Scholarly article on topic 'Cost effectiveness in practice and its effect on clinical outcomes'

Cost effectiveness in practice and its effect on clinical outcomes Academic research paper on "Economics and business"

CC BY-NC-ND
0
0
Share paper
Academic journal
Journal of Cancer Policy
OECD Field of science
Keywords
{Cost-effectiveness / Reimbursement / Access / Outcome}

Abstract of research paper on Economics and business, author of scientific article — Bengt Jönsson, Scott Ramsey, Nils Wilking

Abstract The value of new cancer drugs is maximized when they are used for the right patient in the right way in clinical practice. Clinical trials conducted during drug development are the most important source of information to predict value at the time a drug is introduced in practice. Regulatory approval is an indication of value, which lately has been complemented with an assessment of clinical value for decisions about reimbursement, using the methodology of health technology assessment (HTA). Formal cost-effectiveness studies are an important part of this methodology, aiming to assist decisions about value for money in health care spending. The question is if the addition of a complementary HTA and cost-effectiveness study increases the value realized by the drug in practice compared to how it would be used without these assessments. We review the issues involved in providing an answer by using the introduction of new targeted therapies for metastatic renal cell cancer (mRCC). Specifically, we examine the link between clinical trial data and estimations of cost-effectiveness at drug launch, reimbursement decisions, uptake and use in different countries and evidence about impact on outcome in patient populations for which the new drugs are indicated. We conclude that there is a weak link between the assessments used at drug launch and the value created in clinical practice. We suggest measures that are necessary for the achievement of evidence-based and cost-effective cancer care in clinical practice.

Academic research paper on topic "Cost effectiveness in practice and its effect on clinical outcomes"

Contents lists available at ScienceDirect

Journal of Cancer Policy

journal homepage www.elsevier.com/locate/jcpo

Cost effectiveness in practice and its effect on clinical outcomes*

Bengt Jdnsson3'*, Scott Ramseyb, Nils Wilkingcd

a Stockholm School of Economics, Sweden

b Fred Hutchinson Cancer Research Center, 1100 Fairview Avenue North, M3-B232, Seattle, WA 98109-1024, United States c Karolinska Institutet, Stockholm, Sweden d Skane Oncology Clinic, Lund, Malmo, Sweden

ABSTRACT

The value of new cancer drugs is maximized when they are used for the right patient in the right way in clinical practice. Clinical trials conducted during drug development are the most important source of information to predict value at the time a drug is introduced in practice. Regulatory approval is an indication of value, which lately has been complemented with an assessment of clinical value for decisions about reimbursement, using the methodology of health technology assessment (HTA). Formal cost-effectiveness studies are an important part of this methodology, aiming to assist decisions about value for money in health care spending. The question is if the addition of a complementary HTA and cost-effectiveness study increases the value realized by the drug in practice compared to how it would be used without these assessments.

We review the issues involved in providing an answer by using the introduction of new targeted therapies for metastatic renal cell cancer (mRCC). Specifically, we examine the link between clinical trial data and estimations of cost-effectiveness at drug launch, reimbursement decisions, uptake and use in different countries and evidence about impact on outcome in patient populations for which the new drugs are indicated.

We conclude that there is a weak link between the assessments used at drug launch and the value created in clinical practice. We suggest measures that are necessary for the achievement of evidence-based and cost-effective cancer care in clinical practice.

© 2014 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND

license (http://creativecommons.org/licenses/by-nc-nd/3.0/).

ARTICLE INFO

Article history:

Available online 26 February 2014

Keywords:

Cost-effectiveness

Reimbursement

Access

Outcome

Introduction

Cost effectiveness studies are increasingly used for decisions about reimbursement and appropriate use of new cancer drugs. These studies thus affect access for patients and clinicians to treatments that may be of potential value, and direct the efficient and equitable use of health care resources for cancer.

While the general methodology of such studies is well developed, and applied in a similar fashion in different jurisdictions, the interpretation and use of the findings differ. This is not surprising, since the relative value of a new medicine, reflected in willingness as well as ability to pay for new treatment options, will differ between countries. This results in great variations in the use of new cancer drugs, and most striking is the impact of the per capita income level on the uptake of new therapies [1].

* This paper is part of the Special Issue 'Delivering Affordable Cancer Care in High Income Countries: Papers from a Special Session of Oncology At The Limits, 13th -15th February 2014'.

* Corresponding author. Tel.: +46723985678. E-mail address: hebj@hhs.se (B. Jonsson).

However, the variations in use of new cancer treatments cannot be explained solely by differences in income levels and health care spending. There is substantial variability in the use of new cancer drugs between countries with similar income per capita. These differences can to some extent be explained by differences in coverage decisions [2].

But decisions about coverage do not directly translate into uptake and use of new cancer drugs. National institute for health and care excellence (NICE) in the UK has made 151 individual recommendations on cancer drugs of which 57% have been positive [3]. Still, the actual use of cancer drugs in UK with NICE endorsements is rather low, suggesting a disconnect between recommendations and actual use [4]. The disconnect between coverage decisions is also present in Sweden, where decisions by TLV, The Dental and Pharmaceutical Benefits Agency, is not directly linked to spending in different county councils, resulting in great regional variations in use of new cancer drugs [5].

This paper will review the link between clinical trial data, the cost-effectiveness studies for new cancer drugs, coverage decisions informed by such studies, uptake and use of the drugs, and evidence of impact on outcome in the relevant patient population. The

http://dx.doi.org/10.1016/jjcpo.2014.02.001

2213-5383/© 2014 The Authors. Published by Elsevier Ltd. This is an open access article underthe CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/3.0/).

Evidence from clinical trials

Evidence on outcome in clinical practice

Estimation of cost-effectiveness

Uptake and use in clinical practice

Decisions about reimbursment

6 weeks, more than most patients can afford to pay. Decisions about their use in clinical practice would thus benefit from information about cost-effectiveness [11].

Pazopanib was approved 2009 in the US, and got conditional approval by EMA in 2011 for first line treatment. In a recent direct comparison, pazopanib was shown to be non-inferior to sunitinib [12]. Axitinib is the latest approval in 2012 for RCC, after failure of prior systemic therapy with a cytokine or sunitinib.

The clinical trials and the decisions by FDA and EMA on market authorization indicate survival benefits from the use of the new, targeted therapies. But since the primary clinical end-point in most of the trials has been progression free survival (PFS), and patients who fail on the comparator medicine (or placebo) cross over to the evaluated drug, direct estimates of the magnitude of the overall survival (OS) benefit of these agents are lacking.

Fig. 1. Framework for the analysis.

inclusion of impact on outcome in the analysis reflects the shift in focus from efficacy to relative effectiveness in assessing the value of new medicines [6]. Value for patients is created when anticancer drugs are chosen that provide maximum difference in the chance of benefit and harm, the overall objective for personalized cancer medicine [7]. The framework for the analysis is illustrated in Fig. 1.

Data from clinical trials is the key source for estimation of effectiveness in cost-effectiveness studies. Such studies are used to inform decisions about reimbursement, which aim at guiding the use of the drug in clinical practice. The use in clinical practice will determine outcome and thus the value of the drug. Use in clinical practice may reveal evidence gaps that need new clinical trials to be answered.

We will use data from the introduction of newer targeted drugs to treat mRCC to illustrate the methodological issues and data needs for assessing cost-effectiveness and impact on use and outcome. The paper will conclude with a discussion of the policy implications of the greater emphasis on relative effectiveness and cost-effectiveness in clinical practice as a key goal for health policy.

Renal cell carcinoma and new targeted therapies

Kidney cancer accounts for about 2-3 per cent of all cancers, and the annual global incidence is about 340,000 cases. The incidence is highest in the US and Europe, and low in Africa and Asia. Renal cell carcinoma accounts for about 80% of all kidney cancers. For localized disease, which can be treated with surgery, the prognosis is good. For advanced or metastatic disease the 5 year survival rate is only about 10%, and the annual number of deaths in kidney cancer is 143,000, of which 53,000 is in Europe and 28,000 in the US. Incidence and mortality is twice as high in men as in women, and varies with a factor three between countries in Europe [8].

A number of targeted therapies for mRCC have been introduced since 2005 (Table 1).

Most of the new treatments target tyrosine kinase pathways and block angiogenesis, the formation of new blood vessels, necessary for tumour growth. The arrival of sorafinib and sunitinib in 2005-6 followed by other targeted drugs represents a major step forward compared to the previous standard treatment of interferon-alpha and interleukin-2. For patients with a good Karnofsky performance status and intact organ function, high-dose interleukin-2 has often been used in first line because it can induce durable long-term remissions in approximately 10 per cent of patients. The new drugs offered benefits in terms of both oral administration and improved outcome, but also potentially increased costs. The cost of treatment with the new targeted therapies are in the range of UKP 3-6000 per

Cost-effectiveness studies of new therapies for mRCC

The new, targeted therapies in mRCC are also interesting from an economic perspective due to the challenges offered from their pricing. The costs of therapy range from 5 to 10,000 USD per month, and 100,000 USD or more over lifetime, have been a challenge for reimbursement agencies making recommendations on coverage based on value for money, supported by estimates of cost-effectiveness. This challenge has a methodological aspect, the estimation of cost-effectiveness, and an ethical aspect, the tradeoff between equity and efficiency in making decisions based on value.

Assessment of cost-effectiveness aims at producing estimates of the likely costs and improvement in health from the use of the drugs for defined populations in clinical practice, compared to the present standard of care. Results from clinical trials can be used for modelling the likely consequences, but such predictions will incorporate all the uncertainties in the clinical data as well as uncertainty due to assumptions made in extrapolating these data. Additional issues involved in modelling are the definition of the relevant patient population and the relevant standard of care for comparison, taking into account that the drugs will be used differently in clinical practice than in the clinical trial. An economic evaluation also demands a translation of the gain in median PFS to gain in mean OS, adjusted for quality of life, in order to provide an estimate of the gain in quality-adjusted life years (QALY). The extrapolation from median to mean survival requires estimating the entire survival curve, rather than only the portion above 50%. The tail of the curve often cannot be observed directly, and since it contains observations with long survival and thus a large impact on the mean, estimates of mean survival tend to be uncertain. The adjustment for quality of life has two effects. It incorporates gains in quality of life from increasing the time to progression, but also reduces the value of gains in survival that are spent in states with less than full quality of life (Fig. 2).

Economic evaluations draw upon data from a variety of sources: clinical, epidemiologic, and economic, and the uncertainty around the estimates of cost-effectiveness are often large. It is thus not surprising that there are controversies around both data and methodology for such studies. Since decisions have to be made, the issue is not about the level of uncertainty of the result, but rather if decisions can be improved with, rather than without, economic evaluations.

A review of the cost-effectiveness studies of sunitinib can illustrate the challenges of undertaking cost-effectiveness studies, and the conclusions that have been made [13]. The reviewed studies were published during 2007-2011, including five full text articles, 24 research abstracts and one HTA report.

Table 1

New targeted drugs for mRCC.

FDA approval

EMA approval

Improvement in PSFa

Comments

Sorafinib

Sunitinib

Temsirolimus

Bevacizumab

Pazopanib

Everolimus

Axitinib

2006 2007

2009(RCC)

2006 2007

2009 (RCC)

PFS compared with IFNa (5.7 versus 5.6 months)

when used first-lineb Second line after failure of

IFNa median PFS (5.5 versus. 2.8 months)

PFS versus INF (11versus 5 months)

PFS versus INF (3.8 versus 1.9 months)

PFS versus INF (10.2 versus 5.4 months)

PFS versus INF (9.2 versus 4.2 months)

PFS (4.9 versus 1.9 months)

PFS versus sorafinib (6.7 versus 4.7 months)

Liver cancer Orphan drug

Also GIST

Iv adm. And lymphoma Orphan drug Iv adm. Several other indications

Also breast and pancreas cancer second line Second line only in mRCC. In EU only post sunitinib/or interleukin

a [9]. b [10].

Nearly all studies used a Markov model that simulated disease progression and determined survival and costs for a hypothetical cohort of mRCC patients. The cost-effectiveness models included three or four mutually exclusive health states, among which patients transition, during the modelled time horizon: first line until progression, second line treatment, best supportive care (BSC) and death (due to cancer or other causes).

Out of the 20 congress abstracts concerning first-line mRCC treatment, 16 were based on a same model or similar model structure presented in more detail in the full articles [14-16].

All the articles (n = 5) were either sponsored by pharmaceutical industry or at least one of the authors was an employee of a pharmaceutical company; this was also the case in nearly all abstracts (21 out of 24).

Incremental cost-effectiveness ratios for first-line sunitinib ranged from €4786 to 109,416/QALY and €33,807 to 100,212/life-year gained (LYG) when compared with IFN-alpha. An important driver of cost-effectiveness was the assumption about which drug was used in second line treatment.

Does the review provide practical guidance for decisions based on cost-effectiveness? The answer to this question must be "no". Because of uncertainty in the clinical data, the key questions about how sunitinib effects costs and outcome in terms of (quality adjusted) life years gained is not answered in an unequivocal way. The conclusion, that in most of the reviewed studies sunitinib gives a reasonable cost per QALY is not very helpful. An alternative conclusion could be that the cost-effectiveness studies of sunitinib do not provide clear guidance to help guide use. Thus we cannot expect that those studies have a significant impact on uptake and use. Given lack of direct data estimating OS and quality of life, at

Fig. 2. Present standard of care gives 6 month until progression and 14 moths from progression to death. Quality of life is 0.78 until progression and 0.70 after progression. The total QALY is 0.78*6/12 + 0.70*14/12 = 0.39 + 0.83 = 1.22 QALY. OS is 20 months. New therapy extend the time to progression until 18 months, followed by 8 month in progression until death. OS is increased to 26 months (0.5 LYG). When the survival gain is adjusted forQoLthe gain is 0.35. To this should be added the QoL improvement from a longer time before progression, 12 months at a difference of 0.08, gives 0.08 QALY. Total gain is 0.43 QALY. The total QALY with the new treatment is 18 * 0.78/12 + 8 * 0.70/12 =1.17+ 0.47 = 1.64 QALY.

the price asked, the cost-effectiveness is at best questionable, and thus consistent with both positive and negative decisions.

Several economic studies were developed by/for the pharmaceutical industry based on the same cost-effectiveness model for several jurisdictions. Often, the adaptation to different countries simply redefined the resource use and unit cost information. Because the underlying structure is the same in all cases, and the data on factors most influencing the outcome—the benefit of sunitinib on OS and QoL—is lacking, the large number of studies does not necessarily increase the power of evidence.

Information on the use of healthcare resources may be obtained with different methods and from various sources (e.g., expert opinion, registers, local patients and literature). The estimates can often be very important for the result of economic evaluations. For example, in an article by Remak et al., the cost of best supportive care (BSC) was one of the most sensitive parameters in the analysis and a 20% reduction in its cost doubled the incremental costs per QALY gained [16]. Lack of adequate information on resource use and costs, limited clinical evidence and short follow-up times present major challenges for performing timely economic evaluations of novel cancer treatments.

In essence, there only exist two studies of the cost-effectiveness of sunitinib, the company study PenTAG and the HTA study used for the NICE technology appraisal [11]. Estimates from the PenTAG model suggested that none of the interventions would be considered cost-effective at a willingness-to-pay threshold of 30,000 pounds per quality-adjusted life-year (QALY). Estimates of cost per QALY ranged from 71,462 pounds for sunitinib to 171,301 pounds for bevacizumab plus IFN. Although there are many similarities in the methodology and structural assumptions employed by PenTAG and the manufacturers of the interventions, in all cases the cost-effectiveness estimates from the PenTAG model were higher than those presented in the manufacturers' submissions.

Coverage decisions for new therapies for mRCC

Despite the many problems involved in providing relevant estimates of the cost-effectiveness of the new, targeted therapies in mRCC, such studies are used as a basis for decisions about value for money in coverage decisions. Since there in most cases are no published economic studies to rely on when the product comes to the market, decisions are based on studies undertaken by the company or by the reimbursement agency itself.

NICE has well-developed methods and processes for undertaking economic evaluation for decisions about use of a drug by NHS (the National Health Service) in England and Wales. Sunitinib was recommended in 2009 as a first-line treatment option for people with advanced and/or metastatic renal cell carcinoma who are suitable for immunotherapy and have an Eastern Cooperative Oncology Group (ECOG) performance status of 0 or 1 [17]. The

recommendation includes the condition that the company provides the drug for the first cycle free of charge. This provided two benefits in the light of the uncertainty about cost-effectiveness: it is a hedge against the possibility that the true cost-effectiveness of sunitinib is less favourable than estimated and it provides a low-cost mechanism to get better information during the initial phase of treatment regarding who would respond and who would not respond (presumably those who don't respond in the first month will not receive further doses).

In addition to its guidance on sunitinib stated above, another NICE guidance in 2009 recommended against the use of beva-cizumab, sorafenib and temsirolimus as first-line treatment options for people with advanced and/or metastatic renal cell carcinoma [18]. Sorafenib and sunitinib are not recommended as second-line treatment options for people with advanced and/or metastatic renal cell carcinoma.

Pazopanib was recommended by NICE in 2011 as a first-line treatment option for people with advanced renal cell carcinoma who have not received prior cytokine therapy and have an Eastern Cooperative Oncology Group (ECOG) performance status of 0 or 1 and if the manufacturer provides pazopanib with a 12.5% discount on the list price, and provides a possible future rebate linked to the outcome of the head-to-head COMPARZ trial (Pazopanib versus Sunitinib in Metastatic Renal-Cell Carcinoma), as agreed under the terms of the patient access scheme and to be confirmed when the COMPARZ trial data are made available [19].

Since 2003, most Canadian provinces have participated in the Common Drug Review (CDR), which assesses the clinical and cost-effectiveness of oral agents according to standardized methods. Since 2009, most provinces have also participated in the Joint Oncology Drug Review (iJODR), since 2011 replaced by pCODR (www.pCODR.ca). The assessment of sorafinib by Canadian Expert Drug Advisory Committee (CEDAC) of the CDR, illustrates the use of economic evaluation when the first target drug for mRCC came to the market [20].

The submission in Canada for sorafinib from the drug's manufacturer used a Markov model, with key parameter estimates based on the TARGET trial (Treatment Approaches in Renal Cancer Global Evaluation Trial), which studied sorafinib second line after failure of INFalpha. This model estimate transition probabilities based on the data observed in the trial from PFS to progressive disease or death up until the point of cross-over. The model estimated a difference in overall survival (over a lifetime) of 1.21 years and an incremental cost per life-year gained of $36,046. In its assessment of the evidence, the CDR concluded that, given the early termination of the trial, the overall survival advantage, and hence the true cost-effectiveness of sorafenib, were uncertain. Rather than accepting the manufacturer's projections, it conducted its own analysis, which assumed that once patients entered the progressive disease state, being treated with sorafenib had no further impact on survival. The CDR analysis generated an estimate of overall survival gain of 4.5 months, which was closer to that actually observed in the trial at the time of the most recent analysis of outcome. Using this revised estimate of survival gain, the incremental cost-effectiveness ratio rose to $78,227, more than twice the manufacturer's estimate. Consequently, CEDAC recommended that sorafenib not be listed.

In 2012 pCODR recommended axitinib for use second line for patients that could not tolerate everolimus. Sorafinib was no longer considered a relevant alternative due to lack of data on effectiveness after sunitinib. In 2013 pazopanib was recommended for first line use based on a comparison with sunitinib; similar effectiveness and lower price [21].

In Sweden TLV is assessing prescription drugs for inclusion in the reimbursement scheme. Both sorafenib and sunitinib were accepted for reimbursement on the clinical and economic data

provided by the company. In the decision about sunitinib, some members of the Board did not agree, and asked for more information about cost-effectiveness [22,23]. For everolimus TLV accepted reimbursement for second line treatment of mRCC [24]. TLV accepted the economic model provided by the company, which estimated a cost per QALY of 285,000 SEK (30,000 Euro) for everolimus compared with sorafinib. This indirect comparison was accepted since sorafinib was the most used therapy for this group of patients, despite no evidence of its effectiveness after failure of sunitinib.

In the US, cost-effectiveness analyses are not used to make yes/no funding decisions regarding particular cancer drugs as they are in Europe and Canada. In fact, laws enacted nationally for Medicare enrolees (adults ages 65 and greater) and in most states (affecting commercial insurance plans) mandate coverage of cancer drugs listed as safe and effective according to at least one nationally-recognized compendia or clinical practice guideline (for example, the National Comprehensive Cancer Network). What can vary, however, is the co-pay amount that is levied on specific drugs. The process for determining co-pays can vary widely from insurer to insurer. At one extreme is Premera Blue Cross, a regional insurer in the Pacific Northwest that sets co-pays for new oral cancer drug based on the estimated cost-effectiveness of the therapy relative to the best available alternative. Intermediate are "value based formularies" that use some combination of clinical and economic efficacy to determine co-pay levels. At the other, perhaps more common, extreme are plans that base co-pays on (branded) drug price alone. Thus, in the US all FDA-approved drugs are generally available, and there are few barriers to prescribing them. The very high cost of these new targeted agents, however, means that high co-payments can effectively exclude access to the drug from large segments of the population. To the extent that even modest co-pays can lead to substantial out-of-pocket costs, several studies now suggest that financial distress and bankruptcy are increasingly common problems for cancer patients and their families [25-27].

Uptake and use of new drugs for mRCC

The figures below present the use of sorafenib, sunitinib and temsirolimus in selected countries 2005-2009. Please note that sorafinib is also approved for primary liver cell cancer (Fig. 3).

The most striking result is the very low uptake and use in the UK and the US. France has the highest use, while Germany, Italy and Spain are all rather close to a western EU average. For Switzerland we see a rapid uptake, and then a decline, which may be dependent on substitution with sunitinib or other drugs. The low use of sorafenib in the US may be related to the fact that sunitinib was available first, and that phase II study data for sorafenib at launch did not show an increase in PFS with sorafenib compared with IFNa (5.7 versus 5.6 months) when used first-line, making its role in therapy of untreated patients unclear [10] (Fig. 4).

For sunitinib we see the same fast and high uptake in France, and that Germany here is close to France. Use in Switzerland relatively lower than for sorafenib. UK is in the lower end, but the difference to other countries is small. For US we see a rapid uptake, and then a plateau from 2007 onwards.

France stands out in terms of a more rapid uptake and higher use. With the exception of US and Italy, the use is very marginal until Q2 2009 (Fig. 5).

The low use in UK is supported by data in the report by Mike Richards [28]. According to table A6 in the report the UK usage was 19 and 54 percent of the average use in the 5 big European countries for sorafinib and sunitinib respectivly for the period April 2008 to March 2009. For temsirolimus the UK useage was 17%. Fig. 6

Fig. 3. Use of sorafinib in mg per 100,000 population.

Fig. 4. Use of sunitinib in mg per 100,000 population.

Fig. 5. Use of temsirolimus in mg per 100,000 population.

Health care region|Sweden|

SLL02_131107_Sales1998-2013p7_L01_L02AB_L04TalidOchLenali

Fig. 6. Use of drugs for mRCC in Sweden 2006-2011. SEK per 100,000 population.

|ATC-code4|(Alla)|

SLL02 130109 Sales1998-2012 L01 L02AB L04TalidOchLenali

SEK per incident case of cancer 20101

Northern health care region

South East health care region

Southern health care region

Stockholm-Gotland

2011 Sweden

Uppsala-Orebroregion

Western health care region

Molecule

Temsirolimus Sunitinib ^Sorafenib Pazopanib Everolimus

[Health care region |Year| Fig. 7. Use of drugs for mRCC in the six health care regions of Sweden 2011.

shows the use of five targeted drugs used for treatment of mRCC in Sweden.

Fig. 7 shows the variations within Sweden, indicating significant variations, both in the use of the individual drugs and the use of all drugs combined. More detailed data on use in first line versus second line, and about patient characteristics are necessary in order to understand the magnitude of potential under- and overconsump-tion in different regions/patient populations.

Population evidence of impact of TKIs of outcome

We have identified five studies that examined the effects on survival from the use oftargeted therapies in advanced RCC patients outside of trial settings.

Shah et al. analyzed the Surveillance, Epidemiology, and End Results (SEER) 18 registry database to compare 1- and 3-year relative survival rates among advanced RCC patients during 2001-2009, 2001-2004, and 2006-2009, the 1-year relative survival rates were 27.0 ± 0.8% and 27.1 ± 0.9%, (p value = 1.3) and, the 3-year survival were 10.1 ±0.6% and 9.6 ±0.8%, (p value = 1.42), during 2001-2004 and 2006-2009, respectively. Thus, this population based study suggested that there was no significant improvement in relative survival rates among mRCC patients in the era of targeted agents [29].

In another study from the US, Shek et al. analyzed data from 28,252 patients with RCC in the California Cancer Registry (CCR), a population-based cancer surveillance system, diagnosed between 1998 and 2007 [30]. Inter-era differences in clinical variables—including year of diagnosis, histologic characteristics, age, sex, race, stage, nephrectomy status, overall survival (OS), and cause-specific survival (CSS)—were assessed, and analyzed with univariate and multivariate Cox models. Crude 3-year OS (68.2% vs. 74.6%; 2P..001) and CSS (78.1% vs. 82.3%; 2P..001) were significantly higher in the post-cytokine era. The three strongest predictors for improved survival were localized disease, nephrectomy and histologic type. Insufficient follow-up time in

the post-cytokine era and a higher proportion of localized disease in that era confound the possibility of benefit derived from targeted therapies. Longer follow-up for patients treated in the post-cytokine era is necessary for a more robust comparison of long-term OS.

In a third US overall survival (OS) analysis, the analytic cohort included all patients in the registry diagnosed between January 1, 2007, to May 31, 2011 [31]. Patients were grouped by up to three treatment exposures according to each drug's mechanism of action: vascular endothelial growth factor tyrosine kinase inhibitor (VEGFR TKI), mammalian target of rapamicin inhibitor (mTOR), or no systemic treatment (NSTx, which could include radiation or surgery). OS by exposure sequence was evaluated using Kaplan-Meier, pair-wise comparison, and Cox regression analyses. Median OS was 17.2 months. OS (months) for one exposure was: mTOR 5.4, TKI 18.2, NSTx 18.4; for two exposures: mTOR/TKI 9.3, TKI/mTOR 13.9, TKI/TKI 35.2; and for three exposures: TKI/mTOR/TKI 20.9, TKI/TKI/mTOR33.1. In Cox regression analysis, compared with the referent (TKI), TKI/TKI (hazard ratio 0.53) had a lower risk of death, and mTOR (hazard ratio 2.16) had a higher risk of death. The study concludes that mRCC survival outcomes are different by pattern, with general findings consistent with trial-based expectations in similar patient populations. Real-world data can provide context around patterns of care and impact when experimental trial data are lacking.

Swedish researchers conducted a retrospective register study assessed overall survival (OS) and influential factors on OS in Swedish renal cell carcinoma (RCC) patients. They used three merged national health registers to assess the impact of cytokine (interferon-a and tyrosine kinase inhibitor (TKI; sunitinib or sorafenib) treatment on OS in metastatic mRCC. From 2000 to 2008,8009 patients were diagnosed with RCC and 2753 with mRCC (2002-2008). Median OS in RCC patients diagnosed from 2006 to 2008 compared with 2000-2005 was not reached vs. 47.9 months (p< 0.001), and in mRCC patients diagnosed from 2006 to 2008 compared with 2002-2005, was 12.4 vs. 9.6 months respectively

(p = 0.004). Factors associated with significantly improved OS in RCC were female gender, lower age, and previous nephrectomy, and, in mRCC female gender, previous nephrectomy, and any TKI prescription (Model 1: median-adjusted OS, 19.4 months (TKI patients) vs. 9.7 months (non-TKI patients); hazard ratio, 0.621; p< 0.001). The results suggest that increased nephrectomy rates and the use of TKIs contributed to the improvement seen in mRCC patients. However, it should be noted that only 12% of the patients received a TKI in first line treatment [32].

Soerensen et al. included all Danish patients with mRCC starting first line treatment with immunotherapy, TKIs or mTOR-inhibitors, between 2006 and 2010 in an outcome study. Baseline and outcome data were collected retro-spectively. Between 2006 and 2010, a total of 1073 patients were referred. Of these, 759 patients received first line treatment and 314 received no systemic treatment. The proportion of treated patients increased from 64% in 2006 to 75% in 2010 (p = 0.02). In 2006, 22% received targeted therapy and this increased to 75% in 2010 (p<0.00001). In 2006, 21% of first line patients received second line treatment compared to 41% in 2010 (p = 0.01). From 2006 to 2010, we observed an improved median OS from 11.5 to 15.7 months (p = 0.03. The untreated population of 314 patients had a median OS of 3.1 months from date of metastatic disease with no significant change from 2006 to 2010 [33].

Discussion

This paper reviews the link between clinical trial data, the cost-effectiveness studies for new cancer drugs, coverage decisions informed by such studies, uptake and use of the drugs, and evidence of impact on outcome in the relevant patient population.

The review of economic and clinical evaluations for new targeted therapies for mRCC reveals that there are great uncertainties in the estimates of cost-effectiveness, resulting in uncertain relevance of these studies for decision making following launch. This uncertainty is partly related to the methodological challenges making extrapolation from clinical trial data to clinical practice, since all models used for estimating cost-effectiveness have to make assumptions to predict long term costs and outcome.

The choice of extrapolation methodology can have an important impact on comparative efficacy, costs, and cost-effectiveness. This was shown by Ekman et al. in the illustrative example using data from a pivotal phase III trial of sunitinib vs. IFN-a as first-line mRCC in the US. Cost effectiveness results could vary quite substantially depending on assumptions made. A short time horizon (one year) resulted in an ICER of $120,304 compared to an ICER of $52,571 for the life time horizon assumption. Depending on choice of survival distribution, ICERs varied between $50,000 and 150,000 the pessimistic (stop and drop) scenario and the optimistic (continued benefit) assumption resulted in large differences in ICERs; $114,000 vs. $50,000, respectively [34].

Offering patients in oncology trials the opportunity to cross over to active treatment at disease progression is a commonly used strategy to address ethical issues associated with use of placebo controls, but may lead to statistical challenges in the analysis of overall survival (OS) and cost-effectiveness as cross-over leads to loss of information and dilution of comparative clinical efficacy. This was illustrated by comparing alternative methods for analysing OS data in the presence of cross-over; simple methods (intent-to-treat analysis and censoring data at cross-over) and advanced statistical methods (the inverse probability of censoring weighting [IPCW] and the rank-preserving structural failure time [RPSFT] models, using data from two phase III clinical oncology trials of sunitinib. The study shows that the method for correction may significantly affect the cost-effectiveness ratio, while this was not the case in NICE assessment of sunitinib for mRCC. There the selection of a population with no second line treatment was most important to

bring the cost-effectiveness ratio below the 50,000 GBP per QALY [35].

Statistical methods may to some extent help to overcome the problem that clinical studies for market authorization are not designed to take into account the requirement for health technology assessment. This problem may be reduced with the increasing use of HTA and parallel regulatory/HTA scientific advice during the development process [36]. Progression free survival (PSF) as an outcome measure is also increasingly questioned, also from a regulatory point of view [37]. However, PFS will remain an important endpoint, which can be more useful as an input to estimations of cost-effectiveness if it already at the start of the study is work out how the measure of PFS is linked to OS and quality of life, the two measures used for calculation of the gain in QALY. Reducing the use of crossover is an important factor for improving the value of PFS as an outcome measure [38].

Indirect comparisons play an important role in assessment of cost-effectiveness. The assessment of relative effectiveness of pazopanib by EuronetHTA provided no evidence on differences in effectiveness between the different targeted therapies for mRCC [39]. This conclusion was supported by the results of the direct comparison between pazopanib and sunitinib [12].

With the exception of UK and Canada, there is no evidence that cost-effectiveness estimates has played a key role for reimbursement and use of the new targeted therapies in mRCC. This conclusion is supported from a study comparing the number of reimbursed indications in different countries for a selection of cancer drugs [40]. Most indications were reimbursed in other countries than UK, Canada, Australia and New Zealand.

From one perspective, it is not surprising that formal cost-effectiveness estimate have a limited impact on decisions to reimburse or not reimburse, since the uncertainty surrounding the ICERs is high. On the other hand, decision makers may use the data and the models in these studies to call for restrictions in use while further data are gathered e.g., in the context of an observational study or a pragmatic trial. Value of information analysis can be used to determine whether the investment in further study will yield sufficient reductions in uncertainty to warrant the cost and delay in access entailed in another trial. There are also examples that cost-effectiveness studies are used as a basis for market access agreements including reductions in the prise of the drugs.

While it is possible to notice an impact of the cost-effectiveness ratio on the probability of a positive reimbursement decisions, it is seldom possible to assess the relative contribution in an individual decision. One reason for this is that evidence on clinical efficacy, the probably most important factor for reimbursement, is correlated to estimates of effectiveness, and thus cost-effectiveness. It is also easier to see that cost-effectiveness has an impact when there is a choice between two drugs with similar effectiveness, such as the case with sunitinib and pazopanib for first line treatment of mRCC. Prices will then be the key determinant of cost-effectiveness, and since prices can easily be changed, the impact of cost-effectiveness should be assessed based on data on use, rather than decisions on reimbursement.

Worldwide, economic evaluations appear to have had little impact on the diffusion of targeted therapies into practice. Data on international variations in use indicate that other factors, mainly from the supply side, account for most of the variation. The most important demand side factor is the economic power of the country, measured as income or health care expenditures per capita. Since prices of new drugs vary less than per capita income, it may be interpreted as linked to the cost-effectiveness, as well as the affordability, of new drugs in countries with low per capita income. Data also indicate that how the drugs are paid for in the health care system is important. The high use in France can be explained by the practice of making price volume agreements; in combination

with reimbursing the hospitals for the cost of new drugs outside the DRG based hospital funding.

Our review of available studies of impact on survival for patients with mRCC indicates that there may be an effect, but the magnitude is uncertain. The problems to detect an effect in epidemiological data suggest that the effect is small, and a definitive conclusion is hampered by patient selection issues that cannot be resolved with available data. Further studies are needed to give information on how the new targeted therapies should be optimally used in practice. Questions about optimal use of the drugs in parallel or sequence can be answered with information from data in clinical practice.

Conclusions and policy recommendations

The value of new cancer therapies is based on their use. To achieve an optimal use there are several steps from the initial economic evaluations based on clinical trial data, the reimbursement (coverage) decisions based on these data, and the implementation of these decisions in clinical practice.

Considerable resources have been devoted to the study of cost-effectiveness of new cancer drugs as a basis for decisions about payment and use. Those resources could be much better used if industry and HTA agencies with high competence collaborated at an early stage. That would increase both the quality and the credibility of studies, and make them more useful to guide decisions about reimbursement. The methodology for studies has been debated, and there are opportunities for improvement. Most important from a general methodological point of view is that all costs are accounted for and that the benefit measure incorporates all elements of social value. While there is a discussion if the QALY includes all relevant elements of value, there is an acceptance from all stakeholders that estimations of cost per QALY are relevant for decisions about coverage. Since improvements in survival for cancer patient, it is important that cost per LYG also is calculated and presented in parallel to the cost per QALY. There is also a need for improvements in data used for calculating the QALY gains.

A major problem is when the clinical trial data do not provide clear evidence on outcome, for example due to the use of surrogate end-points and crossover design in the study. Statistical methods for extrapolation may to some extend help to overcome this problem, but it is important that clinical studies are decided taking into account the requirement for health technology assessment. The traditional trial methodology with PSF and crossover has serious shortcomings, when the data should be used for assessment of cost-effectiveness, and there is a need for a new standard. Guidelines for this new standard should be developed in collaboration between regulatory and HTA authorities.

Indirect comparisons are often necessary for assessment of cost-effectiveness. It is not possible to undertake clinical trials for all potentially relevant alternatives in clinical practice. A limitation of all indirect comparisons is that they are based on trial data. The options for using observational studies in clinical practice to answer questions about alternative treatment strategies should be assessed as early as possible. Such studies should be combined with data collection for assessment of relative efficacy and risk-benefit, initiated directly after marketing authorization.

A closer link between studies, decisions about coverage and the actual allocation of resources for new drugs is also needed. The data for studies of the implementation of reimbursement decisions are scarce, and available evidence indicates that there is a gap between decisions and implementation, resulting in an inefficient use of new drugs and thus a loss of value. Linking economic incentives, such as pay for performance is one way to close this gap.

Economic evaluations and coverage decisions are based on uncertain data, and thus the impact on outcome in clinical practice is uncertain. It is therefore necessary to collect data on actual use in to make it possible to study the effects on outcome in clinical practice. Since outcome depends on a number of factors, individual patient data are a necessary requirement for revealing impact on outcome. Such data may also help to inform decisions in clinical practice which are not, and often cannot, be informed by clinical trial data; for example the cost and outcome of use of drugs in combination and/or sequence. These data must include not only drug therapies but also all relevant diagnostic and therapeutic interventions, which are important for optimizing and revealing the value of new the drugs in clinical practice.

Conflict of interest

We have no conflict of interest to report.

References

[1] Wilking N, Jönsson B. A pan-European comparison regarding patient access to cancer drugs; 2005.

[2] Mason A, Drummond M, Ramsey S, Campbell J, Raisch D. Comparison of anticancer drug coverage decisions in the United States and United Kingdom: does the evidence support the rhetoric? J Clin Oncol 2010;28(20):3234-8, http://dx.doi.org/10.1200/jc0.758.

[3] NICE and cancer drugs - the facts; 2013. Available from: http://www.nice. org.uk/newsroom/nicestatistics/niceandcancerdrugsthefacts.jsp

[4] Jonsson B. Being NICE is not the problem! Eur J Cancer 2009;45(7):1100-2, http://dx.doi.org/10.1016/j.ejca.2009.01.035.

[5] Jönsson BWN. Evaluation, implementation and access to new cancer drugs in Sweden. J Cancer Policy 2013 [in press].

[6] Eichler HG, Bloechl-Daum B, Abadie E, Barnett D, Konig F, Pearson S. Relative efficacy of drugs: an emerging issue between regulatory agencies and third-party payers. Nat Rev Drug Discov 2010;9(4):277-91, http://dx.doi.org/10.1038/nrd3079.

[7] Faulkner E, Annemans L, Garrison L, Helfand M, Holtorf AP, Hornberger J, et al. Challenges in the development and reimbursement of personalized medicine-payer and manufacturer perspectives and implications for health economics and outcomes research: a report of the ISPOR personalized medicine special interest group. Value Health 2012;15(8):1162-71, http://dx.doi.org/10.1016/jjval.05.006.

[8] IARC. Globocan 2012. Available from: http://globocan.iarc.fr/old/burden. asp?selection_pop=0&Text-p=China&selection_cancer=290&Text-c=All+ cancers+excl.+non-melanoma+skin+cancer&pYear=8&type=0&window=1 &submit=%C2%A0Execute%C2%3916A0

[9] Coppin C, Kollmannsberger C, Le L, Porzsolt F, Wilt TJ. Targeted therapy for advanced renal cell cancer (RCC): a Cochrane systematic review of published randomised trials. BJU Int 2011;108(10):1556-63, http://dx.doi.org/10.1111/j.464-410X.2011.10629.x.

[10] Escudier B, Szczylik C, Hutson TE, Demkow T, Staehler M, Rolland F, et al. Randomized phase II trial of first-line treatment with sorafenib versus interferon Alfa-2a in patients with metastatic renal cell carcinoma. J Clin Oncol 2009;27(8):1280-9, http://dx.doi.org/10.1200/JC0.2008.19.3342.

[11] Thompson Coon J, Hoyle M, Green C, Liu Z, Welch K, Moxham T, et al. Bevacizumab, sorafenib tosylate, sunitinib and temsirolimus for renal cell carcinoma: a systematic review and economic evaluation. Health Technol Assess 2010;14(2):1-184, http://dx.doi.org/10.3310/hta01402, iii-iv.

[12] Casper J, Schumann-Binarsch S, Kohne CH. Pazopanib versus sunitinib in renal cancer. N Engl J Med 2013;369(20):1969, http://dx.doi.org/10.1056/NEJMc795#SA31311.

[13] Purmonen TT. Cost-effectiveness of sunitinib in metastatic renal cell carcinoma. Expert Rev Pharmacoecon Outcomes Res 2011;11(4):383-93, http://dx.doi.org/10.1586/erp.11.33.

[14] Benedict A, Figlin RA, Sandstrom P, Harmenberg U, Ullen A, Charbonneau C, et al. Economic evaluation of new targeted therapies forthe first-line treatment of patients with metastatic renal cell carcinoma. BJU Int 2011;108(5):665-72, http://dx.doi.org/10.1111/j.464-410X.2010.09957.x.

[15] Chabot I, Rocchi A. How do cost-effectiveness analyses inform reimbursement decisions for oncology medicines in Canada? The example of sunitinib for first-line treatment of metastatic renal cell carcinoma. Value Health 2010;13(6):837-45, http://dx.doi.org/10.1111/j.524-4733.2010.00738.x.

[16] Remak E, Charbonneau C, Negrier S, Kim ST, Motzer RJ. Economic evaluation of sunitinib malate for the first-line treatment of metastatic renal cell carcinoma. J Clin Oncol 2008;26(24):3995-4000, http://dx.doi.org/10.1200/JC0.662 200713.

[17] NICE. NICE technology appraisal guidance 169 Sunitinib for the flrst-line treatment of advanced and/or metastatic renal cell carcinoma; 2009. Available from: http://www.nice.org.uk/TA169

[18] NICE. NICE technology appraisal guidance 178 Bevacizumab (first-line), sorafenib (first- and second-line), sunitinib (second-line) and temsirolimus (first-line) for the treatment of advanced and/or metastatic renal cell carcinoma; 2009. Available from: www.nice.org.uk/TA178

[19] NICE. Pazopanib for the first-line treatment of advanced renal cell carcinoma; 2013. Available from: http://publications.nice.org.uk/pazopanib-for-the-first-line-treatment-of-advanced-renal-cell-carcinoma-ta215

[20] Drummond M, Evans B, LeLorier J, Karakiewicz P, Martin D, Tugwell P, et al. Evidence and values: requirements for public reimbursement of drugs for rare diseases - a case study in oncology. CanJ Clin Pharmacol 2009;16(2):e273-81 [discussion e82-e84].

[21] pCODR. Pan-Canadian Oncology drug review; 2013. Available from: http://www.pcodr.ca/wcpc/portal/Home/FindaReview7_afrLoop» 4536774071444000&lang=en&_afrWindowMode=0&_adf.ctrl-state= 1bhganhue8.91

[22] TLV. Reimbursement decision Nexavar; 2006. Available from: http://www.tlv. se/beslut/beslut-lakemedel/generell-subvention/nexavar-in-i-formanerna/

[23] TLV. Reimbursement decision Sutent; 2006. Available from: http://www. tlv.se/beslut/beslut-lakemedel/generell-subvention/subvention-for-sutent/

[24] TLV. Reimbursement decision Afinitor; 2010. Available from: http://www.tlv. se/beslut/beslut-lakemedel/generell-subvention/afinitor-mot-njurcellscancer-ingar-i-hogkostnadsskyddet/

[25] Bona K, Dussel V, Orellana L, KangT, Geyer R, FeudtnerC, et al. Economic impact of advanced pediatric cancer on families. J Pain Symptom Manage 2013;18(13), 00310-2.

[26] Shankaran V, Jolly S, Blough D, Ramsey SD. Risk factors for financial hardship in patients receiving adjuvant chemotherapy for colon cancer: a population-based exploratory analysis. J Clin Oncol 2012;30(14):1608-14, http://dx.doi.org/10.1200/JC0.2011.37.9511.

[27] Ramsey S, Blough D, Kirchhoff A, Kreizenbeck K, Fedorenko C, Snell K, et al. Washington State cancer patients found to be at greater risk for bankruptcy than people without a cancer diagnosis. Health Aff (Millwood) 2013;32(6):1143-52, http://dx.doi.org/10.1377/hlthaff.2012.1263.

[28] Richards MA. Extent and causes of international variations in drug usage; 2010. Available from: https://www.gov.uk/government/publications/ extent-and-causes-of-international-variations-in-drug-usage

[29] Shah BK, Ghimire KB. Survival trends among patients with advanced renal cell carcinoma (RCC) in the United States. J Clin Oncol 2013;31 [absr 422].

[30] Shek D, Tomlinson B, Brown M, Brunson A, Pan CX, Lara Jr PN. Epidemiologic trends in renal cell carcinoma in the cytokine and post-cytokine eras: a registry analysis of 28,252 patients. Clin Genitourin Cancer 2012;10(2):93-8, http://dx.doi.org/10.1016/j.clgc.01.007.

[31] Harrison MR, Hirsch BR, George DJ, Walker MS, Chen C, Korytowsky B, et al. Real-world outcomes in metastatic renal cell carcinoma: insights from a joint community-academic registry. J Oncol Pract 2013;3:3.

[32] Wahlgren T, Harmenberg U, Sandstrom P, Lundstam S, Kowalski J, Jakobsson M, et al. Treatment and overall survival in renal cell carcinoma: a Swedish population-based study (2000-2008). Br J Cancer 2013;108(7):1541-9, http://dx.doi.org/10.1038/bjc.2013.119.

[33] Sorensen A, Hermann DF, Jensen GG, Spliid NV, Bergan H, Fode EQ, et al. Implementation of targeted therapy in Denmark for patients with metastatic renal cell carcinoma: results from the Danish Renal Cancer Group (DARENCA) study-2; 2012. Available from: http://findresearcher.sdu.dk:8080/ portal/en/publications/implementation-of-targeted-therapy-in-denmark-for-patients-with-metastatic-renal-cell-carcinoma-results-from-the-danish-renal-cancer-group-darenca-study2%283f6fe341-a517-4e54-833c-464adea7%29/export.html0565

[34] Ekman M, Charbonneau C, Ramsberg J, Jonsson L, Sandin R, Jonsson B, Drum-mond M, Weinstein MC. PCN162 extrapolation in trial-based cost-effectiveness modeling: in search of a standard value of health; 2011. Available from: http://www.valueinhealthjournal.com/search/quick

[35] Jönsson L, Ekman SR, Ramsberg M, Charbonneau J, Huang C, Jönsson B, et al. Analyzing overall survival in randomized controlled trials with cross-over and implications for economic evaluation In: ISPOR 13th Annual European Congress Prague, Czech Republic November 2010. 2013.

[36] Elvidge S. EMA's parallel advice workshop bridges regulatory and reimbursement divide. Nat Rev Drug Discov 2013;13(1):8, http://dx.doi.org/10.1038/nrd4219.

[37] Booth CM, Eisenhauer EA. Progression-free survival: meaningful or simply measurable? J Clin Oncol 2012;30(10):1030-3,10.200/JC0.7571.

[38] Hotte SJ, Bjarnason GA. Progression-free survival as a clinical trial endpoint in advanced renal cell carcinoma. Current Oncology 2011;18(Suppl. 2):S11-9.

[39] EUnetHTA WP5, Relative effectiveness of pharmaceuticals. Pilot assessment using the draft HTA Core Model for Rapid Relative Effectiveness Assessment. Version 4, 31 December 2012. Final version http://www.eunethta.eu/ outputs/wp5-ja1-pilot-pazopanib-reportappendix.

[40] Cheema PK, Gavura S, Migus M, Godman B, Yeung L, Trudeau ME. International variability in the reimbursement of cancer drugs by publically funded drug programs. Curr Oncol 2012;19(3):e165-76, http://dx.doi.org/10.3747/co.19.946.