Incremental cost-effectiveness ratio and willingness to pay threshold in a cost-
A cost-effectiveness analysis evaluates a new treatment's value through the incremental cost-effectiveness ratio (ICER), which indicates the cost required for an additional health unit, such as a QALY or LYG (Drummond et al 2005) This analysis compares the new treatment to existing options, focusing on the current standard of care or the most relevant alternative The ICER is calculated using the expected costs and effectiveness of both treatments, with the formula incorporating costs (C_N and C_O) and effectiveness (E_N and E_O) (Briggs 2001) The results are typically represented in a cost-effectiveness plane, divided into four quadrants: NE (more costly and more effective), SE (less costly and more effective), SW (less costly and less effective), and NW (more costly and less effective) In the NW quadrant, the old treatment is preferred, while in the SE quadrant, the new treatment is favored When neither treatment dominates, cost-effectiveness evaluation becomes crucial (Dasbach et al 2010) Often, new treatments are more effective but also more expensive, leading to ICER estimates primarily in the NE quadrant Ultimately, decisions regarding cost-effectiveness hinge on society's willingness to pay (WTP) threshold (Briggs 2001).
The cost-effectiveness plane, as illustrated by Drummond et al (2005), categorizes the incremental cost-effectiveness ratio (ICER) into four quadrants, indicating varying outcomes based on incremental costs and effects The arrows on the plane show the direction of these costs and effects, while the willingness to pay threshold represents the maximum acceptable ICER Alternative notations for the quadrants include NE for north-east, NW for north-west, SE for south-east, and SW for south-west.
The willingness to pay (WTP) threshold indicates the value society is prepared to invest in an additional health unit The cost-effectiveness of a product is ultimately determined by this WTP threshold, which sets the maximum acceptable incremental cost-effectiveness ratio (ICER).
While uncertainty surrounds the ICER estimate, the likelihood of a treatment being deemed cost-effective at a specific willingness-to-pay (WTP) threshold can influence cost-effectiveness decisions Although clear WTP thresholds enhance transparency in decision-making, justifying them in every scenario can be challenging.
2007) Figure 7 illustrates how reimbursement decision would apply in the case ȱȱan explicit WTP cut-off point, or if WTP was considered as a range
Figure 7 Willingness to pay threshold as an explicit point (A), and as a range (B), assuming that reimbursement decision depends only on cost-effectiveness (Devlin and Parkin 2004)
Fixed threshold values for cost-effectiveness are commonly used worldwide, despite being viewed as controversial and arbitrary (Grosse 2008) The National Institute for Health and Care Excellence (NICE) has set a maximum threshold of £30,000 per Quality-Adjusted Life Year (QALY), with higher values proposed for end-of-life treatments (Drummond and Mason 2007) Additionally, the World Health Organization (WHO) suggests that a treatment is highly cost-effective if the cost-effectiveness ratio (CER) is below the per capita gross domestic product (GDP) of the region, and cost-effective if it falls between 1 to 3 times the GDP per capita (WHO 2003) However, caution is advised when comparing these thresholds to those recommended for Incremental Cost-Effectiveness Ratios (ICER).
Grosse (2008) conducted a review to investigate the origins of the widely referenced QALY threshold of USD 50,000, which emerged in the 1990s but has been inaccurately linked to dialysis standards His findings suggest that this figure was selected for its convenience rather than any scientific rationale, and he argues that it is outdated, originating from the early 1990s Furthermore, if adjusted for healthcare inflation, the threshold would currently be approximately USD 200,000 (Tannock et al 2011).
The justification for a strict cut-off point in cost-effectiveness analysis is questionable both theoretically and practically, as empirical evidence for threshold values is insufficient Relying solely on Incremental Cost-Effectiveness Ratios (ICER) and Willingness to Pay (WTP) for reimbursement overlooks essential factors like equity and fairness (Rawlins and Culyer 2004) Threshold values can vary based on health conditions, treatment outcomes, and associated risks, and the rigid application of WTP limits does not reflect real-world scenarios Different societies establish their own cost-effectiveness thresholds, which can be implicit or explicit (Peppercorn et al 2011) Moreover, the same threshold is often applied to both life-years and Quality-Adjusted Life Years (QALYs), complicating the WTP concept due to varying outcome measures As a result, existing thresholds have faced criticism for being too low, too high, or overly ambiguous (Grosse 2008).
Budget impact analysis
Budget impact analysis (BIA) aims to estimate the financial implications of adopting new healthcare interventions within specific settings, considering limited resources It assesses how changes in the mix of drugs and therapies for a particular health condition affect overall treatment spending Unlike cost-effectiveness analyses, which focus on individual drug costs, BIA evaluates the expected changes in treatment practices across the entire population This comprehensive approach provides a clearer picture of the financial impact of new treatments on healthcare budgets.
Figure 8 The basic concept and an illustrative flow of budget impact analysis (Mauskopf et al
Budget impact analyses (BIA) are rarely featured in scientific journals, with a review by Orlewska and Gulácsi (2009) identifying only 34 relevant publications from 2000 to 2008 Of these, 65% focused primarily on budget impact, while 35% included it as a secondary topic The studies were predominantly conducted in Europe (47%) and the USA (41%), with over half (58%) funded by the pharmaceutical industry, and 53% specifically prepared for pharmaceuticals Notably, 65% of these studies were published in the years 2007-2008, highlighting the growing significance and acknowledgment of budget impact issues in healthcare research.
In 2007, the International Society for Pharmacoeconomics and Outcomes Research (ISPOR) established key principles for budget impact analysis, marking a significant step towards international standards in this field Although the need for such guidelines was recognized earlier in the 21st century, many countries still have their own varying national recommendations for conducting budget impact analyses These national guidelines often lack consistency in defining budget impact analysis and typically provide limited insights into crucial factors involved Notably, Finland does not have official guidelines or recommendations regarding budget impact analyses.
Budget impact estimations for new pharmaceuticals have been conducted for years in various countries, but a lack of widely accepted methods has led to inconsistencies and misinterpretations in these analyses (Mauskopf et al 2005) This heterogeneity is evident in the varying methodologies, time horizons, populations, and result reporting across studies For instance, while some studies have time horizons ranging from 100 days to 15 years, nearly half (44%) focus on a 1-year perspective (Orlewska and Gulácsi 2009) Extended follow-up periods increase uncertainty, particularly in the rapidly evolving field of oncology, where a shift from a 1-year to a 4-year perspective can significantly affect results Decision-makers should consider the relevance of the time horizon, as lifetime perspectives often yield unreliable outcomes Additionally, many studies with time horizons over 1 year apply discount rates of 3%, 3.5%, or 5% (Orlewska and Gulácsi 2009), despite current ISPOR guidelines not recommending discounting Furthermore, the impact of uncertainty is seldom assessed in these analyses (Mauskopf et al 2005).
There is a growing need for consistency in budget impact analyses, particularly regarding sensitivity analyses, cost discounting, transparency, and the sources of information used (Orlewska and Gulácsi 2009) While the goal is not to achieve precise figures, the results must be reliable International recommendations aim to foster uniformity and comparability in budget impact analyses at both national and international levels (Mauskopf et al 2007).
General methods used in studies I-V
This section offers an overview of the pharmacoeconomic methods employed in this thesis, with detailed descriptions of specific methods in their respective chapters Three studies (II, III, V) utilized Markov modeling, characterized by three mutually exclusive health states Markov models feature discrete, memoryless stages where patients within a stage are considered homogeneous Time is represented in cycles, allowing a hypothetical patient population to transition between predefined health stages based on these cycles The determination of health stages and cycle length is influenced by the study perspective and the characteristics of the disease.
The model structure and transition probabilities are designed to reflect the natural progression of diseases, recognizing that each disease is unique, which limits the applicability of a single evaluation model across different health conditions The 3-stage model was tailored to meet the specific aims of individual studies, maintaining a consistent basic structure despite variations in technical implementation This framework includes health states such as "No progression," "Progressed disease," and "Dead," which are relevant to various types of cancer These health states are crucial in clinical trials that measure endpoints like time to treatment failure, time to disease progression, progression-free survival, and overall survival Additionally, the model's partitioning related to disease severity impacts treatment costs, shifting from drug-intensive to hospital-intensive expenditures when treatment transitions from active to supportive or palliative care.
Figure 9 Basic framework for the utilized cancer model
Probabilistic sensitivity analyses (PSA) were employed in the cost-effectiveness analyses (II, V) and the budget impact analysis of adjuvant trastuzumab (IV) to account for uncertainty in model parameters In PSA, model inputs vary independently based on predetermined probability distributions (Briggs 2001) Study II utilized WinBUGS software to implement a Bayesian approach (Martikainen 2008) The results of the PSA were presented through cost-effectiveness planes (V), cost-effectiveness acceptability curves (II, V), and affordability curves (IV), with the affordability curve in the budget impact analysis indicating the likelihood of remaining within a specified budget (Sendi and Briggs 2001).
Value of Information (VOI) analysis was utilized in study V's probabilistic model to determine the maximum costs one would be willing to incur to mitigate uncertainty This approach also helps quantify parameter uncertainty, enabling more efficient allocation of research resources (Barton et al 2008).
Incorporating population dynamics into budget impact models is crucial for chronic diseases, particularly in unstable populations and long-term studies This thesis employed state transition models to address these dynamics effectively The cost-effectiveness models utilized closed cohort modeling, tracking hypothetical patients throughout their lifetimes, while the budget impact models adopted open cohort stage-transition frameworks, allowing for the inclusion of new patients based on estimated incidence rates New cases remained in the patient pool until follow-up or death, with treatment effectiveness factored into the models, facilitating a comparison of costs across different disease stages for the treatments evaluated.
A general illustration of patient dynamics utilized in the open cohort models in this thesis is depicted in Figure 10
Figure 10 Illustration of patient dynamics in the utilized budget impact models
Economic evaluations can be conducted from various perspectives, influencing the resources and costs considered The most frequently used perspective in cost-effectiveness analyses is that of society or the healthcare payer While the payer perspective is particularly relevant for budget impact analyses, the societal perspective, though less common, is still a possibility The chosen perspective depends on the study's objectives, leading to variability among the studies included in this thesis.
Figure 11 Different perspectives of economic evaluations (Mogyorosy and Smith 2005)
A model's effectiveness hinges on the quality of its inputs, making the selection of reliable data sources a critical step in the modeling process As illustrated in Figure 12, the required model inputs for cost-effectiveness and budget impact models depend on the specific study requirements and data availability, which means that not all inputs may apply universally due to the absence of certain disease-specific or treatment-specific features The choice of data sources and model inputs often relies on individual decisions Furthermore, the International Society for Pharmacoeconomics and Outcomes Research (ISPOR) has provided guidelines for measuring drug costs and is currently preparing recommendations for best practices in various modeling techniques.
Figure 12 Required model inputs in cost-effectiveness and budget impact analyses The dotted lines represent optional inputs depending on study perspective
5 Specific aspects and methodological challenges of pharmacoeconomics in cancer care
Pharmaceuticals undergo a stringent evaluation process to secure marketing authorization and potential reimbursement Randomized clinical trials serve as the gold standard for demonstrating the efficacy of new therapies, with regulatory authorities like the European Medicines Agency assessing the risk-benefit balance based on these results and additional information However, the regulatory approval process does not address treatment costs or funding, which are managed by individual EU member states In Finland, health economic evaluations are mandatory for new active pharmaceutical ingredients when seeking reasonable wholesale pricing and reimbursement Despite the importance of economic factors, funding decisions also consider clinical issues, equity, uncertainty of results, the innovative nature of the technology, disease characteristics, target population traits, and broader societal costs and benefits.
Entering the Finnish pharmaceutical market involves more than just obtaining marketing authorization, particularly for prescription products used in outpatient care A medication's market penetration can be significantly hindered if it is deemed unsuitable for reimbursement Therefore, successful market entry typically necessitates clear demonstrations of cost-effectiveness and affordability, ensuring that the treatment is financially viable for both healthcare providers and patients.
Figure 13 Pharmaceuticals entry to Finnish market via the European centralized marketing approval process
Limitations related to clinical trials as a data source
Randomized clinical trials, while regarded as the gold standard for efficacy data, have inherent limitations that can impact their real-world applicability They often reflect results under ideal conditions with highly selected patients, which may not represent everyday clinical effectiveness Furthermore, the statistical significance observed between experimental and control interventions does not always equate to clinical significance Many trials utilize surrogate endpoints, such as disease-free survival (DFS) and progression-free survival (PFS), to expedite results, yet these endpoints may not reliably predict overall survival (OS) The uncertainty surrounding the timing of disease progression raises concerns about the validity of PFS as a primary endpoint Nonetheless, PFS has been deemed an acceptable surrogate for OS in certain contexts, such as advanced colorectal cancer and metastatic renal cell carcinoma, where studies have indicated a correlation between differences in progression and overall survival outcomes.
OS Nevertheless, it has been stated that better defined criteria are needed in order to validate surrogate markers for rare diseases (Drummond et al 2009)
Patient crossover in clinical trials, particularly in oncology, poses significant challenges for economic evaluations This phenomenon occurs when patients switch from one treatment arm to another, often due to the observed superiority of one arm over the other Such crossover can obscure the treatment effects and introduce selection bias, as patients who switch may have different prognoses compared to those who remain Specifically, patients who crossover may be sicker and less responsive to standard therapies, complicating the analysis of treatment efficacy While various methods have been proposed to address crossover bias, such as patient regrouping and follow-up censoring, standard statistical techniques often fail to fully correct for these biases Overall survival outcomes are especially susceptible to confounding from crossover and subsequent therapies, making progression-free survival (PFS) a more reliable measure Recent strategies, including prediction equations and external data, aim to mitigate the impact of crossover in economic evaluations.
Extrapolation of survival data
In clinical trials, the follow-up period is typically brief, necessitating the extrapolation of survival estimates, such as Kaplan-Meier survival curves, for effective use in cost-effectiveness analyses (Drummond et al 2005) Furthermore, trials involving rare cancers often struggle to show statistically significant benefits due to the limited patient population available (Chabot et al.).
The extrapolation of survival data can significantly influence cost-effectiveness results, as different modeling approaches may yield varying outcomes (Drummond et al 2005) Minor changes in extrapolated survival can notably affect cost-effectiveness estimates (Hoyle and Henley 2011) For instance, Figure 14 demonstrates how the same empirical data can be extrapolated using different continuous probability distributions An exponential function assumes a constant hazard over time, while a parametric distribution like Weibull allows for a time-dependent hazard based on all data points Consequently, employing different extrapolation methods can lead to diverse results in cost-effectiveness assessments.
Figure 14 Hypothetical example of data which is extrapolated using different continuous probability distributions: A) exponential distribution and B) Weibull distribution The figure is drafted and presented for illustrative purposes
The application of survival analysis in modeling is frequently influenced by data availability; however, patient-level data is often lacking, as noted by Drummond et al.
In survival analysis, mean or median times to an event are often utilized, assuming a constant hazard, with the choice between them depending on data availability and study objectives (Davies et al 2011) Study V illustrates a method to address the challenge of limited patient-level data by manually tracing published survival curves to extract numerical values for patients at risk or censored at various time points When individual patient data is accessible, it should be used for curve fitting in modeling (Hoyle and Henley 2011) However, even with complete patient-level data, complications in extrapolating survival data may persist, as the selected parametric function, which accounts for varying hazards, can yield different interpretations based on underlying assumptions (Thompson Coon et al.).
In 2010, an example demonstrated the fitting of patient-level data through two different scenarios: one that includes all patient data and another that excludes certain outlier patients While visual inspection of the extrapolated curve is important, statistical methods can also be employed to determine the most accurate fitting curve (Hoyle and Henley, 2011) Additionally, it is crucial to assess whether the distribution-based projections are clinically plausible.
Figure 15 An example of possible variability in fitting the same empirical survival data
Concept of social value in cancer
A treatment can be deemed cost-effective even if its incremental cost-effectiveness ratio (ICER) exceeds the typical willingness to pay threshold, as societal factors influence perceptions of acceptable ICER levels For instance, the National Institute for Health and Clinical Excellence (NICE) in England and Wales permits certain treatments to surpass the standard upper ICER limit of £30,000 per Quality Adjusted Life Year (QALY) for patients with short life expectancies, provided they offer at least three additional months of survival Research indicates that society values health benefits for severely ill patients more than those for less severely ill individuals, leading to varied valuation of treatments across different health conditions.
The social value of the disease may affect what is perceived as an acceptable willingness to pay threshold value (Drummond et al 2009) This is illustrated in Figure 16, where circle
In a typical scenario, represented by circle "A," there is a balance between social value and cost-effectiveness, allowing decisions on acceptable incremental cost-effectiveness ratios to be guided by predetermined willingness-to-pay (WTP) levels As a result, cost-effectiveness assessments are generally well-received by the community However, treatments in circle "B," while potentially cost-effective, may not receive public funding due to their perceived low social value Conversely, circle "C" highlights situations where society is more accepting of lower cost-effectiveness, particularly for life-threatening conditions like certain cancers or rare diseases, where effective alternatives are scarce Drummond et al (2009) emphasized the need for reimbursement agencies to adopt a fair and transparent decision-making process.
Traditional economic measures often overlook key elements of social value, highlighting the necessity for alternative metrics that align equity with efficient resource use In certain contexts, such as rare cancers, the social value can be linked to the anticipated budget impact of treatments Orphan drugs with high Incremental Cost-Effectiveness Ratios (ICER) may receive reimbursement and funding due to their limited financial impact, the unmet medical needs they address, and the significant social value associated with these diseases (Drummond et al 2009).
Figure 16 Relationship between social value of the disease and incremental cost-effectiveness ratio (ICER) A=normal situation, B=low social value, C=high social value (Drummond et al
Quality-adjusted life-year as an outcome measure in cancer
Cancer therapy encompasses both curative treatments and supportive care for end-of-life phases, emphasizing improvements in quality of life and survival (Tannock et al 2011) Extended survival is meaningful only when paired with a good quality of life Quality-adjusted life-years (QALYs) serve as a metric that combines the years of life gained through treatment with the quality of life experienced during those years This measure helps capture differences in tolerability, side effects, and other factors influencing health-related quality of life, independent of life length (Arbuckle et al 2002) Despite its established use, there is ongoing debate about whether QALYs fully reflect the health benefits of cancer treatments.
Garau et al (2011) highlighted the methodological limitations of utilizing Quality-Adjusted Life Years (QALYs) in cancer research, emphasizing that generic health-related quality of life measurement tools may not adequately capture the nuances of health status changes in cancer patients Additionally, they pointed out that variations in utility measurement instruments can lead to discrepancies in the calculated QALYs gained, affecting the overall assessment of treatment effectiveness.
The concept of Quality-Adjusted Life Years (QALY) faces significant challenges, particularly in critical care and end-of-life scenarios Techniques used to estimate health state values, such as the time trade-off (TTO) method, may not adequately reflect the complexities of valuing health conditions near death, as individuals must choose between shorter periods of good health or longer periods of poor health Additionally, using valuations from the general population for serious conditions like cancer raises further concerns These limitations highlight the need for careful consideration when applying QALY in evaluating health outcomes.
Adjusting for health-related quality of life does not significantly impact the cost-effectiveness analysis of cancer treatments, as noted by Tengs (2004) A review by Greenberg and Neumann (2011) of cancer-related cost-effectiveness analyses identified 117 studies published before 2010, revealing that around 60% of cost per QALY and 70% of cost per life-year gained were under the USD50,000 threshold The median costs were USD28,451 per QALY and USD26,568 per life-year gained, while the mean costs were USD112,965 and USD111,172, respectively The findings indicate a strong correlation between cost-effectiveness and cost-utility estimates, suggesting that adjustments for health-related quality of life do not alter decision-making in practice.
6 Pharmacoeconomics of metastatic renal cell carcinoma
Review of cost-effectiveness of sunitinib in metastatic renal cell carcinoma (study I)
Introduction
Renal cell carcinoma (RCC) is the most prevalent type of kidney cancer, representing 2-3% of all malignant diseases in adults While RCC can occur at any age, it is primarily diagnosed in individuals over 60 Globally, there are about 209,000 new RCC cases and 102,000 related deaths annually, with both incidence and mortality rates rising with age Additionally, RCC is notably more common in men than in women.
Renal cell carcinoma (RCC) often lacks early warning signs, with many patients showing no identifiable risk factors (Rini et al 2009) Symptoms can be local, such as hematuria, flank or abdominal pain, and a palpable mass, but these occur together in fewer than 10% of cases Additionally, nonspecific symptoms like fever, nausea, and weight loss may present Small local tumors are usually asymptomatic and are often discovered incidentally during abdominal imaging (Motzer et al 1996, Rini et al 2009) Notably, around 25-30% of RCC patients are diagnosed with metastases.
1996, Gupta et al 2008) RCC-related deaths are predominantly due to metastatic disease (Rini et al 2009)
In recent years, the management of renal cell carcinoma (RCC) has evolved significantly due to a deeper understanding of its biological processes, resulting in improved treatments for metastatic disease (mRCC) (Rini et al 2009) Previously, cytokine-based therapies, such as interferon-α and interleukin-2, were the standard care for mRCC, despite their low response rates of 5 to 20% and ongoing debates about their effectiveness since their introduction in the 1980s (Motzer et al 2007a, Rini et al 2009) Traditional chemotherapy and radiotherapy have limited roles in treating mRCC, primarily serving palliative purposes However, the recent launch of several targeted anticancer drugs has significantly enhanced the management of advanced or metastatic RCC, marking a pivotal shift in treatment options.
The introduction of targeted treatments like sunitinib has significantly influenced treatment recommendations and guidelines for metastatic renal cell carcinoma, as highlighted in the cost-effectiveness analysis by Purmonen (2011) published in the Expert Review of Pharmacoeconomics & Outcomes Research.
Sunitinib is a multitargeted receptor tyrosine kinase inhibitor that plays a crucial role in inhibiting tumor growth, neoangiogenesis, and metastatic progression in cancer The recommended dosage for treating metastatic renal cell carcinoma (mRCC) is 50 mg taken orally once daily for four consecutive weeks, followed by a two-week rest period Initially, sunitinib was introduced as a second-line treatment for patients who were intolerant to or had progressed after cytokine therapy, addressing a significant gap in treatment options at that time Prior to sunitinib, patients had limited active treatment alternatives, often relying on cytokine-based therapies followed by various local treatments, known as best supportive care (BSC) The efficacy of sunitinib in cytokine-refractory mRCC has been evaluated in two phase studies.
Sunitinib has shown significant efficacy in treating metastatic renal cell carcinoma (mRCC), with single-arm trials reporting a median progression-free survival of 8.2 months and a median overall survival of 23.9 months (Motzer et al 2006a, Motzer et al 2006b, Motzer et al 2007b) Furthermore, in a Phase III trial, sunitinib demonstrated improved overall survival of 24.4 months compared to 21.8 months, along with enhanced progression-free survival, solidifying its role as a first-line treatment option for mRCC.
(11 vs 5 months) compared to interferon-alpha (IFN-α) (Motzer et al 2007a, Motzer et al
Currently, sunitinib is considered as the new standard of care for the 1 st line treatment of mRCC, in patients with favorable or intermediate prognostic risk (Escudier and Kataja
In the treatment of poor-risk patients with metastatic renal cell carcinoma (mRCC), temsirolimus is recommended as the first-line therapy, with sunitinib as an alternative option For patients who progress after first-line targeted treatment, everolimus is suggested as the next treatment choice Other targeted agents, including sorafenib and bevacizumab (when combined with IFN-α), are also approved for mRCC treatment Additionally, numerous other agents are currently being investigated for their efficacy in this area.
The emergence of new and costly treatments has heightened the necessity for economic evaluations to ensure optimal allocation of limited healthcare resources This study aims to identify all research related to the cost-effectiveness of sunitinib in treating advanced or metastatic renal cell carcinoma (mRCC) The focus of this review is on the cost-effectiveness aspects, while a comprehensive overview of sunitinib in mRCC management can be found in other literature, such as Escudier (2010).
Figure 17 Incidence and mortality per 1 million inhabitants in different age groups among
Finnish RCC patients Mean values from the 10-year period 1995-2004 RCC is defined as cancer of the kidney, excluding cancer of the renal pelvis (Finnish Cancer Registry 2006).
Review of literature
A comprehensive literature search was performed to identify published research articles and congress abstracts related to sunitinib (Sutent) and its economic evaluation in renal cell carcinoma (RCC) and metastatic RCC (mRCC) This search utilized multiple databases, including PubMed, ISI Web of Science, CINAHL, and various health technology assessment registries, without restrictions on publication date or language Key search terms included cost, economic evaluation, utility, QALY, and cost-effectiveness Additionally, abstracts from the International Society for Pharmacoeconomics and Outcomes Research (ISPOR) and the American Society of Clinical Oncology (ASCO) were directly accessed While manufacturer reports and national agency evaluations were not included, formal HTA reports found in the databases were considered Letters, editorials, and comments were excluded from the search results.
In December 2010, a literature search yielded 421 results, leading to the identification of 58 potential titles after initial exclusions Following the removal of duplicates, the final selection comprised 26 congress abstracts, 4 research articles, and one HTA report, which provided evidence for technology appraisals by the National Institute for Health and Clinical Excellence (NICE TA169, NICE TA178) To prevent duplicate reporting, these results were excluded from the current study Additionally, a review on economic evaluations of various mRCC treatments, including sunitinib, was identified (Norum et al.).
In 2011, a comprehensive search was conducted, revealing one new article, while one abstract was excluded due to inaccessibility (Négrier et al 2007) and another for focusing solely on cost without an economic evaluation A manual review of reference lists for all articles and the HTA report was performed, but it yielded no additional abstracts or articles that met the inclusion criteria, including a review article that was ultimately excluded (Norum et al 2010).
Figure 18 Literature search flow chart
A systematic collection of data was conducted from the included studies, focusing on incremental costs, outcomes, and incremental cost-effectiveness ratios (ICER), with efforts made to calculate any missing values for this review Most of the studies analyzed pertained to first-line treatments, with separate sections dedicated to presenting the results of both first and second-line treatments.
Between 2007 and 2010, a significant number of congress abstracts were presented at various meetings, with 16 out of 24 showcased at ISPOR events Additionally, five abstracts were featured at ASCO meetings during the same period, and three were presented at the ESMO 2008 conference Notably, one abstract from the ECCO 2007 meeting remains inaccessible.
2007) From the 16 included ISPOR abstracts, the related congress posters were available in
The analysis included 11 cases sourced from either the congressional website or the authors' own archives Posters were instrumental in providing supplementary information and evaluating the types of cost-effectiveness models used When available, posters served as the primary data source, particularly in instances where discrepancies arose between the abstract and the poster content.
All five articles and 21 out of 24 abstracts disclosed industry funding or employment among authors Notably, one article was sponsored by the manufacturer of temsirolimus, while the others were funded by the sunitinib manufacturer Additionally, one abstract did not report any funding source but utilized the same model as those in manufacturer-sponsored studies.
The methodological quality of the studies included in this review was not assessed, with the author solely responsible for the review process, data extraction, and evaluation of published articles and abstracts/posters Financial results are reported in Euros without any adjustments, using exchange rates from the European Central Bank as of January 18, 2011, which are USD 1.3371, GBP 0.83565, SEK 8.9203, CAD 1.3182, THB 40.768, ILS 4.725, COP 2499.31, and ARS 5.319.
Cost-effectiveness of sunitinib in the first-line treatment of mRCC
6.1.3.1 Methodology and models used in the first-line economic evaluations
The cost-effectiveness of first-line sunitinib for metastatic renal cell carcinoma (mRCC) has been evaluated in three research articles and 20 congress abstracts A Health Technology Assessment (HTA) conducted by the Peninsula Technology Assessment Group (PenTAG) in 2010 specifically assessed sunitinib's effectiveness in mRCC treatment This report included PenTAG's cost-effectiveness estimates, alongside results from the manufacturer's submission to the National Institute for Health and Clinical Excellence (NICE) The estimates provided by PenTAG were derived from their own model as well as the manufacturer's model, which was subsequently adjusted by the assessment group.
All model-based studies employed a Markov model to simulate disease progression and evaluate survival and cost outcomes for a hypothetical cohort of metastatic renal cell carcinoma (mRCC) patients The cost-effectiveness models typically featured three or four health stages, allowing patient transitions throughout the modeled period The most common structure included four mutually exclusive health stages: First-line treatment until progression, Second-line (composite), Best supportive care, and Death (from cancer or other causes), with consistent stage names across first-line articles An alternative Markov model structure with three stages—Progression-free survival, Progressed disease, and Death—was utilized by PenTAG in their estimations Notably, only one study deviated from the model-based approach.
Out of 20 congress abstracts on first-line mRCC treatment, 16 utilized similar model structures detailed in full articles (Remák et al 2008, Chabot and Rocchi 2010, Benedict et al 2011) While these studies showed model similarities, authors may have modified the original cost-effectiveness model to meet national requirements The model, developed in MS Excel, was designed for global applicability, allowing for adaptation to different national health systems (Remák et al 2007a).
Markov models allow patients to occupy a single health stage at any given time, with specific probabilities for transitioning between stages during each cycle Most studies analyzed had a modeled time horizon of 10 years, although some varied between 1 and 10 years, with a consistent cycle length of 6 weeks, reflecting the dosing schedule of sunitinib (4 weeks of treatment followed by a 2-week break) In these models, the population's costs and outcomes are determined by their health stages, with per-cycle costs influenced by healthcare resource utilization and country-specific unit costs Variations in resource use and costs across countries pose challenges for localizing global cost-effectiveness models Expert opinion was the primary source for country-specific resource use in the studies, while some collected data from local patients and a few utilized databases for resource information.
Published clinical trials were used as the primary source of effectiveness in nearly all the studies Phase III trial comparing sunitinib and IFN-α (Motzer et al 2007a, Motzer et al
In 2009, numerous studies highlighted the effectiveness of indirect comparisons for evaluating health outcomes between various substances when direct comparisons are lacking This method enables researchers to infer relationships between substances that have not been tested against one another in the same clinical trial A comprehensive discussion on indirect comparisons can be found in a full-text article by Benedict et al.
In 2011, several treatment options for first-line metastatic renal cell carcinoma (mRCC) were simultaneously compared, using IFN-α as the common comparator (Benedict et al 2011) This approach was also adopted in 13 abstracts, including studies by Remák et al (2008) and Chabot and Rocchi (2010), which assessed the cost-effectiveness of sunitinib versus IFN-α through head-to-head trials to determine treatment efficacy Additionally, three abstracts employed similar comparators and methodologies.
Studies on second-line treatments for cancer have shown varied approaches to incorporating active therapies Chabot and Rocchi (2010) emphasized the need to exclude subsequent tyrosine kinase inhibitors to accurately assess the incremental effect of sunitinib over IFN-α In contrast, Remák et al (2008) adopted a composite second-line treatment strategy that included sunitinib, sorafenib, IL-2, and IFN-α, with two-thirds of patients receiving this treatment after first-line sunitinib Similarly, Benedict et al (2011) reported that 80% of US patients were assumed to receive active second-line treatment regardless of their initial therapy, while only 30% of Swedish patients received second-line treatment after sunitinib The variation in these methodologies highlights the differences in how second-line treatments are integrated across studies Notably, in the NICE manufacturer submission, the analysis did not include second-line drugs, focusing instead on the progression of patients receiving sunitinib or IFN-α followed by best supportive care (Thompson Coon et al 2010).
6.1.3.2 Results from the published cost-effectiveness analyses (first line)
Between 2007 and 2011, multiple studies evaluated the cost-effectiveness of first-line sunitinib, calculating incremental cost-effectiveness ratios (ICER) in terms of cost per life-years (LY), progression-free life-years (PFLY), progression-free months (PFM), and quality-adjusted life-years (QALY) gained ICER provides insight into the additional financial investment required for each unit of health benefit achieved.
The cost-effectiveness of first-line sunitinib has been evaluated in comparison to interferon-alpha (IFN-α), interleukin-2 (IL-2), sorafenib (SFN), temsirolimus (TMS), and the combination of bevacizumab and IFN-α (BEV/IFN) A summary of the findings from these cost-effectiveness analyses is detailed in Table 3.
Incremental cost-effectiveness ratios for 1 st line sunitinib ranged from €4,786 to
The cost-effectiveness analysis revealed that sunitinib had an incremental cost-effectiveness ratio (ICER) of €109,416 per quality-adjusted life year (QALY) and ranged from €33,807 to €100,212 per life year gained (LYG) compared to IFN-α Notably, sunitinib only dominated IFN-α in a single study with a one-year perspective From a US societal viewpoint, excluding indirect costs, the ICER for sunitinib was found to be €13,920 per progression-free life year (PFLY).
Sunitinib has been shown to have an incremental cost-effectiveness ratio (ICER) of €50,269 per life year gained (LYG) and €39,333 per quality-adjusted life year (QALY) compared to interferon-alpha (IFN-α), with a 46% probability of being cost-effective at a willingness to pay (WTP) threshold of approximately €37,400, increasing to 65% at around €74,800 (Remák et al 2008) Additional ICERs reported in the manufacturer's submission to NICE included €25,269/LYG, €54,731 per progression-free life year (PFLY), and €34,160/QALY, indicating a 54% probability of cost-effectiveness at a WTP of £30,000/QALY (~€35,900) (Thompson Coon et al 2010) However, when the model was adjusted by PenTAG, the ICER rose to between €57,500 and €67,000/QALY (Thompson Coon et al 2010) Furthermore, a Canadian study reported ICERs of €100,100/PFLY and €79,525/LYG.
A comparison of sunitinib with IFN-α revealed an incremental cost-effectiveness ratio (ICER) of €109,416 per QALY (Chabot and Rocchi 2010) Additionally, a study conducted from the healthcare provider perspective in Thailand found an ICER of M3.6 Baht per QALY (€88,305) for sunitinib compared to IFN-α, indicating that this cost exceeds the acceptable ICER threshold for a developing country (Topibulpong et al.).
In the realm of renal cell carcinoma (mRCC) treatment, sunitinib has shown a significant cost-effectiveness advantage over other therapies Studies indicate that combination therapy with bevacizumab and IFN-α was consistently favored over sunitinib, with a 99% probability of sunitinib being cost-effective from a Swedish healthcare payer perspective at a willingness to pay (WTP) of SEK500,000 (~€56,050) Sunitinib outperformed sorafenib in seven studies, demonstrating both greater efficacy and lower costs, while in another six studies, it was more costly yet more effective, with incremental cost-effectiveness ratios (ICER) ranging from €24,149 to €52,036 In the U.S., the probability of sunitinib being a cost-effective choice compared to sorafenib and BEV/IFN was 74% at a WTP of USD100,000 (~€74,800) Conversely, temsirolimus was found to be a dominated treatment option against sunitinib in standard mRCC populations, and in studies involving poor prognosis patients, sunitinib was either more effective and less costly or more costly and more effective, with ICER values reported Notably, the cost for an additional QALY gained with temsirolimus compared to sunitinib in poor-risk patients was €21,783, highlighting the complexities of treatment efficacy and cost in mRCC management.
Table 3 Cost-effectiveness studies of first-line sunitinib treatment
Author Year Country Comparator ICER (€/LYG) ICER (€/PFLY) ICER (€/QALY)
Benedict et al 2011 USA SFN domin domin domin
BEV/IFN domin domin domin
Benedict et al 2011 Sweden BEV/IFN domin domin domin
Chabot and Rocchi 2010 Canada IFN-α € 79,525 € 100,100* € 109,416
Thompson Coon et al ** 2010 UK IFN-α € 70,181 N/A € 85,517
Remák et al 2008 USA IFN-α € 50,269 € 13,920 € 39,333
Benedict et al 2009 USA SFN N/A domin domin
Benedict et al 2008a USA SFN € 43,037 € 20,941 € 44,877
BEV/IFN domin domin domin
Benedict et al 2008b USA SFN € 40,229 € 27,201 € 49,121
BEV/IFN domin domin domin
Remák et al 2007b USA IFN-α (model 1) € 50,269 N/A € 39,333
Remák et al 2007a USA IFN-α € 50,269 € 13,920 € 39,333
Remák et al 2009 Sweden BEV/IFN domin domin domin
Sandin et al 2008 Sweden SFN € 19,938 € 13,483 € 24,149
BEV/IFN domin domin domin
Munir et al 2008 Sweden SFN € 19,938 € 13,483 € 24,149
BEV/IFN domin domin domin
Calvo et al 2010 Spain SFN domin domin domin
BEV/IFN domin domin domin
Diaz et al 2008 Spain SFN € 22,577 € 11,398 € 24,272
BEV/IFN domin domin domin
Godoy et al 2009a Colombia IFN-α € 40,211 € 29,110* N/A
Godoy et al 2009b Colombia IFN-α € 40,211 N/A N/A
Caceres et al 2008 Colombia IFN-α (1y)# domin domin domin
Salinas-Escuerdo et al 2009 Mexico SFN domin domin domin
BEV/IFN domin domin domin
Mould-Quevedo et al 2009 Mexico IFN-α € 50,865 € 33,807 N/A
Teich et al 2010 Brazil IFN-α ICER positive+ ICER positive+ N/A
Topibulpong et al 2010 Thailand IFN-α € 66,228 € 49,058 € 88,305
Ondrackova and Demlova 2010 Czech Rep IFN-α € 100,212 € 88,651 N/A
Silverio et al 2009 Portugal TMS (p.r.) N/A N/A €21,783***
Interferon-alpha (IFN-α) and Interleukin-2 (IL-2) are key components in cancer treatment, often used in combination therapies such as sorafenib (SFN) with bevacizumab (BEV) and IFN-α These therapies aim to improve clinical outcomes, measured in life-years gained (LYG), progression-free life-years (PFLY), and quality-adjusted life-years (QALY) Notably, the effectiveness of these treatments can vary, particularly in patients with poor prognosis (p.r.) Understanding these metrics is essential for evaluating the impact of therapies like temsirolimus (TMS) on patient outcomes.
According to published values, TMS is both more effective and more costly compared to sunitinib, while sunitinib is identified as more effective and less costly in certain assessments The analysis also indicates that different time horizons impact the cost-effectiveness outcomes, with sunitinib dominating in scenarios where it offers greater effectiveness at a lower cost.
Cost-effectiveness of sunitinib in the second-line treatment of mRCC
6.1.4.1 Methodology and models used in second-line studies
The literature review revealed five abstracts and two full articles focusing on second-line sunitinib treatment for metastatic renal cell carcinoma (mRCC) Additionally, the included Health Technology Assessment (HTA) report presented cost-effectiveness analyses for second-line treatments submitted to NICE by the manufacturer (Thompson Coon et al., 2010) All second-line studies employing Markov models reported a consistent structure comprising three health states: no new progressions, new progressions, and death (Garduủo-Espinosa et al., 2007; Purmonen et al., 2007; van Nooten et al.).
In 2007, several studies, including those by Paz-Ares et al (2010), utilized a model developed and tested in the USA, with similar evaluation models likely employed in at least two other abstracts (Garduño-Espinosa et al 2007, van Nooten et al 2007) Finnish studies (2nd line) constructed a Markov model using WinBUGS software (Purmonen et al 2007, study II) All model-based studies adopted a lifetime horizon of either 5 or 10 years, with one-month model cycles Notably, one study did not involve a model-based evaluation (Ondrakova and Demlova 2010).
Published clinical trials were utilized to evaluate the efficacy of sunitinib, primarily through single-arm Phase II trials focused on its safety and effectiveness in second-line treatment (Motzer et al 2006a, 2006b, 2007b) However, there is a notable absence of head-to-head trials comparing sunitinib in the second-line setting Consequently, the effectiveness of the comparator, best supportive care (BSC), relied on local patient data, literature, or databases Three studies gathered local patient data to inform current treatment practices, while others sourced comparator data from literature or various databases A key concern in these studies was the comparability of patient populations between clinical trials and real-world clinical practice Additionally, four studies reported incorporating resource use information from a panel of clinical experts.
6.1.4.2 Results from the published cost-effectiveness analyses (second line)
Second-line treatment is initiated for patients who are intolerant to or have not responded to first-line cytokine-based therapies, such as IFN-α or IL-2 Sunitinib emerged as one of the first targeted therapies approved for second-line treatment The cost-effectiveness of second-line sunitinib therapy has been analyzed in five abstracts and two articles published between 2007 and 2010, with best supportive care (BSC) used as the comparator in most studies, reflecting the local treatment practices of that time While BSC was only formally defined in the full articles, it was described as either "current clinical practice including bio-chemo-therapy" or "palliative care without chemotherapy" in the relevant studies.
The evaluation of cost-effectiveness from the healthcare payer perspective in Finland and Spain revealed significant findings In Spain, the Incremental Cost-Effectiveness Ratio (ICER) was €34,196 per Quality-Adjusted Life Year (QALY), with a 95% probability of being cost-effective at a willingness to pay (WTP) threshold of €45,000 (Paz-Ares et al 2010) In Finland, the ICER was reported at €43,698/QALY, demonstrating a 70% probability of cost-effectiveness (study II) Preliminary findings from the Finnish study were presented as an abstract (Purmonen et al 2007) before the full text publication The included research abstracts indicated that the ICER for progression-free life-years ranged from €20,000 to €92,000 (n=5) when sunitinib was compared to best supportive care (BSC) Additionally, the extra cost per QALY gained varied from €10,000 to €43,000 (n=3) In the manufacturer's submission to NICE, the ICERs were approximately €34,800 per life-year gained (LYG).
The cost-effectiveness analysis revealed that sunitinib had a 36% probability of being cost-effective compared to best supportive care (BSC) at a willingness-to-pay (WTP) threshold of €36,000 per quality-adjusted life year (QALY), with an incremental cost-effectiveness ratio (ICER) of €44,900/QALY (Thompson Coon et al 2010) In a non-model-based study, sorafenib was evaluated against a combination of sunitinib (70%) and BSC (30%), showing that the combination was both more effective and more expensive than sorafenib in second-line treatment, resulting in an ICER of €20,000 per progression-free life-year (Ondrakova and Demlova 2010) The findings from the second-line studies are summarized in Table 4, while the incremental costs and QALYs from various jurisdictions are depicted in a cost-effectiveness plane (Figure 20), demonstrating that sunitinib consistently provided additional benefits at higher costs compared to BSC across all studies.
2 nd line treatment of mRCC
Table 4 Cost-effectiveness studies of second-line sunitinib treatment after cytokine failure
Author Year Country Comparator ICER
Paz-Arez et al 2010 Spain BSC € 25,199 € 72,876 € 34,196
Ondrackova and Demlova 2010 Czech Rep SFN** N/A € 19,878 N/A
Aiello et al 2007 Argentina BSC € 7,430 € 21,650 € 10,048
Garduủo-Espinosa et al 2007 Mexico BSC N/A € 54,550* € 26,354 Purmonen et al 2007 Finland BSC € 30,014* € 57,719* € 42,877 van Nooten et al 2007 Belgium BSC € 35,389 € 91,980 N/A
Supportive care in the context of BSC is highlighted, with specific references to treatment options like sorafenib (SFN) and sunitinib, which comprises 70% of the treatment arm alongside 30% BSC Key metrics such as life-year gained (LYG), progression-free life-year (PFLY), and quality-adjusted life-year (QALY) are discussed, with some values estimated from published data.
Cost-effectiveness analyses of sunitinib versus best supportive care (BSC) in the second-line treatment of metastatic renal cell carcinoma (mRCC) reveal significant findings The analysis includes five articles and abstracts that present incremental costs and quality-adjusted life years (QALYs), along with one submission from the manufacturer to the National Institute for Health and Clinical Excellence The geographical regions considered in the analysis are the north-east, north-west, south-east, and south-west.
Discussion
Sunitinib is one of the first targeted anticancer agents used for metastatic renal cell carcinoma (mRCC), and its cost-effectiveness has been evaluated in various studies In first-line treatment, sunitinib demonstrated superior effectiveness compared to multiple comparators, often resulting in cost-savings or additional costs, except in one study where it was less effective and less costly than temsirolimus in a poor-risk population For second-line treatment of cytokine-refractory mRCC, sunitinib was found to be more costly yet more effective compared to best supportive care The incremental cost-effectiveness ratio (ICER) for sunitinib generally fell within an acceptable range, indicating its cost-effective nature However, cost-effectiveness decisions must consider the willingness to pay thresholds in different regions The National Institute for Health and Clinical Excellence (NICE) has endorsed sunitinib as a first-line treatment option for mRCC (NICE TA169, NICE TA178).
The studies reviewed, published between 2007 and 2011, comprised 5 full-text articles, 24 research abstracts, and a health technology assessment (HTA) report Previous reviews have captured only a fraction of the studies identified in this review, which included a comprehensive HTA on the cost-effectiveness of mRCC treatments (Thompson Coon et al 2010) However, this HTA was limited by its literature search from 2007 and 2008, identifying only 3 sunitinib abstracts Another review conducted in 2008 and 2009 (Norum et al 2010) found 1 abstract, 2 articles, and an evaluation report on sunitinib's cost-effectiveness A more recent review published during this article's preparation identified 3 additional articles related to sunitinib's cost-effectiveness in RCC treatment (Shih et al 2011), highlighting the challenges in publishing timely reviews in a rapidly evolving field Despite these findings, the current review faced methodological limitations, including the lack of formal quality evaluation of the included studies and the absence of independent reviewers, which is a standard practice for systematic reviews Additionally, the presence of duplicate results among the articles and abstracts suggests that the actual number of unique studies may be lower than reported (Centre for Reviews and Dissemination 2009, Higgins and Green 2011).
The review highlights that many studies rely on similar cost-effectiveness models developed by the pharmaceutical industry for use across various jurisdictions, often localized by adjusting resource use and unit costs This practice introduces bias, as independent studies using the same models yield comparable outcomes Consequently, the absence of variation from different modeling approaches means that an increased number of studies does not necessarily enhance the evidence's strength Therefore, it is recommended that articles employing global models disclose any prior publications based on the same models and clearly state if similar results have been previously presented.
Health care resource utilization can be assessed through various methods, including expert opinion, registers, local patient data, and literature When adapting cost-effectiveness models for different countries, local clinical experts typically define the resource use associated with the treatments being compared Expert opinion, despite its limitations, often serves as the best available information for estimating the cost-effectiveness of novel products To improve transparency, it is crucial to report estimated resource use throughout different treatment phases Notably, assumptions regarding resource use can significantly influence cost-effectiveness results, which can be analyzed through sensitivity analyses For instance, a study by Remák et al (2008) highlighted that the cost of best supportive care (BSC) was a highly sensitive parameter, with a 20% cost reduction leading to a doubling of incremental costs per QALY gained The cost of BSC was reported as USD16,011 per 6-week cycle, including multiple CT scans In contrast, Benedict et al (2011) reported BSC costs of USD19,467 in the USA and USD2,775 in Sweden, attributing the differences to variations in unit costs and hospitalization rates In their analysis, sunitinib remained dominant even with a 20% reduction in BSC costs Additionally, in Finland, the expected cost of symptomatic care following IFN-α treatment was estimated at €1,500 per month, based on data from 81 mRCC patients.
Patients treated with sunitinib have demonstrated a more favorable health-related quality of life compared to those receiving IFN-α, as indicated by studies utilizing the EQ-5D quality of life instrument This instrument was consistently referenced in full-text articles and the majority of abstracts, suggesting its widespread use in QALY estimations across various studies Consequently, the QALYs reported should be comparable; however, variations in healthcare systems, patient demographics, treatment protocols, and methodological assumptions, such as discount rates, may still impact the overall comparability of these findings.
Timely economic evaluations of novel cancer treatments face significant challenges due to insufficient information on resource utilization and costs, limited clinical evidence, and short follow-up periods The varying adverse event (AE) profiles of targeted treatments highlight the need for a focus on the cost differences associated with managing these AEs A comprehensive cost-effectiveness analysis must incorporate all relevant cost parameters, including those related to adverse event management As new treatment options for metastatic renal cell carcinoma (mRCC) emerge, the importance of AE profiles will likely increase Indirect comparisons of treatment efficacy among different studies have shown variability and conflicting results due to differing methodologies Consequently, uncertainty remains a key factor in cost-effectiveness estimations.
Most studies assessed uncertainty through probabilistic sensitivity analysis, highlighting the need for more accurate estimates of treatment costs and disease burden to improve decision-making and cost-effectiveness evaluations Additionally, evaluating the budgetary impact of new treatments is essential for informing decision-makers about the net costs associated with adoption; however, such analyses are infrequently found in published research The rapid advancement of treatments and the sequencing of targeted therapies may significantly affect both treatment costs and patient survival Therefore, a thorough assessment of the costs and benefits of emerging treatments is crucial before integrating them into clinical practice.
The management of metastatic renal cell carcinoma (mRCC) is evolving with the introduction of new targeted anticancer agents Sunitinib is expected to remain a first-line treatment option, while additional targeted therapies will emerge for subsequent treatment lines However, there is currently insufficient knowledge regarding the optimal treatment strategy As treatment costs and patient numbers rise, cost-effectiveness will become increasingly significant Sunitinib will serve as the standard for comparison with new therapies, leading to a series of cost-effectiveness analyses beyond this review.