across all countries, patients whose asthma was partly controlled or uncontrolled had greater use of asthmarelated medications and medical services than did patients whose asthma was well-controlled. After adjusting for age, sex, and country of residence, total costs for asthma-related medications and health care were greater in patients whose asthma was classified as partly controlled and uncontrolled. Conclusions: Our findings indicate that patients with asthma that are not well-controlled used more health care resources and had greater medical costs in Latin America. Keywords: Global Initiative for Asthma (GINA) guidelines, health care utilization, hospitalization, long-term maintenance medications, oral steroids. Introduction morbidity of patients with asthma because objective physiological measures have been shown to have little correlation with health-related quality of life [7,8]. Instead, the Global Initiative for Asthma (GINA) and the United States National Asthma Education Prevention Program have created guidelines focused on regulating asthma symptoms to maintain normal daily activity levels [9,10]. We recently conducted a US study that demonstrated that asthma that is defined as not wellcontrolled by GINA guidelines is associated with increased rates of adverse outcomes than is asthma that is classified as wellcontrolled [11] and found similar results using data from Latin America [12]. The purpose of this study was to investigate whether partly controlled and uncontrolled asthma in five Latin American countries (Argentina, Brazil, Mexico, Puerto Rico, and Venezuela, which comprise about 65% of the population of Latin America) were associated with increased costs due to asthma medications, as well as asthma-related utilization of health care services such as hospital admissions, emergency room visits, and visits with health care providers, than was asthma classified as well-controlled. Approximately 40 million patients with asthma reside in Latin American countries and although data are sparse, the prevalence of asthma in most of these countries is believed to be increasing [1,2]. In addition, mortality rates from asthma are higher in Latin American countries than in other regions [2,3]. Asthma is also a significant cause of missed school and work in Latin America [4]. Although medications that have been shown to be highly effective at managing asthma have been available since the 1990s [5], most of the patients do not have well-controlled asthma, particularly in Latin American countries [6]. Although few studies have examined the costs of asthma in Latin America, evidence indicates that poorly controlled asthma in Latin America leads to significant economic expenditures, most of which have been attributed to emergency and unscheduled health care sought by patients with severe persistent asthma [6]. Over the past decade, measurements of asthma control, rather than metrics of lung function such as forced expiratory volume in 1 second, have been introduced to characterize Copyright & 2014, International Society for Pharmacoeconomics and Outcomes Research (ISPOR). Published by Elsevier Inc. Conflicts of interest: The authors have indicated that they have no conflicts of interest with regard to the content of this article. * Address correspondence to: Laura S. Gold, Pharmaceutical Outcomes Research and Policy Program, School of Pharmacy, University of Washington, Box 359455, Seattle, WA 98195. E-mail: [email protected]. 2212-1099$36.00 – see front matter Copyright & 2014, International Society for Pharmacoeconomics and Outcomes Research (ISPOR). Published by Elsevier Inc. http://dx.doi.org/10.1016/j.vhri.2014.06.007 26 VALUE IN HEALTH REGIONAL ISSUES 5C (2014) 25–28 Methods Latin America Asthma Insights and Management Survey The Latin America Asthma Insights and Management (LA AIM) survey [12,13] was designed to complement the US AIM survey and was conducted in 2011. Both the US and LA AIM surveys followed the methodology used in the Allergies in America survey, which was designed by physician experts from various regions who reviewed and approved survey questionnaires [14]. Subjects of the survey included in our analyses were 2168 patients with asthma aged 12 years or older identified from a sample of 51,208 households in Argentina, Brazil, Mexico, Puerto Rico, and Venezuela. These countries were selected because preliminary data indicated that the size of the asthma population in these countries was substantive enough to obtain sufficient sample sizes for analysis. Participants were selected randomly using national probability sampling by survey organizations based in each country [15]. Interviews were conducted in person and averaged 35 minutes in length. For subjects aged between 12 and 17 years, the survey was completed by a parent or guardian. The 53 survey questions were designed to assess asthma burden (short- and long-term symptoms, functional impact, and health care utilization such as provider visits and use of asthma medications) and patient beliefs on appropriate use of medications for asthma control. In addition, demographic variables, such as sex, age, education level, and diagnosis of nasal allergies, were queried [15]. Classification of Asthma Control Patients were characterized on the basis of GINA guidelines [9] into three categories of level of asthma control: well-controlled, partly controlled, and uncontrolled (see Table 1 in Supplemental Materials found at http://dx.doi.org/10.1016/j.vhri.2014.06.007). The manifestations of asthma that were used for the categorizations included daytime symptoms, restriction of daily activities, nighttime symptoms, and need for reliever/rescue treatment. We used questions from the surveys that correlated with the GINAdefined asthma manifestations, with the exception of a lung function measurement of less than 80% predicted, because lung function measurements were not queried. Statistical Analysis Respondents to the survey provided data on the number of times they took oral steroids, inhalers for quick relief/rescue, and longterm maintenance medications for the management of asthma symptoms in the previous 12 months. They also reported the number of times they visited the emergency room, their health care providers, or were hospitalized for asthma symptoms in the past 12 months. Finally, respondents reported the number of days they missed school or work in the past year. This article evaluated total costs (e.g., payer, provider, and patient out-ofpocket costs) that could be attributed to each asthma-related medication and health service encounter. In Brazil, costs were determined from Simpro, a Brazilian company that collects and publishes the prices of medical costs that are released by the governmental Agencia Nacional de Vigilância Sanitária [16], and KairosBrasil, an aggregator of pharmaceutical dosing and cost data in Brazil [17]. In Argentina, costs were obtained from consultations with expert physicians and pharmacists. In Mexico, costs were obtained by contacting Hospital Infantil de Mexico Federico Gomez, Hospital Medica Sur, and Farmacia del Ahorro. Because Puerto Rico is an unincorporated territory of the United States and costs of drugs and health services are tied to Centers for Medicare & Medicaid Services reimbursement rates, we used 2011 Centers for Medicare & Medicaid Services reimbursement rates [18], adjusted for the Puerto Rico geographic pricing cost index, to calculate Puerto Rican costs. In Venezuela, costs were obtained from the Hospital de Clinicas Caracas and by visiting local pharmacies. Because costs varied substantially for patients treated in the public versus the private sector, unit costs were weighted by the proportion of patients with access to public/ private health care in each country. To estimate the cost of days of missed work or school, we calculated the average income per day for each country by dividing the yearly average annual income by 240, the approximate number of working days per year. Only those respondents who were employed or younger than 18 years (assuming that a parent had to miss work to care for a child who missed school because of asthma) at the time of the survey were included in the analysis of the relationship between the level of asthma control and missing school or work. To allow comparison of costs between countries, all currencies were converted to US dollars using the exchange rates as of November 1, 2013. Mean annual costs for each medication/ utilization were calculated by multiplying the average unit costs by the number of instances of utilizations in the previous 12 months reported by each AIM survey respondent. Finally, we totaled each participant’s reported costs for all medications and instances of utilization over the previous 12 months to obtain the total mean annual direct costs of asthma care. Frequencies and chi-square tests (or Fisher exact test when the n of any cell was o5) were calculated for demographic variables such as age, sex, and level of education. Next, differences by GINA-defined level of asthma control in rates of utilization of asthma-related medications, use of health care services, and missed days of school and work were examined. We calculated adjusted costs using negative binomial regression, controlling for age and sex. To test the appropriateness of using negative binomial regression, we examined the ratio of each model’s deviance to the degrees of freedom. Because all ratios were approximately 1, we assumed that the models fit the data well [19]. Because variables, particularly costs, tended to vary widely by country of residence, all analyses were stratified on country. Analyses were conducted using SAS for Windows, version 9.3 (SAS Institute, Inc., Cary, NC). This study was determined not to meet the federal regulatory definition of human subjects’ research and hence was exempt from Human Subjects review by the University of Washington Institutional Review Board. Results Demographic characteristics of the respondents to the AIM survey, stratified on level of asthma control, are presented in (see Table 2 in Supplemental Materials found at http://dx.doi.org/ 10.1016/j.vhri.2014.06.007). The greatest proportion of patients with well-controlled asthma were in Brazil (9.3%) and Mexico (9.0%); only 3% of the patients in Venezuela had well-controlled asthma. In Brazil and Puerto Rico, respondents in the older age categories tended to have partly controlled and uncontrolled asthma compared with younger respondents. In all countries, males were more likely than females to have well-controlled asthma but the difference was statistically significant only in Argentina, Brazil, and Puerto Rico. In all countries, greater education was associated with greater proportions of wellcontrolled asthma but this relationship was statistically significant only for respondents from Brazil. Next, we examined the mean number of times that respondents used asthma medications and health care services pertaining to their asthma, stratified on level of asthma control (see Fig. in Supplemental Materials found at http://dx.doi.org/10.1016/j. vhri.2014.06.007). In all countries, patients whose asthma was characterized as uncontrolled had statistically significantly greater frequency of asthma medication (oral steroids, inhalers for quick relief/rescue, and long-term maintenance medicines) use as well as visits to the emergency room and health care providers than did patients whose asthma was well-controlled. Relative to patients with well-controlled asthma, we observed statistically significantly increased hospitalizations in patients with partly controlled and uncontrolled asthma in Argentina, Brazil, Mexico, and Puerto Rico, but in Venezuela, the respondents with well-controlled asthma had the greatest mean number of hospitalizations. In all countries, we observed statistically significant trends of increased missed school or work with decreased level of asthma control. The mean costs, converted to US dollars, of each asthma-related medication and health care service in each country for each level of asthma control are presented in Table 3 in Supplemental Materials found at http://dx.doi.org/10.1016/j.vhri.2014.06.007. Costs were generally highest in Argentina and Puerto Rico and lowest in Brazil, Mexico, and Venezuela. In all countries, we observed significant trends of increasing costs of oral steroids, inhalers for quick relief/ rescue, and long-term maintenance medication for decreasing level of asthma control. We also observed statistically significant trends of increasing costs of emergency health care and health care provider visits and costs of missing school or work in all countries. In all countries except Venezuela, costs of hospitalizations were also statistically significantly greater with decreasing level of asthma control. The greatest absolute differences in cost were seen for missed school/work (e.g., a patient with well-controlled asthma in Argentina cost an average of $12/y of missed school/work compared with $2600/y for a patient with uncontrolled asthma) and for hospitalizations (a patient with well-controlled asthma in Brazil had mean costs of $40/y for hospitalizations compared with $710/y for a patient with uncontrolled asthma). Total adjusted costs for queried asthma-related medications and services and missed school/work showed significant trends by decreasing level of asthma control in all countries (data not shown). Discussion This study demonstrated that Latin American patients whose asthma was characterized as partly controlled or uncontrolled reported using asthma-related medications and health care services and missing school/work at greater rates and had greater costs of care for their asthma than did patients whose asthma was characterized as well-controlled. This study indicates that asthma that is not well-controlled creates a significant cost burden in Latin American countries and successful interventions that could be taken to manage asthma more effectively might result in cost savings for these countries. Examples of such interventions include reducing the costs of asthma medications to increase patient adherence to medication regimens, which would likely decrease costly health care utilizations such as unscheduled provider visits and hospitalizations, improving training of health care providers, and providing information and education to patients and their caregivers to improve selfmanagement of asthma [4]. Although multiple studies in developed countries have shown that these types of interventions are cost-effective (see Labre et al. [20] for review), further research is needed to determine whether they would be cost-effective in Latin American countries. The LA AIM survey was similar in design to the Asthma Insights and Reality in Latin America survey, conducted in 10 Latin American countries (including Argentina, Brazil, Mexico, and Venezuela) in 2003 [6,21]. One publication using data from this survey used 2003 GINA guidelines to classify the level of asthma control by symptom severity using patient-reported asthma manifestations such as daytime and nighttime symptoms, exercise-induced symptoms, and total symptom severity and also found that more severe forms of asthma were associated with significantly greater costs compared with mild intermittent asthma symptoms [6]. In addition, we conducted a similar analysis of the costs of asthma medications and health care utilization among US patients and determined that patients with asthma that was not well-controlled used more health care resources and had substantially greater medical costs than did patients whose asthma was classified as well-controlled (among patients with well-controlled asthma, mean annual asthmarelated costs were $750 compared with $4140 for patients with uncontrolled asthma) [22]. Finally, an article by Sullivan et al. [23] compared costs among patients with controlled and uncontrolled asthma, characterized using guidelines from the Gaining Optimal Asthma Control study [24], and found that costs for patients whose asthma was uncontrolled were more than twice the costs for patients whose asthma was controlled. Although our study included a large population of Latin American patients with asthma, it had several limitations. Because we did not have data on nonrespondents, we cannot be sure that those who agreed to take part in the survey were representative of the Latin American population of patients with asthma and results of this analysis might not be generalizable. In addition, we did not sample in all Latin American countries and notably did not include the populous countries Columbia, Chile, or Peru, so we cannot be sure that our results are generalizable to other Latin American locations. In some countries surveyed, official costs of asthma medications and health services are not available, so we made every effort to obtain accurate estimates of costs by contacting representative health care delivery systems, but actual costs may have varied. Although the wording of the survey questions that we used to define well-controlled, partly controlled, and uncontrolled asthma was similar to that used in the GINA guidelines (see Table 1 in Supplemental Materials found at http://dx.doi.org/10.1016/j.vhri.2014.06.007), it did not correlate perfectly and our study population did not provide information on forced expiratory volume in 1 second so we were not able to use that information in our classifications. Also, our analyses were based on self-reported information, which may have been biased. Finally, although this article reports associations between poorly controlled asthma and health care costs, the data were cross-sectional and thus, we cannot address issues of causality in this study. Using a large population of Latin American patients with asthma, this study demonstrated that the health care costs of asthma that was not well-controlled were substantial in all the five countries that were sampled. Steps to achieve control over asthma symptoms, especially among patients belonging to communities that lack access to quality care, would likely reduce the clinical and economic burden of asthma in Latin America. Acknowledgments We thank Magdalena Sánchez Aguilar, Virginia Blandon, Valeria Garran, Homero A. Monsanto, Hugo E. Neffen, Marvin Rock, and Nancy Smith for their assistance with obtaining cost data. Source of financial support: This work was sponsored by a grant to the University of Washington from Merck & Co., Inc. Supplemental Materials Supplemental materials accompanying this article can be found in the online version as a hyperlink at http://dx.doi.org/10.1016/j. 28 VALUE IN HEALTH REGIONAL ISSUES 5C (2014) 25–28 vhri.2014.06.007 or, if a hard copy of article, at www.valuein healthjournal.com/issues (select volume, issue, and article). R EF E R EN CE S [1] Asher MI, Montefort S, Bjorksten B, et al. Worldwide time trends in the prevalence of symptoms of asthma, allergic rhinoconjunctivitis, and eczema in childhood: Isaac phases one and three repeat multicountry cross-sectional surveys. Lancet 2006;368:733–43. [2] Masoli M, Fabian D, Holt S, et al. The global burden of asthma: executive summary of the GINA dissemination committee report. Allergy 2004;59:469–78. [3] Neffen H, Baena-Cagnani CE, Malka S, et al. Asthma mortality in Latin America. J Investig Allergol Clin Immunol 1997;7:249–53. [4] Fischer GB, Camargos PA, Mocelin HT. The burden of asthma in children: a Latin American perspective. Paediatr Respir Rev 2005;6:8–13. [5] Szefler SJ. Advancing asthma care: the glass is only half full! J Allergy Clin Immunol 2011;128:485–94. [6] Neffen H, Gonzalez SN, Fritscher CC, et al. The burden of unscheduled health care for asthma in Latin America. J Investig Allergol Clin Immunol 2010;20:596–601. [7] Carranza Rosenzweig JR, Edwards L, Lincourt W, et al. The relationship between health-related quality of life, lung function and daily symptoms in patients with persistent asthma. Respir Med 2004;98:1157–65. [8] Farnik M, Pierzchala W, Brozek G, et al. Quality of life protocol in the early asthma diagnosis in children. Pediatr Pulmonol 2010;45:1095–102. [9] Global Initiative for Asthma (GINA) Committee. Global Strategy for Asthma Management and Prevention, Revised 2010. In: (GINA) GIfA, ed. Bethsda, MD: Global Initiative for Asthma (GINA), 2010. [10] National Asthma Education and Prevention Program (NAEPP) Coordinating Committee. Expert Panel Report 3: Guidelines for the Diagnosis and Management of Asthma: National Asthma Education and Prevention Program (NAEPP). Bethesda, MD: National Heart, Lung, and Blood Institute (NHLBI), 2011. [11] Gold LS, Smith N, Allen-Ramey FC, et al. Associations of patient outcomes with level of asthma control. Ann Allergy Asthma Immunol 2012;109(260–5):e2. [12] González-Díaz S, Maspero J, Jardim J, et al. Latin America Asthma Insight and Management (LA AIM): a survey of asthma patients in 5 Latin American locales. World Allergy Organ J 2012;5(Suppl 2): S106. [13] Jardim J, González-Díaz S, Maspero J, et al. Understanding of asthma terminology by patients interviewed in the Latin America Asthma Insight and Management (LA AIM) survey. World Allergy Organ J 2012;5 (Suppl 2):S123. [14] Meltzer EO, Blaiss MS, Naclerio RM, et al. Burden of allergic rhinitis: Allergies in America, Latin America, and Asia-Pacific adult surveys. Allergy Asthma Proc 2012;33(Suppl. 1):S113–41. [15] Maspero JF, Jardim JR, Aranda A, et al. Insights, attitudes, and perceptions about asthma and its treatment: findings from a multinational survey of patients from Latin America. World Allergy Organ J 2013;6:19. [16] SIMPRO. Revista Simpro Hospitalar. Available from: http://www. simpro.com.br/empresas/index.php?search=1. [Accessed December 1, 2012]. [17] KairosBrasil. Kairosbrasil. Available from: http://brasil.kairosweb.com/ producto-alenia-16925. [Accessed February 8, 2013]. [18] Medicare Program: payment policies under the physician fee schedule, five-year review of work relative value units, clinical laboratory fee schedule: signature on requisition, and other revisions to part B for Cy 2012. Final rule with comment period. Fed Reg 2011;76:73026– 73474. [19] Institute for Digital Research and Education University of California at Los Angeles (IDRE UCLA). SAS annotated output: negative binomial regression. Available from: http://www.ats.ucla.edu/stat/sas/output/ sas_negbin_output.htm. [Accessed March 11, 2014]. [20] Labre MP, Herman EJ, Dumitru GG, et al. Public health interventions for asthma: an umbrella review, 1990-2010. Am J Prev Med 2012;42:403–10. [21] Neffen H, Fritscher C, Schacht FC, et al. Asthma control in Latin America: the Asthma Insights and Reality in Latin America (AIRLA) survey. Rev Panam Salud Publica 2005;17:191–7. [22] Gold LS, Yeung K, Smith N, et al. Asthma control, cost and race: results from a national survey. J Asthma 2013;50:783–90. [23] Sullivan SD, Rasouliyan L, Russo PA, et al. Extent, patterns, and burden of uncontrolled disease in severe or difficult-to-treat asthma. Allergy 2007;62:126–33. [24] Bateman ED, Boushey HA, Bousquet J, et al. Can guideline-defined asthma control be achieved? The Gaining Optimal Asthma Control study. Am J Respir Crit Care Med 2004;170:836–44. VALUE IN HEALTH REGIONAL ISSUES 5C (2014) 35–39 Available online at www.sciencedirect.com journal homepage: www.elsevier.com/locate/vhri Assessment of Health Care and Economic Costs Due to Episodes of Acute Pesticide Intoxication in Workers of Rural Areas of the Coquimbo Region, Chile Muriel Ramírez-Santana, MA1,*, Juan Iglesias-Guerrero, MA1, Marianela Castillo-Riquelme, MA2, Paul T.J. Scheepers, PhD3 1 Faculty of Medicine, Public Health Department, Universidad Católica del Norte, Coquimbo, Chile; 2Division of Health Planning, Department of Health Economics, Ministry of Health, Santiago, Chile; 3Department for Health Evidence, Radboud University Medical Center, Nijmegen, The Netherlands AB STR A CT Background: The increase in agricultural activity that Chile experienced in the past 20 years resulted in a boost in the use of pesticides. Despite pesticides’ productivity benefits, they caused health problems such as the increased frequency of episodes of acute poisoning, which constitutes a relevant problem in terms of occupational health. The Chilean authorities require several preventive measures at workplaces, which are not always implemented, increasing the risk of intoxications in farmers. So far in Chile, there are no studies concerning the public health care expenses associated with acute work-related pesticide intoxications. From the societal perspective, there are costs involved if the worker needs to take sick leave and families incur costs to take care of their sick members. Objectives: This study aimed to determine the costs associated with health care services used by people who suffered from work-related acute pesticide poisoning, as well as the economic costs for the families of the workers involved, and finally the costs of these episodes for the employer/industrial sector. Methods: This study considered a 3-year period (January 2009 to December 2011). Three sources of data were reviewed: reported cases at the Regional Health Authority, for the profile of the intoxications; registers of patients attended in public hospitals, for data on costs of health care services; and public information of living conditions nationwide. Results: The overall costs of a single case depend on the severity of intoxication, days of sick leave, and type of health care needed. Most cases (77%) would be ambulatory and would be assisted at an emergency room, with an average cost of US $330 per case. Those cases that might need hospitalization (23%) and, therefore, more days off work have an average cost of US $1158 per case. Taking into account the number of patients reported each year in the country, the cost per annum would be about US $185,000, but considering the underreporting of intoxications and underestimation of costs at the public insurance system fees, this amount could be six to eight times higher (US $1.1 million to US $1.4 million a year). Conclusions: This study is the first attempt to estimate costs related to this occupational and public health problem. There is an opportunity to further improve the cost-benefit balance of preventive measures that relate not only to acute poisonings but also to health problems originating from chronic and low-dose exposure to pesticides (e.g., neuropsychological impairment). Keywords: cost analysis, occupational health, pesticide poisoning. Introduction productivity, but the harmful effects that such applications have on the health of human populations are also well known and have been demonstrated. The Coquimbo region is located in the central/north of Chile, with a total of 720,000 inhabitants. Agriculture is the primary economic activity in this region, with about 45,000 workers [6]. The Regulation on Basic Sanitary and Environmental Conditions in the Workplace in Chile [7] requires safety and preventive measures that should be adopted by the employers at workplaces. These measures focus on providing personal protection equipment, safe storage, restrictions when using hazardous products (pesticides among others), as well as specific reentry times after the use of pesticides. This regulation also establishes According to the ODEPA Employment Bulletin, the workforce in Chile for the period March to May 2012 amounted to 7,613,709 persons, 694,047 of whom were linked to agriculture [1]. Data from the International Labour Organization indicate that Chile exported agricultural products worth US $5.411 million in 2007, primarily fruits (US $3.351 million), wines (US $1.241 million), and seeds (US $230 million) [2–4]. Over the past 20 years, the import of pesticides increased by 469%. The imports of insecticides for agricultural use increased by approximately 20% between 2001 and 2005 [5]. The use of pesticides contribute to the enhancement of agricultural Copyright & 2014, International Society for Pharmacoeconomics and Outcomes Research (ISPOR). Published by Elsevier Inc. Conflict of interest: The authors have indicated that they have no conflicts of interest with regard to the content of this article. * Address correspondence to: Muriel Ramírez-Santana, Faculty of Medicine, Universidad Católica del Norte, Larrondo 1281, Coquimbo, Chile. E-mail: [email protected]. 2212-1099$36.00 – see front matter Copyright & 2014, International Society for Pharmacoeconomics and Outcomes Research (ISPOR). Published by Elsevier Inc. http://dx.doi.org/10.1016/j.vhri.2014.07.006 36 VALUE IN HEALTH REGIONAL ISSUES 5C (2014) 35–39 the permissible limits for biological tolerance values in the case of exposed workers and limits for the environmental exposure to physical and chemical agents. Supreme Decree N1 594 establishes the limit of biological tolerance at 70% of the baseline levels of cholinesterase activity in an individual. This implies that measurements of enzymatic activity must be repeated, a baseline one before exposure and a second one after exposure, thus duplicating the cost of diagnosis. The cost of a single cholinesterase evaluation at the National Institute of Public Health is US $72.98 [8]. The compliance assessment is the task of the Regional Ministerial Secretariats of Health (Sanitary Authorities—Ministry of Health). There is also a national network on acute poisoning surveillance (red de vigilancia epidemiologica de intoxicaciones por Pesticidas [Pesticides Intoxications Network]), created by Supreme Decree Nº 88, dated October 2004 [9]. In the period 2009 to 2011, the total number of poisoning cases notified to the red de vigilancia epidemiologica de intoxicaciones por Pesticidas was 1514. Of these, 932 were work-related (61.6%) and 39% of the intoxications were caused by organophosphate pesticides [10–13]. The real number of poisonings is estimated to be much higher because, in general, the affected workers do not report the incidents (they fear consequences in terms of their employment status) or there are no health care records due to lack of information about the effects of the pesticides and due to underdiagnosis. The total number of poisonings is estimated to be three to four times higher than the official registrations, thus corresponding to around 3000 annual poisoning cases. Acute poisonings among agricultural workers generate a cost to the country (considering the demand for health care services), loss of workforce (and a need for its replacement), and sick leave due to the poisoning, with an average of 6 days for each sick leave [14,15]. Up to date, there is no previously published study regarding the overall costs associated with poisonings due to exposure to pesticides at work. This study aimed to determine the costs associated with health care services used by people who suffered from work- related acute pesticide poisoning, as well as the economic costs for the families of the workers involved, and finally the costs of these episodes for the employer/industrial sector. Methods This cost analysis was conducted from the societal perspective, and it includes health care costs, costs of productivity losses from poisoning episodes, and the costs of seeking and/or undergoing care as they fall on patients. During the investigation, it became clear that not all cases admitted in hospitals were reported to the Epidemiology Unit (Regional Health Authority, Ministry of Health) and vice versa; not all reported cases were attended at the public hospitals; probably many had gone to a primary health care facility in rural areas and/or to the private system when near to a city. Therefore, three sources of data were assessed for the study period January 2009 to December 2011: (1) Information regarding pesticide intoxications reported at the Epidemiology Unit (Ministry of Health), which corresponds to 57 work-related poisonings. From this source, the profile of the poisonings, average days off work, and days of sick leave were known. (2) Information about health care services from the three main public hospitals in the region was reviewed, taking into account people who suffered acute intoxication due to pesticides and who required hospitalization, corresponding to 26 patients. (3) Finally, information from public Web sites and publications related to national living costs and conditions was searched. For more detail, refer to Figure 1. From the hospitals, data to assess the health care costs were taken from the 26 patients’ files, discharges registers, and hospital administration registers. Access to medical files was approved by the directors of the health facilities. For the analysis of health care costs, the study considered only people of working age (women aged 18–60 years and men aged 18–65 Informaon Source 1 Primary Health Care (Rural) Cases of occupaonal pescide poisonings, Coquimbo Region, Chile. 2009-2011. Hospitals (registers in emergency unit and inpaents wards-ICU) •La Serena •Coquimbo •Ovalle Informaon Source 2 Informaon Source 3 Public informaon: Bus fee, food basket, US $ Currency value, Minimum wage, Naonal Health Found fees. N = 26 Reporng Epidemiology Unit Health Authority (MoH) Reportes cases: Total = 92 Work-related = 75 of which Agricultural work = 57 Cost assessment of Health Care Services Hospitalizaon (23%) US $794.1 per case (av.) Ambulatory (77%) US $25.5 per case (av.) Cost esmaon for family /household/ farm Hospitalizaon (23%) US $347.64 per case (av.) N = 57 Populaon profile: •Gender: M 52.6%, F 47.4% •Age average 34.5 years •Exposure: Direct 60%, Noncompliance with regulaon 23%, spray 3,5% •Pescide involved: OPP 42% , Cyanamide 33% •Primary Health Care 87% •Hospital Care 10% •Average Sick leave 7 days Ambulatory (77%) US $304.56 per case (av.) Fig. 1 – Chart of methods, sources of information, sampling strategies, and main results. av., average; F, female; ICU, intensive care unit; M, male; MoH, Ministry of Health; OPP, organophosphate pesticide. 37 VALUE IN HEALTH REGIONAL ISSUES 5C (2014) 35–39 years) whose diagnosis was associated with a work-related poisoning. All cases were seen and treated at public hospitals under coverage by the National Health Fund (Fondo Nacional de Salud [FONASA]), which covers 76.5% of the regional population [16]. Cases from the private insurance system were not considered because private providers refused to give information. Health care costs were estimated using the microcosting methodology, looking into the different kinds of care/services provided to every person treated [17–19]. The cost was measured on the basis of invoiced care services registered at each of the hospitals. Data to determine the package of health care services were completed by the review of medical files. The items that were considered for valuation, when hospitalization was required, were medical consultation at emergency room, daybed at intensive care unit, day-bed in general medical ward, kinesiology sessions, ambulance transport, diagnosis support (laboratory, electrocardiogram, x-ray), and medication. For those who were not hospitalized, medical records showed that no diagnostic support was provided (x-ray, laboratory work, or others); medical consultation at emergency room, medicines, and other pharmaceutical items were included for valuation in the investigation. The costs of the medical items were estimated from the prices defined each year by FONASA, published in the Official Gazette on January 28, 2013. As for the medicines, values were obtained from average prices that every hospital considered when purchasing medicines in the first quarter of the year 2013. For all costs, the considered currency rate was 472.62 Chilean pesos per US $1, as of January 15, 2013 [20]. For the productivity loss at the workplace, the minimum wage was assumed to be US $409/mo [21]. This considers 20 working days a month, assuming a cost per day of US $20.46. Production losses associated with acute pesticide poisoning were assumed to correspond to the days of absence from work on the basis of notified cases at the Epidemiology Unit. Data on the days of sick leave were taken from the medical files of the notified cases. Because all cases were from rural areas, it was assumed that at least one family member would have to travel into the city on a daily basis (no cost for overnight stay was considered). The minimum transport fee was US $3.18 one-way trip (two trips per day were assumed, either for ambulatory cases or for hospitalized cases). Regarding food, the costs were estimated from the value of minimum food basket, that is, about US $80.6/ mo (US $2.7/d), according to the Ministry of Social Development (as of January 2013) [22]. Results Profile of Work-Related Intoxications The description of poisonings is based on reported cases of acute pesticide intoxication due to work-related activities in the study period at the Epidemiology Unit in the Coquimbo region (N ¼ 57). For more details on characteristic of the population, route of exposure, and type of pesticides involved, refer to Figure 1. Health care was provided mainly at the primary level; 86% of the patients were sent back home and only 10.5% were referred to a hospital. For two cases (3.5%), the type of medical support was not registered. Finally, 34 cases registered days off work because of sick leave, with an average duration of 7 days. Cost Assessment on Health Care Data on a total of 26 cases diagnosed with work-related acute poisoning were obtained from the three main public hospitals of the region over the study period. Of the 26 described cases, 20 (77%) sought assistance through ambulatory consultations (emergency room) and 6 (23%) required hospitalization. The length of hospital admission was on average 1 day at the intensive care unit and 4 days at the medical ward, based on medical files registers. For ambulatory cases, the cost was US $25.5 per case. As for the hospitalized cases, the cost per case was US $794. No fatalities were registered during the period of the study. See detailed figures in Table 1. Cost Assessment for Families, Households, or Farmers An average of 7 workdays of medical leave was calculated from reported cases. The costs to the families correspond to a minimum of US $29.52/d. The calculation of family costs per case depended on the severity of the poisoning, the type of health care required, and the period for recovery. For 77% of the cases referred to emergency room (ambulatory cases), the estimated overall cost was calculated to be US $330.05 per case. As for the ones who needed to be hospitalized, one family member was assumed to travel and to have one daily meal in the city during the stay of the patient at the hospital (5 days on average). The same 7 days off for medical reasons were counted as a loss of productivity (in the farm) and loss of family income. The health care cost calculated per case was US $810; adding the productivity Table 1 – Overall cost in US $, per day and per case, of occupational poisonings due to pesticides: Coquimbo region, Chile (2009–2011). Cost of items regarding family care and health care Family care/household/farm Productivity loss Family cost Family lost income Subtotal Health care services Medical consultation at emergency room Hospitalization (1-d ICU þ 4-d medical ward) Ambulance transportation Pharmacy items (medicines and medical supplies) Diagnosis support (laboratory, x-ray, ECG) Subtotal Total cost per case ECG, electrocardiogram; ICU, intensive care unit. Ambulatory Hospitalized Per day Per case 20.46 18.12 20.46 59.05 143.22 18.12 143.22 304.56 22.6 – – 2.98 – 25.49 330.05 Per day 20.46 12.24 20.46 53.16 22.6 550.35 25 127.79 85.25 810.99 1158.63 Per case 143.22 61.2 143.22 347.64 38 VALUE IN HEALTH REGIONAL ISSUES 5C (2014) 35–39 loss and family costs, each severe case costs US $1158. For details, see Table 1. To have a broader view on costs, this study considered estimating the cost on the basis of the reported cases, that is, 75 cases of work-related poisoning notified at a regional level and 933 cases in the country in 3 years [8]. For an overall estimation of cost per year, the number of cases was divided by 3. The overall estimated cost per year would amount to US $14,915 for the Coquimbo region and US $184,947 on a national level. Discussion García [15] reported that for intoxications in the United States, costs depended on the severity of the case because this determines the type of health care needed and the number of days required for recovery. The same situation was found in the present study. Although the costs of health care for ambulatory cases were only US $25, the health care costs of severe cases were 30 times higher. However, the cost for families and productivity loss were similar in any of the situations. Nevertheless, the expenses of these cases on families correspond to a substantial part of the minimum income (50%). The costs calculated for hospitalization (average of US $810/d) are similar to the costs published by García [15] regarding hospital health care for pesticide intoxication in the United States during the 1990s (US $1000/d). The expenses for ambulatory cases are much lower compared with costs calculated for the United States (US $25 and 630 per case, respectively). A possible explanation is the low cost rate for medical consultations in primary health care in Chile, where diagnostic support was not provided in any of the studied cases. When looking at the medicines prescribed for ambulatory cases, no atropine was used and the administration of antidote (oxime) was reported only in two cases. If those medicines were prescribed more frequently, the cost of consultations would probably be twofold higher. As was explained, the costing analysis of health care services used unit costs that correspond to reference fees supplied by FONASA, which are known to be underestimated and therefore do not reflect the cost of the services in economic terms [23]. Also, the costing analysis includes productivity losses in terms of people being absent from work for being ill and/or seeking and undergoing health care. At a societal level, these costs may be an overestimation of the real costs because, considering unemployment, these productivity losses can be mitigated by recruiting other workers. If precautionary measures were observed, most of the poisonings would be prevented. Although preventive measures are enforced by law as an employer obligation, this is only for workers who have signed a contract. Unfortunately, in agricultural settings, temporary work occurs with no contracts as well as involves family farming (self-employed). It is well known that public resources to implement preventive programs and education in agricultural areas where workers are not part of any insurance system are scarce. Besides using protective equipment and clothing, one of the recommended preventive measures for acute intoxication is health surveillance by biomonitoring the workers. Plasma detection of cholinesterase activity should be conducted two times (as was explained before). Simple test evaluations of cholinesterase levels would be expected to be more cost-effective than the cost generated by the poisonings. The cost of assessing a single patient is two times the cost of cholinesterase, which corresponds to US $146 [12], which is about 20% of the cost of one of the less severe cases (ambulatory care). Considering that the estimated incidence would be one case per 500 workers, US $70,000 would be needed to diagnose a single case. The question arises: is cholinesterase detection an effective diagnostic test for poisoning? Scientific evaluations of this test show that the applicability is not as good as expected [24], with wide variation in the test results among people tested. More clinical diagnosis based on signs and symptoms, plus exposure information, could represent a cost-effective alternative solution to an expensive laboratory test, but physicians must have the knowledge to properly identify cases and treat them. The study has limitations: First, the Coquimbo region contributes with only about 3% of the cases that are reported on a national scale. Therefore, the studied population could not be considered to be representative of the entire nation. Second, the cost analysis is based on fees that poorly represent the real costs as is described in the study by Camilo et al. [23] about FONASA costs, in which costs appear to be on average 84% higher than established rates, showing differences ranging from 77% to 634%. Therefore, it is likely that estimated costs per case of the present study are underestimated. Regarding the approximate overall yearly costs for the Coquimbo region and Chile, they were calculated on the basis of reported cases. The estimated nationwide total cost was about US $185,000, which would be three to four times higher, considering that underreporting is suspected. Given all these considerations, estimations of cost would be probably six to eight times higher, reaching US $ 1.1 million to US $1.4 million a year. Finally, besides costs and consequences of acute poisoning, there are well known neuropsychological effects of chronic exposure to pesticides [25–27]. The social and economic impact of such long-term health effects is difficult to study but could contribute significantly to the overall public health problem. This study is the first step for further investigations that could assess cost-benefits of preventive measures, considering both acute and chronic health outcomes in workers exposed to pesticides. Source of financial support: The authors have no other financial relationships to disclose. R EF E R EN C ES [1] Oficina de Estudios y Políticas Agrarias (ODEPA), Ministerio de Agricultura. Base de Datos Estadísticos. Available from: www.odepa.cl/ estadisticas/macrorubros/agricultura. [Accessed June 20, 2012]. [2] Instituto Nacional de Estadísticas (INE). Compendio Estadístico 2011. Available from: www.ine.cl/publicaciones/compendio estadístico 2011. [Accessed June 23, 2012]. [3] VII Censo Agropecuario y Forestal. INE (Instituto Nacional de Estadísticas) (2007). Available from: http://www.censoagropecuario.cl/. [Accessed June 20, 2014]. [4] Cambios Estructurales en la Agricultura Chilena. Análisis Intercensal 1976, 1997, 2007. Santiago, Chile: Instituto Nacional de Estadística, 2007. [5] Servicio Agrícola y Ganadero (SAG), Sub-Departamento de Plaguicidas y Fertilizantes (www.sag.gob.cl). Declaración de ventas de plaguicidas 2004. Chile: Ministry of Agriculture, 2006. [6] Síntesis de estadísticas mensuales, Región de Coquimbo, Abril 2010. Available from: http://www.inecoquimbo.cl/archivos/files/pdf/Boletin/ 2010/5/Boletin_31_05_2010.pdf. [Accessed May 13, 2013]. [7] Decreto Supremo N1 594, Aprueba reglamento sobre condiciones sanitarias y ambientales básicas en los lugares de trabajo, publicada en el Diario Oficial el 15 de Septiembre de 1999. [8] Instituto de Salud Pública de Chile. List of laboratory test and fees. Available from: http://www.ispch.cl/prestaciones? field_codigo_value=&title=colinesterasa. [Accessed May 16, 2013]. [9] Decreto Supremo 88, Norma Técnica de Vigilancia de Intoxicaciones Agudas por Plaguicidas, REVEP, Art. 2, publicado en Diario Oficial el 19 de Marzo del año 2007. Available from: http://epi.minsal.cl/epi/ 0notransmisibles/cancer/Decreto-88.pdf. [Accessed June 20, 2014]. [10] Vallebuona C. Red Nacional de Vigilancia Epidemiológica de Intoxicaciones Agudas por Plaguicidas REVEP. Boletín Electrónico Mensual de Vigilancia Epidemiológica Ministerio de Salud 2008;60:11. [11] Clelia Vallebuona Stagno, REVEP. Available from: http://epi.minsal.cl/ epi/html/vigilan/revep/revepwebfinal.htm. [Accessed July 7, 2012]. [12] Boletin electronico mmensual de Vigilancia Epidemiologica. Ministerio de Salud, Chile. Available from: http://epi.minsal.cl/vigilanciaepidemiologica/situacion-intoxicaciones-agudas-porplaguicidas/ informes-revep/. [Accessed June 20, 2014]. [13] Mena C, Bettini M, Cerda P, et al. [Epidemiology of intoxications in Chile: ten years of registry] [in Spanish]. Rev Med Chil 2004;32:493–9. VALUE IN HEALTH REGIONAL ISSUES 5C (2014) 35–39 [14] Bustamante UM, Campos TR. Contaminación por Plaguicidas en la Región del Maule, Chile. Panorama Socioeconómico 2004: Available from: http://www.redalyc.org/articulo.oa?id=39902804. [Accessed July 20, 2013]. [15] García JE. Intoxicaciones Agudas con Plaguicidas: Costos Humanos y Económicos. Revista Panamericana Salud Pública 1998;4:383–7. [16] Statistics Gazzette, National Health Found – FONASA, 2010–2011. [17] Guía Para el Uso de Evidencia de Costo Efectividad en Intervenciones en Salud. Minsal, Chile: División de Planificación Sanitaria, Departamento Economía de la Salud, Subsecretaría de Salud Públicas, 2011. [18] Rony L.A. Análisis de costos en evaluaciones económicas en salud: Aspectos introductorios. Rev Méd Chil. Available from: http://www. scielo.cl/scielo.php?script=sci_arttext&pid=S0034-98872010001000006& lng=es. [Accessed July 7, 2012]. [19] Reveco R, Vallejos C, Orellana J, Pavez M. Estudio Costo Efectividad de Intervenciones en Salud, Informe II: Costos, Centro Excelencia Capacitación, Investigación y Gestión para la Salud Basada en la Evidencia (CIGES). Temuco, Chile: Facultad de Medicina, Universidad de la Frontera, 2009. [20] Central Bank of Chile. Statistic database. Available from: http://www. bcentral.cl/bde/index.htm. [Accessed June 20, 2012]. 39 [21] Minimum wage 2012. Available from: http://www.tusalario.org/chile/ main/salario/salario-minimo. [Accessed October 7, 2013]. [22] Costo de canasta básica de alimentos. Observatorio Social. Available from: http://observatorio.ministeriodesarrollosocial.gob.cl/layout/doc/ ipc/13.02%20COSTO%20CBA%20ENERO_511a8479baa53.pdf. [Accessed June 20, 2012]. [23] Cid C, and cols. Estudio de costos de prestaciones de salud. Estudio realizado por la Pontificia Universidad Católica de Chile para FONASA. Santiago, Chile: FONASA, 2011. [24] Nigg HN, Knaak JB. Blood cholinesterases as human biomarkers of organophosphorus pesticide exposure. Rev Environ Contam Toxicol 2000;163:29–111. [25] Roldán-Tapia L, Leyva A, Laynez F, Sánchez Santed F. Chronic neuropsychological sequelae of cholinesterase inhibitors in the absence of structural brain damage: two cases of acute poisoning. Environ Health Perspect 2005;113:762–7. [26] Abou-Donia MB. Organophosphorus ester-induced chronic neurotoxicity. Arch Environ Health 2003;58:484–97. [27] Roldán-Tapia L, Parrón T, Sánchez-Santed F. Neuropsychological effects of long-term exposure to organophosphate pesticides. Neurotoxicol Teratol 2005;27:259–66. VALUE IN HEALTH REGIONAL ISSUES 5C (2014) 40–47 Available online at www.sciencedirect.com journal homepage: www.elsevier.com/locate/vhri Costo Utilidad de Colágena Polivinil-Pirrolidona en el Tratamiento de Osteartrosis de Rodilla en México Ximena Burbano-Levy, MD1,2,*, Diana Cardona, QF, PhD3, Elsa Palacios, Act1, Israel Rico Alba, MD, PhD1, León Zapata, QF1 1 Guia Mark, S.A de C.V. México D.F, México; 2Zilonis Health, Inc. Boca Ratón, Fl, USA; 3Línea de Investigación en Epidemiología y Economía de la Salud, Grupo Promoción y Prevención Farmacéutica, Facultad de Química Farmacéutica, Universidad de Antioquia, Medellín, Colombia. AB STR A CT Objective: To assess the cost-utility of collagen-polyvinyl pyrrolidone (collagen-PVP), a new drug developed and manufactured in Mexico, compared to non-steroidal anti-inflammatory drugs in the treatment of patients with knee osteoarthritis (OA) in Mexico. Methods: A probabilistic Markov model with a lifetime horizon was designed to evaluate the costs and outcomes (QALYs) of collagen-PVP compared to NSAIDs in patients with OA. Data from randomized clinical trials conducted in México were obtained as input for probabilities. Utility values were elicited from international publications and costs information and service utilization was obtained from official publications of the Instituto Mexicano de Seguro Social - IMSS. Cost per qualityadjusted life-years was calculated by using the Mexican Healthcare payer perspective. A discount rate of 5% for both cost and outcomes was used. Sensitivity analyzes were performed with variations in the parameters of gender, age, and price of collagen-PVP. Results: Patients treated with collagen-PVP showed a gain of 6.62 QALYs compared with 5.36 in patients with NSAIDs. The cost-utility ratio was higher in the NSAID group (MEX$ 170,010.76/QALY) compared to collagen-PVP (MEX $ 130,953.63/QALY). The incremental cost-utility ratio was MEX $ 35,194.16. Conclusions: This is the first economic evaluation of a new medication developed in Mexico and conducted partially with local data. The results of this study allow concluding that collagen-PVP is deemed as the dominant alternative compared with NSAIDs in the treatment of knee osteoarthritis. Keywords: cost-effectiveness, cost-utility, colagen, osteoarthoris. Introducción De otra parte, aunque únicamente el 15% de los pacientes diagnosticados con osteoartrosis de rodilla (OAR) requieren tratamiento [6], la revisión por parte de la OMS en el 2003 señala de manera importante la carga económica de la enfermedad, en particular por costos indirectos, relacionados con pérdida de productividad, incapacidad permanente, uso crónico de servicios de salud (incluyendo servicios de fisioterapia), medicamentos y costos asociados al tratamiento quirúrgico [7]. A estos factores se adicionan los riesgos asociados al uso crónico de antinflamatorios no esteroideos (AINEs) y a la exposición a cirugías de una población con mayor probabilidad de enfermedades crónicas, tales como obesidad, diabetes e hipertensión, lo que incrementa los riesgos para el paciente y los costos del sistema. En 1996, Lozano y cols [8], estimaron que la OA era la novena causa de años de vida perdidos por muerte prematura y años vividos con discapacidad en mujeres mayores de 60 años. Además, las personas mayores de 60 años perdieron en promedio 48 días laborales por todas las causas de discapacidad, y de estas el 12% era debido a osteoartritis y artritis. La artrosis representa un daño articular persistente que condiciona la pérdida de la adecuada función articular. En este Osteoartrosis (OA) es la enfermedad reumática más prevalente en el mundo. De acuerdo con los datos presentados por la Organización Mundial de la Salud (OMS), en el 2003, la OA afectaba clínicamente a 9.6% de los hombres y 18% de las mujeres en el mundo [1,2]. En un estudio de prevalencia en Yucatán, el 5.8% de la población encuestada fue diagnosticada con esta patología [3]. Posteriormente, una de las más recientes revisiones bibliográficas de las características epidemiológicas de la enfermedad en México, publicada en el 2011, señala una prevalencia del 2% en los menores de 45 años, 30% entre los 45 y 65 años y del 68% en los mayores de 65 años [4]. Sin embargo, a partir de los datos de la pirámide poblacional del año 2000, la OMS estimó para México, un número aproximado de 8 millones de habitantes con OA [1]. Las proyecciones de la OMS, sugieren un incremento mayor en la prevalencia de la OA dado el aumento en la esperanza de vida y factores de riesgo asociados a la enfermedad tales como la obesidad y sedentarismo [5]. Este hecho toma mayor importancia cuando en México y en general en los países en desarrollo el acceso a tratamientos como artroplastia y reemplazo de rodilla es restringido. Copyright & 2014, International Society for Pharmacoeconomics and Outcomes Research (ISPOR). Published by Elsevier Inc. Conflicts of interest: The authors have indicated that they have no conflicts of interest with regard to the content of this article. * Correspondencia: Ximena Burbano-Levy. MD. Zilonis Health, Inc. Telefono: þ1-561-8436767. E-mail: [email protected]. 2212-1099$36.00 – see front matter Copyright & 2014, International Society for Pharmacoeconomics and Outcomes Research (ISPOR). Published by Elsevier Inc. http://dx.doi.org/10.1016/j.vhri.2014.06.006 VALUE IN HEALTH REGIONAL ISSUES 5C (2014) 40–47 fenómeno, la degradación articular juega un papel predominante; sin embargo, otros fenómenos como la degradación del hueso subcondral o la adecuada producción de líquido sinovial representan el punto inicial de la patogenia. En las fases finales de este proceso se presenta la destrucción acelerada de la matriz por enzimas procedentes de los condrocitos y células sinoviales, lo cual acompañado de alteraciones en el sistema de reparación del cartílago, desencadena el deterioro de la articulación. Así pues, en el deterioro de la articulación participan mecanismos tanto enzimáticos como inflamatorios del tejido sinovial que tornan crónica la patología. En los últimos años se ha utilizado colágena polivinilpirrolidona (colágena-PVP), producto de desarrollo Mexicano, para el tratamiento de enfermedades fibrosantes, por sus efectos en la regulación de la Interleucina (IL-β), Factor de Necrosis Tumoral Alfa (TNF), PDGF, moléculas de adhesión VCAM-1 y ELAM-1 [9–14]. Hasta la fecha sus aplicaciones en el proceso de cicatrización de tejidos, incluido el óseo y la modulación en enfermedades como la artritis reumatoide ha significado importantes beneficios terapéuticos y un uso eficiente de los recursos con generación de ahorros en la atención de estos pacientes [15]. Recientemente Furuzawa y cols, 2012 [16] publicaron un ensayo clínico aleatorizado doble ciego del uso de colágena-PVP en el tratamiento de pacientes con OA después de un lavado artroscópico. En el estudio clínico se aplicaron 6 dosis intraarticulares de 2 mL de colágena-PVP durante 3 meses, con un seguimiento de 3 meses posteriores a la aplicación. La efectividad fue medida por la reducción en la intensidad del dolor en la escala visual análoga (VAS) e incremento en los puntajes de WOMAC (Western Ontario and McMaster Universities Osteoarthritis Index) y Lequesne. Así mismo se evaluaron los resultados de estudios paraclínicos y los parámetros clínicos de inflamación. Los resultados muestran una reducción del 2.9 (SD 0.6) y 3.1 (SD 0.6) a los 3 y 6 meses en la escala de VAS, respectivamente, siendo estadísticamente significativa la diferencia al contrastarla contra los resultados obtenidos en el grupo placebo. En cuanto al índice de Lequesne existió una reducción del puntaje en el 43 y 51% en las dos mediciones efectuadas y una reducción del WOMAC de 28% y 46% a los 3 y 6 meses. Por otra parte y de igual interés, se reportó un 83% de reducción en el consumo de AINEs. Adicional a los datos acerca del beneficio clínico, la generación de información relacionada con las consecuencias económicas del uso de colágena-PVP en OAR es necesaria, dado el reto que tienen los sistemas de salud para asignar adecuadamente los recursos financieros de los cuales disponen. El objetivo de este estudio fue estimar la relación costo-utilidad de colágeno-PVP en pacientes con OAR, a nivel de la institución pública para atención en salud más representativa de México. Métodos La elección de la técnica a utilizar para el análisis económico, se fundamentó en el efecto a largo plazo y la disminución de la progresión de la enfermedad con el uso de colágena-PVP en pacientes con OAR. Para tal efecto, y parcialmente con base en los estudios de Soto y cols [17] y Furuzawa y cols [9,16] se diseñó un modelo de Markov para evaluar los costos y resultados (años de vida ajustados por calidad – QALYs) del tratamiento con colágena-PVP (1.5 ml/semana x 6 semanas) comparado con AINEs (Diclofenaco Sódico 250 mg/día) en pacientes con diagnóstico de OAR. Datos de otras publicaciones fueron utilizados para desarrollar la simulación del comportamiento de los costos y efectividad el tratamiento [18–22]. El modelo fue desarrollado en MS Excel (Microsoft Corporation, Redmond, WA, U.S.A.). 41 Selección del comparador La selección del comparador, se basó en las recomendaciones de tratamiento de las guías del IMSS y la Secretaría de Salud de México [23–26]. Estas guías clínicas recomiendan el uso de paracetamol como primera línea de elección [23,24,27,28] y el uso de visco-suplementación en casos donde el tratamiento farmacológico no ha sido exitoso o está contraindicado [23,24,27]. La segunda línea de tratamiento la conforman los AINEs, recomendados cuando los efectos de paracetamol no son satisfactorios y hay presencia de derrame sinovial [23,24,27,28]. En el presente modelo, el AINEs de elección es naproxeno 250 mg/día, medicamento disponible en el cuadro básico del IMSS. De otra parte, actualmente el uso de COX-2, está indicado en el tratamiento de OA, pero su manejo ha sido altamente controversial. Las guías clínicas indican que existe evidencia de que el uso a largo plazo de los inhibidores de la COX-2 se asocian a infarto agudo al miocardio e infarto cerebral, por ende se recomiendan como tratamiento de segunda línea, por periodos cortos y con vigilancia periódica. Adicionalmente estos medicamentos están contraindicados en pacientes que utilizan de forma concomitante anticoagulantes orales [28]. Por lo anteriormente descrito, este último grupo de medicamentos no se utilizó como comparador en el presente estudio. Finalmente, el uso de corticoesteroides orales para el tratamiento de OA, está descartado en la Guía de práctica clínica del IMSS [28] y la comparación con corticoides intraarticulares no fue considerada, porque de acuerdo a la Colaboración Cochrane, su efecto en OA no ha sido comprobado en periodos mayores a 4 semanas [29]. Estructura del Modelo El modelo de Markov, representa los estados de transición en los cuales puede incurrir la población estudiada. Los pacientes con diagnóstico de OAR ingresan al modelo de Markov en el estado denominado pacientes con diagnóstico de OA 4 de 50 años. En cada ciclo de Markov (1 año de duración) los pacientes pueden desarrollar síntomas y ser asignados a una de las dos opciones terapéuticas planteadas por el modelo (colágena-PVP o AINEs). En el curso natural de la enfermedad, la sintomatología puede incrementar hasta requerir tratamiento quirúrgico con artroscopia y/o reemplazo articular. Los pacientes recibiendo tratamiento médico pueden desarrollar síntomas concomitantes, relacionados con úlcera gastrointestinal, sangrado, anemia secundaria y muerte. Así mismo, aquellos sujetos sometidos a procedimientos quirúrgicos pueden presentar remisión o persistencia de los síntomas (falla de cirugía) que puede resultar en incapacidad permanente. La Figura 1 describe el diagrama de transiciones entre los estados descritos. Los estados absorbentes, en los cuales los pacientes no experimentan ninguna transición a otro estado, están reflejados en la misma gráfica. La transición a un estado previo también es posible, por ejemplo en los pacientes que evolucionan satisfactoriamente con remisión de su sintomatología. Como fuente de datos para la comparación con AINEs, se tomaron estudios que evaluaron la eficacia y seguridad de colágena polivinilpirrolidona o AINEs contra placebo [18]. En estos estudios se ha reportado que colágena polivinilpirrolidona aplicado intra-articularmente evidenció una disminución del dolor en un 55%. Por el contrario estudios con AINEs revelan que 12% de los pacientes presenta mejoría clínica, por la alta prevalencia de efectos secundarios y consecuente baja adherencia al medicamento [31]. La probabilidad de permanecer sintomático en el grupo de colágena-PVP de acuerdo al estudio de Furuzawa y cols es de 11% [16]. En cuanto a efectos secundarios es ampliamente estudiado el efecto en la mucosa gastrointestinal de los AINEs con complicaciones severas tales como sangrado y perforación de úlceras 42 VALUE IN HEALTH REGIONAL ISSUES 5C (2014) 40–47 Pacientes con OAR > 50 años Asintomáticos Sintomáticos Úlcera Péptica no complicada Artroscopia Hemorragia GI Falla cirugía Reemplazo Articular Incapacidad Permanente Muerte Fig. 1 – Modelo de Markov – Estados de Transición. OAR, osteo artrosis de rodilla. gastroduodenales. Estudios poblacionales han mostrado incidencia de eventos secundarios gastrointestinales hasta del 39% [20]. El estudio de colágena-PVP de Furuzawa y col. en México demostró una disminución significativa en el consumo de AINEs en los pacientes que recibieron tratamiento con colágena-PVP. Sin embargo, 17% de los pacientes en colágena-PVP continuaron el uso de AINEs [16]. De importancia para el modelo económico, es la presencia de complicaciones debidas a sangrado gastrointestinal descritas en diversos metanálisis y revisiones de la literatura en un rango de 10 a 30% de prevalencia [21]. El modelo calcula con base en el riesgo relativo una probabilidad de 20% de síntomas gastrointestinales en pacientes en colágena-PVP y de 39% en los pacientes con AINEs. La presencia de sangrado gastrointestinal es de 15% con AINEs y asume la misma probabilidad para los pacientes en colágena-PVP [21]. El grado de incapacidad fue calculado con base en los datos de estudio de Montoya y colaboradores para colágena –PVP y Pacheco, 2002 para AINEs [7,18]. Los datos de probabilidades dependientes de los medicamentos evaluados utilizados en el modelo se encuentran en la Tabla 1. Otros valores que fueron considerados en el modelo, se obtuvieron de estudios clínicos multicéntricos, que refieren el curso natural de la enfermedad o información relacionada con los tratamientos quirúrgicos posteriores a tratamiento clínico o reemplazo articular en OA de rodilla (Tabla 2). Medidas de Efectividad Múltiples estudios a nivel mundial han evaluado la calidad de vida, utilidades y QALYs en OA, generando datos válidos y replicables. Los valores de utilidades incorporados en el modelo para el cálculo de QALYs y sus fuentes de origen se encuentran en la Tabla 3. Costos y utilización de servicios Con el fin de emplear datos nacionales, la fuente de costos y utilización de servicios fue el listado de grupos relacionados de diagnóstico (GRD) publicados por la Dirección de Prestaciones Médicas, División de Economía de la Salud del IMSS [31–35], el Diario Oficial 2012 [36] y la información publicada a través del portal de internet del IMSS [37]. El modelo de Markov, asocia el costo de cada estado de transición a la probabilidad de presentación del evento. El costo para el estado “incapacidad permanente” se obtuvo de la publicación de Brosseau y cols [38], en la que se recomienda que el tratamiento por incapacidad causada por OA debe consistir en 24 sesiones de terapia en 10 semanas. Igualmente, se asoció el uso de paracetamol por ser el medicamento de menor costo así como el protector gástrico (omeprazol) de forma permanente, en dosis de 20 mg/día. Los GRDs y costos utilizados en el modelo se resumen en la Tabla 4. El costo final por grupo es calculado en la simulación, y se adiciona el costo acumulado por cada ciclo para obtener el costo total del grupo analizado. En la Tabla 5 se describe el costo de tratamiento con CPVP durante tres meses. Horizonte temporal Considerando el efecto a largo plazo de la OA, la progresión de la enfermedad y la incapacidad que se genera con el transcurso del tiempo, el modelo evalúa el comportamiento de la población de estudio durante el tiempo de sobrevida de los 2000 pacientes incorporados en la simulación. El horizonte temporal del modelo es de 34 años, con ciclos de un año de duración, que es el tiempo de sobrevida de los pacientes que ingresaron en la evaluación. Las tasas de mortalidad por grupo de edad y género 43 VALUE IN HEALTH REGIONAL ISSUES 5C (2014) 40–47 Tabla 1 – Probabilidades dependientes del agente terapéutico. CPVP Evento Disminución Dolor Síntomas Gastrointestinales Perforación y sangrado gastrointestinal Cirugía Artroscópica Sintomáticos con Discapacidad Permanente Discapacidad Permanente por tiempo de sobrevida Probabilidad 0.55 0.20 0.15 AINEs Probabilidad 0.12 0.39 0.15 Referencia [30] [20] [21] 0.08 0.11 Referencia 18 Cálculo del modelo Asume igual probabilidad de AINEs 19 18 0.14 0.01 0.33 18 0.76 [39] No descrito en la literatura. Supuesto del modelo [7] AINEs, antinflamatorios no esteroideos.; CPVP, colágena polivinilpirrolidona. son tomadas en cuenta cuando se cuantifica la mortalidad por ciclo. Tasa de Descuento La tasa de descuento utilizada en el presente análisis fue de 5% tanto para los costos como para los valores de utilidad. El costo de medicamentos fue ajustado por inflación basado en los Índices de Precio del sector salud del año 2012. Análisis de Sensibilidad El modelo probabilístico toma en cuenta los efectos de la incertidumbre en cada iteración de la simulación. El efecto de la incertidumbre incluido en el modelo está en el rango de más ó menos 1 Desviación Standard (DS) por encima de los valores obtenidos de la literatura. El efecto de la aleatorización se aplicó a los valores de utilidades (QALYs) dando así la oportunidad de calcular el costo utilidad con diferentes rangos de estos valores y evaluar la confiabilidad del resultado del análisis. En el modelo de Markov, se condujo el análisis de sensibilidad probabilístico usando simulación de Montecarlo, en el cual los parámetros variaron de acuerdo a la distribución de los QALYs. El proceso incluyó la repetición de 1000 ensayos. La distribución escogida para el análisis fue una distribución Beta por la característica de los parámetros incluidos en el modelo (escala de 0 a 1). Adicionalmente se realizaron análisis de sensibilidad con variaciones en los parámetros de género, edad, y precio de colágena-PVP, Figura 2. Con el fin de incluir en el análisis, el efecto de la incertidumbre generada por la variación de los valores de la utilidades (QALYs), el modelo se desarrolló como un modelo probabilístico. Bajo esta técnica el resultado final (RCUI) se calcula cada vez con base en valores aleatorios de utilidades, con el fin de corroborar la robustez del modelo, y poder concluir que a pesar de las variaciones en estos valores el resultado conserva el efecto de dominancia, independientemente de los valores de utilidad. Resultados Con base en los estimados de prevalencia de OAR en México, se simuló una cohorte hipotética de 2000 pacientes. La edad promedio de la cohorte fue de 67 años y la distribución por género fue similar. Costos En el grupo en tratamiento médico con colágena-PVP, el costo por el tiempo de sobrevida de la corte por el tratamiento de OA de rodilla fue MEX $ 866,913.03 cuando se aplicó la tasa de descuento del 5%. En el grupo de AINEs, el costo del tratamiento alcanzó una cifra de MEX $ 911, 257.67 con una diferencia de MEX $ 44,344.64 a favor de colágena-PVP. Costo-Utilidad La medición de QALYs ganados en cada grupo fue analizada por el tiempo simulado en el estudio. Los pacientes recibiendo tratamiento con colágena-PVP presentaron una ganancia de 6.62 QALYS comparados con 5.36 en los pacientes con AINEs. La relación costo utilidad (RCU) fue mayor en el grupo de AINEs (MEX $170,010.76/QALY) cuando se compara con el RCU de colágena – PVP (MEX $130,953.63). La razón costo utilidad incremental (RCUI) representa la cantidad adicional o los ahorros que se generan con una u otra alternativa. Esta razón sugiere que con el tratamiento con AINEs, el costo para obtener un QALY adicional es de MEX $ 35,194.16 (Tabla 6 ). Tabla 2 – Probabilidades de otros eventos no relacionados con el medicamento de escogencia. Evento Muerte por Sangrado Gastrointestinal Falla Cirugía - Discapacidad Permanente Falla Cirugía - Persistencia Síntomas Reemplazo articular Éxito Artroscopia Reemplazo Total de Rodilla. RTR Probabilidad Referencia 0.10 [22] 0.44 [29] 0.14 [40] 0.01 0.09 0.09 [41] [19] [42] Tabla 3 – Utilidades (QALYs). Estadío Sintomáticos Asintomáticos Úlcera Péptica no complicada Hemorragia Gastrointestinal (GI) Artroscopia Falla Cirugía Reemplazo articular Incapacidad Permanente Muerte Utilidad Fuente 0.51 0.86 0.38 0.31 0.72 0.75 0.75 0.56 0.00 [43,44] [43,44] [21] [21] [45] [45] [46] [47] [44] 44 VALUE IN HEALTH REGIONAL ISSUES 5C (2014) 40–47 Tabla 4 – Costos de acuerdo a los GRD. GRD* Código Úlcera péptica no complicada debida a Esofagitis, gastroenteritis y alteraciones misceláneas digestivas sin CCM. Artroscopia Reemplazo articular (GRD: Revisión de cadera o reemplazo de rodilla sin CC/CCM) Hemorragia GI con CCM Muerte debida a Úlcera Péptica complicada Incapacidad Permanente Número Costo (2012) M392 [48] 1 $17,779.00 Q489 [49] Q466 [50] M377 [51] M380 [52] 1 1 1 1 $46,695.00 $271,645.00 $44,051.00 $89,265.00 $102,011.60 Clave IMSS Paracetamol 01000001040201 Naproxeno Omeprazol Ranitidina 01000034070102 01000051820001 01000012330001 MEDICAMENTOS Presentación Tableta 500 mg. 250gr 40 mg 150 mg Costo unitario 2012 $9.00 $38.70 $116.00 $2.44 GRD, grupos relacionados de diagnóstico. * Fuente: Dirección de Prestaciones Médicas. División de Economía de la Salud Análisis de sensibilidad sobre parámetros con incertidumbre El modelo por sus características probabilísticas, utiliza rangos en los valores de utilidades que varían aleatoriamente en cada iteración. El valor de RCUI no se ve alterado con los cambios en los valores de las utilidades (1 DS) lo que demuestra la consistencia de los datos. Otros parámetros que se incluyeron en los análisis de sensibilidad, fueron: edad, género y precio de colágena-PVP. La Figura 2 evidencia los cambios en las variables que representan incertidumbre, con un impacto reflejado en incremento en los ahorros generado por colágena-PVP. Adicionalmente se condujo un análisis de sensibilidad probabilístico (ASP) con 1000 iteraciones, para estimar el impacto simultáneo de todos los parámetros del modelo. La Figura 3, representa los resultados del ASP, en donde el costo incremental se relaciona con las utilidades (QALYs) incrementales. De acuerdo a esta figura, una vez se incluyen todos los parámetros de incertidumbre en el modelo, la respuesta de costo-utilidad de colágena-PVP en comparación con AINEs, persiste como costo-efectiva (predominancia de los valores en el cuadrante Noreste). Resultados de análisis vs. El PIB mexicano Los resultados finales del modelo se expresan como índice de costo-utilidad: costo incremental por QALYs (años de vida ganados ajustados por calidad). El umbral de referencia establecido por el Consejo General de Salubridad (CSG) para México define que el costo-efectividad incremental para un insumo debe estar por debajo de 1 PIB per cápita/QALY para permitir que el mismo sea aceptado para su cobertura en las instituciones de salud. La Organización Mundial de la Salud (2011) establece que para la región a la que pertenece México (Américas B) el PIB per cápita es de $9,489 dólares americanos (USD), por lo que, una estrategia costo-efectiva es aquélla cuyo valor se encuentra por debajo de MEX $ 130,485.67 (1 PIB ). El referente de 1 PIB/cápita es utilizado como guía para determinar el costo-efectividad cuando medidas de disponibilidad a pagar no se encuentra disponibles para una determinada patología en un sistema de salud, como sucede en México. Según el análisis presentado en este informe, colágenaPVP es una alternativa costo ahorradora en el tratamiento de OA de rodilla en México (MEX $ 35,194.16/QALY). Discusión Este es el primer estudio económico diseñado para evaluar una nueva alternativa terapéutica desarrollada en México para el tratamiento de OA. La evaluación económica sugiere que colágena – PVP (225 mg IA) es un tratamiento de OA clínica y económicamente superior a los AINEs (naproxeno 250 mg) y paracetamol medidos por costos y QALYs ganados. Este medicamento Tabla 5 – Costo tratamiento durante tres meses. [16,37,53]. CPVP Costo Medicamentos CPVP, ampolla de 1.5 mm Ultrasonido*,† Terapia Física‡ Número Costo Tratamiento 3 meses 6 6 12‡ $635.00 $266.00 $789.00 $3,810.00 $1,596.00 $9,468.00 $14,874.00 CPVP, colágena polivinilpirrolidona. * La aplicación interarticular de CPVP debe guiarse por ultrasonido. † Diario oficial 2012 [36] ‡ Guia Clínica del IMSS, 2009 [53,54] 45 VALUE IN HEALTH REGIONAL ISSUES 5C (2014) 40–47 Análisis de Sensibilidad Variable $(46,547.87) Disminución Precio CPVP hasta 50% $(44,516.81) Incremento Precio CPVP hasta 50% $(47,682.41) Género Masculino > 55 años $(49,825.44) Género Femenino > 55 años $(66,186.03) Género Masculino 50 años $(49,271.82) Género Femenino 50 años $(132,456.00) Población General 75 - 85 años $(98,765.00) Población General 65 - 74 años $(79,186.00) Población General 55 - 64 años $(52,170.00) Población General 50 - 54 años $(35,194.16) Caso Base RCUI Fig. 2 – Análisis de Sensibilidad – Diagrama de tornado para la RCUI. CPVP, colágena polivinilpirrolidona; RCUI, costo utilidad incremental. demostró su dominancia en el análisis de costo-utilidad, por periodos de un año y por el tiempo de sobrevida de la cohorte hipotética de 2000 pacientes (34 años). El costo del tratamiento con colágena – PVP, podría percibirse como mayor, dado el precio del medicamento, así como el procedimiento requerido para su aplicación. Sin embargo, en adición a los costos de los medicamentos, el grupo de AINEs presenta un costo adicional por el tratamiento y manejo de síntomas gastrointestinales, así como por progresión de la enfermedad que puede requerir reemplazo articular en hasta 9% de la población afectada [42]. Estas diferencias explican los ahorros generados por colágena-PVP, al evitar las complicaciones GI gracias a la disminución del consumo de AINEs y baja progresión de la enfermedad debida a la potencial regeneración del cartílago articular entre otras [9,16,18]. Para la presente evaluación económica se diseñó un modelo de Markov probabilístico. Con esta aproximación la comparación de la eficacia y seguridad de los medicamentos fue integrada con el análisis de costo-utilidad en un solo contexto. La ventaja de este enfoque es que permite incluir las probabilidades clínicas y sus desenlaces en términos monetarios y de calidad de vida. El modelo probabilístico de otra parte, permite tener en cuenta el efecto de la incertidumbre. Los datos de seguridad de AINEs y colágena – PVP fueron obtenidos de estudios clínicos controlados en OA y los riesgos de complicaciones gastrointestinales de AINEs fueron exhaustivamente extraídos de estudios clínicos multicéntricos controlados [20–22]. La información de QALYs, de otra parte, es difícil de transferir de un país a otro. Sin embargo ante la ausencia de datos locales, se ha aceptado el uso de cifras internacionales [55]. De manera ventajosa para la presente evaluación, múltiples estudios alrededor del mundo han considerado la efectividad y los QALYs en OA, debido tal vez a los altos índices de morbilidad y el impacto de esta enfermedad en la calidad de vida. Los datos obtenidos para el modelo se incorporaron en el mismo, con base en su consistencia a lo largo de múltiples publicaciones. Nuevamente el carácter probabilístico del modelo permite obtener resultados con variaciones en los rangos de los valores de utilidades, que le dan mayor fortaleza al resultado final de la evaluación, al no limitarse al análisis de datos estáticos. En relación con los costos, la obtención de información de las publicaciones oficiales del IMSS permite tener un resultado basado en datos locales, reales y específicos. Los costos de las complicaciones GI asociadas al uso de AINEs, se obtuvieron de los GRD del IMSS, CDM : 06 - Enfermedades y trastornos del aparato digestivo, Tipo M: 377 [34] y 380 [31]. Los tratamientos artroscópicos para el manejo de OA son múltiples y pueden incluir lavado de la articulación, desbridamiento del menisco, extracción de líquido sinovial, extracción de fragmentos de cartílago articular. El modelo agrupa estos procedimientos en el GRD del IMSS, CDM: 08 - Enfermedades y trastornos del sistema musculoesquelético y tejido conectivo, Tipo Q: 489 [33] y 466 [32], para cálculos de los costos. Las limitaciones del estudio han sido claramente detectadas. Al igual que en la mayoría de los estudios económicos, los datos no pueden ser generalizados a otras jurisdicciones por las variaciones en los costos tanto de los tratamientos agudos como de los costos de las consecuencias a largo plazo. Adicionalmente, la ausencia de datos específicos en probabilidades de discapacidad permanente causada por OA para la población mexicana puede considerarse una limitación importante. Con el fin de evitar sesgos causados por los datos extraídos de la literatura, se han utilizado en el modelo de colágena-PVP los valores altos de los rangos reportados en la literatura. En el modelo de AINEs se utilizaron los valores centrales de discapacidad permanente reportados. La metodología y la búsqueda exhaustiva de evidencia para el análisis económico, permiten afirmar que el presente estudio Tabla 6 – Resultados análisis costo-utilidad. Alternativa Costo Esperado Costo Incremental Efectividad (QALYs) Utilidad/ Efectividad Incr. RCU RCUI CPVP AINEs $866,913.03 $911,257.67 6.62 5.36 1.26 $130,953.63 $170,010.76 $(35,194.16) $44,344 AINEs, antinflamatorios no esteroideos; RCUI, costo utilidad incremental. 46 VALUE IN HEALTH REGIONAL ISSUES 5C (2014) 40–47 $500,000.00 1 PIB = $ 130,485 $400,000.00 $300,000.00 $200,000.00 $100,000.00 $$(100,000.00) $(200,000.00) $(300,000.00) $(400,000.00) $(500,000.00) -2.00 -1.50 -1.00 -0.50 0.00 0.50 1.00 1.50 2.00 Fig. 3 – RCUI: Análisis de Sensibilidad Probabilístico. RCUI, costo utilidad incremental. arroja suficiente información válida y reproducible que hace posible la generalización de los resultados en el ámbito del Instituto Mexicano de Seguridad Social y del Sistema de Salud de México en general. Conclusiones En el contexto Mexicano, colágena-PVP evidencia un RCUI claramente dominante. En los países en desarrollo el valor equivalente a 1 PIB se considera el límite para determinar el costo efectividad de las tecnologías sanitarias. Colágena-PVP, se presenta como un alternativa costo efectiva en el tratamiento de OA comparada con el tratamiento estándar recomendado en las guías clínicas nacionales. Se requieren futuros estudios para realizar la comparación con otras opciones terapéuticas tales como ácido hialurónico y otros agentes utilizados para viscosuplementación. Fuentes de Financiamiento: Trabajo financiado por Aspid, S.A. de C.V, México. Los conceptos expresados en este articulo son propios de los autores sin compromiso legal con Aspid. R EF E R EN CI AS [1] Brooks P. Inflammation as an important feature of osteoarthritis. Bulletin of the World Health Organization. [Online] 2003. Available from: http://www.who.int/bulletin/volumes/81/9/PHC.pdf. [Accessed November 6, 2013]. [2] Woolf A, Pfleger B. Burden of major musculoskeletal conditions. Bulletin of the World Health Organization 2003;81:646–56. [3] Álvarez J, Nuño B, Alcocer J. Enfermedades reumáticas y discapacidad laboral en población adulta rural. Rev Med IMSS 2005;43:287–92. [4] Bori G. Principales enfermedades reumáticas en México: aspectos epidemiológicos e impacto socioeconómico. Rev Mex Reumatol 2000;15:183–201. [5] Wolfe F, Zhao S, Lane N. Preference for non-steroidal antiinflammatory drugs over acetaminophen by rheumatic disease patients: a survey of 1799 patients with osteoarthritis, rheumatoid arthritis and fibromyalgia. Arthritis Rheum 2000;43:378–85. [6] Hannan MT, Felson DT, Pincus T. Analysis of the discordance between radiographic changes and knee pain in osteoarthritis of the knee. J Rheumatol 2000;27:1513–7. [7] Pacheco D, Radrigán F. "REUMA" Reumatología para médicos de Atención Primaria, Soc. Chil. de Reumatología y Liga Panamericana de Asociaciones de Reumatología, Noviembre 2002. [8] Lozano-Ascencio R, Frenk-Mora J, González-Block MA. El peso de la enfermedad en adultos mayores, México 1994. Salud Pública Mex 1996;38:419–29. [9] Furuzawa J, Rojas E, Valverde M, et al. Cellullar and humoral responses to collagenpolyvinylpyrrolidone administrated during short and long periods in humans. Can J Physiol & Pharmacol 2003;81:1029–35. [10] Furuzawa-Carballeda J, Alcocer-Varela J, Diaz de León L. Collagen-_ PVP decreases collagen turnover in synovial tissue cultures from rheumatoid arthritis patients. Ann N Y Acad Sci 1999;878:598–602. [11] Krötzsch-Gómez FE, Furuzawa-Carballeda J, Reyes-Márquez R, et al. Cytokine expression is downregulated by collagenpolyvinylpyrrolidone in hypertrophic scars. J Invest Dermatol 1998;111:828–34. [12] Colín A, Salgado R, Zamira A, et al. Inducción del tejido de granulación por pasta de Lassar vs colágena-polivinilpirrolidona en úlceras por insuficiencia venosa. Cir Plast 2004;14:5–13. [13] Furuzawa CCZ, García ARL, Furuzawa CJ. Estudio piloto sobre la eficacia del uso de la colágena-polivinilpirrolidona (clg-PVP) en la regeneración ósea en cirugía endodóncica. Rev Odont Mex 2005;9:191–6. [14] Ruiz- Eng R, Montiel A, De la Rosa-Pérez R, et al. Colágenopolivinilpirrolidona, nueva opción para tratar secuelas de la mastectomía radical en mujeres con cáncer mamario. Informe preliminar. Cir Cir 2010;78:310–4. [15] Mendieta H. Reducción de costos de atención médica con el empleo de colágena-polivinilpirrolidona. Rev Espec Medico-Quirúrgicas 2002;7:32–4. [16] Furuzawa J, Lima G, Llorente L, et al. Polymerized-type I collagens downregulates inflammation and improves clinical outcomes in patients with symptomatic knee osteoarthritis following arthroscopic lavage: a randomized, double-blind, and placebo-controlled clinical trial. ScientificWorldJournal 2012;2012:342854. [17] Soto H, Cordoba R, Pizarro M, et al. Cost-utility analysis of condroitin sulphate in the mexican treatment of osteoarthritis (OA) of knee in mexican patients. Value Health 2011;14:A129. [18] Montoya CA. Evolución clínica de los pacientes con gonartrosis tratados mediante la aplicación de colágeno polivinilpirrolidona intraarticular. Orthotips 2012;8:96–102. [19] Magaña MC, Ontiveros E, Ramírez A. Uso del colágeno polimerizado tipo I en el tratamiento no quirúrgico de la gonartrosis, experiencia en el IMSS. Orthotips 2012;8:103–10. [20] Rubio-Terrés C. Evaluación económica del uso de condroitín sulfato y antiinflamatorios no esteroideos en el tratamiento de la artrosis. Datos del estudio VECTRA. Rev Reumatol Clin 2010;6:187–95. [21] Maetzel A, Krahn M, Naglie G. The cost-effectiveness of celecoxib and rofecoxib in patients with osteoarthritis or rheumatoid arthritis. Ottawa: Canadian Coordinating Office for Health Technology Assessment; 2002. Technology report no. 23. Available from: http:// cadth.ca/media/pdf/109_cox2_tr_e.pdf. [Accessed January 20, 2013]. [22] Spiegel BM, Targownik L, Dulai GS, Gralnek IM. The cost-effectiveness of cyclooxygenase-2 selective inhibitors in the management of chronic arthritis. Ann Intern Med 2003;138:795–806. [23] Secretaria de Salud. México. Guía de práctica clínica para el diagnóstico y tratamiento de la osteoartrosis de rodilla. Evidencias y Recomendaciones. IMSS-079-08. 2008. Available from: http://www.cenetec.salud.gob.mx/ VALUE IN HEALTH REGIONAL ISSUES 5C (2014) 40–47 [24] [25] [26] [27] [28] [29] [30] [31] [32] [33] [34] [35] [36] [37] [38] descargas/gpc/CatalogoMaestro/079_GPC_Osteoartrosisrodilla/ OSTEORTROSIS_ER_CENETEC.pdf. [Accessed August 12, 2012]. Secretaria de Salud. México. Guía de práctica clínica para el diagnóstico y tratamiento de la osteoartrosis de rodilla. Guia de Referencia Rapida. IMSS-079-08. 2008. Available from: http://www. cenetec.salud.gob.mx/descargas/gpc/CatalogoMaestro/ 079_GPC_Osteoartrosisrodilla/OSTEORTROSIS_R_CENETEC.pdf. [Accessed August 12, 2012]. Secretaria de Salud.Guía de práctica clínica. Tratamiento alternativo en pacientes con osteoartrosis de rodilla grado I-II. Guia de Referencia Rapida /GRR_IMSS_329_10. [Online] 2009. [Cited: Ago 2, 2012.]. Available from: http://www.cenetec.salud.gob.mx/descargas/gpc/CatalogoMaestro/ 329_IMSS_10_Osteoartrosis_Rodilla_Grado_I_II/GRR_IMSS_329_10.pdf. [Accessed August 13, 2012]. Secretaria de Salud. Guía de práctica clínica. Tratamiento alternativo en pacientes con osteoartrosis de rodilla grado I-II. Evidencias y Recomendaciones /EyR_IMSS_329_10. [Online] 2009. [Cited: Ago 2, 2012.]. Available from: http://www.cenetec.salud.gob.mx/descargas/ gpc/CatalogoMaestro/329_IMSS_10_Osteoartrosis_Rodilla_Grado_I_II/ EyR_IMSS_329_10.pdf. [Accessed July 2, 2014]. IMSS. Guía de Practica Clínica Diagnóstico y Tratamiento osteoartrosis de rodilla. GER.IMSS-079-08. 2008. Available from: http://www.imss.gob.mx/ profesionales/guiasclinicas/Documents/GER_OsteoartrosisdeRodilla.pdf. [Accessed August 12, 2013]. IMSS. Guía de Practica Clínica Diagnóstico y Tratamiento osteoartrosis de rodilla. GRR.IMSS-079-08. 2008. IMSS. Guía de Practica Clínica Diagnóstico y Tratamiento osteoartrosis de rodilla. GER.IMSS-079-08. Bellamy N, Campbell J, Welch V, et al. Intraarticular corticosteroid for treatment of osteoarthritis of the knee. Cochrane Database of Systematic Reviews. In: The Cochrane Library, Issue 11, Art. No. CD005328. DOI: 10.1002/14651858.CD005328.pub2. Stam W, Jansen J, Taylor S. Efficacy of etoricoxib, celecoxib, lumiracoxib, non-selective NSAIDs, and acetaminophen in osteoarthritis: a mixed treatment comparison. Open Rheumatol J 2012;6:6–20. IMSS. CDM : 06 - Enfermedades y trastornos del aparato digestivo, Tipo M: 380. Úlcera péptica complicada con CCM. [Online] [Cited: Ago 2, 2012.]. Available from: http://www.imss.gob.mx/profesionales/grd/ GRD_380.pdf. [Accessed August 12, 2013]. IMSS, CDM : 08 - Enfermedades y trastornos del sistema musculoesquelético y tejido conectivo, Tipo Q: 466. Revisión de cadera o reemplazo de rodilla sin CC/CCM. [Online] [Cited: Ago 27, 2012.]. IMSS, CDM : 08 - Enfermedades y trastornos del sistema musculoesquelético y tejido conectivo, Tipo Q: 489. Procedimientos de rodilla sin diagnóstico principal de infecciones sin CC/CCM. [Online] [Cited: Ago 27, 2012.] Available from: http://www.imss.gob.mx/ profesionales/grd/GRD_489.pdf. [Accessed November 15, 2012]. IMSS. CDM : 06 - Enfermedades y trastornos del aparato digestivo, Tipo M: 377. Hemorragia gastrointestinal con CCM. [Online] http://www. imss.gob.mx/profesionales/grd/GRD_377.pdf. IMSS. CDM : 06 - Enfermedades y trastornos del aparato digestivo, Tipo M: 392. Esofagitis, gastroenteritis y alteraciones misceláneas digestivas sin CCM. [Online] [Cited: Ago 27, 2012.]. Available from: http://www. imss.gob.mx/profesionales/grd/GRD_392.pdf. [Accessed August 23, 2012]. IMSS. Diario Oficial. ACUERDO ACDO.SA3.HCT.280312/78.P.DF, dictado por el H. Consejo Técnico. 9 de mayo de 2012. [Online] 2012. [Cited: Dic 1, 2012.]. IMSS. IMSS compro. [Online] 2012. [Cited: Jul 1, 2012.] Available from: http://compras.imss.gob.mx/. [Accessed September 2, 2013]. Brosseau, L, et al. "Intensity of Exercise for the Treatment of Osteoarthritis." Cochrane Database System Review 2 (2003): CD004259. National Center for Biotechnology Information. National Library of Medicine. 21 Nov. 2008 oPMID: 128045104. [Online]. 47 [39] Quintana JM, Arostegui I, Escobar A, et al. Prevalence of knee and hip osteoarthritis and the appropriateness of joint replacement in an older population. Arch Intern Med 2008 28;168:1576–84. [40] Beswick AD, Wylde V, Gooberman-Hill R, et al. What proportion of patients report long-term pain after total hip or knee replacement for osteoarthritis? A systematic review of prospective studies in unselected patients. BMJ Open 2012 Feb 22;2:e000435. [41] Leskinen J, Eskelinen A, Huhtala H, et al. The incidence of knee arthroplasty for primary osteoarthritis grows rapidly among baby boomers: a population-based study in Finland. Arthritis Rheum 2012;64:423–8. [42] Wai EK, Kreder HJ, Williams JI. Arthroscopic débridement of the knee for osteoarthritis in patients fifty years of age or older: utilization and outcomes in the Province of Ontario. J Bone Joint Surg Am 2002; 84-A:17–22. [43] Zhang W, Moskowitz RW, Nuki G, et al. OARSI recommendations for the management of hip and knee osteoarthritis, Part I: Critical appraisal of existing treatment guidelines and systematic review of current research evidence. Osteoarthritis Cartilage 2007;15:981e1000. [44] Tso P, Walker K, Mahomed N, et al. Comparison of lifetime incremental cost: utility ratios of surgery relative to failed medical management for the treatment of hip, knee and spine osteoarthritis modelled using 2year postsurgical values. Can J Surg 2012;55:181–90. [45] Lubowitz JH, Appleby D. Cost-effectiveness analysis of the most common orthopaedic surgery procedures: knee arthroscopy and knee anterior cruciate. Arthroscopy 2011;27:1317–22. [46] Dakin H, Gray A, Fitzpatrick R, et al. Rationing of total knee replacement: a cost-effectiveness analysis on a large trial data set. BMJ Open 2012;2:e000332. [47] Norman-Taylor FH, Palmer CR, Villar RN. Quality-of-life improvement compared after hip and knee replacement. J Bone Joint Surg Br 1996;78:74–7. [48] IMSS. CDM : 06 - Enfermedades y trastornos del aparato digestivo, Tipo M: 392. Esofagitis, gastroenteritis y alteraciones misceláneas digestivas sin CCM. [Online] 2008. [Cited: Ago 27, 2012.]. Available from: http:// www.imss.gob.mx/profesionales/grd/GRD_392.pdf. September 2, 2013. [49] IMSS, CDM : 08 - Enfermedades y trastornos del sistema musculoesquelético y tejido conectivo, Tipo Q: 489. Procedimientos de rodilla sin diagnóstico principal de infecciones sin CC/CCM. [Online] 2008. [Cited: Ago 27, 2012.] Available from: http://www.imss.gob.mx/ profesionales/grd/GRD_489.pdf. [Accessed September 2, 2013]. [50] IMSS, CDM : 08 - Enfermedades y trastornos del sistema musculoesquelético y tejido conectivo, Tipo Q: 466. Revisión de cadera o reemplazo de rodilla sin CC/CCM. [Online] 2008. [Cited: Ago 27, 2012]. [51] IMSS. CDM : 06 - Enfermedades y trastornos del aparato digestivo, Tipo M: 377. Hemorragia gastrointestinal con CCM. [Online] 2008. [Cited: 1 2012, Dic.]. Available from: http://www.imss.gob.mx/profesionales/grd/ GRD_377.pdf. [Accessed August 12, 2013]. [52] IMSS. CDM : 06 - Enfermedades y trastornos del aparato digestivo, Tipo M: 380. Úlcera péptica complicada con CCM. [Online] 2008. [Cited: Ago 2, 2012.]. Available from: http://www.imss.gob.mx/profesionales/grd/ GRD_380.pdf. [Accessed September 2, 2013]. [53] IMSS. Guía de práctica clínica. Tratamiento alternativo en pacientes con osteoartrosis de rodilla grado I-II. Referencia Rapida /GRR_IMSS_329_10. [Online] Oct 2009. [Cited: Jul 30, 2012.] Available from: http://www.imss.gob.mx/profesionales/guiasclinicas/ Documents/329GRR.pdf. [Accessed September 2, 2013]. [54] IMSS. Guía de práctica clínica. Tratamiento alternativo en pacientes con osteoartrosis de rodilla grado I-II. Evidencias y Recomendaciones /EyR_IMSS_329_10. Oct 2009. Available from: http://www.imss.gob.mx/ profesionales/guiasclinicas/Documents/329GER.pdf. [Accessed September 2, 2013]. [55] National Institute for Health and Care Excellence. Incorporating health economics in guidelines and assessing resource impact. Mar 2005. Available from: http://www.nice.org.uk/niceMedia/pdf/ GDM_Chapter8_0305.pdf. [Accessed September 2, 2013]. VALUE IN HEALTH REGIONAL ISSUES 5C (2014) 48–57 Available online at www.sciencedirect.com journal homepage: www.elsevier.com/locate/vhri Cost-Effectiveness Analysis of Atorvastatin versus Rosuvastatin in Primary and Secondary Cardiovascular Prevention Populations in Brazil and Columbia Joaquín F. Mould-Quevedo, PhD1,*, Magda Vianey Gutiérrez-Ardila, DMD, MSc(c)2, Jaime Eduardo Ordóñez Molina, MD, PhD3, Brett Pinsky, MPH4, Nicolás Vargas Zea, MBA, MSc2 1 Pfizer, New York, NY, USA; 2Pfizer S.A.S., Bogota, Colombia; 3Center for the Evaluation of Health Technologies, CES University, Medellin, Colombia; 4OptumInsight, Eden Praire, MN, USA (Brett Pinsky is a former employee of OptumInsight) AB STR A CT Background: Latin America has witnessed a marked increase in cardiovascular (CV) disease, the leading cause of death in many countries. The benefits of lipid-lowering therapy to reduce CVrelated events are widely accepted. Clinical evidence suggests that rosuvastatin is associated with slightly greater reductions in lowdensity lipoprotein cholesterol levels than is atorvastatin at comparable doses. Rosuvastatin, however, is often priced at a premium. Objective: Our objective was to examine the cost-effectiveness of using atorvastatin versus rosuvastatin in reducing CV events in Brazil and Colombia using real-world prices. Methods: A global Markov cohort model of primary and secondary CV prevention was developed and adapted to Brazilian and Colombian settings. The risks and costs of major CV events and efficacy, adherence, and costs of statins were considered. Total gains in life-years, quality-adjusted life-years, major CV events avoided, and costs over the lifetime horizon were estimated. Several dose comparisons were considered. Results: In the Colombian analyses, differences in drug costs between therapies were considerable while outcomes were similar. The incremental cost per quality-adjusted life-year gained for rosuvastatin versus atorvastatin was more than $700,000 and $200,000 in primary and secondary prevention, respectively. Brazilian analyses found lower incremental cost-effectiveness ratios for rosuvastatin at some dose comparisons due to similar pricing between statins. Sensitivity analyses revealed that changes in treatment efficacy and adherence had the largest impact on results. Conclusions: In primary and secondary CV prevention, the efficacy advantage of rosuvastatin was minimal, while its acquisition cost was higher, particularly in Colombia. The incremental cost-effectiveness ratios were, therefore, generally in favor of atorvastatin being the cost-effective option. Keywords: atorvastatin, Brazil, cardiovascular, Colombia, costeffectiveness, rosuvastatin, statins. Introduction cerebrovascular disease are roughly 1.4 to 2 times those in the United States [7]. The increase in CV disease in Latin and South America is thought to be due in part to major changes associated with economic development and urbanization, the reduction in communicable disease, and the high prevalence of major risk factors such as obesity and smoking [3,8–10]. For example, a survey in Colombia found that 46% of those surveyed were overweight or obese and 24% of women and 20% of men had high low-density lipoprotein cholesterol (LDL-C) levels [11]. This higher burden of CV disease and CV-related death can have substantial economic and socioeconomic consequences, especially in countries with limited capacity to manage disease. The benefits of lipid-lowering therapy to reduce CV events and mortality are widely accepted [12,13]. Because of the Latin America has a high burden of cardiovascular (CV) disease. The percentage of deaths attributable to CV disease in Latin America and the Caribbean has been estimated at more than 30% [1,2] and in 2000 was the leading cause of death in women aged 15 to 69 years in the region and the second leading cause in men of the same age [1,2]. It is estimated that from 1990 to 2020, deaths from CV illness will increase by 145% in Latin America compared with 28% to 50% in developed countries [3]. In Brazil, CV disease is responsible for 32% of total mortality [4] and is the main cause of death in those older than 40 years [5]. There are also concerns about CV disease developing at younger ages throughout the country [6]. In Colombia, adjusted mortality rates due to ischemic heart disease and Copyright & 2014, International Society for Pharmacoeconomics and Outcomes Research (ISPOR). Published by Elsevier Inc. Conflict of interest: Joaquín F. Mould-Quevedo, Magda Vianey Gutiérrez Ardila, and Nicolás Vargas Zea are employees of Pfizer and have financial interests in Pfizer in the form of stock. Jaime Eduardo Ordóñez and Brett Pinsky have no conflicts to report. * Address correspondence to: Joaquín F. Mould-Quevedo, Pfizer, EMEP Outcomes Research and Epidemiology, 235 East 42nd Street, New York, NY 10017, USA. E-mail: joaquin.mould@pfizer.com. 2212-1099$36.00 – see front matter Copyright & 2014, International Society for Pharmacoeconomics and Outcomes Research (ISPOR). Published by Elsevier Inc. http://dx.doi.org/10.1016/j.vhri.2014.07.007 49 VALUE IN HEALTH REGIONAL ISSUES 5C (2014) 48–57 effectiveness of statins in treating elevated cholesterol levels, their use has increased over the years [14] for both primary and secondary prevention of CV events [15]. In Brazil, statins are considered a first choice for treating elevated LDL-C levels [16] and have been shown to be effective in reducing LDL-C levels immediately after acute myocardial infarction [17]. When multiple therapeutic choices are available, payers, physicians, and patients must consider not only efficacy and safety but also costs. Atorvastatin and rosuvastatin are two of the most commonly prescribed statins in much of Latin America, each with substantial clinical support. Clinical evidence tends to demonstrate that rosuvastatin is associated with a slightly greater reduction in LDL-C levels than is atorvastatin at comparable doses [18]. In most countries (including Brazil and Colombia), rosuvastatin, however, is priced at a premium. The objective of this analysis was to assess the cost-effectiveness of using atorvastatin versus rosuvastatin in reducing CV events in both primary and secondary prevention populations in Brazil and Colombia. Methods Cost-Effectiveness Model Overview A global Markov cohort model was developed in Microsoft Excel by i3 Innovus Research, Inc., to estimate the cost-effectiveness of statin therapies in populations at risk for CV disease. The global model was designed to examine the effect of the use of statins on reductions in LDL-C levels separately in primary and secondary CV prevention. For the current analysis, the model was adapted to the Brazilian and Colombian settings. Here, we provide a detailed overview of the global model structure along with a description of the adaptations made for these settings. Model Structure For the primary prevention model, the clinical starting point was the decision to treat patients with atorvastatin or rosuvastatin. Subsequent branching of the decision tree reflects the risk of a major CV event (MCVE) or death (Fig. 1A). An MCVE is a composite of myocardial infarction, stroke (ischemic or hemorrhagic), revascularization (coronary artery bypass graft or percutaneous transluminal coronary angioplasty), and CV-related death. The model has three health states: 1) no previous CV event, 2) history of MCVE, and 3) death (Fig. 1B, shown with arrows indicating possible movements between states). Once a patient experiences a nonfatal CV event in the primary prevention model, he or she moves to the secondary prevention model in which he or she is at an increased risk of death and subsequent CV events. Transitions between health states are assumed to occur in 1-year intervals, and the baseline transition probabilities (i.e., the risk of MCVE and death) are based on average starting age, mean LDL-C level, and sex of the cohort. Treatment efficacy is modeled as a No Major CV Event – see model states atorvastatin Death Major CV Event – see model states Patients initiating on a statin therapy No Major CV Event – see model states rosuvastatin Death Major CV Event – see model states No prior CV event History of CV event Dead Primary Prevenon Stable CVD One CV event Two or more CV events Dead Secondary Prevenon Fig. 1 – (A) Treatment decision tree, primary prevention, and (B) Markov health states, primary prevention and secondary prevention models. CV, cardiovascular; CVD, cardiovascular disease. 50 VALUE IN HEALTH REGIONAL ISSUES 5C (2014) 48–57 decrease in event risk associated with reductions in LDL-C levels. Event-specific disutility and costs are applied at the time of each event, carried forward for the duration of the model, and summed across all patients in each treatment arm for the calculation of cost-effectiveness ratios. Patients are followed until they die. The secondary prevention model is similar, except for the possible health states, which are 1) stable CV disease (no events have occurred during the patient’s time in the model), 2) one nonfatal CV event, 3) two or more nonfatal CV events, and 4) death (Fig. 1B). Target Populations and Treatment Comparisons The baseline risk of MCVEs and mortality for the population of interest is defined by the starting age of the cohort, the mean LDL-C level, and the proportion of males versus females. These reflect averages of the baseline characteristics from the trials in the Cholesterol Treatment Trialists' Collaboration (CTTC) metaanalysis [19], sorted according to primary (r3.5% annual risk of MCVEs in the control arm) and secondary (43.5% annual risk of MCVEs in the control arm) prevention and weighted according to trial population size. For the reference case analysis, the primary prevention population was defined as being 33% women with a mean age of 62 years and having a baseline LDL-C level of 137 mg/dL. For the secondary prevention population, these values were 25% women, 64 years, and 143 mg/dL, respectively. For the reference case analysis for primary prevention in Colombia, atorvastatin 10 mg (A10) was compared with rosuvastatin 5 mg (R5), and atorvastatin 20 mg (A20) was compared with rosuvastatin 10 mg (R10). In the Brazilian analysis, A20 was compared with both R10 and rosuvastatin 20 mg (R20). For secondary prevention, the comparisons were atorvastatin 40 mg (A40) versus R20 and atorvastatin 80 mg (A80) versus rosuvastatin 40 mg (R40) for Colombia and A40 versus R40 for Brazil. Model Parameters and Data Sources The risk of first and second MCVEs (also from the CTTC metaanalysis [19]) were calculated by using baseline trial values for untreated patients (trials with r3.5% annual baseline risk of MCVE per year for primary prevention, trials with 43.5% annual baseline risk of MCVE per year for secondary prevention) and then adjusting for the impact of each statin using the expected statin-specific reduction in LDL-C level and the hazard ratio of a 1 mmol/L drop in the LDL-C level, which has been estimated as 0.21 [19]. After year 5, patients in each treatment arm were assumed to have an equal risk of MCVE due to an absence of longer-term data on reductions in LDL-C levels. Rates after the second MCVE were extrapolated on the basis of the increase in risk between previous MCVEs (Table 1). The distribution of MCVEs was assumed to be 29% myocardial infarction, 19% stroke, 37% revascularization, and 15% death [20]. For those patients experiencing an MCVE in the model, the 1year postevent mortality probabilities for both models were estimated from the literature [21–23]. For those starting the model and for those who survive at least 1 year after an MCVE, adjusted ageand sex-specific all-cause mortality rates were based on national estimates for Brazil [24] and Colombia [25]. For the Brazilian model, the mortality inputs were calculated proportionally, considering the informed values from the original file, assuming a proportion of 50% for each gender at the original model for both genders’ proportions. Brazilian mortality data were censored at age 80 years, so mortality for those aged 80 to 100 years was calculated by growth tendency from the interval from age 70 to 79 years. Treatment efficacy represents a percent reduction in LDL-C level based on data published by Schneck et al. [18]. Specifically, the expected percent reduction for each statin dose was as follows: A10 ¼ 38%, A20 ¼ 43%, A40 ¼ 48%, A80 ¼ 54%, R5 ¼ 42%, R10 ¼ 47%, R20 ¼ 52%, and R40 ¼ 57%. Adherence was assumed to have a proportionate impact on reductions in LDL-C levels and on costs. The primary prevention model assumed that first-year adherence was 55% and in each subsequent year was 33%. For the secondary prevention model, first- and subsequent-year adherence was assumed to be 70% and 48%, respectively [26]. Information about any adherence difference for both comparators was not available. The cost of statin therapy for the Brazilian analysis was obtained from the Brazilian National Agency for Sanitary Vigilance maximum price for government sales list, including taxes to the applicable cases [27], while for the Colombian analysis, the costs represent the institutional price reported by the Sistema de Información de Precios de Medicamentos (National Information System Drug Price) [28] (Table 1). To estimate the costs for MCVEs, the Colombian analysis utilized information from a health provider with national coverage in Colombia, and prices correspond to tariffs of 2011 adjusted to 2012 values with the Colombian consumer price index for 2011. For the Brazilian analysis, event costs reflect information available from the Brazilian Healthcare Public System Hospital data Information System [29]; values represent the average bill for one hospitalization event, which does not include high-cost materials and medications nor emergency services. To adjust for emergency services, clinical protocols [30,31] from the Brazilian Ministry of Health were used. Procedures and examination costs were retrieved from the SUS Management Information System of the Procedures Medications and Materials List 2013 values [32]. Materials and medication costs were obtained from the SUS Healthcare Prices Database as the 2012 lowest purchase values [33] (Table 1). Baseline utility values used to estimate the number of qualityadjusted life-years (QALYs) associated with each treatment are presented in Table 1. QALYs were calculated by summing the product of utility weights associated with each health state and the time spent in the health state. Patients who experienced MCVEs incurred cumulative disutility based on the type and event number (Table 1). Death was assumed to result in a utility of zero. Cost-Effectiveness Analysis For each reference case analysis, the number of QALYs, life-years, MCVEs, and costs over the lifetime time horizon were summed by the model. Incremental cost-effectiveness ratios (ICERs) were estimated as the cost per QALY gained, cost per life-year saved, and cost per CV event avoided. ICERs were estimated by rankordering the treatment strategies by increasing cost and then comparing the more costly strategy with the less costly strategy by dividing the additional cost by the additional benefit. A therapy is considered to be dominated by its comparator— referred to as simple dominance—when it is both more costly and less effective. All costs and health outcomes were discounted at a 3.0% annual rate for the Colombian analysis and a 5.0% rate [34] for the Brazilian analysis, and reported in 2012 USD. When necessary, Brazilian costs were converted to USD [35]. The analysis was conducted from a payer perspective. A series of deterministic sensitivity analyses (DSAs) were performed by varying key model parameters through plausible ranges while keeping other parameters constant. The parameters that were allowed to vary within 25% of the base-case estimate included the expected percent drop in LDL-C level, the risk reduction associated with 1 mmol/L drop in LDL-C level, the 1-year post-CV event mortality rates, mortality multipliers 2 or more years after an event, adherence rates, disutility values for MCVEs, and the cost associated with each MCVE and statin. A second-order probabilistic sensitivity analysis (PSA) was also conducted on each reference case scenario to assess uncertainty around selected base-case parameter values. Using Monte-Carlo methods, the uncertainty in these parameters was characterized by probability distributions 51 VALUE IN HEALTH REGIONAL ISSUES 5C (2014) 48–57 Table 1 – Model parameters. Risk of MCVE % Primary prevention Untreated Treated Secondary prevention Untreated Untreated (patients with 2þ events) Treated Treated (patients with 2þ events) 1.83 1.41 5.68 18.01 4.48 14.48 Cost/unit Brazil Colombia NA 1.17 1.72 NA NA 1.18 1.93 2.06 0.92 1.74 2.92 3.67 2.88 3.43 4.88 6.17 1,523 720 1,914 1,523 7,441 3,267 9.320 10,048 Male 0.861 0.840 0.802 0.782 Female 0.837 0.811 0.771 0.724 * Statin therapy A10 A20 A40 A80 R5 R10 R20 R40 Major CVE† MI Stroke Revascularization CV death Utility values for both Brazil and Colombia Baseline utility values [39] Age 50–59 y Age 60–69 y Age 70–79 y Age 80–100 y MCVE disutility values‡ [40] Primary prevention MI Stroke Revascularization Secondary prevention MI Stroke Revascularization Stable CVD First event 0.0345 0.0483 0.0180 Second event 0.1255 0.1393 0.1090 First event 0.1255 0.1393 0.1090 0.018 (applied to everyone starting model) Third event 0.1185 0.1323 0.1020 Second event 0.1185 0.1323 0.1020 A10, atorvastatin 10 mg; A20, atorvastatin 20 mg; A40, atorvastatin 40 mg; A80, atorvastatin 80 mg; CV, cardiovascular; CVD, cardiovascular disease; MCVE, major cardiovascular event; MI, myocardial infarction; NA, not applicable/available; R5, rosuvastatin 5 mg; R10, rosuvastatin 10 mg; R20, rosuvastatin 20 mg; R40, rosuvastatin 40 mg; USD, US dollars. * 2012 USD. † Cost per event, 2012 USD. ‡ Utility associated with death (all-cause or CV related) assumed to be equal to zero. (beta or gamma) around each parameter’s midpoint estimate. The standard error was assumed to be 20% of the midpoint value, unless there were data available regarding the standard errors, in which case those data were used. The parameters included in the PSA were adherence rates, therapy efficacy, the risk reduction for a 1 mmol/L drop in LDL-C level, mortality multipliers after a CV event (first year and 2 or more years after), disutility values associated with MCVEs, and the cost of each MCVE and statin. Results The total discounted costs for each analysis are displayed by treatment group in Fig. 2, disaggregated by statin and MCVE treatment costs. In both scenarios (i.e., primary and secondary prevention) in the Colombian analysis, the drug costs for rosuvastatin were considerably higher than for atorvastatin and drove the differences seen in total costs. In the Brazilian analysis, drug costs also represented the largest portion of total costs, but in the analysis comparing R10 with A20 (primary prevention) and R40 with A40 (secondary prevention), drug costs were similar, and as a result, so were total costs; in the primary prevention analysis comparing A20 with R20, the drug cost for rosuvastatin was markedly higher, resulting in a higher total cost for the R20 cohort. The cost-effectiveness results for the reference case analyses are presented in Table 2. In all analyses for both countries, the incremental differences in QALYs, life-years, and MCVE per 52 VALUE IN HEALTH REGIONAL ISSUES 5C (2014) 48–57 Fig. 2 – Discounted total cohort (n ¼ 100) costs for (A) primary and (B) secondary prevention. CVD, cardiovascular disease. patient over the lifetime horizon displayed in Table 2 are minimal, ranging from 0 to 0.02 per patient. As a result of the minimal clinical benefits, the analyses with large differences in drug costs produced large ICERs. In the primary prevention analyses for Colombia, the incremental costs per QALY gained were $700,202 (A20 vs. R10) and $811,376 (A10 vs. R5), with incremental costs per life-year gained over $590,000 and the cost per MCVE avoided ranging from $3.8 million to $4.4 million; ICERs in the secondary prevention analyses were lower, ranging from approximately $189,000 to as high as $964,000. In the Brazilian analysis, the largest difference in drug costs was noted in the comparison of A20 with R20 in primary prevention, resulting in the largest ICERs. The incremental costs per QALY gained, life-year saved, and MCVE avoided were $135,120, $114,723, and $540,236, respectively. However, in the primary prevention analysis comparing A20 with R10, the total drug costs were similar, resulting in a cost per QALY gained of only $4,335, a cost per life-year gained of $3,681, and a cost per MCVE avoided of just above $17,000. In the Brazilian secondary prevention model, similar drug costs also produced modest ICERS: costs per QALY and life-year gained were both less than $20,000 while the cost per MCVE avoided was less than $40,000 (Table 2). As demonstrated by the DSA, the variables that had the largest impact on the results were the efficacy of each therapy in reducing the LDL-C level and the adherence rates. Fig. 3A displays a tornado diagram for the Colombian primary prevention DSA for A20 versus R10 and demonstrates that the ICERs in these scenarios vary greatly— each treatment is dominated by the other at some point—depending on the efficacy and adherence, while changes in cost and MCVE risk have a lesser effect. In the DSA for the other Colombian primary prevention analysis (A10 vs. R5), results were similar, except that the 1-year adherence rate of rosuvastatin emerged as having the largest impact on the cost per QALY gained (not shown). For the Brazilian primary prevention analyses, the DSA for A20 versus R10 revealed that treatment efficacy has the largest impact, yet did not alter dominance (Fig. 3B). The DSA for A20 versus R20 was similar, though the impact of adherence rates was much more considerable, similar to the results shown for Colombia (Fig. 3A). As with the Colombian models, the cost of therapy and the risk of MCVE had less of an impact on the results. DSA results for 53 VALUE IN HEALTH REGIONAL ISSUES 5C (2014) 48–57 Table 2 – Total and incremental cost and outcomes per patient for primary and secondary prevention. Prevention level and treatment dose Primary prevention A10 R5 Difference A20 R10 Difference Secondary prevention A40 R20 Difference A80 R40 Difference Colombia Total ICERs Cost ($) QALYs Lifeyears CV events* Cost/QALY gained ($) Cost/life-year saved ($) Cost/CV event avoided ($) 4,438 8,749 4,312 6,234 9,953 3,719 13.41 13.42 0.01 13.42 13.42 0.01 16.81 16.82 0.01 16.82 16.83 0.01 0.30 0.30 0.01 0.30 0.30 0.00 Reference 811,376 – Reference 700,202 – Reference 685,005 – Reference 591,086 – Reference 4,422,698 – Reference 3,819,276 – 10,633 13,621 2,988 11,748 15,575 3,826 8.63 8.64 0.01 8.65 8.67 0.01 11.12 11.13 0.02 11.14 11.16 0.02 0.80 0.79 0.00 0.79 0.79 0.00 Reference 201,667 – Reference 258,168 – Reference 188,591 – Reference 241,654 – Reference 752,378 – Reference 963,614 – Brazil Total Primary prevention A20 R10 Difference A20 R20 Difference Secondary prevention A40 R40 Difference ICERs Cost ($) QALYs Lifeyears CV events* Cost/QALY gained ($) Cost/life-year saved ($) Cost/CV event avoided ($) 2,061 2,076 15 2,061 3,223 1,162 9.12 9.12 0.00 9.12 9.13 0.01 11.29 11.29 0.00 11.29 11.30 0.01 0.19 0.19 0.00 0.19 0.19 0.01 Reference 4,335 – Reference 135,120 – Reference 3,681 – Reference 114,723 – Reference 17,322 – Reference 540,236 – 2,606 2,957 351 5.69 5.72 0.02 7.22 7.24 0.02 0.52 0.51 0.01 Reference 16,771 – Reference 16,620 – Reference 36,783 – A10, atorvastatin 10 mg; A20, atorvastatin 20 mg; A40, atorvastatin 40 mg; A80, atorvastatin 80 mg; CV, cardiovascular; ICER, incremental costeffectiveness ratio; QALY, quality-adjusted life-year; R5, rosuvastatin 5 mg; R10, rosuvastatin 10 mg; R20, rosuvastatin 20 mg; R40, rosuvastatin 40 mg. * Difference ¼ events avoided. secondary prevention in each country were similar to those of primary prevention (not shown). Results of the PSA for the analyses comparing A20 versus R10 in Colombia and Brazil in primary prevention are displayed in Fig. 4. For the Colombian analysis, at a willingness-to-pay threshold of $100,000 per incremental QALY, R10 had less than a 3% chance of being the optimal therapy, and at approximately $680,000 per QALY, this probability is 50%. The PSA curves for A10 versus R5 in primary prevention and A40 versus R20 and A80 versus R40 in secondary prevention in Colombia had similar profiles, with atorvastatin having a high probability of being the cost-effective therapy at very low willingness-to-pay values; as the willingness-to-pay threshold increased, the probability of rosuvastatin becoming the cost-effective treatment improved, reaching a probability of 50% at approximately $890,000, $220,000, and $270,000 per QALY gained, respectively (not shown). In the Brazilian analyses, the PSA curves had similar profiles to those of the Colombian curves; however, the thresholds at which rosuvastatin had equal probability to atorvastatin of being the cost-effective therapy were considerably lower; the willingness-to-pay values to obtain an additional QALY were approximately $13,000 (A20 vs. R10, primary prevention [Fig. 4]), $140,000 (A20 vs. R20, primary prevention; data not shown), and $16,000 (A40 vs. R40, secondary prevention; data not shown). Discussion The goal of this cost-effectiveness analysis was to determine whether the additional reduction in LDL-C level associated with rosuvastatin translated into enough of a reduction in major CV events to offset the additional cost compared with equivalent doses of atorvastatin in both primary and secondary prevention populations in Brazil and Colombia. The clinical benefits of rosuvastatin in each analysis were minimal, whether measured in QALYs, life-years gained, or major CV events avoided. This fact resulted in large ICERs when coupled with a large differential in drug acquisition cost, and the differences in the total costs associated with each therapy were almost exclusively driven by the differences in costs for the drugs themselves. 54 VALUE IN HEALTH REGIONAL ISSUES 5C (2014) 48–57 Fig. 3 – Tornado diagrams, primary prevention, for (A) Colombia A20 versus R10 and (B) Brazil A20 versus R10. A20, atorvastatin 20 mg; CVD, cardiovascular disease; LDL, low-density lipoprotein; LDL-C, low-density lipoprotein cholesterol; QALY, quality-adjusted life-year; R10, rosuvastatin 10 mg. In the Colombian analyses, the ICERs per QALY gained for primary prevention analyses were more than $700,000 for both dose comparisons and for secondary prevention were more than $200,000. The DSA demonstrated that changes in efficacy and adherence—which would, in turn, produce changes in the clinical benefit associated with each therapy—have the largest impact on the ICERs, considering the long-term nature of the model and its impact on direct costs for each comparator. The PSAs indicated that willingness-to-pay thresholds of more than $600,000 per additional QALY were necessary for rosuvastatin to have a 50% chance of being the optimal primary prevention therapy. At the high-dose comparisons for secondary prevention, the thresholds VALUE IN HEALTH REGIONAL ISSUES 5C (2014) 48–57 55 Fig. 4 – CE acceptability curves, primary prevention, for (A) Colombia A20 versus R10 and (B) Brazil A20 versus R10. A20, atorvastatin 20 mg; CE, cost-effectiveness; QALY, quality-adjusted life-year; R10, rosuvastatin 10 mg. were in excess of $200,000. These results suggest that rosuvastatin is not a cost-effective treatment option. For the Brazilian analysis comparing A20 with R10, total costs were very similar, resulting in an ICER of only $4000 per QALY gained. Although the comparison of A20 with R20 resulted in a much larger ICER, some might not consider A20 and R20 to represent equivalent statin doses. The ICER obtained in the secondary prevention analysis (A40 vs. R40) was also much lower in Brazil than in Colombia at approximately $17,000 per QALY gained. As with the Colombian analyses, the DSAs for the Brazilian analyses indicated that changes in efficacy and adherence would have a large impact on the resulting ICER. As shown in the PSAs, a willingness-to-pay threshold of only $13,000 per QALY gained is required for R10 to have a 50% probability of being the optimal therapy in primary prevention versus A20. Several limitations are worth noting. First, the model used in this analysis examines only the impact of lowering the LDL-C level on cost-effectiveness. Also, several inputs are taken from the CTTC meta-analysis, not directly from Latin American populations, and therefore may not be representative of Brazilian or 56 VALUE IN HEALTH REGIONAL ISSUES 5C (2014) 48–57 Colombian populations. For example, previous literature has demonstrated poor adherence to statin therapy in Brazilian populations [36,37], often due to the cost. Because the CTTC meta-analysis provided a hazard ratio for CV risk linked to LDL-C level based on a large amount of clinical evidence, our model used this new data in a novel approach to estimate costeffectiveness in CV prevention populations. It is well established, however, that the risk of CV events is also affected by changes in total cholesterol and high-density lipoprotein cholesterol levels, among a number of other factors such as smoking status and systolic blood pressure [38]. Given that such risk factors were not examined in the CTTC meta-analysis, our model was limited to considering the LDL-C level only. Finally, prices of statin therapy as well as treatment costs for MCVE are always changing; therefore, these results may not apply in specific health care institutions within both markets. Similar values for reduction in LDL-C levels were reported in a previous meta-analysis according to included statins and daily dose [41]. A recent head-to-head comparison [42] shows that there is no statistical difference between comparators by reducing LDL-C and LDL particles; similar findings were reported from an open-label study [43]. Clinical benefits are similar for both comparators; the difference in acquisition costs is the critical variable for the results. In conclusion, in most analyses, the efficacy advantage of rosuvastatin in primary and secondary CV prevention is relatively small compared with the differential in acquisition cost, especially in the analyses in the Colombian setting. The results of our cost-effectiveness analysis suggest that the additional efficacy of rosuvastatin does not support its extra cost in these scenarios. These similar results in terms of outcomes, QALYs, life-years, and CV events suggest atorvastatin as an option that generates lower costs to the payer than does rosuvastatin. Source of financial support: These findings are the result of work supported by Pfizer, Inc. The views expressed in this article are those of the authors, and no endorsement by Pfizer is intended or should be inferred. R EF E R EN CE S [1] Barceló A. Cardiovascular diseases in Latin America and the Caribbean. Lancet 2006;368:625–6. [2] Korenfeld Y, Mendoza-Bastidas C, Saavedra L, et al. Current status of cardiac rehabilitation in Latin America and the Caribbean. Am Heart J 2009;158:480–7. [3] Rubinstein AL, Irazola VE, Poggio R, et al. Detection and follow-up of cardiovascular disease and risk factors in the Southern Cone of Latin America: the CESCAS I study. BMJ Open 2011;1:e000126. [4] Polanczyk CA, Ribeiro JP. Coronary artery disease in Brazil: contemporary management and future perspectives. Heart 2009;95:870–6. [5] Cesarino EJ, Vituzzo AL, Sampaio JM, et al. Assessment of cardiovascular risk of patients with arterial hypertension of a public health unit. Einstein (Sao Paulo) 2012;10:33–8. [6] Gil GP, Dessotte CA, Schmidt A, et al. Clinical evolution of patients hospitalized due to the first episode of acute coronary syndrome. Rev Lat Am Enfermagem 2012;20:830–7. [7] Pan American Health Organization/World Health Organization (PAHO/ WHO) HIaAPHH. Health Situation in the Americas: Basic Indicators 2012. Washington, DC, 2012. Available from: http://ais.paho.org/chi/ brochures/2012/BI_2012_SPA.pdf?ua=1. [Accessed July 22, 2013]. [8] Avezum A, Braga J, Santos I, et al. Cardiovascular disease in South America: current status and opportunities for prevention. Heart 2009;95:1475–82. [9] Lanas F, Avezum A, Bautista LE, et al. Risk factors for acute myocardial infarction in Latin America: the INTERHEART Latin American study. Circulation 2007;115:1067–74. [10] Smith SC Jr. Risk factors for myocardial infarction in Latin America: sobrepeso y obesidad. Circulation 2007;115:1061–3. [11] Bautista LE, Orostegui M, Vera LM, et al. Prevalence and impact of cardiovascular risk factors in Bucaramanga, Colombia: results from the Countrywide Integrated Noncommunicable Disease Intervention [12] [13] [14] [15] [16] [17] [18] [19] [20] [21] [22] [23] [24] [25] [26] [27] [28] [29] [30] [31] [32] [33] [34] [35] Programme (CINDI/CARMEN) baseline survey. Eur J Cardiovasc Prev Rehabil 2006;13:769–75. Brunzell JD, Davidson M, Furberg CD, et al. Lipoprotein management in patients with cardiometabolic risk: consensus conference report from the American Diabetes Association and the American College of Cardiology Foundation. J Am Coll Cardiol 2008;51:1512–24. Expert Panel on Detection, Evaluation, and Treatment of High Blood Cholesterol in Adults. Executive Summary of The Third Report of The National Cholesterol Education Program (NCEP) Expert Panel on Detection, Evaluation, And Treatment of High Blood Cholesterol In Adults (Adult Treatment Panel III). JAMA 2001;285:2486–97. Ma J, Sehgal NL, Ayanian JZ, Stafford RS. National trends in statin use by coronary heart disease risk category. PLoS Med 2005;2:e123. Taylor F, Huffman MD, Macedo AF, et al. Statins for the primary prevention of cardiovascular disease. Cochrane Database Syst Rev 2013;1:CD004816. III Diretriz Brasileira sobre Dislipidemias e Prevenção da Aterosclerose, do Departamento de Aterosclerose, da Sociedade Brasileira de Cardiologia 2001. Available from: www.cardiol.com.br. [Accessed July 22, 2013]. Sposito AC, Santos SN, de Faria EC, et al. Timing and dose of statin therapy define its impact on inflammatory and endothelial responses during myocardial infarction. Arterioscler Thromb Vasc Biol 2011;31:1240–6. Schneck DW, Knopp RH, Ballantyne CM, et al. Comparative effects of rosuvastatin and atorvastatin across their dose ranges in patients with hypercholesterolemia and without active arterial disease. Am J Cardiol 2003;91:33–41. Murphy SA, Cannon CP, Wiviott SD, et al. Effect of intensive lipidlowering therapy on mortality after acute coronary syndrome (a patient-level analysis of the Aggrastat to Zocor and Pravastatin or Atorvastatin Evaluation and Infection Therapy-Thrombolysis in Myocardial Infarction 22 trials). Am J Cardiol 2007;100:1047–51. Cholesterol Treatment Trialists’ (CTT) Collaboration. Baigent C, Blackwell L, Emberson J, et al. Efficacy and safety of more intensive lowering of LDL cholesterol: a meta-analysis of data from 170,000 participants in 26 randomised trials. Lancet 2010;376:1670–81. Dennis MS, Burn JP, Sandercock PA, et al. Long-term survival after firstever stroke: the Oxfordshire Community Stroke Project. Stroke 1993;24:796–800. Pell JP, Walsh D, Norrie J, et al. Outcomes following coronary artery bypass grafting and percutaneous transluminal coronary angioplasty in the stent era: a prospective study of all 9890 consecutive patients operated on in Scotland over a two year period. Heart 2001;85:662–6. Rosengran ASC, Koster M, Hammar N, et al. Sex differences in survival after myocardial infarction in Sweden: data from the Swedish National Acute Myocardial Infarction register. Eur Heart J 2001;22:314–22. Instituto Brasileiro de Geografia e Estatistica IBGE B. In Tabua completa de mortalidade ambos os sexos. 2011. Available from: http://www.ibge. gov.br/home/estatistica/populacao/tabuadevida/2011/. [Accessed July 24, 2013]. National Administrative Department of Statistics - DANE. Deaths by Sex and Age Group, Total National 2012. Available from: https://www. dane.gov.co/files/investigaciones/fichas/ingles/Methodology%20of% 20Vital%20Statistics.pdf. [Accessed July 24, 2013]. Straka RJ, Mamdani M, Damen J, et al. Economic impacts attributable to the early clinical benefit of atorvastatin therapy—a US managed care perspective. Curr Med Res Opin 2007;23:1517–29. ANVISA ANdVS. Lista de Precos de Medicamentos. National Information System of Prices of Medicines - SISMED. Republic of Colombia: Ministry of Health and Social Protection, 2012. Sistema de Informacoes Hospitalares SIH. Morbidade Hospitalar do SUS - por local de interacao, Brasil. Available from: http://tabnet.datasus. gov.br/cgi/deftohtm.exe?sih/cnv/mibr.def. [Accessed July 24, 2013]. Brazilian Ministry of Health. Protocol for Acute Coronary Syndrome. Available from: http://www.saude.pr.gov.br/arquivos/File/HOSPSUS/ protocolo_sindrome_coronariaMS2011.pdf. [Accessed July 31, 2013]. Brazilian Ministry of Health. Protocol for Acute Cerebrovascular Emergency. Available from: http://bvsms.saude.gov.br/bvs/publicacoes/ manual_rotinas_para_atencao_avc.pdf. [Accessed August 30, 2013]. Sistema de Gerenciameto da Tabela de Procedimentos MeOdS. Available from: http://sigtap.datasus.gov.br/tabela-unificada/app/sec/ inicio.jsp. [Accessed August 27, 2013]. Portal da Saùde. Banco de Preços en saude. Available from: http:// portalsaude.saude.gov.br/index.php/cidadao/principal/banco-deprecos-em-saude. [Accessed August 12, 2013]. Brasil MdS, Secretaria de Ciencia, Tecnologia e Insumos Estrategicos, Departmento de Ciencia e Tecnologia, Diretrizes Metodologicas para Estudos de Avaliacao Economica de Tecnologias para o Ministerio da Saude. 2013. Coinmill.com. Brazilian real (BRL) and United States dollar (USD) currency exchange rate conversion calculator. Available from: http:// coinmill.com/BRL_USD.html#BRL=10048. [Accessed July 22, 2013]. VALUE IN HEALTH REGIONAL ISSUES 5C (2014) 48–57 [36] Luz PL, Nishiyama M, Chagas AC. Drugs and lifestyle for the treatment and prevention of coronary artery disease: comparative analysis of the scientific basis. Braz J Med Biol Res 2011;44:973–91. [37] Mansur AP, Mattar AP, Tsubo CE, et al. Prescription and adherence to statins of patients with coronary artery disease and hypercholesterolemia. Arq Bras Cardiol 2001;76:111–8. [38] National Cholesterol Education Program Expert Panel on Detection Evaluation, and Treatment of High Blood Cholesterol in Adults. Third Report of the National Cholesterol Education Program (NCEP) Expert Panel on Detection, Evaluation, and Treatment of High Blood Cholesterol in Adults (Adult Treatment Panel III) final report. Circulation 2002;106: 3143–21. [39] Hammer J, Lawrence WF, Anderson JP, et al. Report of nationally representative values for the noninstitutionalizd US adult population for 7 heath-related quality-of-life scores. Med Decis Making 2006;26:391–400. 57 [40] Sullivan PW, Lawrence WF, Ghushchyan V. A national catalog of preference-based scores for chronic conditions in the United States. Med Care 2005;43:736–49. [41] Law MR, Wald NJ, Rudnicka AR. Quantifying effect of statins on low density lipoprotein cholesterol, ischaemic heart disease, and stroke: systematic review and meta-analysis. BMJ 2003;326:1423. [42] Saku K, Zhang B, Noda K. PATROL Trial Investigators. Randomized head-to-head comparison of pitavastatin, atorvastatin, and rosuvastatin for safety and efficacy (quantity and quality of LDL): the PATROL trial. Circ J 2011;75:1493–505. [43] Milionis HJ, Rizos E, Kostapanos M, et al. Treating to target patients with primary hyperlipidaemia: comparison of the effects of ATOrvastatin and ROSuvastatin (the ATOROS study). Curr Med Res Opin 2006;22:1123–31. VALUE IN HEALTH REGIONAL ISSUES 5C (2014) 58–64 Available online at www.sciencedirect.com journal homepage: www.elsevier.com/locate/vhri Análisis Costo-Utilidad Del Uso De Adalimumab, Etanercept E Infliximab Para El Tratamiento De La Artritis Psoriásica En Uruguay Nicolás González-Vacarezza, Pharm, MSc1,*, Eduardo Gehling Bertoldi, MD, MSc2, Ana Deminco, Pharm1, Graciela González, MD3, Miguel Martínez Asuaga, MD4 1 División Evaluación Sanitaria, Ministerio de Salud Pública, Montevideo, Uruguay; 2Universidade Federal do Rio Grande do Sul, Instituto de Avaliação de Tecnologia em Saúde (IATS), Porto Alegre, RS, Brasil; 3Instituto Nacional de Reumatología, Administración de Servicios de Salud del Estado, Montevideo, Uruguay; 4Cátedra de Dermatología, Facultad de Medicina, Universidad de la República, Montevideo, Uruguay AB STR A CT Objective: To support an evidence-based decision regarding the inclusion of the anti TNFʼs adalimumab, etanercept or infliximab to the National Formulary of Uruguay for the treatment of psoriatic arthritis, in patients who did not respond to first-line treatment. Methods: We perform a cost-utility evaluation using a Markov model, to estimate the incremental costs and quality-adjusted life years (QALY) gained with each of these biologic drugs compared with palliative care only. The model considered the perspective of the National Health System, with a time horizon of 40 years. Results: Under base-case assumptions, infliximab results in an incremental cost-effectiveness ratio (ICER) of United States Dollars (USD) 47,294 compared to palliative care only. Adalimumab was extendedly dominated and etanercept was dominated. However, infliximab requires intravenous administration which could significantly reduce access to treatment, thus producing inequities for patients from different regions. The probabilistic sensitivity analysis shows that all biologics had null probabilities of being cost-effective considering a willingness-to-pay (WTP) threshold of one time the gross domestic product per capita of Uruguay per QALY. Conclusions: In our analysis from the perspective of the National Health System, the use of biologic drug treatments for psoriatic arthritis results in an ICER above the WTP ceiling proposed by the World Health Organization. Reduction of drug prices could bring more favorable cost-effectiveness results. Keywords: arthritis, cost-effectiveness, cost-utility, psoriatic. Introducción sería de aproximadamente 30.317 sujetos. Sin embargo, la falta de datos a nivel nacional hace que esta estimación pueda diferir significativamente de la realidad, considerando además el alto nivel de subdiagnóstico que presenta esta patología. Las repercusiones para los pacientes con artritis psoriásica tienen un amplio espectro, tanto a nivel dérmico como musculo esquelético las cuales repercuten en las actividades diarias del individuo, su vida laboral y social, generando una calidad de vida inferior a otras enfermedades crónicas comunes, pudiéndose producir en algunos casos deformidades y daño en las articulaciones [7–8]. Incluso la percepción del paciente respecto a la afectación física y mental puede ser superior respecto a enfermedades como cáncer, artritis, hipertensión, enfermedades cardíacas, diabetes y depresión [9–10]. El tratamiento de la artritis psoriásica en primera línea consta de analgésicos antiinflamatorios no esteroideos (AINEs) en la menor dosis y tiempo posible, si bien la evidencia que sustenta su eficacia es limitada. El efecto se produce en la reducción de La artritis psoriásica es una enfermedad inflamatoria crónica de origen autoinmune, la cual presenta una afectación significativa de calidad de vida de los pacientes, principalmente cuando la actividad es de moderada a severa [1–2]. Si bien debe considerarse que existe un subdiagnóstico importante, aproximadamente un 3% de la población de Estados Unidos padece psoriasis en placa, de los cuales un 17% presentan actividad de la enfermedad moderada a severa [3]. Dentro de estos pacientes, se estima que un 6-39% manifiestan además artritis psoriásica, porcentaje que varía según la población evaluada y el método utilizado [4–5]. Específicamente, un estudio poblacional europeo informó una prevalencia de 30% de artritis psoriásica en pacientes con psoriasis en placa [6]. No existen estudios epidemiológicos en población uruguaya, por lo que si consideramos una prevalencia de 3% para psoriasis en placa y que de éstos pacientes un 30% podrían además presentar artritis psoriásica, la población teórica con esta patología en nuestro país Copyright & 2014, International Society for Pharmacoeconomics and Outcomes Research (ISPOR). Published by Elsevier Inc. Conflict of interest: The authors have indicated that they have no conflicts of interest with regard to the content of this article. * Corresponding Author: Nicolás González-Vacarezza MSc, Health Assessment Division, Ministry of Public Health, 1892 18 de Julio Avenue, PO 11200, Montevideo, Uruguay. Tel.: (598)-2402-8032 ext. 127; fax: (598)-2402-5836. E-mail: [email protected].. 2212-1099$36.00 – see front matter Copyright & 2014, International Society for Pharmacoeconomics and Outcomes Research (ISPOR). Published by Elsevier Inc. http://dx.doi.org/10.1016/j.vhri.2014.09.001 VALUE IN HEALTH REGIONAL ISSUES 5C (2014) 58–64 síntomas a nivel de las articulaciones, sin efecto sobre la lesiones en la piel [11]. Si bien el inicio del tratamiento con fármacos modificadores de la enfermedad (FARMEs) de síntesis no está totalmente claro, los pacientes con enfermedad activa a pesar del uso de AINEs y con mal pronóstico deberían iniciar tratamiento con éstos fármacos, siendo el metotrexato el de primera elección. Otros FARMEs de síntesis que pueden ser utilizados son sulfasalazina, leflunomida y ciclosporina A. Sin embargo, es importante indicar que los FARMEs parecen no ser eficaces para tratar la entesitis e implicancias axiales, ni tampoco para reducir el daño estructural. Adicionalmente, la terapia concomitante con inyección de glucocorticoides en enfermedad localizada puede producir mejoras. Sin embargo, si bien el uso de corticoides sistémicos son una opción terapéutica, su uso prolongado puede resultar en eventos adversos de importancia y generar una exacerbación de las manifestaciones cutáneas [11]. Generalmente, cuando un paciente no muestra una respuesta al tratamiento con FARMEs de síntesis durante 3-6 meses o presenta intolerancia, se recomienda iniciar tratamiento con fármacos biológicos inhibidores del factor de necrosis tumoral alfa (anti-TNF), como adalimumab, etanercept o infliximab. Los anti-TNF han demostrado ser eficaces respecto a la afectación a nivel de piel y articulaciones, así como también para prevenir el daño evaluado mediante radiografía. Además, han demostrado mejorar la entesitis y la afección axial. Dado que hasta el momento no se han reportado ensayos clínicos controlados que comparen anti-TNFs entre ellos, no es posible conocer con buen nivel de evidencia la superioridad de uno respecto a otro [11]. El Sistema de Salud de Uruguay cubre los tratamientos con AINEs y FARMEs de síntesis, no estando incluidos por el momento los anti-TNFs. Esto deja una necesidad clínica insatisfecha respecto a alternativas terapéuticas que logren disminuir y prevenir el daño estructural, entesitis e implicancias axiales, y por lo tanto mejoren la calidad de vida de los pacientes que padecen artritis psoriásica. Por esto fue realizada una evaluación costo-utilidad del uso de adalimumab (HUMIRAs), etanercept (ENBRELs) e infliximab (REMICADEs) para el tratamiento de la artritis psoriásica moderada a severa, desde la perspectiva del Sistema Nacional Integrado de Salud. Métodos Fue realizada una búsqueda de evaluaciones económicas que evalúen el uso de de adalimumab, etanercept e infliximab para el tratamiento de la artritis psoriásica. Para la búsqueda de evaluaciones económicas fueron utilizadas las bases Centre for Reviews and Dissemination de la University of York, La Biblioteca Cochrane y MedLine. La búsqueda identificó una evaluación económica de buena calidad que cumple con los criterios antes definidos [12], por lo que fue adaptada al contexto de Uruguay. Este estudio determinó la costo-utilidad del uso de adalimumab, etanercept e infliximab en el tratamiento de la artritis psoriásica, en pacientes que no respondieron a al menos dos FARMEs, desde la perspectiva del Sistema Nacional de Salud del Reino Unido. Para adaptar este modelo al contexto de Uruguay, se procedió al desarrollo completo del mismo considerando aspectos nacionales de la práctica clínica, costos e información epidemiológica. Diseño del modelo Dado que la artritis psoriásica es una enfermedad crónica que evoluciona durante toda la vida del paciente, fue desarrollado en el software TreeAge Pro 2011 (TreeAge Software, Inc., Williamstown, MA) un modelo de Markov considerando que los pacientes cursan la enfermedad por estados de salud bien definidos y mutuamente excluyentes, los cuales son evaluados cada tres 59 meses. El modelo asume una cohorte homogénea de pacientes, los cuales inician tratamiento con adalimumab, etanercept, infliximab o tratamiento de soporte. Respecto al abandono del tratamiento con medicamentos biológicos, el modelo considera la respuesta en articulaciones es evaluada trimestralmente por la escala PsARC (Psoriatic Arthritis Response Criteria) como el criterio de decisión para continuar o discontinuar la terapia. Al inicio deben producirse dos evaluaciones negativas (no logra respuesta PsARC) correspondientes a 3 y 6 meses de terapia para que se indique abandonar el tratamiento [11], criterio que es compartido por los expertos clínicos consultados. Vinculado a lo anterior, los pacientes que no logran una respuesta PsARC en articulaciones pero presentan una respuesta en piel de PASI90 continúan el tratamiento biológico, siendo un criterio aceptado por expertos clínicos de nuestro país. Luego de los 6 meses se consideró que los pacientes respondedores presentan una tasa de abandono por cualquier causa para los tratamientos biológicos de 0,165 por año. Esta tasa de abandono fue estimada de cuatro estudios de registro de pacientes con artritis psoriásica, lo que representa de mejor forma la práctica clínica real [12]. Todos los pacientes que dejan de recibir el medicamento biológico pasan a tratamiento soporte no considerándose la posibilidad del uso secuencial de anti-TNFs, dado que el objetivo de esta evaluación es determinar cuál de estos medicamentos es la estrategia más costo efectiva. El modelo incluyó nueve estados de salud posibles considerando una respuesta positiva o negativa en la escala PsARC, cuatro intervalos de respuesta en la reducción de la escala Psoriasis Area Severity Index (PASI) de 0-49%, 50-74%, 75-89%, 90-100%, y muerte (Figura 1). Población objetivo Pacientes con diagnóstico de artritis psoriásica sin restricción de edad ni comorbilidades, que no hayan respondido o sean intolerantes a al menos dos FARMEs de síntesis. Las características de la población considerada en el modelo fue la informada en los ensayos clínicos de los tres medicamentos, la cual no difiere significativamente de la población objetivo de esta evaluación según la opinión de expertos. En resumen, el modelo asume una población de 47-53 años de edad, 7 años de duración de la enfermedad, no respondedores a al menos dos FARMEs y con actividad moderada de la enfermedad. Se asume la mortalidad de la población general dado que no hay evidencia que demuestre que los tratamientos biológicos aumenten la expectativa de vida. Por lo tanto, se espera que esta variable no influya significativamente en la comparación respecto al tratamiento estándar. Perspectiva del estudio La evaluación fue realizada desde la perspectiva del Sistema Nacional de Salud de Uruguay. Intervenciones comparadas En Uruguay se encuentran incluidos tratamientos farmacológicos que incluyen AINEs y FARMEs, siendo metotrexato el de primera elección pero pudiendo ser utilizados otros como sulfasalazina, leflunomida y ciclosporina A, y glucocorticoides orales e inyectables. Esto resulta que los diferentes pacientes, según la severidad de la enfermedad, respuesta o intolerancia al tratamiento, reciben diferentes combinaciones de las farmacoterapias antes mencionadas. Los pacientes que no responden a dichos tratamientos reciben principalmente cuidado paliativo sintomático, el cual es considerado el comparador para esta evaluación. 60 VALUE IN HEALTH REGIONAL ISSUES 5C (2014) 58–64 1 1 1 1 2 2 2 2 Figura 1 – Esquema simplificado de parte de la estructura del modelo. Los números 1 y 2 indican que esas ramas del modelo se repiten en los nodos indicados. PASI, Psoriasis Area Severity Index; PASI 75, logra una reducción del 75% en el Psoriasis Area Severity Index respecto al valor basal; PsARC, Psoriatic Arthritis Response Criteria. Por tanto, la evaluación costo-utilidad comparó cada anti-TNF adicionado al tratamiento de soporte o en monoterapia respecto al tratamiento de soporte. El modelo consideró que la administración de los tres anti-TNFs se realiza como indican las respectivas informaciones de prescripción de los productos. Horizonte temporal Los ensayos clínicos de los tres anti-TNFs incluyeron pacientes en el rango de edades de 43 a 50 años [12–13], por lo que considerando una expectativa de vida en Uruguay de 76 años [14] el horizonte temporal fue definido en 40 años para el caso base. Sin embargo, en la práctica clínica un paciente que en forma persistente (6 meses) no presenta respuesta a nivel de articulaciones deja de ser tratado con anti-TNF. Por lo tanto, dado el relativo bajo tiempo de permanencia en tratamiento con estos medicamentos fue incluido en el análisis de escenarios un horizonte temporal de 10 años. Tasa de descuento La tasa de descuento utilizada fue de 5% en el caso base, si bien en el análisis de sensibilidad incluyeron valores entre 0% y 10% [15]. Los descuentos fueron aplicados a los costos y utilidades futuras. Variables y resultados de eficacia La respuesta en articulaciones periféricas fue considerada en el modelo mediante la escala PsARC. Esta escala evalúa las articulaciones inflamadas y con dolor, actividad global de la enfermedad evaluada por el médico y evaluación global por el paciente (ambas en escala de 0 a 5). La respuesta es definida por una reducción mayor o igual a 30% en las articulaciones inflamadas y/o con dolor, y una reducción de un punto en la evaluación de la enfermedad por el paciente o por el médico. Una respuesta global se define como una mejora en dos de los cuatro ítems, uno de los cuales tiene que ser de articulaciones, sin empeorar ninguno de los otros [16]. La respuesta en piel fue incluida en el modelo mediante la escala PASI, la cual evalúa la severidad y superficie de las lesiones, siendo por ejemplo PASI50 una respuesta que implica una mejora del 50% respecto al nivel basal [17]. El modelo consideró que las respuestas PsARC y PASI son dependientes para los pacientes tratados con medicamentos biológicos mediante un coeficiente de correlación (Tabla 1), y no se encuentran relacionados para el caso del tratamiento de soporte. Por lo tanto fueron calculadas las probabilidades combinadas de las posibles respuestas PsARC – PASI para cada medicamento biológico. Se utilizó como medida de beneficios en salud los años de vida ajustados por calidad (AVAC), calculando las utilidades a partir de los resultados del cuestionario Health Assessment Questionnaire (HAQ) y la escala PASI. El HAQ es un cuestionario que permite evaluar principalmente el nivel de incapacidad del paciente para realizar tareas cotidianas, por lo que se vincula mayormente a la respuesta en articulaciones. Los pacientes que no responden al tratamiento soporte les fue asignado valores de HAQ correspondientes la evolución natural de la enfermedad para esta población. Sin embargo, la mejora medida por la escala HAQ en una parte de los pacientes que reciben el tratamiento soporte puede deberse al contexto del ensayo clínico, lo que puede ser discordante con la práctica clínica real, dado que estos pacientes tienen como criterio de inclusión ser no respondedores. Por lo tanto, fue evaluado el escenario en el cual todos los pacientes tratados con tratamiento soporte tienen una evolución en la escala HAQ de acuerdo a la evolución natural de la enfermedad, sin importar la respuesta en la escala PsARC. Fueron considerados los resultados de eficacia y seguridad informados por Rodgers et al. [12], dado que al momento de realizar esta evaluación económica no fueron identificados ensayos controlados adicionales que justificaran actualizar los metaanálisis. La estrategia de búsqueda para actualizar la evidencia se indica en el Anexo (see Appendix in Supplemental Materials found at http://dx.doi.org/10.1016/j.vhri.2014.09.001). Costos Los costos trimestrales fueron calculados para cada estado de salud, considerando únicamente los costos directos de atención para el Sistema de Salud. Para esto fue elaborada una lista de todos los posibles servicios e insumos (incluyendo los medicamentos) que utiliza un paciente con artritis psoriásica, considerando la frecuencia de uso en esa población y la frecuencia temporal. Los 61 VALUE IN HEALTH REGIONAL ISSUES 5C (2014) 58–64 Tabla 1 – Variables incluidas en el modelo para el caso base. Variable HAQ inicial PASI inicial Horizonte temporal (años) Tasa de descuento anual (%) Intercepción función utilidad Cambio en utilidad por unidad de HAQ Cambio en utilidad por unidad de HAQ en pacientes que no responden al tratamiento soporte. Cambio en utilidad por unidad de PASI Probabilidad PsARC(þ) adalimumab Probabilidad PsARC(þ) etanercept Probabilidad PsARC(þ) infliximab Probabilidad PASI 50(þ) adalimumab Probabilidad PASI 50 (þ) etanercept Probabilidad PASI 50 (þ) infliximab Probabilidad PASI 75 (þ) adalimumab Probabilidad PASI 75 (þ) etanercept Probabilidad PASI75 (þ) infliximab Probabilidad PASI 90 (þ) adalimumab Probabilidad PASI 90 (þ) etanercept Probabilidad PASI 90 (þ) infliximab Coeficiente de correlación PsARC – PASI Probabilidad trimestral de abandono del tratamiento biológico Mortalidad Costo trimestral por paciente PsARC(þ) y PASI 75(þ) (USD) Costo trimestral por paciente PsARC(þ) y PASI 75(-) (USD) Costo trimestral por paciente PsARC(-) y PASI 75(þ) (USD) Costo trimestral por paciente PsARC(-) y PASI 75(-) (USD) Media 1,05 7,5 40 5,0 0,897 Fuente Rodgers et al. Rodgers et al. Rodgers et al. Directrices MERCOSUR Rodgers et al. [12] [12] [12] [15] [12] 0,298 Rodgers et al. [12] 0,0174 Rodgers et al. [12] 0,004 Rodgers et al. [12] 0,587 Rodgers et al. [12] 0,713 Rodgers et al. [12] 0,795 Rodgers et al. [12] 0,738 Rodgers et al. [12] 0,403 Rodgers et al. [12] 0,913 Rodgers et al. [12] 0,477 Rodgers et al. [12] 0,177 Rodgers et al. [12] 0,769 Rodgers et al. [12] 0,257 Rodgers et al. [12] 0,0737 Rodgers et al. [12] 0,557 Rodgers et al. [12] 0,435 Rodgers et al. [12] 0,04125 Rodgers et al. [12] Tabla 394,9 420,0 948,7 974,8 Instituto Nacional de Estadística de Uruguay [14] Ministerio de Salud Pública de Uruguay Ministerio de Salud Pública de Uruguay Ministerio de Salud Pública de Uruguay Ministerio de Salud Pública de Uruguay continued on next page Tabla 1 – Continuación Variable Media Costo trimestral adalimumab (USD) 4.686 Costo trimestral etanercept (USD) 6.178 Costo trimestral infliximab (USD)* 3.904 Costo inicio tratamiento infliximab (USD)† 2.403 Costo administración intravenosa (USD) 149,2 Fuente Fondo Nacional de Recursos de Uruguay Fondo Nacional de Recursos de Uruguay Fondo Nacional de Recursos de Uruguay Fondo Nacional de Recursos de Uruguay Ministerio de Salud Pública de Uruguay HAQ, Health Assessment Questionnaire; PASI, Psoriasis Area Severity Index; PASI 50/75/90(þ), logra una reducción del 50/75/ 90% en el Psoriasis Area Severity Index respecto al valor basal; PASI 50/75/90(-), no logra una reducción del 50/75/90% en el Psoriasis Area Severity Index respecto al valor basal; PsARC (þ), logra una respuesta en articulaciones medida por el Psoriatic Arthritis Response Criteria; PsARC (-), no logra una respuesta en articulaciones medida por el Psoriatic Arthritis Response Criteria; USD, dólares de Estados Unidos con cotización de venta de 21,85 pesos uruguayos. * Costo promedio por ciclo trimestral. † Costo de inicio de tratamiento con infliximab. costos de servicios asistenciales y de medicamentos de síntesis fueron obtenidos del Instituto Nacional de Reumatología y de la Unidad Centralizada de Adquisiciones del Ministerio de Economía y Finanzas (UCA). Si bien los datos de la UCA son precios a los cuales la Administración de Servicios de Salud del Estado compra determinados servicios y productos, estos generalmente no difieren significativamente respecto a otras instituciones y respecto a los costos informados por el Instituto Nacional de Reumatología. Por otro lado, estos costos provienen del subsector público del Sistema de Salud el cual atiende a aproximadamente el 40% de los usuarios del país. De todas formas, en el análisis de sensibilidad fue evaluada la influencia de variaciones en estos costos respecto a los resultados de la evaluación económica. La determinación de la frecuencia de utilización de productos y servicios de salud por esta población de pacientes fue determina mediante entrevista a expertos. Estas entrevistas fueron estructuradas para determinar los costos que insume un paciente con respuesta PsARC, uno que no logra la respuesta PsARC, uno con respuesta PASI75 y por último un paciente que no logra una respuesta PASI75. Luego, cada bloque de costos se suma para obtener el costo de un paciente con una determinada combinación de respuesta en articulaciones y piel. Esto asume únicamente dos niveles de costos para respuesta en piel, si bien el modelo discrimina en más niveles de respuesta para el cálculo de las utilidades. Estos costos incluyen los análisis de laboratorio, estudios de imagenología, consultas médicas, servicio de fisiatría, atención en salud mental, asistente social, internación en cuidados normales, atención domiciliaria y el consumo general de medicamentos de síntesis. Finalmente los costos fueron ponderados por la frecuencia de utilización en el tiempo y por la proporción de pacientes que los utilizan, para cada estado de salud considerando ciclos trimestrales. 62 VALUE IN HEALTH REGIONAL ISSUES 5C (2014) 58–64 Los costos de los tres medicamentos biológicos fueron obtenidos de los precios de compra del Fondo Nacional de Recursos, el cual es una Institución pública encargada de administrar la cobertura de prestaciones en salud de alto costo para todo el Sistema de Salud. Los costos o precios de más de un año de antigüedad fueron actualizados por el índice de precios al consumo (IPC) para la división salud. Los precios obtenidos de la UCA se actualizaron mediante ajuste paramétrico, el cual contempla la variación del IPC y la cotización del dólar americano (USD). Análisis de sensibilidad Se realizó un análisis de sensibilidad probabilístico para el cual fueron definidas las distribuciones de cada parámetro del modelo. Estos parámetros fueron utilizados para correr 2.500 simulaciones de Monte Carlo, las cuales evaluaron el impacto en los resultados de las variaciones de los costos, la incertidumbre de las utilidades y de las probabilidades de transición. Se realizó un análisis de sensibilidad determinístico mediante diagrama de tornado donde se variaron los precios de los medicamentos biológicos, los costos asistenciales, las utilidades, la tasa de descuento y las probabilidades de transición. Validación Los resultados del presente modelo fueron comparados con los publicados por Rodgers et al. [12], dada la similitud en la estructura, la población definida y los datos de eficacia y seguridad. Sin embargo, debe considerarse que este modelo presenta algunas modificaciones respecto al de Rodgers et al., las cuales son explicadas en el diseño del modelo y en la discusión de este artículo. Además, la tasa de respuesta y la tasa de abandono de los tratamientos biológicos fueron comparadas con información no publicada del Instituto Nacional de Reumatología de Uruguay respecto a cohortes de pacientes que reciben estos tratamientos, para valorar la concordancia con la práctica clínica. Resultados En el caso base, etanercept resulta en una estrategia dominada dado que infliximab produce un mayor beneficio a un menor costo. Adalimumab resulta en una estrategia con dominancia extendida respecto al tratamiento soporte e infliximab. Infliximab es la estrategia más efectiva con una RCEI respecto al tratamiento soporte de USD 47.294 (Tabla 2). Fue evaluado el escenario en el cual todos los pacientes tratados con tratamiento soporte tienen una evolución de la escala HAQ de acuerdo a la evolución natural de la enfermedad, sin importar la respuesta en articulaciones. En este escenario etanercept sigue siendo una estrategia dominada, adalimumab continua presentando una dominancia extendida e infliximab sigue siendo la estrategia más costo-efectiva con una RCEI de USD 41.073. Fue explorado otro escenario posible, el cual asume un horizonte temporal de 10 años. Si bien este acortamiento del horizonte temporal reduce las utilidades y costos totales para las cuatro estrategias, los resultados se mantienen prácticamente iguales a los obtenidos para las RCEI con un horizonte temporal de 40 años. Los análisis de diagrama de tornado muestran que los precios de los medicamentos biológicos seguidos de las variaciones de utilidades en las respuestas al tratamiento, son los parámetros que más influyen en las RCEI. En el caso de no aplicarse una tasa de descuento anual a los beneficios y costos, la RCEI para infliximab respecto al tratamiento soporte pasa a ser de USD 42.609. Adicionalmente se realizó un análisis de sensibilidad probabilístico donde fue calculado para las tres estrategias la probabilidad de ser costo-efectiva respecto al tratamiento soporte, considerando tres niveles aceptables a pagar (del inglés willingness to pay), de uno, dos y tres veces el producto interno bruto (PIB) per cápita de Uruguay por AVAC adicional, correspondiendo a aproximadamente USD 14.451, 28.902 y 43.353 respectivamente para el año 2012 [18]. El análisis de sensibilidad considerando límites deseables a pagar de uno y tres veces el PBI per cápita, las probabilidades de que infliximab sea una estrategia costoefectiva son de 0% y 92,7% respectivamente. El beneficio monetario neto (NMB, del inglés net monetary benefit) resulta negativo para las tres estrategias considerando los tres niveles aceptables a pagar. Por ejemplo, para un límite aceptable a pagar de un PIB per cápita por AVAC adicional, los NMB resultan en USD -60.789, -32.843 y -73.478 para adalimumab, infliximab y etanercept respectivamente. Discusión y conclusiones Los resultados muestran que para el caso base o para cualquiera de los escenarios ensayados en esta evaluación, la inclusión de cualquiera de los tres anti-TNFs a la cobertura resulta en una RCEI por encima del límite propuesto por la Organización Mundial de la Salud para considerar una estrategia como altamente costo efectiva, desde la perspectiva del Sistema Nacional de Salud. Si bien infliximab es la estrategia más efectiva respecto a las otras alternativas evaluadas, el análisis de sensibilidad probabilístico muestra que la probabilidad de que la RCEI se encuentre por debajo de un PIB per cápita por AVAC adicional es de cero. Este valor límite ha sido generalmente considerado como una referencia a nivel internacional para definir una estrategia como altamente costo efectiva [19], lo que quizás se vincule con la riqueza promedio que produce una persona en un año en un determinado país. En este sentido, el Reino Unido ha establecido como límite aceptable a pagar £ 20.000 a 30.000 por AVAC adicional (aproximadamente entre USD 30.000 y 45.000) [20], lo que se aproxima en promedio al valor de un PIB per cápita (USD 38.510 en 2012) [18]. Sin embargo, es necesario que cada país considere diferentes límites aceptables a pagar según sus mecanismos propios de inclusión de nuevas tecnologías, otras tecnologías ya incluidas en la cobertura, entre otros factores. Relacionado a lo anterior, es importante considerar que en patologías donde hasta el momento no hay cubierto ningún tratamiento de alto costo, es esperable obtener resultados de RCEI altos debido a un incremento significativo del costo respecto al tratamiento soporte o estándar. En cambio, para patologías en las cuales ya se cubre al menos un medicamento de alto costo, la introducción de un nuevo medicamento de este tipo generalmente produce una RCEI significativamente menor respecto a igual AVAC incrementales. Esto puede generar un sesgo en la inclusión de este tipo de tecnologías a la cobertura, dado que ante dos estrategias con similares AVACs incrementales, es más Tabla 2 – Resultado de la evaluación costo-utilidad para el caso base. Estrategia Costo (USD) AVAC RCEITS (USD) Tratamiento soporte Adalimumab Infliximab Etanercept 67.696 109.556 113.145 137.424 5,072 5,629 6,033 5,865 0 Dominancia extendida 47.294 Dominancia absoluta AVAC, años de vida ajustados por calidad; RCEI, relación costoutilidad incremental; USD, dólares de Estados Unidos con cotización de venta de 21,85 pesos uruguayos por dólar. VALUE IN HEALTH REGIONAL ISSUES 5C (2014) 58–64 probable incluir la que trata una patología donde ya existe un tratamiento de alto costo cubierto. Los resultados obtenidos en esta evaluación concuerdan con lo publicado en el estudio de Rodgers et al. [12] respecto a que el tratamiento con infliximab es la estrategia más efectiva. Sin embargo, el presente modelo muestra valores menores en los años de vida ajustados por calidad totales para las cuatro estrategias respecto al modelo de Rodgers et al. Esto se debe a que nuestro modelo considera para el caso base, que cuando un paciente pierde la respuesta al tratamiento vuelve a presentar valores de HAQ que corresponden a la evolución natural de la enfermedad en tratamiento soporte en ese determinado momento, lo que difiere del modelo realizado por el grupo de la Universidad de York. Una explicación más amplia respecto a los diferentes criterios posibles sobre la evolución del HAQ puede encontrarse en la bibliografía [12]. La evaluación de Rodgers et al. [12] muestra que para el Sistema Nacional de Salud del Reino Unido etanercept es la estrategia más costo efectiva con una RCEI de £ 18.000, infliximab presenta una RCEI de £ 44.000 respecto a etanercept, y adalimumab resulta con dominancia extendida respecto a los cuidados paliativos y etanercept. Estos resultados difieren respecto al presente estudio muy probablemente debido a las diferencias en el precio de los medicamentos biológicos. Esto queda evidenciado en que el orden creciente del costo de cada estrategia difiere en ambos modelos. La discusión de los resultados de la evaluación económica con expertos clínicos sugiere que si bien infliximab es la estrategia con menor RCEI, la administración intravenosa de este medicamento puede generar limitaciones en el acceso al tratamiento en pacientes que acuden a determinados servicios de salud donde no hay disponible un hospital de día, o en regiones donde algunos pacientes puedan tener dificultades de asistir en forma regular a la administración del medicamento. Esta situación puede resultar en inequidades entre diferentes grupos de pacientes en el acceso al tratamiento. En el caso de adalimumab y etanercept, los mismos son medicamentos de administración subcutánea por lo que incluso pueden ser autoadministrados por el paciente, lo que confiere una ventaja significativa en cuanto generar un acceso más equitativo a los tratamientos. Otro aspecto importante a considerar, es que parte de los pacientes que se tratan con fármacos biológicos pueden presentar una respuesta en piel significativa sin lograr una respuesta en articulaciones. Es posible considerar que aproximadamente un 6,5% y un 18,3% de los pacientes pueden presentar una reducción del 75% o del 50% en la escala PASI respectivamente, sin tener respuesta en articulaciones (valores obtenidos del modelo). Según lo define la guía de práctica clínica de la European League Against Rheumatism [11], la ausencia persistente de respuesta a nivel de articulaciones es un motivo para indicar el abandono del tratamiento biológico. En pacientes que presentan lesiones en piel de relevancia clínica, con el consecuente impacto negativo en su calidad de vida, la indicación de abandonar el tratamiento por no lograr una respuesta en articulación según la escala PsARC pero habiendo experimentado un nivel de respuesta relevante en piel puede representar un dilema ético. Este aspecto no impacta significativamente en los resultados de la evaluación económica, sin embargo debería ser considerado al momento de definir los criterios de cobertura de un eventual tratamiento biológico. Otro aspecto a considerar es que esta evaluación no considera costos sociales. Existe cierto consenso a nivel internacional sobre la necesidad de incluir la perspectiva social en las evaluaciones económicas [21], sin embargo existen algunas limitantes que hacen que no sea la perspectiva más frecuentemente utilizada [22], como la inclusión de los costos de productividad en los modelos económicos [23]. Incluso, diferentes guías de evaluaciones económicas recomiendan por el momento utilizar la 63 perspectiva del financiador únicamente o como caso base [24– 26]. Sin embargo, es importante tomar en cuenta que para este tipos de patologías puede existir una subvaloración del impacto de estos tratamientos a nivel social cuando no se incluye esta perspectiva. Según nuestro conocimiento al momento de redactar este trabajo, esta es la primera publicación de un estudio costoutilidad realizada para apoyar la decisión de inclusión de medicamentos a la cobertura del Sistema de Salud de Uruguay. Agradecimientos Se agradece a los Doctores Luciane N. Cruz, Steffan F. Stella, André Ferreira, Rodrigo Ribeiro y Carisi Anne Polanczyk del Instituto de Avaliação de Tecnologia em Saúde (IATS) de Porto Alegre, Brasil; y a la Dra. Alejandra Croci y el Ing. Rafael Alonso del Ministerio de Salud Pública de Uruguay, por sus contribuciones a la evaluación económica. Source of financial support: This evaluation was supported by the Ministry of Public Health of Uruguay and the National Agency of Research and Innovation of Uruguay. The views expressed in this paper are those of the authors, and no official endorsement by the Ministry of Health of Uruguay is intended or should be inferred. Supplemental Materials Supplemental material accompanying this article can be found in the online version as a hyperlink at http://dx.doi.org/10.1016/j. vhri.2014.09.001 or, if a hard copy of article, at www.valuein healthjournal.com/issues (select volume, issue, and article). R EF E R EN C I A S [1] Krueger G, Koo J, Lebwohl M, et al. The impact of psoriasis on quality of life: results of a 1998 National Psoriasis Foundation patientmembership survey. Arch Dermatol 2001;137:280–4. [2] Finlay AY, Coles EC. The effect of severe psoriasis on the quality of life of 369 patients. Br J Dermatol 1995;132:236–44. [3] Kurd SK, Gelfand JM. The prevalence of previously diagnosed and undiagnosed psoriasis in US adults: results from NHANES 2003-2004. J Am Acad Dermatol 2009;60:218–24. [4] Shbeeb M, Uramoto KM, Gibson LE, et al. The epidemiology of psoriatic arthritis in Olmsted County, Minnesota, USA, 1982-1991. J Rheumatol 2000;27:1247–50. [5] Leonard DG, O'Duffy JD, Rogers RS. Prospective analysis of psoriatic arthritis in patients hospitalized for psoriasis. Mayo Clin Proc 1978;53:511–8. [6] Dubertret L, Mrowietz U, Ranki A, et al. European patient perspectives on the impact of psoriasis: the EUROPSO patient membership survey. Br J Dermatol 2006;155:729–36. [7] Taylor WJ. Impact of psoriatic arthritis on the patient: through the lens of the WHO International Classification of Functioning, Health, and Disability. Curr Rheumatol Rep 2012;14:369–74. [8] Gladman DD. Early psoriatic arthritis. Rheum Dis Clin North Am 2012;38:373–86. [9] Rapp SR, Feldman SR, Exum ML, et al. Psoriasis causes as much disability as other major medical diseases. J Am Acad Dermatol 1999;41:401–7. [10] Stern RS, Nijsten T, Feldman SR, et al. Psoriasis is common, carries a substantial burden even when not extensive, and is associated with widespread treatment dissatisfaction. J Investig Dermatol Symp Proc 2004;9:136–9. [11] Gossec L, Smolen JS, Gaujoux-Viala C, et al. European League Against Rheumatism recommendations for the management of psoriatic arthritis with pharmacological therapies. Ann Rheum Dis 2012;71:4–12. [12] Rodgers M, Epstein D, Bojke L, et al. Etanercept, infliximab and adalimumab for the treatment of psoriatic arthritis: a systematic review and economic evaluation. Health Technol Assess 2011;15:i-xxi, 1–329. 64 VALUE IN HEALTH REGIONAL ISSUES 5C (2014) 58–64 [13] Etanercept, infliximab and adalimumab for the treatment of psoriatic arthritis. NICE technology appraisal guidance 199, National Institute of Health and Care Excellence (NICE), 2010. [14] Instituto Nacional de Estadística. Disponible en: http://www.ine.gub. uy/socio-demograficos/vitales2008.asp. [Accessed October 31, 2014]. [15] Directrices para la evaluación económica de tecnologías en salud. Resolución N1 06/09, MERCOSUR, Montevideo, 2009. [16] Clegg DO, Reda DJ, Mejias E, et al. Comparison of sulfasalazine and placebo in the treatment of psoriatic arthritis. A Department of Veterans Affairs Cooperative Study. Arthritis Rheum 1996;39: 2013–2020. [17] Langley RG, Ellis CN. Evaluating psoriasis with Psoriasis Area and Severity Index, Psoriasis Global Assessment, and Lattice System Physician's Global Assessment. J Am Acad Dermatol 2004;51:563–9. [18] Banco Mundial. Disponible aquí: http://www.bancomundial.org/. [Accessed October 31, 2014]. [19] Organización Mundial de la Salud, Choosing Interventions that are Cost-Effective (CHOICE), Disponible en: http://www.who.int/choice. [Accessed October 31, 2014]. [20] Devlin N, Parkin D. Does NICE have a cost-effectiveness threshold and what other factors influence its decisions? A binary choice analysis. Health Econ 2004;13:437–52. [21] Jönsson B. Ten arguments for a societal perspective in the economic evaluation of medical innovations. Eur J Health Econ 2009;10:357–9. [22] Roy S, Madhavan SS. Making a case for employing a societal perspective in the evaluation of Medicaid prescription drug interventions. Pharmacoeconomics 2008;26:281–96. [23] Krol M, Brouwer W, Rutten F. Productivity costs in economic evaluations: past, present, future. Pharmacoeconomics 2013;31:537–49. [24] Guidelines for the economic evaluation of health technologies. [3rd Edition], Canadian Agency for Drugs and Technologies in Health, Ottawa, 2006. Disponible en: http://www.cadth.ca/. [Accessed October 31, 2014]. [25] Guidelines for the Economic Evaluation of Health Technologies in Ireland. Health Information and Quality Authority, Ireland, 2010. Disponible en: http://www.hiqa.ie/. [Accessed October 31, 2014]. [26] Choices in Methods for Economic Evaluation. Department of Economics and Public Health Assessment. Haute Autorité de Santé, France, 2012. Disponible en: www.has-sante.fr. [Accessed October 31, 2014]. VALUE IN HEALTH REGIONAL ISSUES 5C (2014) 65–72 Available online at www.sciencedirect.com journal homepage: www.elsevier.com/locate/vhri Cost-Effectiveness of Biologic Agents in the Treatment of Moderate-to-Severe Psoriasis: A Brazilian Public Health Service Perspective Bruno Salgado Riveros, MSc1, Patrícia Klarmann Ziegelmann, PhD2, Cassyano Januário Correr, PhD1,* 1 Pharmaceutical Sciences Department, Federal University of Parana, Curitiba, Brazil; 2Department of Statistics, Federal University of Rio Grande do Sul, Porto Alegre, Brazil AB STR A CT Background: Psoriasis is a chronic disease that affects public health and budget payers. In Brazil, biologic therapy for psoriasis is mostly provided by means of lawsuit with no strategy for efficient allocation of resources. Objective: This study aimed to identify which of the available biologic alternatives for psoriasis is the most efficient from the perspective of the Brazilian Public Health Service (SUS). Methods: Direct costs and efficacy were expressed in Brazilian currency (real [R$]; US $1 ¼ R$1.97) and Psoriasis Area Severity Index 75 (PASI75), respectively. The Markov model process included 12 cycles of 3 months each, comprising 3 years of horizon. Adalimumab (80 mg at week 0 followed by a maintenance dose of 40 mg at week 1 and then every other week), etanercept (50 mg twice weekly for 12 weeks followed by a maintenance dose of 25 mg weekly), infliximab (5 mg/ kg at weeks 0, 2, and 6 and then every 8 weeks), and ustekinumab (45 mg at weeks 0 and 4 and then every 12 weeks) were assessed. Oneway and horizon sensitivity analyses were performed. Moreover, probabilistic sensitivity analysis was applied to evaluate model robustness. The final result was interpreted as the cost for each patient who achieved and maintained PASI75 for at least 3 years. Results: Adalimumab was the most cost-effective biologic therapy (R$120,981.45/PASI75) for moderate-to-severe psoriasis, followed by ustekinumab (R$126,336.67/PASI75), etanercept (R$225,074.71/PASI75), and infliximab (R$377,656.28/PASI75). One-way sensitivity analysis determined that the acquisition cost of biologics was the most sensitive parameter of the model. Horizon analysis suggests that the result was the same when the horizon was varied from 1 year to a lifetime. Probabilistic sensitivity analysis showed that adalimumab has 80% to 10% probability of being the most cost-effective biologic considering a willingness-to-pay value ranging from R$50,000 to R$500,000, whereas ustekinumab presented a probability of 20% to 90% for the same range. Conclusions: From the pharmacoeconomics point of view, adalimumab 80 mg at week 0 followed by a maintenance dose of 40 mg at week 1 and then every other week should be the first-line therapy for patients with plaque psoriasis concomitant or not to psoriatic arthritis or nail psoriasis. This study does not have the potential to evaluate the impact of incorporating a specific biologic agent on the final budget. Its goal is to point out which of the technologies is the most efficient, that is, the one that adds more value to the financial resource invested. Keywords: biological agents, cost effectiveness, drug therapy, pharmacoeconomics, psoriasis. Introduction than 10% of body surface area (BSA) affected by the disease, 2) a score of 10 or more for the Dermatology Life Quality Index (DLQI), or 3) Psoriasis Area Severity Index (PASI) as patients with moderate-tosevere psoriasis. Some authors consider those with a PASI value of 20 as suffering from a clinically severe condition [7]. In cases of mild psoriasis, topic treatment is generally effective [14]. In cases of moderate-to-severe psoriasis, systemic treatment is based on phototherapy, methotrexate, acytretin, or cyclosporine. For patients who do not respond to any of these therapeutic options or develop adverse reactions, biologic agents are an option [15]. In the Brazilian Public Health System (SUS), the clinical protocol for psoriasis does not indicate the best approach regarding the use of biologics. One of the reasons for this might be that there is a lack of economic evaluations that consider the SUS Psoriasis is a chronic autoimmune disease that affects mainly the skin. Its prevalence around the world varies between 0.6% and 4.8% [1]. There are different phenotypes for this disease, with plaque psoriasis (or psoriasis vulgaris) being the most common and affecting 80% of all patients with such a clinical condition [2]. Concomitant phenotypes are possible, such as psoriatic arthritis and plaque psoriasis (40%), with or without nail psoriasis (35%–50%) [3]. Other morphological combinations are less common, but possible as well. Treatment is based on disease severity (mild, moderate, or severe). There is no consensus in the way to classify it, but most guidelines [4–13] suggest the “rule of 10” as an acceptable tool. The aforementioned clinical approach considers patients with 1) more Copyright & 2014, International Society for Pharmacoeconomics and Outcomes Research (ISPOR). Published by Elsevier Inc. Conflict of interest: The authors have indicated that they have no conflicts of interest with regard to the content of this article. * Address correspondence to: Cassyano Januário Correr, nº 632, Pref. Lothario Meissner Avenue, Curitiba 81540050, Brazil. E-mail: [email protected] 2212-1099$36.00 – see front matter Copyright & 2014, International Society for Pharmacoeconomics and Outcomes Research (ISPOR). Published by Elsevier Inc. http://dx.doi.org/10.1016/j.vhri.2014.09.002 66 VALUE IN HEALTH REGIONAL ISSUES 5C (2014) 65–72 perspective [7]. Moreover, biologics for the treatment of psoriasis remain unavailable in the SUS [16], making lawsuit the only way for patients to access such expensive treatment. Therefore, we aimed to identify the most cost-effective biologic agent for moderate-to-severe psoriasis according to the perspective of the SUS. Methods This is a cost-effectiveness analysis in which costs were expressed in real (R$, Brazilian currency) and efficacy in PASI75 response (PASI75). The exchange rate between real and US dollar was US$ 1 ¼ R$1.97 at the time of the study. This outcome corresponds to an improvement of 75% to 100% in the basal PASI score. Because the chosen outcome corresponds to efficacy and not effectiveness, it is important to highlight that data extracted from clinical trials were obtained in a controlled environment and not in a real-world scenario. The result was interpreted as the amount of money spent for a patient who achieve and maintain PASI75 for at least 3 years. The adopted perspective is that of the SUS. A Markov model process with 12 cycles of 3 months each was built to assess the scenario of patients with moderate-to-severe psoriasis, eligible for treatment based on biologics, following Brazilian Consensus of Psoriasis [6]. This pharmacoeconomic study is part of a broader project that involved systematic reviews of clinical efficacy and safety [17] and patient-related outcomes. Moreover, a mixed treatment comparison for these three outcomes and a benefit-risk multicriteria decision analysis were carried out. These studies are under review in scientific journals. Population Patients with moderate-to-severe psoriasis treated within the SUS who had an indication to start a biologic agent were our targeted population. Efficacy data of each biologic agent were obtained from the literature [18–22]. Thus, our results are applicable to patients with characteristics described in Table 1, which corresponds to the weighted average of the population evaluated in each clinical trial. Technologies Assessed The evaluated biologic agents were the ones approved by the National Health Surveillance Agency (ANVISA) for marketing up to the end of 2012, and selected dosages were the ones indicated by Brazilian Consensus of Psoriasis [6]. Thus, adalimumab (80 mg at week 0 followed by a maintenance dose of 40 mg at week 1 and then every other week), etanercept (50 mg twice weekly for 12 weeks followed by a maintenance dose of 25 mg weekly), infliximab (5 mg/kg at weeks 0, 2, and 6 and then every 8 weeks), and ustekinumab (45 mg at weeks 0 and 4 and then every 12 weeks) were assessed. Markov Model The proposed model, which consisted of four health states, was based on Woolacott et al. [23]: (Fig. 1): PASI75—patients who achieved an improvement of 75% to 100% in their basal PASI score. PASI50-75—patients who achieved an improvement of 50% to 75% in their basal PASI score. Failure—patients who did not achieve an improvement of 50% to 75% in their basal PASI score nor achieved better scores, patients who achieved an improvement of 50% to 75% in their basal PASI score but after 12 weeks did not improve their response to PASI75, or patients who developed an adverse event preventing the maintenance of biologic therapy. Death—includes all death cases regardless of cause. Each Markov cycle corresponds to 12 weeks, and the study time horizon was 3 years. Discounting of 5% [24] was applied following Brazilian statements. The outcome was assessed considering the number of patients with PASI75 health state at the end of the model. The first 12 weeks of treatment is not shown in the model. It was, however, represented in cycle 0, and costs were expressed as initial costs. Thus, _stage¼0 corresponds to a period between 12 and 24 weeks after treatment initiation. Because all models are a simplified way to understand a complex situation, all of them have assumptions [25]. The present model assumes the following: 1. After therapeutic failure with any biologic agent, patient did not use any other biologic. 2. Temporary interruptions of biologics were not considered in this model. 3. Patients who achieved PASI75 interrupted biologic therapy only if they a. got a clinical response worse than 50% of improvement or b. developed adverse reaction or any adverse event that increased the risks over the benefits. 4. Patients with an improvement of 50% to 75% in their basal PASI score for more than 12 weeks had their biologics interrupted. 5. Only the clinical efficacy of biologics was taken into account, regardless of association with topic or systemic drugs or phototherapy treatment. Probabilities Table 1 – Characteristics of the population with psoriasis from which data about efficacy were extracted. Characteristic Age (y) Men (%) Patients with PsA (%) Disease duration (y) PASI score DLQI score Mean ⫾ SD 44.8 67.8 29.7 19.5 19.6 11.7 ⫾ ⫾ ⫾ ⫾ ⫾ ⫾ 1.31 2.14 3.54 1.18 1.9 0.66 DLQI, Dermatology Life Quality Index; PASI, Psoriasis Area Severity Index; PsA, psoriatic arthritis. Data of PASI75 were extracted from the literature to serve as foundations for transition probabilities (Table 2). The selected randomized controlled trials (RCTs) were the ones that 1) assessed the same dose regimen as us, 2) showed low risk of bias by means of Cochrane Collaboration’s tool, 3) presented long-term results (at least 1 year of follow-up), and 4) had a number of participants weighing more than 500 lb. Probabilities related to the short-term treatment were retrieved from an network meta-analysis involving the four biologics assessed [18]. From the second year of treatment, efficacy data were extrapolated from the last known result. This assumption was based on literature findings [19,20,22]. Death probability was extracted from the Life Table published by the Brazilian Institute of Geography and Statistics (IBGE) [26]. 67 VALUE IN HEALTH REGIONAL ISSUES 5C (2014) 65–72 Fig. 1 – Markov model for patients with moderate-to-severe psoriasis treated with biologics. PASI, Psoriasis Area Severity Index. Considering the absence of data for patients treated with infliximab after 1 year, a linear regression based on available data was developed and the linear equation was predicted considering unknown responses. External validation and graphs are available in full detail in Appendix 1 in Supplemental Materials found at http://dx.doi.org/10.1016/j.vhri.2014.09.002. First, extracted data (epidemiological values) had to be calibrated, and then inserted in the model. The calibration method aimed to predict transition probabilities among health states to respect both epidemiological evidence and natural history of the disease. Probabilities applied to the model and other details are available in Appendix 2 in Supplemental Materials found at http://dx.doi.org/10.1016/j.vhri.2014.09.002. According to the International Society of Pharmacoeconomics and Outcomes Research, external validation is essential to establishing the credibility of any model [27]. Thus, we validated it considering the epidemiologic data regarding PASI75 response (Table 2) and the reproducibility of these results by the model. In other words, we sought to identify whether the model was capable of retrieving the same PASI75 response as shown in the RCT [19–22]. Costs Direct costs were assessed from the perspective of the SUS, including costs for biologics, conventional therapy, drug administration, laboratory and imaging tests, hospitalization, consultations, and adverse event management. First, the main costs were identified, then measured, and finally valued. Costs varied according to the technology involved, cycle, and health state. In Table 2 – PASI75 efficacy data extracted from the literature. Cycle Months _stage Drug Adalimumab Etanercept Infliximab Ustekinumab 0 1 2 0–3 3–6 0 6–9 1 0.58* 0.52* 0.8* 0.69* 0.67† 0.60‡ 0.82§ 0.76¶ 0.64† 0.60‡ 0.74§ 0.72¶ 3 9–12 2 0.6† 0.63‡ 0.61§ 0.64¶ Note. Italic values correspond to data obtained through linear regression. PASI, Psoriasis Area Severity Index. * Reich et al. [18]. † Gordon et al. [20]. ‡ Tyring et al. [22]. § Reich et al. [21]. ǁ Calculated by the author. ¶ Kimball et al. [19]. 4 5 6 12–15 3 15–18 4 18–21 5 0.59† 0.6‡ 0.51ǁ 0.61¶ 0.59† 0.55‡ 0.40ǁ 0.61¶ 0.57† 0.55‡ 0.30ǁ 0.61¶ 7 21–24 6 0.54† 0.52‡ 0.20ǁ 0.61¶ 68 VALUE IN HEALTH REGIONAL ISSUES 5C (2014) 65–72 Appendix 3 in Supplemental Materials found at http://dx.doi.org/ 10.1016/j.vhri.2014.09.002, all these data were thoroughly described. The sources used for valuing each cost component were the SIGTAP [28] (a database of procedures performed by the SUS) and the CMED [29] (the Regulation Chamber Drug Market in Brazil). For biologic agents, the costs were retrieved from Sistema Brasileiro de Informação sobre Medicamentos (Brazilian System of Informations about Medications), a database for drugs obtained through lawsuit. Adalimumab 40-mg vial costs R$ 1,792.56, whereas etanercept 25 and 50 mg, infliximab 100 mg, and ustekinumab 45 mg cost R$1093.22 and R$571.22, R$1896.32, and R$7527.15, respectively. In addition, for adverse events management and hospitalizations, we extrapolated and adjusted findings from cost-of-illness studies [30–32]. First, we identified how much a given health problem costs yearly. Then, we updated this value to the present days by means of Índice Nacional de Preços ao Consumidor Amplo (National Consumer Price Index Board) and we adjusted its annual cost for each Markov cycle. Results Table 3 presents base-case results of a cost-effectiveness model with a 3-year time horizon using direct costs in the SUS perspective. Adalimumab is the most cost-effective technology because each patient who achieves and maintains PASI75 for at least 3 years cost R$120,981.45. It is followed by ustekinumab (R $126,336.67/PASI75), etanercept (R$225,074.71/PASI75), and infliximab (R$377,656.28/PASI75). Our findings suggest that adalimumab is dominant over etanercept and infliximab and show an incremental cost-effectiveness ratio (ICER) of R$169,283.28/incremental PASI75 between adalimumab and ustekinumab. Considering the threshold for each incremental PASI75 response, the lowest paid value for each response (R$120,981.45/PASI75), the ICER between ustekinumab and adalimumab suggests adalimumab (80 mg at week 0 followed by a maintenance dose of 40 mg at week 1 and then every other week) as the most cost-effective technology. Efficacy Sensitivity Analysis Findings The outcome considered is PASI75. It is interpreted as an improvement of 75% to 100% in the basal score measured by the PASI questionnaire, which assesses the area, erythema, induration, and desquamation of psoriasis injuries around the body [33]. The model was built to use the percentage of patients at PASI75 health state at the end of 3 years as the measure of efficacy. Figure 2 shows a Tornado diagram expressed in ICER considering adalimumab as the common comparator. Dotted lines correspond to the ICER value in the base-case scenario. One-way sensitivity analysis pointed out the acquisition cost of biologics (c_ADA, c_ETA, c_INF, C_UST) as the most sensitive parameter of the model. None of the other variables was able to alter the final interpretation of base-case results. Important information regarding one-way sensitivity analysis is the identification of threshold values. Decreasing the acquisition cost of ustekinumab by 16.7% makes this technology dominant over all other drugs, whereas a reduction of 5.3% matches efficiency (cost-effectiveness ratio) with adalimumab. The acquisition cost for etanercept should decrease 50% to have the same efficiency as well. There were no scenarios in which infliximab was as efficient as adalimumab. The impact of the time horizon on the cost-effectiveness ratio was assessed varying the number of cycles of the model. Scenarios from 1 year up to the lifetime period were simulated. Figure 3 shows the final result up to 5 years. It is seen that costeffectiveness ratios for adalimumab and ustekinumab are very similar through the years, whereas etanercept and infliximab increase their ratio compared with other biologics. Analysis with a longer time horizon maintained that tendency. In the first 6 months of treatment, infliximab is the most effective treatment, with 82% of PASI75 response [21] (see Table 2). In this time horizon, adalimumab is the most cost-effective treatment (R$27,847.92/PASI75) followed by ustekinumab (R$30,559.99/ PASI75), infliximab (R$41,648.74/PASI75), and etanercept (R$63,572.69/PASI75). A reduction in the acquisition cost of ustekinumab, infliximab, and etanercept by 11%, 36%, and 67%, respectively, would change their cost-effectiveness ratio to that shown by adalimumab. Probabilistic sensitivity analysis was undertaken as recommended by the Brazilian Guideline for Economic Evaluations [24]. Sensitivity Analysis Following the statements provided by Brazilian Guideline for Economic Evaluations [24], one-way sensitivity analysis and probabilistic sensitivity analysis were performed. Moreover, the impact of time horizon was assessed in the final result. All the parameters of the model were submitted to one-way sensitivity analysis. The range used for each parameter corresponds to the 95% confidence intervals. Time-horizon analysis considered scenarios ranging from 1 year up to the lifetime period; thus, probabilities from 2 years were extrapolated to the lifetime period. In addition, probabilistic sensitivity analysis was performed considering all variables of the model. Costs were distributed using gamma distributions in which hyperparameters for biologic agents were calibrated so that lowest and highest prices registered in Sistema Brasileiro de Informação sobre Medicamentos corresponded to 0.95 probability interval. Probability parameters were distributed using beta (nodes with two branches) or Dirichlet (nodes with more than two branches) distributions. The hyperparameters for these distributions were calibrated considering the number of participants in each study by which PASI responses were extracted. A total of 10,000 iterations were applied. Hyperparameters of each distribution as well as the ranges used for one-way analysis are available in Appendix 4 in Supplemental Materials found at http://dx.doi.org/ 10.1016/j.vhri.2014.09.002. Table 3 – Results from 3 y of treatment with biologics. Biologic agent Cost (R$) Effectiveness (PASI75) Cost-effectiveness (R$/PASI75) Incremental cost-effectiveness ratio Adalimumab Infliximab Ustekinumab Etanercept 64,422.62 74,813.71 75,663.03 116,678.73 0.5325 0.1981 0.5989 0.5184 120,981.45 377,656.28 126,336.67 225,074.71 – Dominated 169,283.28 Dominated PASI, Psoriasis Area Severity Index. VALUE IN HEALTH REGIONAL ISSUES 5C (2014) 65–72 69 treatment being the most cost-effective (or efficient) alternative for a range of willingness-to-pay (WTP) values. On varying the WTP value from R$50,000 to R$500,000, the probability of adalimumab being the most cost-effective treatment varies from 80% to 10%, respectively. For the same WTP range, the probability of ustekinumab being the most cost-effective treatment varies from 20% to 90%, respectively. When the WTP is R$240,000, both biologics have 50% probability of being the most cost-effective treatment. For the assessed WTP range, etanercept and infliximab have probabilities close to 0%. The ICER between adalimumab and ustekinumab showed a mean of R$197,000/incremental PASI75, and the most likely values are between R$100,000 and R $200,000/incremental PASI75. Nevertheless, the aforementioned ICER shows a probability of 62% to be above the considered WTP. Discussion This is the first pharmacoeconomic analysis that assessed biologics in the treatment of moderate-to-severe psoriasis in the Brazilian Public Health Service (SUS). In this context, the following five issues deserve special attention, and, thus, discussion was carried out by considering the following. Why Not Cost-Utility Analysis? Until the present date, pharmacoeconomic studies assessing biologic agents in the long-term treatment of moderate-tosevere psoriasis have not considered PASI75 response as an outcome. This is the first study that considered PASI75 response as an outcome. Previous studies used cost-utility analysis [34–41] or cost-effectiveness for assessing short-term treatment [42,43]. Studies that have used Markov models were based on the one proposed by Woolacott et al. [23], which consists of health states defined by PASI response, in which a value of the DLQI is attributed to each health state and then changed by the EuroQol five-dimensional questionnaire (EQ-5D) utility through a linear equation. It is already known that psoriasis has a high-level impact on quality of life [44]. Thus, cost-utility analyses comparing biologic agents are of great relevance to analyze psoriasis treatment in a broader way. Norlin et al. [45], however, demonstrated that PASI response and the EQ-5D instrument have low correlation (r ¼ 0.25; P o 0.001), and, thus, suggest that defining utility values by means of PASI response seems to have a high level of bias. Moreover, the Markov model proposed by Woolacott et al. may not be feasible for cost-utility analysis assessed by the EQ-5D. Considering this new evidence, we chose to develop a cost-effectiveness analysis considering PASI75 as a clinical outcome because it is the tool used most often to assess psoriasis. Patients Benefited from This Study Fig. 2 – Tornado diagram expressed in terms of ICER impact. (A) Etanercept (ETA) vs. adalimumab (ADA). (B) Infliximab (INF) vs. adalimumab. (C) Ustekinumab (UST) vs. adalimumab. ICER, incremental cost-effectiveness ratio. A cost-effectiveness acceptability curve was obtained from the simulation whereby distributions were applied for all parameters in the model (Fig. 4). This graph shows the probability of a given People with moderate-to-severe psoriasis treating their disease in the SUS were the target population for our study. Because efficacy data were extrapolated by clinical trials up to now, their characteristics remained unknown. These results should ideally be applied to patients with clinical particularities given in Table 1. Thus, children with psoriasis are not covered by our results. Our team developed a psoriasis cost-of-illness model and determined the characteristics of patients with moderate-to-severe psoriasis using biologics [32]. The profile of these patients is different from that of patients included in RCTs, especially regarding age and time of disease. We understand, however, that these patients are representative of the real ones benefited when the SUS start dispensing such biologic drugs. For the moment, patients with psoriasis get access to these drugs by the SUS; they will start the treatment earlier, and so their profile will tend to be similar to that of patients included in RCTs. 70 Cost Effectiveness ratio (R$/PASI75) VALUE IN HEALTH REGIONAL ISSUES 5C (2014) 65–72 500000 450000 400000 350000 300000 250000 200000 150000 100000 50000 0 Adalimumab Etanercept Infliximab Ustekinumab 0.5 0.75 1 1.25 1.5 1.75 2 3 4 5 Fig. 3 – Time-horizon analysis from 1 to 5 years and its impact on the cost-effectiveness ratio. The horizontal line represents the time horizon expressed in years. Study Horizon Because psoriasis is a chronic disease, its treatment lasts for lifetime. So, the ideal scenario to evaluate it would be a model capable of predicting changes and interruptions during the treatment, as well as its effectiveness for many years. The main limitation to such a model is the lack of data to support long-term transition probabilities. The horizon defined in different studies varies. Knight et al. [37], Marcellusi et al. [38], and Lloyd et al. [41] defined 10 years, extrapolating the last known data through the next cycles up to the end of the study. Verma and Dharmarajan [46] defined 5 years, Villacorta et al. [40] 3 years, Ahn et al. [47] 1 year, and Ferrándiz et al. [42] 6 months. Our choice of 3 years was sustained by the available evidence [19,20,22] and its feasibility with the real world. In other words, it is frequently seen that patients treat psoriasis with biologic agents up to 3 years, but not 10 years, for instance. It is important to highlight that a 3-year time horizon is unable to capture costs, benefits, and adverse events that occurred after these 3 years of treatment. Our horizon analysis shows the same tendency for a very broad horizon range. This evidence provides more reliability for our results, though there are some biases around the results from 3 years to lifetime because probabilities were extrapolated to make this analysis possible. Sensitivity Analysis Although one-way sensitivity analysis pointed out the acquisition cost of biologics as a sensitive parameter for the model, we consider the final results robust. It can be explained by the nature of this critical parameter. The uncertainty around it can be known and measured because a decision maker knows exactly how much will each vial of a biologic agent cost. Other parameters such as discounting, hospitalization, and adverse event costs and shortand long-term probabilities were not able to change the final result. Important data provided by this analysis were the required costs for each biologic agent vial to change the found results (described in the Results section). These data can be useful for stakeholders when defining a specific price for a determined market. Regarding probabilistic sensitivity analysis, 10,000 iterations were performed and the interpretation was based on the cost-effectiveness acceptability curve. Because the perspective’s WTP is unknown, the cost-effectiveness acceptability curve provides a range of values, defined by us from R$50,000 to R$ 500,000. In scenarios with limited financial resources—WTP values up to R$240,000—adalimumab had the highest probability to be the most efficient. By considering greater values, ustekinumab is the best option. Considering the least costeffectiveness ratio (R$120,981.45/PASI75) as the WTP for each incremental patient who achieves and maintains PASI75 for at least 3 years, adalimumab (80 mg at week 0 followed by a maintenance dose of 40 mg at week 1 and then every other week) is the most likely biologic agent to be cost-effective. Moreover, probabilistic sensitivity analysis showed that the ICER between adalimumab and ustekinumab is more likely to be above the proposed WTP value. First-Line Therapy Definition Choosing a biologic agent to be used as first-line therapy for psoriasis is not easy. After all, many variables such as Fig. 4 – Cost-effectiveness acceptability curve. Willingness to pay for each incremental case of PASI75 varies between R$50,000 and R$500,000. VALUE IN HEALTH REGIONAL ISSUES 5C (2014) 65–72 comorbidities and severity affect this decision [48,49]. The most sensible analysis is dividing the population with psoriasis requiring biologics into subgroups. Although this study was based on patients with moderate to severe psoriasis, the population profile included in clinical trials [18–22] allowed us to infer about some subgroups. Patients with severe psoriasis, defined by some authors as those with a PASI value of more than 20, benefit from treatment based on infliximab (5 mg/kg at weeks 0, 2, and 6 and then every 8 weeks) because of its quick response [50]. Findings from this study show that its cost-effectiveness ratio is close to that of the best option, adalimumab (80 mg at week 0 followed by a maintenance dose of 40 mg at week 1 and then every other week) for more than 6 months of treatment. In this subgroup, we recommend starting infliximab 5 mg/kg and switching to adalimumab (80 mg at week 0 followed by a maintenance dose of 40 mg at week 1 and then every other week) as soon as PASI75 is achieved. Besides the pharmacoeconomic aspect, we believe in this modification therapy because secondary failure (failure after therapeutic response) with infliximab has the potential to impair the efficacy of other anti–TNF-α agents [48], such as adalimumab and etanercept. Psoriatic arthritis is presented in 6% to 10% of the patients with psoriasis and in 40% in those who have severe disease [3]. For patients who have plaque psoriasis concomitant to joint involvement, we suggest adalimumab (80 mg at week 0 followed by a maintenance dose of 40 mg at week 1 and then every other week) as first-line therapy because it is the most efficient treatment among anti–TNF-α agents when the outcome is PASI score. Atteno et al. [51] assessed the efficacy of anti–TNF-α agents for psoriatic arthritis by means of the American College of Rheumatology 20 outcome. It was found that 72% of the patients treated with etanercept (50 mg twice weekly for 12 weeks followed by a maintenance dose of 25 mg weekly) achieved an improvement of 20%, measured by the American College of Rheumatology 20 outcome, compared with 70% of the patients treated with adalimumab (80 mg at week 0 followed by a maintenance dose of 40 mg at week 1 and then every other week) and 75% of the patients treated with infliximab (5 mg/kg at weeks 0, 2, and 6 and then every 8 weeks) [51]. Considering these slight differences among the results, we believe that adalimumab (80 mg at week 0 followed by a maintenance dose of 40 mg at week 1 and then every other week) has the best cost-effectiveness ratio when plaque psoriasis is associated with psoriatic arthritis. Ustekinumab is contraindicated for these patients because this biologic (anti–IL-12/23) does not have satisfactory efficacy in psoriatic arthritis [52,53]. Nail psoriasis is rarely presented alone but in combination with some other phenotypes (50% of the cases) [5]. In cases in which plaque psoriasis is diagnosed with or without nail psoriasis, and there is indication to start using a biologic agent, we suggest adalimumab (80 mg at week 0 followed by a maintenance dose of 40 mg at week 1 and then every other week) as first-line therapy. Patients with hepatitis B or C are discouraged to use anti–TNFα agents because of the risk of the disease being reactivated [54]. Thus, from the clinical and pharmacoeconomic points of view, ustekinumab (45 mg at weeks 0 and 4 and then every 12 weeks) should be the first-line therapy. German [4] and British [8] guidelines suggest that etanercept may not be involved in hepatitis C reactivation. We recommend, however, this biologic as secondline therapy because of our pharmacoeconomic findings. When antinuclear factor or antibodies are positive, it is not recommended to start treatment based on anti–TNF-α agents because there is evidence of the use of such biologic agents leading to drug-induced lupus erythematosus [55]. So, the biologic chosen in these patients is ustekinumab (45 mg at weeks 0 and 4 and then every 12 weeks). 71 Psoriasis can assume many phenotypes, such as plaque (or vulgar), nail, erythrodermic, guttate, inverse, and pustular psoriasis. For most of the patients, more than one phenotype can occur concomitantly, including joint involvement called psoriatic arthritis [2]. PASI75 response is the main outcome for psoriasis because it is the criterion standard used to evaluate plaque psoriasis, which is the most common phenotype affecting 80% [2] of the patients. This outcome, however, cannot properly evaluate all the phenotypes. Although in clinical trials there were individuals with other phenotypes, we recommend these results for those with plaque psoriasis, with or without concomitant psoriatic arthritis and nail psoriasis, in which the main objective of the treatment is to control the disease in the skin. It is important to highlight that for some cases, such as pustular psoriasis, anti–TNF-α agents are contraindicated [8]. Conclusions From the pharmacoeconomic point of view, in Brazilian SUS, adalimumab (80 mg at week 0 followed by a maintenance dose of 40 mg at week 1 and then every other week) should be the firstline therapy among the biologics used for the treatment of patients with moderate to severe psoriasis. We recommend adalimumab (80 mg at week 0 followed by a maintenance dose of 40 mg at week 1 and then every other week) for patients with plaque psoriasis with or without concomitant psoriatic arthritis, or nail psoriasis, whereas those with contraindication to anti– TNF-α agents should be treated with ustekinumab (45 mg at weeks 0 and 4 and then every 12 weeks). This study does not have the potential to evaluate the impact of incorporating a specific biologic agent on the budget. Its goal was to point out which of the assessed technologies is the most efficient, that is, the one that adds more value to the financial resources invested. Supplementary Materials Supplemental material accompanying this article can be found in the online version as a hyperlink at http://dx.doi.org/10.1016/j. vhri.2014.09.002 or, if a hard copy of article, at www.valuein healthjournal.com/issues (select volume, issue, and article). R EF E R EN C ES [1] Kurd SK, Richardson SK, Gelfand JM. Update on the epidemiology and systemic treatment of psoriasis. Expert Rev Clin Immunol 2007;3: 171–85. [2] Soung J, Lebwohl M. Clinical presentation. In: Gordon KB, Ruderman EM, eds., Psoriasis and Psoriatic Arthritis: An Integrated Approach (1st ed.). Speyer, Germany: Springer, 2005. [3] Villaseñor-Park J, Wheeler D, Grandinetti L. Psoriasis: Evolving treatment for a complex disease. Cleve Clin J Med 2012;79:413–23. [4] Nast A, Boehncke W-H, Mrowietz U, et al. S3 - Guidelines on the treatment of psoriasis vulgaris (English version): update. J Dtsch Dermatol Ges 2012;10(Suppl. 2):S1–95. [5] National Institute for Health and Care Excellence. CG153 Psoriasis: full guideline. 2012. Available from: http://guidance.nice.org.uk/CG153/ Guidance/pdf/English. [Accessed August 10, 2012]. [6] Sociedade Brasileira de Dermatologia. Consenso Brasileiro de Psoríase. 2012. [7] DAE/SAS/MS. Consulta Pública No 09 de 28 de agosto de 2012. Brasilia, DF: Ministério da Saúde, 2012. [8] Smith CH, Anstey AV, Barker JNWN, et al. British Association of Dermatologists’ guidelines for biologic interventions for psoriasis 2009. Br J Dermatol 2009;161:987–1019. 72 VALUE IN HEALTH REGIONAL ISSUES 5C (2014) 65–72 [9] SIGN. Diagnosis and management of psoriasis and psoriatic arthritis in adults. Scottish Intercollegiate Guidelines Network. Available from: http://sign.ac.uk/guidelines/fulltext/121/index.html. [Accessed November 10, 2014]. [10] Puig L, Carrascosa JM, Daudén E, et al. Spanish evidence-based guidelines on the treatment of moderate-to-severe psoriasis with biologic agents [in Spanish]. Actas Dermosifiliogr 2009;100:386–413. [11] Papp K, Gulliver W, Lynde C, et al. Canadian guidelines for the management of plaque psoriasis: overview. J Cutan Med Surg 2009;15:210–9. [12] Menter A, Gottlieb A, Feldman SR, et al. Guidelines of care for the management of psoriasis and psoriatic arthritis, section 1: overview of psoriasis and guidelines of care for the treatment of psoriasis with biologics. J Am Acad Dermatol 2008;58:826–50. [13] Hsu S, Papp KA, Lebwohl MG, et al. Consensus guidelines for the management of plaque psoriasis. Arch Dermatol 2012;148:95–102. [14] Martins G. Tratamento Tópico da Psoríase. Consenso Brasileiro de Psoríase. (1st ed.). Rio de Janeiro, Brazil: Sociedade Brasileira de Dermatologia, 2009. [15] Azulay-Abulafia L, Felix O. Imunobiológicos na Psoríase. Consenso Brasileiro de Psoríase. (1st ed.). Rio de Janeiro, Brazil: Sociedade Brasileira de Dermatologia, 2009. [16] Ministério da Saúde. Relatório de Recomendaçao da CONITEC: Medicamentos Biológicos para o tratamento da Psoríase moderada a grave em adultos. Brasilia, DF: Ministério da Saúde, 2012. [17] Correr CJ, Rotta I, Teles T de S, et al. Efficacy and safety of biologics in the treatment of moderate to severe psoriasis: a comprehensive metaanalysis of randomized controlled trials. Cad Saude Publica. Escola Nacional de Saúde Pública. Fundação Oswaldo Cruz 2013;29(Suppl): s17–31. [18] Reich K, Burden AD, Eaton JN, Hawkins NS. Efficacy of biologics in the treatment of moderate to severe psoriasis: a network meta-analysis of randomized controlled trials. Br J Dermatol 2012;166:179–88. [19] Kimball AB, Gordon KB, Fakharzadeh S, et al. Long-term efficacy of ustekinumab in patients with moderate-to-severe psoriasis: results from the PHOENIX 1 trial through up to 3 years. Br J Dermatol 2012;166:861–72. [20] Gordon K, Papp K, Poulin Y, et al. Long-term efficacy and safety of adalimumab in patients with moderate to severe psoriasis treated continuously over 3 years: results from an open-label extension study for patients from REVEAL. J Am Acad Dermatol 2012;66:241–51. [21] Reich K, Nestle FO, Papp K, et al. Infliximab induction and maintenance therapy for moderate-to-severe psoriasis: a phase III, multicentre, double-blind trial. Lancet 2005;366:1367–74. [22] Tyring S, Gordon KB, Poulin Y, et al. Long-term safety and efficacy of 50 mg of etanercept twice weekly in patients with psoriasis. Arch Dermatol 2007;143:719–26. [23] Woolacott N, Hawkins N, Mason A, et al. Etanercept and efalizumab for the treatment of psoriasis: a systematic review. Health Technol Assess 2006;10:1–233, i–iv. [24] Brasil. Diretrizes Metodológicas: estudos de avaliação econômica de tecnologias em saúde (1st ed.). Brasília, DF: Ministério da Saúde, 2009. [25] Philips Z, Bojke L, Sculpher M, et al. Good practice guidelines for decision-analytic modelling in health technology assessment: a review and consolidation of quality assessment. Pharmacoeconomics 2006;24:355–71. [26] IBGE. Tábua da Vida. Rio de Janeiro, RJ, Brazil: Instituto Brasileiro de Geografia e Estatística, 2010. [27] Eddy DM, Hollingworth W, Caro JJ, et al. Model transparency and validation: a report of the ISPOR-SMDM Modeling Good Research Practices Task Force–7. Value Health 15:843–850. [28] SIGTAP - Sistema de Gerenciamento da Tabela de Procedimentos. Medicamentos e OPM do SUS. Available from: http://sigtap.datasus.gov. br/tabela-unificada/app/sec/inicio.jsp. [Accessed August 9, 2012]. [29] CMED. Lista de Preços e Medicamentos - Preços Fábrica e Preço Máximo ao Consumidor. 2012. Available from: http://portal.anvisa.gov.br/wps/ wcm/connect/5dd6198044cfe3389465fd60074bb479/LISTA +CONFORMIDADE_2014-07-21.pdf?MOD=AJPERES. [Accessed November 10, 2004]. [30] Costa JG, Santos AC, Rodrigues LC, Barreto ML. Tuberculose em Salvador: custos para o sistema de saúde e para as famílias Tuberculosis in Salvador, Brazil: costs to health system and families. Rev Saude Publica 2005;39:1–7. [31] Araujo DV, Tavares LR, Veríssimo R, et al. Custo da Insuficiência Cardíaca no Sistema Único de Saúde. Arq Bras Cardiol 2005;84:422–7. [32] Garcia MM. Custo da Doença e perfil dos pacientes dos pacientes com psoríase moderada a grave atendidos no Hospital de Clínicas da Universidade Federal do Paraná. Curitiba, Brazil: Universidade Federal do Paraná 2013. [33] Fredriksson T, Pettersson U. Severe psoriasis–oral therapy with a new retinoid. Dermatologica 1978;157:238–44. [34] Colombo GL, Di Matteo S, Peris K, et al. A cost-utility analysis of etanercept for the treatment of moderate-to-severe psoriasis in Italy. Clinicoecon Outcomes Res 2009;1:53–9. [35] De Portu S, Del Giglio M, Altomare G, et al. Cost-effectiveness analysis of TNF-alpha blockers for the treatment of chronic plaque psoriasis in the perspective of the Italian health-care system. Dermatol Ther 2010;23(Suppl. 1):S7–13. [36] Heinen-Kammerer T, Daniel D, Stratmann L, et al. Cost-effectiveness of psoriasis therapy with etanercept in Germany. J Dtsch Dermatol Ges 2007;5:762–8. [37] Knight C, Mauskopf J, Ekelund M, et al. Cost-effectiveness of treatment with etanercept for psoriasis in Sweden. Eur J Health Econ 2012;13:145–56. [38] Marcellusi A, Gitto L, Giannantoni P, et al. Modeling the costeffectiveness of ustekinumab for moderate to severe plaque psoriasis in US. Value Health 2012;15:A570. [39] Pan F, Brazier NC, Shear NH, et al. Cost utility analysis based on a headto-head phase 3 trial comparing ustekinumab and etanercept in patients with moderate-to-severe plaque psoriasis: a Canadian perspective. Value Health 2011;14:652–6. [40] Villacorta R, Hay JW, Messali A. Cost effectiveness of moderate to severe psoriasis therapy with etanercept and ustekinumab in the United States. Pharmacoeconomics 2013;31:823–39. [41] Lloyd A, Reeves P, Conway P, et al. Economic evaluation of etanercept in the management of chronic plaque psoriasis. Br J Dermatol 2009;160:380–6. [42] Ferrándiz C, García A, Blasco AJ, Lázaro P. Cost-efficacy of adalimumab, etanercept, infliximab and ustekinumab for moderate-to-severe plaque psoriasis. J Eur Acad Dermatol Venereol 2012;26:768–77. [43] Greiner R-A, Braathen LR. Cost-effectiveness of biologics for moderateto-severe psoriasis from the perspective of the Swiss healthcare system. Eur J Dermatol 2009;19:494–9. [44] Langley RGB, Krueger GG, Griffiths CEM. Psoriasis: epidemiology, clinical features, and quality of life. Ann Rheum Dis 2005;64(Suppl. 2): ii18–23; discussion ii24–5. [45] Norlin JM, Steen K, Persson U, Schmitt-Egenolf M. Analysis of three outcome measures in moderate to severe psoriasis: a registry-based study of 2450 patients. Br J Dermatol 2012;166:797–802. [46] Verma S, Dharmarajan SYY. Modeling the cost-effectiveness of ustekinumab for moderate to severe plaque psoriasis in US. Value Health 2010;13:A147. [47] Ahn CS, Gustafson CJ, Sandoval LF, et al. Cost effectiveness of biologic therapies for plaque psoriasis. Am J Clin Dermatol 2013;14:315–26. [48] Goodfield M. Sequential use of biologics in moderate-to-severe plaque psoriasis: case studies illustrating an individualized approach. Br J Dermatol 2012;167(Suppl.):21–6. [49] Leman J, Burden AD. Sequential use of biologics in the treatment of moderate-to-severe plaque psoriasis. Br J Dermatol 2012;167 (Suppl.):12–20. [50] National Institute for Health and Care Excellence. TA134 Psoriasis Infliximab: Guidance. London, UK: National Institute for Health and Care Excellence, 2008. [51] Atteno M, Peluso R, Costa L, et al. Comparison of effectiveness and safety of infliximab, etanercept, and adalimumab in psoriatic arthritis patients who experienced an inadequate response to previous diseasemodifying antirheumatic drugs. Clin Rheumatol 2010;29:399–403. [52] Gan EY, Chong W-S, Tey HL. Therapeutic strategies in psoriasis patients with psoriatic arthritis: focus on new agents. BioDrugs 2013;27:359–73. [53] McInnes IB, Kavanaugh A, Gottlieb AB, et al. Efficacy and safety of ustekinumab in patients with active psoriatic arthritis: 1 year results of the phase 3, multicentre, double-blind, placebo-controlled PSUMMIT 1 trial. Lancet 2013;382:780–9. [54] Rustin MHA. Long-term safety of biologics in the treatment of moderate-to-severe plaque psoriasis: review of current data. Br J Dermatol 2012;167(Suppl.):3–11. [55] Marzano AV, Vezzoli P, Crosti C. Drug-induced lupus: an update on its dermatologic aspects. Lupus 2009;18:935–40. VALUE IN HEALTH REGIONAL ISSUES 5C (2014) 1–6 Available online at www.sciencedirect.com journal homepage: www.elsevier.com/locate/vhri La Compleja Relación Entre Posición Socioeconómica, Estatus Migratorio y Resultados de Salud Báltica Cabieses, PhD1,2,* 1 Facultad de Medicina, Universidad del Desarrollo Clínica Alemana, Santiago, Chile; 2Department of Health Sciences, University of York, York, UK AB STR A CT Introduction: The relationship between socioeconomic position (SEP), migration and health is dynamic, difficult to predict, multifactorial and poorly studied in the Latin American region. Moreover, there are high levels of uncertainty about reasons, types, and consequences of migration to an individual, family, community, country of origin and the receiving country. Objective: To discuss the evidence about the complex relationship between SEP, migration and health outcomes. Results: According to international evidence available, migration has a direct impact on health outcomes in a population and, in turn, on public health policy decisions in each locality. The available evidence on this issue affects multiple parts of the social sciences. The "healthy migrant" effect is not consistently observed among immigrant populations, particularly after adjusting for PSE. Moreover, the immigrant population tends to assimilate in terms of risk factors and morbidity to the local population after about 10 years of stay in the foreign country. Migration has consequences for international relations, economic productive capacity of a country, inequality, demographic changes and health outcomes, to mention a few. Conclusion: There is now the opportunity to generate better and more evidence longitudinal population-based around the relationship between SEP, migration status and health. This will contribute to reduce uncertainty about the health status of immigrants that is required for decision- making in public health in Chile and the region. Keywords: Chile, health outcomes, Latin America, social determinants of health, socioeconomic position. Introducción mediada por dimensiones como clase social, género y etnia, ocupación, nivel educacional e ingreso. Reducir las inequidades en salud ha sido un eje prioritario en el trabajo desarrollado por la OMS y en múltiples países, incluido Chile [10]. Este país ha liderado la inclusión explicita de este objetivo en la reforma de salud y ha sido uno de los primeros en Latinoamérica en atender su relevancia para la salud de la población [11]. Otros ejemplos han sido los dos últimos documentos de Objetivos Sanitarios, de las décadas 2001-2010 y 20112020, junto a otros reportes ministeriales y académicos que dan cuenta del amplio cuerpo de evidencia que existe a nivel mundial y el creciente conocimiento que se ha desarrollado en Chile en esta temática [12]. El presente artículo de opinión y análisis tiene como propósito dar a conocer la compleja relación entre posición socioeconómica, migración y resultados de salud. Esta relación es dinámica, difícil de predecir, multifactorial y pobremente estudiada en la región de Latinoamérica. Más aún, existen altos grados de incertidumbre en torno a razones, formas, y consecuencias de migrar en un individuo, su familia, su comunidad y país de origen y el país que lo recibe. Esto tiene directas y profundas consecuencias en los resultados de salud de una población, y a su vez, en las decisiones políticas de salud pública de cada localidad. La Los determinantes sociales se definen como las condiciones sociales en las cuales las personas desarrollan su vida y, por distintos mecanismos, tienen un impacto en su salud [1,2]. El concepto de inequidad en salud se refiere a todas aquellas diferencias en salud que son evitables e injustas [3,4]. Gran parte de las inequidades en salud entre distintos grupos de la población son explicables por estos determinantes sociales [5–7]. Por lo tanto, por concepto las inequidades son posibles de revertir, para lo cual es necesario contar con políticas públicas que logren reducir la fuerte asociación que existe entre condiciones sociales de vida y la salud en la población. La Comisión de Determinantes Sociales de la Salud (DSS) de la Organización Mundial de la Salud ha desarrollado un marco conceptual, el cual identifica y categoriza los diversos DSS identificando de qué forma interactúan y generan las inequidades en salud [8,9]. Se distinguen dos categorías: DSS estructurales y DSS intermedios. El DSS intermediario por su parte, se refiere a la interacción entre salud y factores psicosociales, biológicos y de condiciones materiales de vida. También incluye los efectos del sistema de salud. Los DSS estructurales incluyen el nivel macro / contextual (político y cultural) y la posición socioeconómica (PSE), & 2014 Published by Elsevier Inc. on behalf of International Society for Pharmacoeconomics and Outcomes Research (ISPOR). Conflicts of interest: The author has indicated that she has no conflicts of interest with regard to the content of this article. Corresponding Author: Báltica Cabieses, Facultad de Medicina, Universidad del Desarrollo, Avenida Las Condes 12587 Lo Barnechea Santiago zip code 7590943 Chile. Tel/Fax: þ56 2 23279516. E-mail: [email protected]; [email protected]. 2212-1099/$36.00 – see front matter & 2014 Published by Elsevier Inc. on behalf of International Society for Pharmacoeconomics and Outcomes Research (ISPOR). http://dx.doi.org/10.1016/j.vhri.2013.11.001 2 VALUE IN HEALTH REGIONAL ISSUES 5C (2014) 1–6 evidencia disponible en este tema afecta a múltiples partes: tomadores de decisión en políticas de salud, educación, grupos vulnerables y derechos humanos, entre otros; e interesa a profesionales académicos, investigadores, clínicos y teóricos de cualquier disciplina de las ciencias sociales. La migración tiene consecuencias en relaciones internacionales, capacidad productiva económica de un país, desigualdad, variaciones demográficas, por mencionar algunos aspectos. Salud es, para efectos de este manuscrito, el reflejo perfecto de la complejidad inherente al proceso de migración y su necesario desarrollo en Chile y Latinoamérica. A continuación se desarrolla una discusión crítica de esta evidencia para luego realzar, de acuerdo a la opinión del autor, la urgencia de desarrollar mayor investigación en esta área en la región. Lo que ya sabemos en cuanto a la relación entre posición socioeconómica y salud: la brecha, la gradiente y el ciclo vital Estudios internacionales han explorado los posibles mecanismos explicativos que están a la base de las diferencias en los resultados de salud según la posición socioeconómica de los individuos [13,14]. La PSE es una variable ampliamente utilizada en investigación en salud [15]. Es un concepto multidimensional y no existe un gold standard para su medición. Se reconocen tres indicadores clásicos: ingreso, educación y ocupación. Otras variables estrechamente relacionadas a posición social son etnicidad, estatus migratorio, género y clase social. La existencia de diferencias prevenibles e injustas en salud según PSE es una realidad indiscutible [2,16,17]. Existen grandes diferencias en las tasas de morbimortalidad en los distintos grupos sociales, tanto en Chile y Latinoamérica como en la amplia literatura disponible en este tema [18–21]. En Chile por ejemplo, entre 1998 y 2006 la esperanza de vida a los 20 años aumentó en 1,5 años, pero este incremento no fue homogéneo según nivel educacional. Respecto de las inequidades en morbimortalidad según PSE, existen dos fenómenos de relevancia. Primero, la brecha en la morbimortalidad entre aquellos en mejor y peor posición social en el país, que genera por ejemplo un exceso de muertes en la población infantil y adulta chilena cada año. Estos resultados son consistentes en diversos problemas de salud en el país, independiente del indicador de PSE que se utilice. De esta forma, la reducción de la brecha en salud según posición social es urgente en Chile, pues se mantiene estable e incluso se amplía a lo largo del tiempo. El segundo fenómeno corresponde a la existencia de la fina gradiente social en salud, esto es que no solo los menos educados tienen peores resultados de salud y el resto de la población está igualmente protegida [22]. Por el contrario, existe una asociación continua o escalonada entre PSE y morbimortalidad [23]. Esta fina gradiente ha sido observada de manera consistente en la literatura y sugiere establecer estrategias que mejoren la situación de salud de manera proporcional a cada grupo afectado [24–27]. En otras palabras, la población en su conjunto se ve deteriorada cuando presenta desigualdades sociales e inequidades en salud y no sólo aquellos más vulnerables en el estrato más bajo de PSE. Es importante además observar la relación entre posición social y salud desde una perspectiva de ciclo vital [28,29]. La perspectiva de curso de vida, desarrollada en especial por medio del estudio de grandes cohortes poblacionales en países más desarrollados, indica que la prevención de factores de riesgo y problemas de salud en la adultez exige modificar sus desencadenantes desde la vida temprana [30]. No es accidental que la población adulta chilena en desventaja sea la más propensa a enfermar, ni tampoco lo es que sus hijos lo serán en 20 años más [31,32]. La probabilidad de enfermar de cualquier chileno en la adultez está, al menos parcialmente, pre-determinada por la posición social de su familia de origen [33–35]. Importante evidencia demuestra el peso de la posición social sobre la salud poblacional en el curso de vida, cuya asociación es consistente a través de contextos históricos, geografía, poblaciones y diseños de investigación. Más aún, existe un creciente grupo de investigadores en el mundo interesados en develar la relación intergeneracional de estos procesos de salud y enfermedad. La cuna de origen determina desde el día de nuestra concepción, y con cierto grado de incertidumbre, nuestra esperanza de vida, enfermedades más probables de desarrollar y como consecuencia nuestro bienestar. Dichos patrones tienden a repetirse a lo largo de las generaciones, en especial en sociedades altamente jerárquicas, con gran desigualdad socioeconómica, y con poca movilidad social (esta última generalmente asociada a pobres estrategias políticas de redistribución del ingreso, de seguro social y de apoyo a la educación en los grupos menos favorecidos). En una era globalizada y marcada por la búsqueda de la igualdad y la equidad, la falta de oportunidades balanceadas para que cada individuo en su sociedad florezca en todas sus potencialidades y lleve una vida saludable feliz, es uno de los mayores fracasos y desafíos pendientes que tenemos en salud pública y políticas públicas en todos los países del mundo. El fenómeno migratorio en Latinoamérica y en Chile Existen diversos grupos que por sus características particulares de vulnerabilidad, presentan importantes inequidades en los resultados en salud, y que deben ser considerados para un abordaje especial. Uno de ellos corresponde a personas inmigrantes. La inmigración es un reconocido determinante social [36]. Las condiciones que rodean al proceso de migración hacen que esta población sea vulnerable y que su movimiento transfronterizo tenga gran impacto en salud pública [37–40]. A nivel mundial, se estima que 200 millones de personas migran cada año [41,42]. En América Latina y el Caribe, a unos 25 millones de personas (alrededor del 4% de la población total) habían emigrado a otro país en el 2011 [43]. En general, los EE.UU. es el destino preferido de los migrantes de América Latina y el Caribe, siendo la búsqueda de oportunidades de trabajo una de las principales razones de estos movimientos. También hay una creciente migración dentro de la región latinoamericana, la llamada migración Sur-Sur, sobre todo el movimiento de personas que viven en países de menor desarrollo hacia países cercanos más desarrollados dentro de la región [44]. Chile es un país de mediano ingreso con un Producto Interno Bruto (PIB) per cápita algo superior a $16.000 (USD). Cuenta con una población de poco más de 16 millones de habitantes y en las últimas décadas ha experimentado grandes cambios económicos y demográficos, una mejora progresiva de las condiciones de salud de la población, la disminución de la mortalidad infantil y de mortalidad general y el aumento de la esperanza de vida [45]. Hoy en día, el estado de salud de la población chilena es muy similar a algunos países de ingreso alto y mejor que muchos otros países de América Latina y el Caribe [46–48]. Hay múltiples razones para el estado de salud relativamente bueno de la población chilena en comparación con otros países de similar desarrollo económico. Desde principios del siglo XX Chile ha desarrollado importantes iniciativas de salud pública para el mejoramiento de la salud de su población, en primer lugar se centró en la mortalidad materno-infantil y las epidemias infecciosas y, más recientemente, en las enfermedades crónicas y el cáncer. Estos esfuerzos sistemáticos en materia de salud pública y protección social han dado importantes frutos en salud. Sin embargo, no todos los grupos socioeconómicos se han beneficiado de estos avances en la misma medida [49]. Existen profundas VALUE IN HEALTH REGIONAL ISSUES 5C (2014) 1–6 diferencias en el estado de salud de la población chilena al comparar por PSE, previsión, región, sexo, edad y otros factores [18,19]. En contraste con otros países de la región de América Latina, como Argentina o Brasil, Chile es un país fundamentalmente emisor de migrantes en lugar de ser un país receptor de inmigrantes. El 2007 por ejemplo, más de 857 781 chilenos vivían fuera del país, con una proporción de 3 chilenos fuera del país por 1 inmigrante viviendo en Chile (razón emigrante: inmigrante de 3: 1) [50,51]. Sin embargo, la inmigración en Chile ha aumentado en los últimos años, llegando a alrededor del 2% de la población total en el año 2007 (total estimado de 258 a 350.000 personas). Esta corresponde a la tasa de inmigrantes más alta observada desde 1950 en Chile. Durante las dos últimas décadas, América del Sur y otros países asiáticos han aumentado su tasa de inmigración regional. Las últimas cifras gubernamentales indican que en la actualidad Chile está experimentando un “nuevo patrón migratorio” [43]. Este nuevo patrón consiste en una importante mayoría de inmigrantes provenientes de otros países de América Latina, en busca de oportunidades laborales. También ha ido aumentando la inmigración femenina en nuestra región, entre ellos Chile, para trabajar en los servicios manuales y domésticos. Nueva evidencia acerca de la salud de inmigrantes: construyendo puentes de evidencia científica para el bienestar de la población latinoamericana en Chile y el mundo La evidencia internacional indica que existen diversos factores que afectan de manera directa e indirecta la salud de la población inmigrante y de sus descendientes, tales como: las razones que determinaron la migración; condiciones de vida, familiares y de salud previas a migrar y durante la migración; la llegada a la nueva sociedad; experiencias de discriminación y estigma; desconocimiento del funcionamiento de los diversos sistemas sociales; soledad y abandono, y pobres condiciones laborales, entre otros [52–55]. Por otra parte, las barreras percibidas de acceso a la salud por parte de inmigrantes incluyen dificultades de lenguaje, diferencias culturales en la apreciación de la salud y enfermedad, experiencias previas de discriminación y maltrato, y fundamentalmente, desconocimiento y/o desconfianza del sistema de salud del país en que residen. Al mismo tiempo, inmigrantes latinoamericanos en otros continentes como Europa presentan elevados porcentajes de ausencia de previsión de salud, optando al “pago de bolsillo” en caso de necesidad [56–58]. Existen importantes logros alcanzados en Chile respecto de la población inmigrante. Cualquier residente independiente de su situación legal, incluso indocumentada, puede acceder a los beneficios de la provisión de salud que señala la Ley de atención universal en Chile, en casos de emergencias, del control prenatal y de salud infantil. En este sentido, destacan el Programa para la inmigrante embarazada, el Programa para el inmigrante menor de edad, y el Programa de atención gratuita para inmigrantes peruanos viviendo en pobreza [50,51]. Al mismo tiempo, existen acuerdos internacionales especiales, como por ejemplo con el Perú, que facilitan la adaptación de cualquier inmigrante de ese país a Chile (ej. Acuerdo de seguridad social para inmigrantes peruanos en Chile). Por otra parte, cualquier inmigrante documentado (esto es, con al menos un miembro del hogar con contrato laboral y pasaporte al día) puede inscribirse en el consultorio correspondiente a su sector comunal y formar parte del sistema público de salud [50,51]. No obstante, cerca de un 20% de quienes reportaron ser inmigrantes en la encuesta nacional de caracterización socioeconómica del 2006 (CASEN) no tienen previsión, lo cual duplica el porcentaje de la población nacida en Chile [56–58]. En Chile residen aproximadamente 290.000 personas nacidas en el extranjero, equivalente al 2% de la población 3 aproximadamente [50,51]. Sus principales países de origen son latinoamericanos, y aunque un tercio de ellos han residido en el país por menos de un año, el promedio de residencia en Chile alcanza los 11 años. Recientes estudios de base poblacional en Chile destacan la compleja situación socioeconómica de los inmigrantes en Chile. Se realza en particular la relativa mejor situación socioeconómica de la población inmigrante en su conjunto en comparación a la nativa chilena, aun cuando presentan importantes variaciones de PSE [53,54,59]. Esto es, el quintil de ingreso más alto (de acuerdo a ingreso total autónomo del hogar auto-reportado en encuestas poblacionales) reporta un ingreso promedio más alto que el grupo equivalente de la población nacida en Chile. No obstante, el quintil de ingreso más pobre del grupo inmigrante recibe menor ingreso promedio que el mismo quintil de la población nacida en Chile. Si esto se traduce a la medición de la brecha en el ingreso (la conocida razón 20:20, esto es la comparación entre el promedio de ingreso del quintil más rico y el quintil más pobre), el año 2006 la población chilena tenía una brecha de 12 a 13 veces en el ingreso promedio de los más ricos versus los más pobres, mientras que la población inmigrante presentaba una brecha de 23 veces de diferencia entre los mismos grupos de ingreso [55,56,60]. Por otra parte y de manera similar a lo observado en la literatura internacional, la población inmigrante en Chile muestra el efecto del “migrante sano”, vale decir que este grupo presenta una menor prevalencia de problemas crónicos de salud que la población chilena nativa [54]. No obstante, existe una clara gradiente según PSE del inmigrante, lo que explica que los inmigrantes de baja PSE no presentan el efecto de “migrante sano”. Más aun, inmigrantes viviendo en baja PSE en Chile, aun siendo considerablemente más jóvenes que la población local, presentan tasas de morbilidad similar a la población nacida en Chile. Esto muestra ser consistente independiente del problema de salud observado (agudo o crónico) pero especialmente importante en inmigrantes que llevan largo tiempo en Chile (más de una década). La información sobre la salud de los inmigrantes en Chile y Latinoamérica es escasa, especialmente en los indocumentados [60]. Hoy se desconocen importantes factores relacionados a la salud, como razones de migración, PSE y de salud antes de migrar, efectos en segundas y terceras generaciones de inmigrantes, y variaciones en estatus contractual y legal en el tiempo. Esto cobra especial relevancia al considerar la heterogeneidad demográfica y económica de este grupo [59]. Es necesario desarrollar más estudios cuantitativos y cualitativos focalizados a la población inmigrante, idealmente de manera longitudinal y de representación nacional. Sólo de esta manera se podrá orientar de manera efectiva a estrategias políticas que nos ayuden a todos a proteger la salud de nuestra población, independiente de su edad, género, raza o estatus migratorio. Aporte de la reciente investigación en salud para la toma de decisiones políticas en inmigrantes en Chile Pese al intenso trabajo de diversos investigadores, organizaciones gubernamentales e instituciones internacionales en torno a las características socio-demográficas de los inmigrantes en Chile y la región, es aún muy poco lo que sabemos acerca de la vida y bienestar de los inmigrantes de Latinoamérica. Sabemos de su situación en un instante del tiempo por medio de grandes encuestas poblacionales (como la CASEN desde el año 2006 que incluyó una pregunta sobre estatus migratorio en Chile), conocemos aspectos generales del movimiento migratorio en la región por medio de registros de gobiernos, y pobremente nos acercamos a sus condiciones de salud y necesidades a través de registros de los servicios de salud de cada país. 4 VALUE IN HEALTH REGIONAL ISSUES 5C (2014) 1–6 No obstante, aún quedan preguntas sustanciales por responder, tales como (1) ¿cuáles son las características sociodemográficas y de salud de los inmigrantes antes de su arribo al nuevo país?; (2) ¿qué factores afectan su decisión de inmigrar?; (3) ¿cuál es la historia migratoria de cada inmigrante que llega a Chile, cuáles son sus expectativas en el nuevo país y bajo qué circunstancias volvería a migrar?; (4) ¿cómo varía su estatus migratorio en el tiempo (documentado o indocumentado, por ejemplo) y cómo esto afecta sus oportunidades laborales, bienestar familiar y salud?; (5) ¿qué factores determinan su percepción de integración a la nueva sociedad y cómo se podría facilitar esta experiencia de manera efectiva?; y (6) ¿de qué manera los hijos de inmigrantes experimentan su vida en su país y cómo el estatus migratorio de sus padres podría afectar sus oportunidades de florecer como ser humano, su salud y su bienestar a lo largo de su vida? Se requieren nuevas y más investigaciones que permitan aproximarse al fenómeno migratorio en Latinoamérica y Chile. Tal como fue mencionado anteriormente, estas iniciativas deberían estar idealmente orientadas a movilizar políticas desde la autoridad sanitaria involucrando a todos los sectores y con un enfoque de largo plazo. Ver resumen de desafíos a considerar en esta temática en la Figura 1. Conclusión En la opinión del autor y conforme a la evidencia disponible en Chile y Latinoamérica sobre salud de inmigrantes, se enfrenta hoy la necesidad de modificar los desencadenantes de problemas de salud, no solo en la población adulta actual sino en la futura. En este sentido, es urgente reconocer que una política que posicione y priorice la reducción de desigualdades en la PSE (reducción de las diferencias en años de educación e ingreso económico entre chilenos) es esencial para modificar la salud y conductas individuales a lo largo del tiempo. Un foco antipobreza no sería necesariamente suficiente si se considera la importancia de la gradiente socioeconómica en salud, ampliamente investigada en el extranjero. La incorporación de este concepto debe movilizar políticas desde la autoridad sanitaria, pero también debe involucrar al mundo académico, al sector privado y al intersector, a través de un ejercicio serio, participativo y con un enfoque de planificación a largo plazo. Es el país en su conjunto quien debe transformarse para lograr un verdadero cambio en la salud de toda su población incluyendo a aquellos que han inmigrado a nuestro país. Existe hoy la oportunidad como sociedad no sólo de generar mejor y más evidencia en torno a la relación entre PSE, estatus migratorio y salud, sino también de transformar nuestras formas de entender y acercarnos a fenómenos tan complejos como este. Es necesario mantener una mirada abierta y altamente crítica hacia la forma en la cual generamos nuestras preguntas de investigación y producimos nueva evidencia en Chile y en la región de Latinoamérica. Evidencia longitudinal y mixta es urgente para lograr representar de mejor manera las trayectorias de vida de los inmigrantes y sus descendientes en Chile y la región. Así también podremos representar con menor grado de incertidumbre las necesidades de salud de inmigrantes y se podrá planificar de mejor manera en términos de recursos de protección social y de salud. Solamente así lograremos desarrollar marcos conceptuales y teóricos que verdaderamente nos permitan observar de una manera más genuina, los complejos y múltiples determinantes de la salud de la población inmigrante en cualquier lugar del mundo y en especial en Chile. 1. La posición socioeconómica (PSE) es un constructo social complejo y multidimensional. Mediciones únicas con indicadores clásicos (por ejemplo ingreso) pueden no reflejar la totalidad de este constructo lo que aumenta el riesgo de endogenicidad y su débil interpretabilidad para la toma de decisiones en salud 2. La migración tiene distintas definiciones y teorías. Existe además un moderado grado de evidencia acerca de sus consecuencias en salud en Latinoamérica y Chile. Debemos avanzar hacia diseños longitudinales que den cuenta de variaciones en PSE y salud en inmigrantes en Chile y la región 3. Prácticamente no se sabe de la salud de quienes viven indocumentados pues este grupo es muy difícil de reclutar en estudios de base poblacional. Estrategias de reclutamiento específicas para ellos son urgentes en encuestas poblacionales y otros diseños de investigación que produzcan mejor nivel de evidencia científica 4. La poca evidencia disponible en Chile acerca de PSE, migración y salud no es aun adecuadamente utilizada para la generación de políticas de salud específicas para este grupo. Investigadores en la región deben dar cuenta de la evidencia existente en torno a este tema y buscar mecanismos efectivos de transferencia de este conocimiento a tomadores de decisiones Fig. 1 – Algunos desafíos pendientes en la reducción de la incertidumbre en torno a la evidencia disponible acerca de posición socioeconómica, migración y salud. VALUE IN HEALTH REGIONAL ISSUES 5C (2014) 1–6 Source of financial support: The authors have no other financial relationships to disclose. R EF E R EN CI AS [1] Tarlov AR. The coming influence of a social sciences perspective on medical education. Acad Med 1992;67:724–31. [2] Wilkinson RG. Social class differences in infant mortality. BMJ 1992;305:1227–8. [3] Phillips K, Muller-Clemm W, Ysselstein M, Sachs J. Evaluating health inequity interventions: applying a contextual (external) validity framework to programs funded by the Canadian Health Services Research Foundation. Eval Program Plann 2013;36:198–203. [4] Asada Y. A framework for measuring health inequity. J Epidemiol Commun H. [Research Support, N.I.H., Extramural Research Support, Non-U.S. Gov't Research Support, U.S. Gov't, P.H.S.]. 2005 Aug;59:700–5. [5] Marmot M. Policy overview. Closing the gap in a generation. Perspect Public Health 2009;129:16–7. [6] Marmot M. Fair Society, Healthy Lives. London, UK: The Marmot Review, 2010. [7] Marmot M, Friel S, Bell R, Houweling TA, Taylor S. Closing the gap in a generation: health equity through action on the social determinants of health. Lancet 2008;372:1661–9. [8] Irwin A, Valentine N, Brown C, et al. The commission on social determinants of health: tackling the social roots of health inequities. PLoS Med 2006;3:e106. [9] Solar O, Irwin A. Social determinants, political contexts and civil society action: a historical perspective on the Commission on Social Determinants of Health. Health Promot J Austr 2006;17:180–5. [10] Sutherland WJ, Bellingan L, Bellingham JR, Blackstock JJ, Bloomfield RM, Bravo M, et al. A collaboratively-derived science-policy research agenda. Plos One. [Research Support, Non-U.S. Gov't] 2012;7:e31824. [11] Jadue L, Marin F. Determinantes sociales de la salud en Chile. En la perspectiva de la equidad. Available. First ed. Santiago, Chile: Universidad del Desarrollo, MINSAL, 2005. [12] Ministry of Health in Chile. Elige vivir sano: objetivos sanitarios de la decada 2011-2020. Santiago, Chile: Ministry of Health, 2011. [13] Beckman A, Merlo J, Lynch J, et al. Country of birth, socioeconomic position, and healthcare expenditure: a multilevel analysis of Malmo, Sweden. J Epidemiol Community Health 2004;58:145–9. [14] Galobardes B, Lynch J, Smith GD. Measuring socioeconomic position in health research. Br Med Bull 2007;81–82:21–37. [15] Cabieses B, Zitko P, Pinedo R, et al. How has social status been measured in health research? A review of the international literature. Rev Panam Salud Publica 2011;29:457–68. [16] Wilkinson G. Social class and mental disorders. Brit J Hosp Med 1990;43:321. [17] Breeze E, Jones DA, Wilkinson P, et al. Area deprivation, social class, and quality of life among people aged 75 years and over in Britain. Int J Epidemiol 2005;34:276–83. [18] Arteaga O, Astorga I, Pinto AM. [Inequalities in public health care provision in Chile]. Cad Saude Publica 2002;18:1053–66. [19] Arteaga O, Thollaug S, Nogueira AC, Darras C. Information for health equity in Chile. Rev Panam Salud Publica 2002;11:374–85. [20] Mardones F. Inequality of health care for the elderly in Chile. Rev Med Chile 2004;132:865–72. [21] Melo PZ, Valdes BC. Socioeconomic determinants of disability in Chile. Disabil Health J 2011;4:271–82. [22] Hammarstrom A, Stenlund H, Janlert U. Mechanisms for the social gradient in health: results from a 14-year follow-up of the Northern Swedish Cohort. Pub Health 2011;125:567–76. [23] Theodossiou I, Zangelidis A. The social gradient in health: the effect of absolute income and subjective social status assessment on the individual's health in Europe. Econ Human Bio 2009;7:229–37. [24] Sabbah W, Watt RG, Sheiham A, Tsakos G. Effects of allostatic load on the social gradient in ischaemic heart disease and periodontal disease: evidence from the Third National Health and Nutrition Examination Survey. J Epidemiol Commun H 2008;62:415–20. [25] Arendt JN, Lauridsen J. Do risk factors explain more of the social gradient in self-reported health when adjusting for baseline health? Euro J Pub Health 2008;18:131–7. [26] Vallgarda S. Social inequality in health: dichotomy or gradient? A comparative study of problematizations in national public health programmes. Health Pol 2008;85:71–82. [27] Icks A, Moebus S, Feuersenger A, et al. Widening of a social gradient in obesity risk? German national health surveys 1990 and 1998. Eur J Epidemiol 2007;22:685–90. [28] Nicolau B, Marcenes W. How will a life course framework be used to tackle wider social determinants of health? Community Dent Oral Epidemiol 2012;40(Suppl. 2):33–8. 5 [29] Bond J. Pathways to optimal health: a life course framework for adolescents. Am J Pub Health 2012;102:1817. [30] Halfon N, Hochstein M. Life course health development: an integrated framework for developing health, policy, and research. Milbank Quarterly 2002;80:433–79, iii. [31] Bourne PA. A theoretical framework of the good health status of Jamaicans: using econometric analysis to model good health status over the life course. N Am J Med Sci 2009;1:86–95. [32] Misra DP, Grason H. Achieving safe motherhood: applying a life course and multiple determinants perinatal health framework in public health. J Womens Health Issues 2006;16:159–75. [33] Mishra GD, Anderson D, Schoenaker DA, et al. InterLACE: a new international collaboration for a life course approach to women's reproductive health and chronic disease events. Maturitas 2013. http:// dx.doi.org/http//dx.doi.org/10.1016/j.maturitas.2012.12.011. [Epub ahead of print]. [34] Jackson JS, Knight KM, Rafferty JA. Race and unhealthy behaviors: chronic stress, the HPA axis, and physical and mental health disparities over the life course. Am J Public Health year;100:933–9. [35] Barondess JA. Toward reducing the prevalence of chronic disease: a life course perspective on health preservation. Perspect Biol Med 2008;51:616–28. [36] Davies AA, Mosca D, Frattini C. Migration and health service delivery. World Hosp Health Serv 2010;46:5–7. [37] Akhavan S, Bildt C. Health in relation to unemployment and sick leave among immigrants in Sweden from a gender perspective. Ethnic Health 2004;9(Suppl.):S93–4. [38] Alver K, Meyer HE, Falch JA, Sogaard AJ. Bone mineral density in ethnic Norwegians and Pakistani immigrants living in Oslo–The Oslo Health Study. Osteoporos Int 2005;16:623–30. [39] Azerkan F, Zendehdel K, Tillgren P, et al. Risk of cervical cancer among immigrants by age at immigration and follow-up time in Sweden, from 1968 to 2004. Int J Cancer 2008;123:2664–70. [40] Beiki O, Allebeck P, Nordqvist T, Moradi T. Cervical, endometrial and ovarian cancers among immigrants in Sweden: importance of age at migration and duration of residence. Eur J Cancer 2009;45: 107–18. [41] United Nations. World Migration Managing Migration - Challenges and Responses for People on the Move 2003. Available from: http://www.un. org. [Accessed March 17, 2014]. [42] United Nations. Trends in total migrant stock: the 2005 revision. Department of Economic and Social Affairs. Washington DC2006. Available from: http://www.un.org/esa/population/publications/. [Accessed March 17, 2014]. [43] International Organization for Migration. World Migration report 2011. Communicating effectively about migration World Migration report [serial on the Internet]. 2011. Available from: http://publications. iom.int/bookstore/free/WMR2011_English.pdf. [Accessed March 17, 2014]. [44] Alvarado I., Sanchez H. Migration in the Latin America and the Caribbean. A view from the ICFTU/ORIT [serial on the Internet]. 2002. Available from: http://www.ilo.org/public/english/dialogue/actrav/ publ/129/19.pdf. [Accessed March 17, 2014]. [45] Albala C, Vio F. Epidemiological transition in Latin America: The case of Chile. Public Health 1995;109:431–42. [46] Atena E, editor. Chile: Sistema de Salud en Transición a la Democracia. Santiago: J. Jiménez org; 1991. [47] Bastias G, Pantoja T, Leisewitz T, Zarate V. Health care reform in Chile. CMAJ 2008;179:1289–92. [48] Bitran R, Escobar L, Gassibe P. After Chile's health reform: increase in coverage and access, decline in hospitalization and death rates. Health Aff (Millwood) 2010;29:2161–70. [49] Burrows J. Inequalities and healthcare reform in Chile: equity of what? J Med Ethics 2008;34:e13. [50] Departamento de Extranjeria y Migracion. Desarrollo del fenómeno de las migraciones en Chile. Evolución de la gestión gubernamental desde 1990. Ministerio del Interior. Available from: http://www.extranjeria. gov.cl/filesapp/migraciones.pdf. [Accessed March 17, 2014]. [51] Departamento de Extranjeria y Migracion. Preguntas frecuentes acerca de la inmigración en Chile. Secretaria y Admistracion General, Sistema de Gestion de la Calidad [serial on the Internet]. 2010. Available from: http://www.extranjeria.gob.cl/filesapp/FAQ%20ISO.pdf. [Accessed March 17, 2014]. [52] Bodenmann P, Vaucher P, Wolff H, et al. Screening for latent tuberculosis infection among undocumented immigrants in Swiss healthcare centres; a descriptive exploratory study. Bmc Infect Dis 2009;9:34. [53] Cabieses B, Tunstall H. Access to, need and use of healthcare by international immigrants in Chile. In: F. GJaT, editor. Migration, Health and Inequalities. First ed. London, UK: Zed Press, 2013. [54] Cabieses B, Tunstall H, Pickett KE. The health of international immigrants in Chile: what is beyond the apparent healthy migrant effect? Rev Med Chile 2013; (in press). 6 VALUE IN HEALTH REGIONAL ISSUES 5C (2014) 1–6 [55] Cabieses B, Pickett KE, Tunstall H. Comparing sociodemographic factors associated with disability between immigrants and the Chileanborn: are there different stories to tell? Int J Environ Res Public Health 2012;9:4403–32. [56] Cabieses B, Tunstall H. Socioeconomic vulnerability and its association with access to health care among international immigrants in Chile. In: J Gideon FT, editor. Migration, health and inequality. London: Zed; 2012. [57] Cabieses B, Tunstall H, Pickett K. Understanding the socioeconomic status of international immigrants in Chile through hierarchical cluster analysis: a population-based study. International Migration 2012; in press. [58] Cabieses B, Tunstall H, Pickett KE, Gideon J. Understanding differences in access and use of healthcare between international immigrants to Chile and the Chilean-born: a repeated cross-sectional populationbased study in Chile. Int J Equity Health 2012;11:68. [59] Cabieses B, Tunstall H, Pickett K, Gideon J. Changing patterns of migration in Latin America: how can research develop intelligence for public health? Pan American Journal of Public Health 2013; (under review). [60] Cabieses B, Pickett KE, Tunstall H. What are the living conditions and health status of those who don't report their migration status? A population-based study in Chile. Bmc Public Health 2012;12: 1013. VALUE IN HEALTH REGIONAL ISSUES 5C (2014) 7–13 Available online at www.sciencedirect.com journal homepage: www.elsevier.com/locate/vhri Guidance Document: Global Pharmacoeconomic Model Adaption Strategies C. Daniel Mullins, PhD1,*, Nneka C. Onwudiwe, PharmD, PhD, MBA1, Gabriela Tannus Branco de Araújo, MSc, MBA2, Wen Chen, PhD3, Jianwei Xuan, PhD4, Aleš Tichopád, PhD5, Shanlian Hu, MD, MSc6 1 Department of Pharmaceutical Health Services Research, School of Pharmacy, University of Maryland, Baltimore, MD, USA; 2Axia. Bio Consulting, São Paulo, Brazil; 3Center for Pharmacoeconomics Evaluation and Research, School of Public Health, Fudan University, Shanghai, China; 4Medical Development Group, Emerging Markets, Pfizer, New York, NY, USA Inc.; 5CEEOR – Central and Eastern European Outcomes Research, Prague, Czech Republic; 6Shanghai Health Development and Research Center, School of Public Health, Fudan University, Shanghai, China AB STR A CT Objective: The purpose of this guidance was to assist in the adaptation of pharmacoeconomic models originally developed in one country and intended for use in another. The intent was to produce user-friendly recommendations and a checklist for adapting a global model to treat a specific disease state. This guidance will allow model developers to tailor existing models so that they are “locally applicable,” while maintaining the scientific integrity of the original pharmacoeconomic model and will benefit formulary decision makers and other stakeholders involved in evaluating pharmacoeconomic studies. Methods: A working group of experts from various countries participated in the Global Pharmacoeconomic Model Guidance development to discuss the adaptation of pharmacoeconomic models. A systematic review of studies adapting pharmacoeconomic models and translation across countries was conducted and recommendations were made for adaptation. The working group interviewed internal and external stakeholders to solicit best practices for model adaptation and developed a draft set of key principles and general recommendations for global adaptation. Results: The working group provided a set of 16 recommendations for adapting pharmacoeconomic models for local decision makers. The recommendations span various aspects of estimating or modeling both the costs and effectiveness of pharmacoeconomic models as well as guidance for ensuring local acceptability. Conclusions: These recommendations and the related principles not only will provide pharmacoeconomic models that are meaningful to local decision makers but also will improve the consistency and credibility of pharmacoeconomic model adaptations. The guidance may also help those who will build the original models to design them with the flexibility to allow pharmacoeconomic model adaptations as described in this document. Keywords: cost-effectiveness analysis, health technology assessment, pharmacoeconomic model. Introduction Canada, and Australia because of the applicability of the description, as well as the safety and effectiveness of the technology [3]. The adaptation of a pharmacoeconomic model across different countries to support region-specific economic evaluation of pharmaceuticals requires the originally developed model to structurally adapt to the economic and clinical characteristics of the intended country. Ensuring the reliability (i.e., reproducibility) of measurements across different geographical regions requires comparing and/or adjusting data from clinical trials, observational studies, claims databases, case registries, public health statistics, and surveys to estimate the economic impact of the uptake and use of a particular pharmaceutical in the intended country of interest. The concept of “pharmacoeconomic model adaptation” raises the issue of “transferability” across geographical regions. The transferability of pharmacoeconomic models refers to the adaptation of Economic modeling is widely used in economic evaluation of pharmaceuticals (cost-effectiveness analysis, cost-utility analysis, cost-benefit analysis, and budget impact analysis) to evaluate the health care costs and health outcomes of alternative courses of action in the presence of scarce resources in terms of both their cost and consequences. A number of countries faced with increasing pressure to make use of health care resources use economic evaluations to guide their reimbursement of pharmaceuticals [1,2]. For example, Latin America and Caribbean stakeholders need to adapt existing pharmacoeconomic models for the local region. They need to consider coverage and reimbursement, as well as clinical decision making. These stakeholders prefer to adapt health technology assessment reports from Europe, the United States, Copyright & 2014, International Society for Pharmacoeconomics and Outcomes Research (ISPOR). Published by Elsevier Inc. Conflict of interest: J. Xuan was an employee of Pfizer. All other authors received a consulting fee from Pfizer for participating in this project. * Address correspondence to: C. Daniel Mullins, Pharmaceutical Health Services Research Department, University of Maryland School of Pharmacy, 220 Arch Street, 12th Floor, Baltimore, MD 21201. E-mail: [email protected] 2212-1099$36.00 – see front matter Copyright & 2014, International Society for Pharmacoeconomics and Outcomes Research (ISPOR). Published by Elsevier Inc. http://dx.doi.org/10.1016/j.vhri.2014.02.006 8 VALUE IN HEALTH REGIONAL ISSUES 5C (2014) 7–13 clinical effectiveness and cost-effectiveness data across geographical regions [4]. The transferability of economic evaluation results requires the use of a general “knockout criteria” to determine whether the model can be transferred to the decision country [2]. To ensure the reliability of the pharmacoeconomic model, the analyst must then determine which part(s) of the model needs to be adapted to reproduce the model in a different geographic region. To determine which parts of the model need adaptation, there are several transferability factors to consider. The factors that create challenges for developing a model for adaptation include differences in the epidemiology of the disease, mortality rates, disease severity, demographic characteristics, risk factors, available treatment options, discount rates, absolute or relative prices, and differences in practice patterns [5–8]. These factors can be broadly grouped into methodological, health care system, and population characteristics [7]. The transferability of the above factors or of the entire pharmacoeconomic model will depend on the type of economic modeling, data availability, and the need for modeling-based adjustments. For example, if the goal of the pharmacoeconomic model is to measure the economic burden of a disease or diseases on a particular society in monetary terms, it is important that costs and effects accruing in future years be discounted to their present value using widely accepted rates because the time horizon for therapies in certain conditions is long [5,9]. This is especially true for chronic diseases—such as heart disease, cancer, and diabetes—in which the course of the disease is persistent or long-lasting in nature. However, because the perspective of the decision maker is usually the societal perspective for resource allocation decisions, the choice of a discount rate for economic evaluation may not reflect the societal preference for the intended country. When there is variability in the discount rate, the appropriate societal discount rate should be chosen on the basis of the perspective of the analysis and on some theoretical approach, especially when the analytic result is sensitive to the discount rate [6]. Although the choice of the discount rate is an important topic in the context of health economic evaluations, we must also underscore the importance of relative prices. It is recommended that pharmacoeconomic models include all relevant direct health care costs in the evaluation, including indirect costs when appropriate, which will depend on the aim of the study, treatment comparator, the perspective of the evaluation, and the guidelines of the jurisdiction [10]. Unit cost prices of pharmaceuticals and/or medical services should be from the jurisdiction of interest, but due to possible differences in relative or absolute prices, the data on resource use may need to be adapted to the jurisdiction of interest [3]. Currently, there is no consistent guidance on how to address the transferability of economic data for evaluation or on how to adjust for such differences in prices between jurisdictions [3,11]. Addressing the differences in relative prices is very important for determining what happens to the transfer of economic data from one country to another because these differences can lead to different interpretations of cost-effectiveness data in the jurisdiction of interest, especially if there are substantial differences in relative prices [3,11,12]. The comparison of prices across jurisdictions has been the subject of careful investigation of whether markets are truly integrated [13], a term used to describe how much different markets are to each other. There is evidence to suggest that countries/jurisdictions within geographic proximity, similar health care structure, and/or similar political economy will likely have costeffectiveness results that are generalizable [10,14,15]. This idea of prices of similar products to be equal across countries is especially true for countries within the European Monetary System, which operate under a unified currency, the euro. The use of purchasingpower-parity exchange rates or market exchange rates in economic modeling may be necessary when there exist some widely varying price structures between countries; the latter is likely to provide inaccurate estimates of relative incomes and outputs [16]. Difference in medical practice patterns across geographic regions is another important factor to consider when transferring cost-effectiveness data to another jurisdiction. These differences in medical practice between countries would produce differences in resource input, utilization of services, and expenditure among neighboring jurisdictions [17]. Therefore, practice variations between countries/jurisdictions are likely to cause uncertainty in the apparent effectiveness of the health service and thereby make the transferability of cost-effectiveness estimates from one country to another impractical unless adjustments can be made [11,18]. Adjustments for differences in medical consumption on relatively homogenous groups can be done by correcting for the difference either upward or downward [11]. In determining the transferability of clinical and economic data, pharmacoeconomic models must also address another important factor known as the case mix. A case mix is composed of subgroups of patients possessing similar demographic characteristics, clinical attributes, and output utilization patterns [19]. A case mix–based payment system assumes that within diagnosis-related groups there is little variability in clinical attributes and processes of care; therefore, the cost-effectiveness results can be transferable across jurisdictions with similar case mix. Case-mix differences can account for higher medical cost [20], differences in medical treatment practices [21], and variations in treatment outcomes [22] in certain jurisdictions. Variation in treatment outcome is termed “heterogeneity of treatment effects” and identifying potential heterogeneity of treatment effects is necessary to aid in the design of pharmacoeconomic model for adaptation. The type or mix of patients treated may vary substantially between countries, which can affect the cost-effectiveness of an intervention. Therefore, if the heterogeneity of treatment effects in certain jurisdictions is likely, then the use of statistical methods may be needed to adjust for observed differences, and thus allow for more (less) specific therapeutic recommendations in the jurisdiction of interest [23–26]. Methods A working group was convened consisting of experts from various countries who participated in a Global Pharmacoeconomic Model Guidance development to discuss the adaptation of pharmacoeconomic models, originally developed in one country for use in another country. A review of studies adapting pharmacoeconomic models and translation across countries was conducted. The working group discussed controversies surrounding “translation” across countries and recommendations to consider for adaptation. Before preparing the draft report, the working group interviewed internal and external stakeholders responsible for conducting modeling studies to solicit best practices for model adaptation. The Global Pharmacoeconomic Model Guidance working group developed a draft set of key principles and general recommendations for global adaptation. The working group met by phone 4 times and used a Delphi approach via e-mail to obtain consensus on the final set of recommendations. Each working group member was also asked to obtain input from two or three additional experts from his or her region. Based on solicited feedback on these draft recommendations, a set of final recommendations and corresponding rationale was developed. Results The research results described in this guidance define acceptable standards and explains best practices for the transferability of economic and clinical data before submitting an economic evaluation for reimbursement. It takes into account the accepted hierarchies in the levels of evidence (see Fig. 1) and also provides pragmatic recommendations. A checklist (see Fig. 2) is provided at VALUE IN HEALTH REGIONAL ISSUES 5C (2014) 7–13 9 Levels of Evidence for Pharmacoeconomic Decision making Sources of Efficacy Data Sources of Cost Data Higher 1.Clinical trials 1. Naonal esmates 2. Observaonal studies 2. Regional/local esmates 3. Insurance databases 3. Single or group of hospitals or instuons Strength of Evidence 4. Case registries 4. Esmates from similar geographical regions Lower 5. Public health stascs 5. Expert opinion 6. Surveys 7. Unpublished data Fig. 1 – Levels of evidence for efficacy/effectiveness and cost data. the end of the document to guide those who implement pharmacoeconomic model adaptions. Recommendation 1: When adapting a pharmacoeconomic model for local decision makers, follow recommendations for good research practices for conducting pharmacoeconomic studies and pharmacoeconomic modeling, such as those provided by ISPOR. Rationale: Local adaptation of a pharmacoeconomic model assumes that the original model is methodologically sound. Otherwise, the adaptation will retain the same flaws that were inherent in the original pharmacoeconomic model. Implementation: Before adapting a model, the original model should be vetted by at least three experts for structure and scientific integrity if that was not done during the original pharmacoeconomic model development process. Recommendation 2: Familiarize yourself with national, regional, local, or individual payer or technology appraisal agency pharmacoeconomic guidelines and use recommendations and suggested best practices when adapting a global model for individual payer decision makers. Rationale: National or local pharmacoeconomic guidelines developed by jurisdictional experts reflect the best judgment on pharmacoeconomic practices and strategies of pharmacoeconomics in the context of their jurisdiction. Many of these guidelines are consensus documents developed by multiple stakeholders and experts working within the country even if they are not officially endorsed by a government entity. After extensive research and evaluation, the opinion leaders in the country propose their expert opinion on the most effective pharmacoeconomic practices for the country, and such experts often influence the appraisal or coverage decision processes. Thus, such guidelines should be followed when they exist, even when they are in conflict with other recommendations contained in this guidance. A sensitivity analysis, however, may be warranted to reflect the recommendations contained herein to provide scientific validity and because guidance documents may evolve over time. Implementation: Before submitting an economic evaluation for reimbursement, determine whether there are guidelines that represent a consensus view among economists and experts in the intended country of interest. If no such guidance exists, consider regional guidelines and/or recruiting a local expert and/ or key opinion leader from the region to review the adaptation process and resulting pharmacoeconomic model to ensure credibility and applicability. Recommendation 3: Determine the perspective of the adaptation, which may be a societal national payer, a regional payer, or an institutional perspective. The perspective will determine other aspects of pharmacoeconomic model adaptation. Rationale: The societal perspective is commonly used in costeffectiveness analysis, but the choice of the study perspective should consider the decision maker regarding the use of new technologies in a particular country. Implementation: In the absence of specific guidance from the local decision maker, provide both the societal perspective and a narrow focus on direct medical costs only. If desirable, also include an intermediate perspective (e.g., direct medical costs and productivity costs). Recommendation 4: Select the comparators that reflect the current treatments and treatment algorithms (e.g., dose and duration) that are most likely to be replaced by the new therapy. Rationale: The choice of the appropriate comparator is very critical because the cost-effectiveness of an intervention is a relative measure of the cost and effectiveness of any alternative, which is dependent on the choice of the comparator. Implementation: The choice of the comparator must be appropriate to the research question and should reflect current practice or the most widely used therapy/therapies in the jurisdiction of interest. Occasionally, current practice may differ by jurisdiction, so the payer or decision maker in the jurisdiction of interest may specify which alternatives should be compared. Recommendation 5: Always use cost data from the specific country. When various data sources exist, the following data hierarchy should guide which data source is most reliable and applicable for pharmacoeconomic model adaptations for cost or resource utilization estimation: 1) national estimates, 2) regional/ local estimates, 3) single or group of hospitals or institutions, and 4) expert opinion. Rationale: The geographical separations of markets, differences in clinical practice patterns, different payment systems, and various incentives from drug manufacturers are some of the various factors identified as preventing prices of similar products to be equal across countries or variability in resource use between locations. The international markets are not as integrated as domestic markets, so it is important that national estimates data for resource consumption be considered first. Implementation: If cost data from the specific country are not available, apply a standard cost per procedure across all participating sites to derive a dollar value for costs of care. 10 VALUE IN HEALTH REGIONAL ISSUES 5C (2014) 7–13 No. 1 Recommendation Implementation Conduct good research practice for Pharmacoeconomic studies Use recommended economic appraisal guidelines and required reporting and appraisal standards The original model should be vetted for structure and scientific integrity. 3 Determine perspective of economic appraisal 4 Select available treatment options (comparators) In the absence of specific guidance from local decision maker, use both the societal perspective and a narrow focus on direct medical costs only. If desirable, include intermediate perspective. Use current practice or the most widely used therapy/therapies in the jurisdiction of interest. 5 Consider the source of cost data Identify and quantify resource use and costs 2 6 7 8 9 Consider clinical practice patterns and guidelines Use country/region specific epidemiologic data Explain and justify use of estimated treatment effect Refer to recommended economic appraisal guidelines. If no such guidance exists, consider recruiting a local expert and/or key opinion leader from the region to assure credibility and applicability. If cost data from the specific country is not available, apply a standard cost per procedure. Include relevant direct and indirect costs associated with the treatment. An activity-based costing method can generate a more accurate product costs. When using decision analytic modeling, incorporate clinical practice patterns/guidelines of the intended country/jurisdiction of interest. If country/region specific epidemiologic data are not available, use random-effect meta-analysis models and transition probabilities where necessary. Use the average treatment effect from a multinational trial. Conduct a sensitivity analysis using treatment effect based upon patients from the specific country or region. Use local health state preferences and utilities whenever they are available; Use the average of published ones if local utilities are not available. If a revalidation is required/desired, include forward translation, back translation, and pretesting of the instrument. Expert opinion represents lower levels of evidence. Whenever expert opinion is used, multiple experts should be involved. Use the Delphi method for consensus. For data elements that are nontransferable, the model structure, data used as inputs to models, and model validation are important when assessing the quality of models. See http://www.ispor.org/taskforces/GRPModelingTf.asp for more information. 10 Use health state preferences/utilities that are applicable to the region 11 Utilize expert opinion sparingly and appropriately 12 Use modeling to address nontransferable elements 13 Utilize quality-adjusted life years(QALYs) 14 Determine and justify discount rate 15 Source and justification of each data element in PE model Translate findings for The perspective, the recommendations concerning the desired perspective evaluation of resource use/costs, the choice of the comparator, and the valuation of costs should be considered before considering the transferability and reproducibility. 16 Yes No Determine threshold to enable transfer and applicability of QALYs across jurisdictions unless local guidelines recommend a different metric or approach. Use local guidance for discount rate. If none exist, use a “real riskless” discount rate of 3% and conduct sensitivity analysis. To reflect an evidence-based approach to PE modeling, systematic reviews of the literature should be conducted. Fig. 2 – Checklist of pharmacoeconomic model adaptation strategies: A tool for decision making. Whenever expert opinion is used, multiple experts should be involved and ideally a Delphi process should be used to achieve consensus. Recommendation 6: Cost data should include direct medical costs as well as other components of costs (e.g., lost productivity and overhead costs) when available from the specific country. VALUE IN HEALTH REGIONAL ISSUES 5C (2014) 7–13 Costs should be disaggregated so that payer decision makers are able to determine where the components come from and which costs are the major drivers. Rationale: Many factors affect medical care costs, and the disaggregation of cost into individual cost drivers can uncover the major source of variation in medical care costs from one jurisdiction/region from another. Implementation: In cost estimation procedures, assigning cost to cost out products may require including other components of costs on a cause-and-effect-basis. An activity-based costing method with multiple cost pools, activity drivers, and allocation bases can generate more accurate product costs. Recommendation 7: Incorporate clinical practice patterns that reflect how the disease or condition is treated within the jurisdiction. In doing so, be sure to address whether treatment occurs at a hospital, an outpatient clinic, or a physician office. Also address the number and types of drugs, other health technologies, and services received, including whether general practitioners and/or specialist physicians are involved in treatment. Rationale: Differences in medical practice patterns between countries can produce differences in resource input, utilization of services, and expenditure among neighboring jurisdictions, thereby causing uncertainty in the apparent cost-effectiveness of the health service. Implementation: Understand and incorporate the patterns of clinical practice or clinical guidelines in the intended country/ jurisdiction of interest when using statistical techniques such as Markov analysis, decision trees, and discrete event simulation to extrapolate results from one jurisdiction to another. Recommendation 8: Use country-specific incidence or prevalence and other local epidemiologic data (e.g., clinical parameters, mortality, morbidity, and comorbidity) and transition probabilities when adapting a pharmacoeconomic model, even if treatment patterns are similar to the practice of medicine in the original model’s country. Rationale: Even if the baseline risk of disease is the same across countries, the progression of disease could differ, so local transition probabilities between health states may differ from those within the original pharmacoeconomic model. Implementation: Use Markov transition models to model disease progression. Transition probabilities should reflect the current health state of persons in the intended country of interest and not the health states of those in the original model. If country-specific incidence or prevalence rates and other local epidemiologic data are not available, use random-effect metaanalysis models to combine the evidence data and then convert into transition probabilities if necessary; in such instances, sensitivity analysis is recommended. Recommendation 9: In contrast to other aspects of a pharmacoeconomic model, the treatment effect generally is highly transferable and, therefore, often does not need to be altered during country adaptation. Therefore, if adaptations are made, justify why the treatment effect is tailored and provide a valid source for the local data on the effect of treatment. Rationale: Unlike estimates of baseline risk, estimates of treatment effect are generally considered to have high transferability when there is little to no difference in patient characteristics, comparators, and treatment practice patterns. Implementation: In addition to using the average treatment effect from a multinational trial, the treatment effect based on patients from the specific country or region should be provided as sensitivity analysis unless the sample size is extremely small. Recommendation 10: Health state preferences/utilities are not transferable; therefore, a study-specific evaluation is needed to obtain local values for such data elements. Unless such estimates are obtained through a valid patient-reported outcome (PRO) instrument (including linguistic validation), it is preferable to use a published or otherwise validated estimate from another jurisdiction. 11 Rationale: Health state utilities quantify an individual’s preferences for different health outcomes and so the use of an existing validated PRO instrument does not necessarily enable the transfer of preferences/utilities across different countries/ jurisdictions to predict preferences associated with actual experienced health states. Therefore, to reflect a more valid preference associated with an experienced health state in the country/ jurisdiction of interest, translation and cultural adaptation of existing instrument is recommended. The average of published utilities, however, can be used if local utilities are not available. Implementation: Use local health state preferences and utilities whenever they are available; however, if local utilities are not available, use the average of published ones along with sensitivity analysis. If a revalidation of a PRO is required/desired, the process should include forward translation, back translation, and pretesting of the instrument to assess the PRO. Recommendation 11: Country-specific data are generally preferred over local expert opinion when available. The balance between country-specific data and expert opinion is a spectrum. The more data that are available, the less expert opinion is necessary, while a lack of data must be supplemented with expert opinion. Expert opinion should be used to evaluate the quality of the data and the ability for it to be transferred to the specific country. Rationale: Clinical decision making and recommendations based on expert opinion represent lower levels of evidence, indicating little to no objective empirical evidence. Implementation: As stated in Recommendation 5, whenever expert opinion is used, multiple experts should be involved and ideally a Delphi process should be used to achieve consensus. Recommendation 12: Modeling can be used when key data elements are found to be nontransferable and should incorporate the demographic characteristics of the local jurisdiction. Rationale: If the cost-effectiveness results of a treatment are not easily transferable from location to location, then randomeffects meta-analysis models may be used to pool estimates and some form of formal modeling approach is needed. Implementation: ISPOR and the Society for Medical Decision Making had a Joint Modeling Good Research Practices Task Force. The current draft of the series of reports is available online at http://www.ispor.org/taskforces/GRPModelingTf.asp. Recommendation 13: Despite concerns about the estimation and transferability of utilities, quality-adjusted life-years (QALYs) remain an accepted “second best” metric and should be used as a primary output of pharmacoeconomic models across jurisdictions unless local guidelines recommend a different metric or approach. Rationale: Population values and preferences could vary, but the QALY metric represents a generic outcome measure that can be applied more generally. Implementation: The transfer and applicability of QALYs across jurisdictions will be largely driven by the threshold value for the QALY, the use of an utility instrument that ensures that concepts within an instrument are equal between original and target jurisdiction (language, time, and context), and the appropriateness of the instrument for assessing utility. Recommendation 14: Use a discount factor that reflects the appropriate discount rate for the local jurisdiction. Rationale: The perspective of the decision maker is usually the societal perspective for resource allocation decisions. The choice of a discount rate for economic evaluations should reflect the societal preference for the intended country because an analytic result can be sensitive to the discount rate. Implementation: If there is local guidance for a specific discount rate, use that rate. If there is no local guidance, use a “real riskless” discount rate of 3% and sensitivity analysis using 5% as well as a reasonable range of discount rates (e.g., 0%–10%). 12 VALUE IN HEALTH REGIONAL ISSUES 5C (2014) 7–13 Recommendation 15: Provide a source and justification for each data element or assumption that is included in the adaptation so that the pharmacoeconomic model adaptation clearly reflects an evidence-based approach to pharmacoeconomic modeling; also, justify why parameters that are kept from the original model are not changed. Rationale: The transparency of methods and justification of key parameters in the model will not only ensure transferability and reproducibility of results but would also influence the allocation of health resources by decision makers. Implementation: Conduct a systematic literature review to provide inputs for the economic model. Regularly, sources such as “previous unpublished study” or “presentation on a local conference” are used to retrieve input data. In particular in small countries with missing national literature sources, such inputs may be inherited for years. Recommendation 16: Provide a translation of findings that addresses the perspective of the local decision maker so that results are believable and meaningful. Rationale: The guidelines for health economic evaluations generally point out that the perspective should be the societal perspective. The choice of the analytical aspects of the study should be within the scope of the perspective. Implementation: Guidelines for health economic evaluations may vary between jurisdictions and so the perspective of the evaluation, the recommendations concerning the evaluation of resource use/ costs, the choice of the appropriate treatment comparator, and the valuation of costs should be considered before considering the transferability and reproducibility of the study results. Conclusions The following key guiding principles should be kept in mind to guide pharmacoeconomic model adaptations across geographic jurisdictions. Principle 1: Before developing a local adaptation of a model, the model and availability of data should be evaluated using general knockout criteria or a checklist of critical transferability factors. A minimum standard of quality must be met before the local adaptation can even be considered for transferability [7,27]. Principle 2: Throughout this document, the term “local” jurisdiction is used to describe the jurisdiction or institution for which a pharmacoeconomic model is being adapted and, as such, could reflect a country, a region within a country, a specific national, regional, or local payer, a single institution, or a group of institutions. Assess the variability of the factors that can potentially affect the transferability of data from one “local” jurisdiction to another and perform probabilistic or multivariate sensitivity analysis [8,11,28,29]. Principle 3: Transferability is a spectrum, with “fully transferable” data describing a model that is completely transferable across a group of countries or jurisdictions with similar health care environments and pharmacoeconomic practices. “Nontransferable” describes a situation in which there is a critical failure that makes the entire model nontransferable either because of the model structure or data availability in the local jurisdiction. Within a study deemed to be transferable, each data element also exists on a spectrum of transferability [6,8,30]. Principle 4: A well-constructed model that is designed with adaptation in mind often is transferable across a wide area of countries. Frequently, the problem with adaptation is not so much the structure of the model so much as the lack of data availability for the adaptation process. The model structure is more problematic when specific comparators or components of treatment are “hardwired” into the model structure; when that is the case, the model cannot be adapted to local decision makers for whom the comparator treatments or aspects of care do not apply [14,31]. Principle 5: Following the decision on whether part(s) or the whole model is deemed to be transferable, transparency of model assumptions, characteristics, and limitations for data elements that are considered transferable will allow economic evaluations to be more informative for the intended user in the decisionmaking process [10,32]. Principle 6: Before adaptation, the individual(s) who is responsible for tailoring the model must be trained on the model, ideally by the developer of the original model. This ensures that the local adaptation maintains the general structure and key components and properties of the pharmacoeconomic model [22,33]. Principle 7: Local adaptation is best informed with input from local decision makers or their proxies/surrogates in advance of model adaptation [34–36]. Principle 8: Countries differ in terms of whether they view pharmacoeconomic models to be transferable from countries that are either within their geographic proximity and/or have a similar health care structure. Political, social, and economic factors also affect whether a pharmacoeconomic model is transferable across jurisdictions. Before pooling data, the transferability between jurisdictions should be viewed as acceptable by the intended user of the pharmacoeconomic model. For example, transferability is viewed as reasonably acceptable among countries in Latin and South America that have similar size and economies and generally not acceptable among Asian countries. In Europe, transferability of effectiveness parameters is acceptable across Western and parts of Eastern Europe, whereas costs are not [21,37,38]. Principle 9: Countries that have limited experience with pharmacoeconomic evaluation and limited data availability are more likely to deem economic evaluations and data elements to be transferable. They also, however, are more likely to focus on affordability and so the pharmacoeconomic model may be less impactful than a budget impact model [24,39]. Principle 10: Despite advances in the use of pharmacoeconomic models and improvements in pharmacoeconomic modeling and adaptation, there is continued need for knowledge dissemination and efforts to increase awareness of local decision makers, which may include education of clinicians as well as policymakers because clinicians affect the policymakers’ decisions [24,40,41]. Source of financial support: This work was supported by Pfizer, Inc. Authors received consulting income to participate in the development of the guidance document. Professor Mullins also discloses that he has received other consulting income and grant support from Pfizer. The views expressed in this article are those of the authors, and no official endorsement by Pfizer is intended or should be inferred. R EF E R EN C ES [1] Harris A, Buxton M, O’Brien B, Drummond FR. Using economic evidence in reimbursement decisions for health technologies: experience of 4 countries. Expert Rev Pharmacoecon Outcomes Res 2001;1:7–12. [2] Welte R, Feenstra T, Jager H, Leidl R. A decision chart for assessing and improving the transferability of economic evaluation results between countries. Pharmacoeconomics 2004;22:857–76. [3] Pichon-Riviere A, Augustovski F, Garcia Marti S, et al. Transferability of health technology assessment reports in Latin America: an exploratory survey of researchers and decision makers. Int J Technol Assess Health Care 2012;28:180–6. VALUE IN HEALTH REGIONAL ISSUES 5C (2014) 7–13 [4] Barbieri M, Drummond M, Willke R, et al. Variability of cost-effectiveness estimates for pharmaceuticals in Western Europe: lessons for inferring generalizability. Value Health 2005;8:10–23. [5] Fuchs VR. Economic Aspects of Health. A Conference Report/National Bureau of Economic Research. Chicago: University of Chicago Press, 1982. [6] Krahn M, Gafni A. Discounting in the economic evaluation of health care interventions. Med Care 1993;31:403–18. [7] Goeree R, He J, O’Reilly D, et al. Transferability of health technology assessments and economic evaluations: a systematic review of approaches for assessment and application. Clinicoecon Outcomes Res 2011;3:89–104. [8] Koopmanschap MA, Touw KC, Rutten FF. Analysis of costs and costeffectiveness in multinational trials. Health Policy 2001;58:175–86. [9] Ramsey SD, Sullivan SD. The burden of illness and economic evaluation for COPD. Eur Respir Rev 2003;41(Suppl.): 29 s–35. [10] Hjelmgren J, Berggren F, Andersson F. Health economic guidelines– similarities, differences and some implications. Value Health 2001;4:225–50. [11] Rizzo JD, Powe NR. Methodological hurdles in conducting pharmacoeconomic analyses. Pharmacoeconomics 1999;15:339–55. [12] O’Brien BJ. A tale of two (or more) cities: geographic transferability of pharmacoeconomic data. Am J Managed Care 1997;3(Suppl.):S33–9. [13] Chen N. The Behaviour of Relative Prices in the European Union: A Sectoral Analysis. CEPR Discussion Paper No. 3320. Warwick: University of Warwick, Department of Economics, 2002. [14] Wild D, Grove A, Martin M, et al. Principles of good practice for the translation and cultural adaptation process for patient-reported outcomes (PRO) measures: report of the ISPOR Task Force for Translation and Cultural Adaptation. Value Health 2005;8:94–104. [15] Chang WC, Midodzi WK, Westerhout CM, et al. Are international differences in the outcomes of acute coronary syndromes apparent or real? A multilevel analysis. J Epidemiol Community Health 2005;59:427–33. [16] Nordhaus W. Alternative measures of output in global economicenvironmental models: purchasing power parity or market exchange rates? Energy Econ 2007;29:349–72. [17] Wennberg J, Gittelsohn. Small area variations in health care delivery. Science 1973;182(4117):1102–8. [18] James BC, Hammond ME. The challenge of variation in medical practice. Arch Pathol Lab Med 2000;124:1001–3. [19] Fetter RB, Shin Y, Freeman JL, et al. Case mix definition by diagnosisrelated groups. Med Care 1980;18(2 Suppl.):1–53: (iii). [20] Frick AP, Martin SG, Shwartz M. Case-mix and cost differences between teaching and nonteaching hospitals. Med Care 1985;23:283–95. [21] Laupacis A, Feeny D, Detsky AS, Tugwell PX. How attractive does a new technology have to be to warrant adoption and utilization? Tentative guidelines for using clinical and economic evaluations. CMAJ 1992;146:473–81. [22] Garrison LP Jr, Neumann PJ, Erickson P, et al. Using real-world data for coverage and payment decisions: the ISPOR Real-World Data Task Force Report. Value Health 2007;10:326–35. [23] Hollenbeck BK, Miller DC, Taub DA, et al. The effects of adjusting for case mix on mortality and length of stay following radical cystectomy. J Urol 2006;176:1363–8. 13 [24] Tarlov AR, Ware JE Jr, Greenfield S, et al. The Medical Outcomes Study: an application of methods for monitoring the results of medical care. JAMA 1989;262:925–30. [25] DerSimonian R, Laird N. Meta-analysis in clinical trials. Control Clin Trials 1986;7(3):177–88. [26] Brennan A, Akehurst R. Modelling in health economic evaluation. What is its place? What is its value? Pharmacoeconomics 2000;17:445–59. [27] Bekkering GE, Kleijnen J. Procedures and methods of benefit assessments for medicines in Germany. HEPAC 2008;9(Suppl. 1): 5–29. [28] Augustovski F, Iglesias C, Manca A, et al. Barriers to generalizability of health economic evaluations in Latin America and the Caribbean region. Pharmacoeconomics 2009;27:919–29. [29] Goeree R, Burke N, O’Reilly D, et al. Transferability of economic evaluations: approaches and factors to consider when using results from one geographic area for another. Curr Med Res Opin 2007;23:671–82. [30] Manca A, Willan AR. ‘Lost in translation’: accounting for betweencountry differences in the analysis of multinational cost-effectiveness data. Pharmacoeconomics 2006;24:1101–19. [31] Drummond M, Barbieri M, Cook J, et al. Transferability of economic evaluations across jurisdictions: ISPOR Good Research Practices Task Force Report. Value Health 2009;12:409–18. [32] Smith T.W., Tierce J.C. Designing and developing budget impact models suited for global adaptation. ISPOR CONNECTIONS. 2006. Available from: http://www.ispor.org/news/articles/aug06/designing.asp. [Accessed July 14, 2013]. [33] Weinstein MC, O’Brien B, Hornberger J, et al. Principles of good practice for decision analytic modeling in health-care evaluation: report of the ISPOR Task Force on Good Research Practices–Modeling Studies. Value Health 2003;6:9–17. [34] Rawlins MD, Culyer AJ. National Institute for Clinical Excellence and its value judgments. BMJ 2004;329:224–7. [35] Wonderling D, Sawyer L, Fenu E, et al. National Clinical Guideline Centre cost-effectiveness assessment for the National Institute for Health and Clinical Excellence. Ann Intern Med 2011;154:758–65. [36] Baltussen R, Ament A, Leidl R. Making cost assessments based on RCTs more useful to decision-makers. Health Policy 1996;37: 163–83. [37] Greiner W, Schöffski O, Graf vd, Schulenburg J-M. The transferability of international economic health-economic results to national study questions. HEPAC Health Econ Prev Care 2000;1(2):94–102. [38] Pang F. Design, analysis and presentation of multinational economic studies: the need for guidance. Pharmacoeconomics 2002;20:75–90. [39] Wordsworth S, Ludbrook A. Comparing costing results in across country economic evaluations: the use of technology specific purchasing power parities. Health Econ 2005;14:93–9. [40] Fernandez LA, Martin JM, del Castillo JD, et al. Sources of influence on medical practice. J Epidemiol Community Health 2000;54:623–30. [41] Hutubessy R, Chisholm D, Edejer TT. Generalized cost-effectiveness analysis for national-level priority-setting in the health sector. Cost Effec Resour Alloc 2003;1:8.