Value in Health Regional Issues Vol. 5 (focusing on Latin America)

Transcripción

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
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.
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
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.
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.
[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.
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.
* 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.
2
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,
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
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
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
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
Diagnostic aids
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
Diagnostic aids
Fig. 4 – CBCISS update requests approved. CBCISS, Health Care Formulary and Supply Catalog. Source: Database of the
Commission’s Technical Secretariat [12].
6
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.
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
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.
15
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
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
once-daily injection for
insulin detemir (IU/kg)
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
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
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.
[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.
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].
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.
21
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
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 $108,870 for drug preparation by
an ICU nurse. Variation within these mean costs is higher for
compounding center preparation than for premix drugs, due to
the number of variables involved in the respective processes.
This also applies to MINIBAG and nursing staff preparation,
compared with premix drugs.
According to the model, the rate of errors that lead to harm
(expressed as the number of errors per 10,000 administered
medications) is 1 for premix drugs, 2 for compounding center
preparations, 4 for MINIBAG Plus, and 17 for nurse drug delivery
(which is equivalent to 1 harming error per every 600 administered medications).
A similar calculation for deaths attributable to a drug administration error, per every 100,000 IV medications administered,
resulted in 6 deaths for premix drugs, 21 for compounding center
preparations, 42 for MINIBAG Plus, and 168 for nurse staff
preparation delivered with buretrol.
23
Table 2 – Selected variables used in the model, with wide intervals used for the sensitivity analysis.
Variable
Compounding center preparation
Compounding center application
Bedside application by nursing staff
MINIBAG Plus application
Premix drugs application
Professional nurse salary (h)
Cost of dopamine
Day in ICU
Cost of mild error
Cost of moderate error
Cost of severe error
Base case ($)
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
Market prices
Weighted mean of market price
Costs of local ICU
Expert panel estimation
ICU, intensive care unit.
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
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.
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
* Address correspondence to: Laura S. Gold, Pharmaceutical Outcomes Research and Policy Program, School of Pharmacy, University of
Washington, Box 359455, Seattle, WA 98195.
26
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
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.
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
* Address correspondence to: Muriel Ramírez-Santana, Faculty of Medicine, Universidad Católica del Norte, Larrondo 1281, Coquimbo,
Chile.
36
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%)
Cost esmaon for family
/household/ farm
Hospitalizaon (23%)
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%)
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
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
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.
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.
[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.
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.
* Correspondencia: Ximena Burbano-Levy. MD. Zilonis Health, Inc. Telefono: þ1-561-8436767.
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
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
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
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
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)
$(79,186.00)
$(52,170.00)
$(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
$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/
[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.
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.
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].
M: 377. Hemorragia gastrointestinal con CCM. [Online] http://www.
imss.gob.mx/profesionales/grd/GRD_377.pdf.
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].
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].
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].
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
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.
49
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
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
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
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
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 $40,000
(Table 2).
As demonstrated by the DSA, the variables that had the
largest impact on the results were the efficacy of each therapy
in reducing the LDL-C level and the adherence rates. Fig. 3A
displays a tornado diagram for the Colombian primary prevention DSA for A20 versus R10 and demonstrates that the ICERs in
these scenarios vary greatly— each treatment is dominated by
the other at some point—depending on the efficacy and adherence, while changes in cost and MCVE risk have a lesser effect. In
the DSA for the other Colombian primary prevention analysis
(A10 vs. R5), results were similar, except that the 1-year adherence rate of rosuvastatin emerged as having the largest impact
on the cost per QALY gained (not shown). For the Brazilian
primary prevention analyses, the DSA for A20 versus R10
revealed that treatment efficacy has the largest impact, yet did
not alter dominance (Fig. 3B). The DSA for A20 versus R20 was
similar, though the impact of adherence rates was much more
considerable, similar to the results shown for Colombia (Fig. 3A).
As with the Colombian models, the cost of therapy and the risk of
MCVE had less of an impact on the results. DSA results for
53
Table 2 – Total and incremental cost and outcomes per patient for primary and secondary prevention.
Prevention level and
treatment dose
Primary prevention
A10
R5
Difference
A20
R10
Difference
A40
R20
Difference
A80
R40
Difference
Colombia
Total
ICERs
Cost
($)
QALYs
Lifeyears
CV
events*
Cost/QALY
gained ($)
Cost/life-year
saved ($)
Cost/CV event
avoided ($)
4,438
8,749
4,312
6,234
9,953
3,719
13.41
13.42
0.01
13.42
13.42
0.01
16.81
16.82
0.01
16.82
16.83
0.01
0.30
0.30
0.01
0.30
0.30
0.00
Reference
811,376
–
Reference
700,202
–
Reference
685,005
–
Reference
591,086
–
Reference
4,422,698
–
Reference
3,819,276
–
10,633
13,621
2,988
11,748
15,575
3,826
8.63
8.64
0.01
8.65
8.67
0.01
11.12
11.13
0.02
11.14
11.16
0.02
0.80
0.79
0.00
0.79
0.79
0.00
Reference
201,667
–
Reference
258,168
–
Reference
188,591
–
Reference
241,654
–
Reference
752,378
–
Reference
963,614
–
Brazil
Total
Primary prevention
A20
R10
Difference
A20
R20
Difference
A40
R40
Difference
ICERs
Cost
($)
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
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
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
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].
[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.
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
* 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]..
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
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
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
unidad de HAQ en
pacientes que no
responden al tratamiento
soporte.
unidad de PASI
Probabilidad PsARC(þ)
adalimumab
etanercept
infliximab
Probabilidad PASI 50(þ)
adalimumab
Probabilidad PASI 50 (þ)
etanercept
infliximab
adalimumab
etanercept
Probabilidad PASI75 (þ)
infliximab
adalimumab
etanercept
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)
paciente PsARC(þ) y PASI
75(-) (USD)
paciente PsARC(-) y PASI
75(þ) (USD)
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
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.
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.
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
[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/.
[19] Organización Mundial de la Salud, Choosing Interventions that are
Cost-Effective (CHOICE), Disponible en: http://www.who.int/choice.
[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].
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
* Address correspondence to: Cassyano Januário Correr, nº 632, Pref. Lothario Meissner Avenue, Curitiba 81540050, Brazil.
E-mail: [email protected]
66
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
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
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$63,572.69/PASI75). A reduction in the acquisition cost of
ustekinumab, infliximab, and etanercept by 11%, 36%, and 67%,
respectively, would change their cost-effectiveness ratio to that
shown by adalimumab.
Probabilistic sensitivity analysis was undertaken as recommended by the Brazilian Guideline for Economic Evaluations [24].
Following the statements provided by Brazilian Guideline for
Economic Evaluations [24], one-way sensitivity analysis and
probabilistic sensitivity analysis were performed. Moreover, the
impact of time horizon was assessed in the final result.
All the parameters of the model were submitted to one-way
sensitivity analysis. The range used for each parameter corresponds to the 95% confidence intervals. Time-horizon analysis
considered scenarios ranging from 1 year up to the lifetime
period; thus, probabilities from 2 years were extrapolated to the
lifetime period. In addition, probabilistic sensitivity analysis was
performed considering all variables of the model. Costs were
distributed using gamma distributions in which hyperparameters
for biologic agents were calibrated so that lowest and highest
prices registered in Sistema Brasileiro de Informação sobre
Medicamentos corresponded to 0.95 probability interval. Probability parameters were distributed using beta (nodes with two
branches) or Dirichlet (nodes with more than two branches)
distributions. The hyperparameters for these distributions were
calibrated considering the number of participants in each study
by which PASI responses were extracted. A total of 10,000
iterations were applied. Hyperparameters of each distribution
as well as the ranges used for one-way analysis are available in
Appendix 4 in Supplemental Materials found at http://dx.doi.org/
10.1016/j.vhri.2014.09.002.
Table 3 – Results from 3 y of treatment with biologics.
Biologic agent
Cost (R$)
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.
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)
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.
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.
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.
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
[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.
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 decisionmaking 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).
2
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
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
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.
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
[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.
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,
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]
8
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
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
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.
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
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.
[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.

Value in Health Regional Issues Vol. 5 (focusing on Latin America)

Transcripción

Documentos relacionados

After the rally

Sts. Peter and Paul Catholic Church Sanctuary of St. Toribio Romo

Saint Vito - Most Holy Trinity - Parish of St. Vito and Most Holy Trinity

| CALIFORNIA WINE DRAPES HERE! THE GREATEST SHOW ON

Summary of WPD activities 2013

Summary of WPD activities 2012

Summary of WPD activities 2011