SEARCH

SEARCH BY CITATION

Keywords:

  • cost-utility analysis;
  • economics;
  • modeling;
  • pharmaceutical pricing;
  • product life cycle

ABSTRACT

  1. Top of page
  2. ABSTRACT
  3. Introduction
  4. Study Data and Methods
  5. Results
  6. Discussion
  7. Conclusions
  8. References

Objective:  Pharmacoeconomic analyses typically project the expected cost-effectiveness of a new product for a specific indication. This analysis develops a dynamic life-cycle model to conduct a multiindication evaluation using the case of trastuzumab licensed in the United States for both early-stage and metastatic (or late-stage) human epidermal growth factor receptor 2 (HER2)-positive breast cancer therapy (early breast cancer [EBC]; metastatic breast cancer [MBC]), approved in 2006 and 1998, respectively.

Methods:  This dynamic model combined information on expected incremental cost-utility ratios for specific indications with an epidemiologically based projection of utilization by indication over the product life cycle—from 1998 to 2016. Net economic value was estimated as the cumulative quality-adjusted life years (QALYs) gained over the life cycle multiplied by a societal valuation of health gains ($/QALY) minus cumulative net direct treatment costs. Sensitivity analyses were performed under a range of assumptions.

Results:  We projected that the annual number of EBC patients receiving trastuzumab will be more than three times that of MBC by 2016, in part because adjuvant treatment reduces the future incidence of MBC. Over this life cycle, the estimated overall incremental cost-effectiveness ratio (ICER) was $35,590/QALY with a total of 432,547 discounted QALYs gained. Under sensitivity analyses, the overall ICER varied from $21,000 to $53,000/QALY, and the projected net economic value resulting from trastuzumab treatment ranged from $6.2 billion to $49.5 billion.

Conclusions:  Average ICERs for multiindication compounds can increase or decrease over the product life cycle. In this example, the projected overall life-cycle ICER for trastuzumab was less than one half of that in the initial indication. This dynamic perspective—versus the usual static one—highlights the interdependence of drug development decisions and investment incentives, raising important reimbursement policy issues.


Introduction

  1. Top of page
  2. ABSTRACT
  3. Introduction
  4. Study Data and Methods
  5. Results
  6. Discussion
  7. Conclusions
  8. References

Most economic evaluations involving cost-effectiveness analyses of new pharmaceutical products are performed shortly following product launch for a specific indication covered in the license. They are based on mathematical models that project clinical and economic outcomes from results of phase III randomized clinical trials for a typical patient for the expected horizon of clinical impact. Indeed, such models are often a key element in dossiers submitted to public and private payers for purposes of coverage and reimbursement.

In the past decade, we have witnessed the licensing of many new, innovative advances in biologic agents that work in multiple cancer types and indications. When these products are launched for the initial indication, manufacturers will set the price by which cost-effectiveness is judged. From a payer's perspective, this evaluation is logical and useful as they see each indication as a separate “purchase” on behalf of their beneficiaries. However, this process has some limitations and implications from a broader societal and global perspective that are underappreciated. First, the aggregate economic value delivered by a new medicine will ultimately be determined by the different types and number of patients using it over its life cycle, and this may include totally new indications (often at different doses). Second, not only does the cost-effectiveness of a product vary among individual patients, but also the economic value will typically vary systematically across indications. Because the price per milligram of a pill often cannot be varied across indications, the actual cost-effectiveness achieved will generally vary. This creates a dilemma for manufacturers working within a reimbursement environment that simultaneously fixes the price at launch for the duration of the product life cycle and evaluates the product by applying a cost-effectiveness threshold: should potential future indications have some impact on the proposed initial product price?

In oncology, for example, both for reasons of safety and for risk–benefit, it is common to first conduct clinical trials in the most severely ill patients (e.g., patients with metastatic disease who have failed first- or second-line therapies). If efficacy is demonstrated at reasonable tolerability in this situation, then the therapy can be tested at earlier stages of disease progression, for example, moving from later stage, metastatic use to treatment at early diagnosis, called “adjuvant” use (e.g., in combination with surgery). This process of drug development and testing usually takes 8–12 years for the initial indication, and several more years for each subsequent indication [1]. Once a therapy is approved in an early-stage, adjuvant setting, the characteristics of the patients who progress to metastatic disease and those individuals who are newly diagnosed with metastases may be quite different from patients originally diagnosed with metastatic disease at the beginning of the product life cycle or during the registration clinical trials. Early therapy will affect the later incidence of metastatic disease as well as the treatment of adjuvant patients. Moreover, before a product is approved for use in the adjuvant setting, there can also be changes in the standard of care in the metastatic setting.

This analysis takes a broader long-term perspective, asking what is the overall cost-effectiveness across multiple indications throughout the product life cycle and what is the aggregate economic value delivered. The objective was to develop a dynamic life-cycle modeling (DLM) approach, and to apply it to a case example. We evaluated a targeted cancer therapy, trastuzumab (Herceptin®, Genentech, Inc., South San Francisco, CA), a monoclonal antibody approved for treatment of human epidermal growth factor receptor 2 (HER2)-positive breast cancer in combination with chemotherapy. This analysis considers both the treatment for metastatic HER2-positive breast cancer (approved by the Food and Drug Administration in 1998) and the more recently approved (2006) indication for adjuvant treatment of early-stage HER2-positive breast cancer. We chose this example because trastuzumab has a more recently approved second indication, national-level epidemiological data were available, and data from suitable indication-specific cost-effectiveness analyses were available.

Study Data and Methods

  1. Top of page
  2. ABSTRACT
  3. Introduction
  4. Study Data and Methods
  5. Results
  6. Discussion
  7. Conclusions
  8. References

Background and Model Overview

A DLM approach combines information on expected incremental cost-effectiveness ratios (ICERs) with an epidemiologically based projection of utilization by indication over the product life cycle. The effective product life cycle for biologics is not fixed given the current lack of a regulatory pathway in the United States for so-called “biosimilars” or “follow-on biologics” and the potential entrance of competing branded biologics in the same class. For purposes of this exercise, the “end” of the product life cycle is assumed to be 10 years after the launch in adjuvant treatment (i.e., through 2016). This DLM projection thus has three major components: a forecast of the volume of trastuzumab use over the product life cycle—from initial launch in 1998 to 2016, an estimate of the average cost-effectiveness (i.e., the average ICER) in metastatic treatment over this period, and an estimate of the average ICER in adjuvant treatment following approval in this indication eight years after the initial launch.

Projecting Disease Incidence

Approximately 20% to 25% of patients with breast cancer will have HER2-positive disease, which is associated with a poor prognosis [2,3]. Trastuzumab is a monoclonal antibody that targets HER2, and is approved for the treatment of HER2-positive metastatic breast cancer either as first-line therapy in combination with paclitaxel or as second- or third-line monotherapy. For this analysis, we relied on previously published and publicly available data to project the economic impact of trastuzumab.

To our knowledge, there are no published projections of the volume of long-term trastuzumab use in the United States. The approach we used projects an increasing volume of the use of trastuzumab from 1998 to 2016 based on estimates of the anticipated annual incidence of metastatic breast cancer and early-stage breast cancer in the United States. These estimates were derived from Surveillance Epidemiology and End Results (SEER) data from the National Cancer Institute, and assumed that 25% of patients tested are HER2-positive [4]. In the years prior to approval of the adjuvant indication, the volume of utilization was based solely on metastatic breast cancer patients. After approval of the adjuvant indication in November 2006, the volume of use was projected to shift to include women receiving adjuvant therapy for EBC as well as women receiving first-line metastatic treatment, whether initially diagnosed, or for recurrent disease. Furthermore, given that data from the joint analysis of the National Surgical Adjuvant Breast and Bowel Project B-31 and the North Central Cancer Treatment Group N9831 trials demonstrated that the addition of trastuzumab to a standard adjuvant regimen reduced the risk of recurrence by 52% (P < 0.001) and improved survival by 33%, there is an anticipated reduction in the number of future metastatic breast cancer patients [5].

Projecting Volume of Trastuzumab Use

The projection of the volume of use over the product life cycle was based on demographic projections, epidemiological estimates, and assumptions about use rates among candidate (i.e., HER2-positive) patients. The female population was divided into five age groups: less than 21, 21–39, 40–54, 55–64, and more than 64. Data on US census age-specific subpopulations through 2016 were aggregated to within these age groups [6]. An analysis of the SEER registry for the period 1999–2001 provided the incidence rates of breast cancer used in the analysis: these are shown in Table 1 for both EBC and newly diagnosed MBC patients. An assumption was made that 25% of the previously diagnosed early patients in SEER would later suffer a recurrence. An analysis of recent SEER data was used to estimate the share of EBC patients who are node-negative low-risk, node-negative high-risk, and node positive. Based on these shares, the use rate in the base case for HER2-positive patients was assumed to be 60% for all patients in the metastatic setting and 70% for node-positive and high-risk node-negative patients in the adjuvant setting. This also reflects the impact of a number of factors, including that not all women will be tested, some will have comorbidities that limit trastuzumab use, some may not have adequate insurance coverage, and some will choose not to receive treatment.

Table 1.  Input parameters for volume of use projection
Input parametersBase-caseRangeSource
  • *

    Both HER2-positive and HER2-negative.

  • SEER, Surveillance Epidemiology and End Results; HER2-positive, human epidermal growth factor receptor positive.

Incidence rates (per 100,000 women) for early breast cancer* by age:  SEER 1999–2001
 <21 years0.02 
 21–39 years23.2 
 40–54 years161.5  
 55–64 years317.1 
 65+ years372.0 
Incidence rates (per 100,000 women) for newly diagnosed metastatic disease*:  SEER 1999–2001
 <21 years0 
 21–39 years1.1 
 40–54 years6.6  
 55–64 years15.1 
 65+ years19.4 
Incidence of metastatic disease among women diagnosed previously25% Assumption
Incidence of HER2-positive breast cancer among women diagnosed with breast cancer25% Slamon et al. [2]
Early breast cancer distribution at diagnosis:  SEER
 Node-negative, low-risk17%  
 Node-negative, high-risk50%  
 Node positive33%  
Trastuzumab utilization rate (%) by status:  Assumptions
 Node-negative, low-risk0% 
 Node-negative, high-risk70%50% to 90% 
 Node positive70%50% to 90% 
 Metastatic60%50% to 90% 

Indication-Specific Cost-Utility Ratios

Several studies report on the cost-effectiveness of trastuzumab in either the metastatic [7,8] or adjuvant settings [9–11] using Markov models to assess the incremental lifetime costs and QALYs of the addition of trastuzumab in metastatic and early-stage adjuvant breast cancer. The MBC models were based on the data from the trials reported in Slamon et al. (2001) [12]. The EBC models were based on the data from the trials reported in Romond et al. (2006) [5]. We developed the base case MBC and EBC cost-effectiveness ratios for trastuzumab based on the models described above.

The number of QALYs gained and the incremental lifetime cost per QALY gained were estimated from the studies cited above that used Markov models, relying in varying degrees of clinical trial data and published aggregate results. These assumptions are summarized in Table 2 for both metatstatic and adjuvant treatment. The Markov models included costs for: 1) HER2 testing (using immunohistochemistry and/or fluorescence in situ hybridization); 2) trastuzumab therapy based on average wholesale prices for medication costs and Medicare reimbursement rates for procedures and resources; 3) patients with EBC were assumed to receive one year of trastuzumab in the adjuvant setting and were treated with trastuzumab until disease progression in the metastatic setting; 4) adverse event monitoring; and 5) treatment of adverse events. In the adjuvant setting, costs for treating metastatic disease were also included for those patients projected to progress over time. The indication-specific outcome measures for the model were QALYs and were based on reported projections from literature and clinical efficacy reported in the studies cited above. Clinical outcomes included survival, probability of recurrence for EBC, time to progression for MBC, and adverse events including the incidence of cardiac dysfunction.

Table 2.  Assumptions regarding the cost-effectiveness of trastuzumab in metastatic and early breast cancer
Inputs*Costs and outcomes of treatment for early breast cancerCosts and outcomes of treatment for metastatic breast cancer
  • *

    All costs and outcomes discounted at 3% annual rate.

  • Garrison et al. [9].

  • Based on current drug costs, survival estimates from Hornberger et al. [8] and utility weights from Elkin et al. [7].

  • QALY, quality-adjusted life years; ICER, incremental cost-effectiveness ratio.

Total costs—no trastuzumab$28,749$40,000
Total costs with trastuzumab$73,672$87,728
QALY: no trastuzumab10.080.70
QALY: trastuzumab11.781.26
Difference  
 Cost$44,923$47,728
 QALYS1.700.56
Incremental cost/QALY gained (ICER)$26,417$85,676

The base case estimates for the costs and QALYs for trastuzumab in the metastatic and adjuvant indications are presented in Table 2. There were multiple sources available for the cost-effectiveness of trastuzumab in the metastatic setting. The assumed MBC ICER represents a combination of two studies, using utility weights from Elkin et al. [7] and survival estimates from Hornberger et al. [8], but also with updated drug cost. As Hornberger et al. had access to the original trial data and used propensity scoring to adjust for crossovers following progression, their survival estimates were used. For the cost-effectiveness of trastuzumab in the adjuvant setting, all data elements were based on the work of Garrison et al. [9], whose estimated ICER falls between two other recently published studies [10,11] and for which we had access to full model, which is necessary for these calculations.

Dynamic Life-Cycle Cost-Utility

We examined life-cycle cost-effectiveness using several measures. The primary measure was an overall life-cycle cost-effectiveness ratio that was discounted to 1998 based on the projected costs and QALYs for all patients receiving trastuzumab between 1998 and 2016. This overall, cumulative ICER was computed by multiplying the mean QALYs gained and mean costs separately by the estimated number of patients with HER2-positive breast cancer in the respective adjuvant and metastatic settings in each year.

QALYS gained and costs were discounted at 3% and were cumulated separately and then divided to calculate the overall life-cycle ICER.

  • image

Where:

  • MBC_QALYs = mean discounted MBC_QALY per patient × Σ (MBC patients over the 19-year period);

  • EBC_QALYs = mean discounted EBC_QALY per patient × Σ (EBC patients over the 19-year period);

  • MBC_Costs = mean discounted MBC_Cost per patient × Σ (MBC patients over the 19-year period); and

  • EBC_Costs = mean discounted EBC_Cost per patient × Σ (EBC patients over the 19-year period)

We also calculated this figure on a “history to date” basis to examine how it changes over time. For comparison, we calculated two alternative “naive” estimates to understand how a proper accounting over the life cycle would compare. These two measures were a lifetime weighted average and annual weighted average, both without discounting to 1998 and based only on the mean ICERs and volume of use. Proper calculation of the overall life-cycle ICER requires specific information on both the numerator and denominators for each of the two indications.

Aggregate Economic Value

We defined net economic value as the potential “social surplus” as defined in economics as the sum of consumer surplus and producer surplus. Essentially, it is the amount by which aggregate societal willingness to pay for benefits exceed the costs of providing them. This requires an assumption about the willingness to pay for a QALY gained. As there is no consensus about a specific value, we use a range of $50,000–$150,000, reflecting the variation of what analysts have used in practice [13,14]. The gross life-cycle economic value is defined as the discounted QALYs gained multiplied by the mean threshold value T that society places on a QALY:

  • image

Where T is the threshold value society places on a QALY and is varied between $50,000/QALY and $150,000/QALY.

  • ndMBC = discounted Σ MBC_QALYs over the 19-year period

  • ndEBC = discounted Σ EBC_QALYs over the 19-year period

The net economic value was thus defined as the difference between this gross willingness to pay and the incremental costs of providing trastuzumab over the entire period, that is, the projected sales of trastuzumab minus the net of other treatment costs (discounted at 3%).

Impact of Adjuvant Use on Metastatic Use

Adjuvant use of trastuzumab is expected to reduce the downstream use of trastuzumab in the metastatic indication. Projected ICERs for EBC must include an assumption about whether future MBC patients will receive trastuzumab. For example, Garrison et al. [9] assumed this for their base case. This raises a question of potential double-counting in the epidemiological projections. In the life-cycle incidence forecasts, we have distinguished between newly diagnosed MBC and previously diagnosed MBC patients. This could lead to an overstatement of both the costs of trastuzumab and the benefits in terms of QALYs gained. We attempted to estimate the potential size of this bias by performing the life-cycle calculation without including the previously diagnosed patients, assuming they were fully reflected in the CE ratio for adjuvant patients.

Sensitivity Analyses

Our approach to sensitivity analyses was to vary deterministically four critical drivers: the uptake of trastuzumab in metastatic and adjuvant use, and the mean ICERs in metastatic and adjuvant use. The former ranged from 50% to 90%, while the mean ICER in metastatic use was varied from $70,000 to $115,000, and was varied from $15,000 to $40,000 in adjuvant use. These ranges should encompass the bulk of the uncertainty generated by all of the underlying variables in each of these components.

Results

  1. Top of page
  2. ABSTRACT
  3. Introduction
  4. Study Data and Methods
  5. Results
  6. Discussion
  7. Conclusions
  8. References

As shown in Figure 1, by 2016, the number of patients treated annually with adjuvant trastuzumab was projected to be approximately three times the number of patients treated with trastuzumab in the metastatic setting. Over the entire period, an estimated 319,000 US women with HER2-positive breast cancer are projected to receive adjuvant trastuzumab and 161,000 are projected to receive treatment for metastatic breast cancer. Between 1998 and 2016, the cumulative net cost of trastuzumab was projected to be $15.4 billion and the cumulative QALYs gained were projected to be 432,547. The cumulative incremental cost-utility ratio was $35,590 per QALY gained. Changes in this cumulative ratio over time are reflected in Figure 2. As this is below most commonly cited thresholds, trastuzumab use appears to have a net surplus. When these QALYs were valued at varying rates of societal willingness to pay from $50,000 to $150,000, the projected gross economic value of trastuzumab treatment due to QALY gains ranged from $21.6 billion to $64.9 billion. Subtracting the cumulative net cost of trastuzumab, the economic value to society is projected to be between $6.2 billion to $49.5 billion Or viewed alternatively, the proportional reward to the manufacturer represents as much as 71% of the social surplus generated to as little as about 24%, depending on the societal willingness to pay for QALYs.

image

Figure 1. Projected numbers of patients treated with trastuzumab for early breast cancer (EBC) and metastatic breast cancer (MBC).

Download figure to PowerPoint

image

Figure 2. Cumulative, discounted life-cycle incremental cost-effectiveness ratio (ICER) for trastuzumab (1998$).

Download figure to PowerPoint

The overall ICER results compared to the indication-specific estimates are depicted in Figure 3. Given the change in volumes of use for the two indications, this estimate is lower than either naïve estimate—a mean annual ICER for 2016 of $40,652 per QALY or a mean cumulative (nondiscounted) ICER of $46,262 per QALY.

image

Figure 3. Indication-specific and overall life-cycle incremental cost-effectiveness ratios (ICERs).

Download figure to PowerPoint

In the sensitivity analyses, as shown in Table 3, the overall life-cycle ICER varied from $32,914 to $42,129 as the utilization rate varied, and the overall life-cycle ICER varied from $21,210 to $52,842 as the indication-specific ICERs were varied.

Table 3.  Sensitivity of life-cycle ICER to use rates (as a percentage of incident population) and to variations in indication-specific ICERs (base case in italics)
Sensitivity to use rates as percentage of incident population
 MBC
EBC 50%60%70%90%
90%$32,914$33,849$34,751$36,465
70%$34,471$35,590$36,662$38,674
60%$35,567$36,808$37,989$40,188
50%$37,008$38,398$39,710$42,129
Sensitivity to variations in indication-specific ICERs
 MBC ICER
  1. Source: Authors' calculations.

  2. ICER, incremental cost-effectiveness ratio; MBC, metastatic breast cancer; EBC, early breast cancer.

EBC ICER $70,000$85,676$115,000
$15,000$21,210$21,658$22,191
$26,417$34,398$35,590$37,053
$40,000$47,602$49,917$52,842

Excluding metastatic patients who were previously diagnosed to adjust for potential double-counting reduces the overall life-cycle ICER by 18% to $29,357, and reduces total QALYs gained by 10% to 384,675. This also decreases the gross life-cycle economic value from $21.6 billion to $19.2 billion at the low end, and from $64.9 billion to $57.7 billion at the high end. However, this probably overstates the reduction somewhat as Garrison et al. [9] conservatively assume that all patients who initially received adjuvant trastuzumab—as well as those who did not—receive trastuzumab after metastatic progression.

Discussion

  1. Top of page
  2. ABSTRACT
  3. Introduction
  4. Study Data and Methods
  5. Results
  6. Discussion
  7. Conclusions
  8. References

We developed a DLM approach and applied the methodology to the breast cancer drug trastuzumab. Specifically, we found that the cumulative, life-cycle ICER trastuzumab was $35,590, with a range of $21,000–$53,000 in sensitivity analyses. The gross life-cycle value was estimated to be between $21.6 billion and $64.9 billion, with a corresponding net economic value of $6.2 billion–$49.5 billion. The overall life-cycle ICER for trastuzumab is less than one half of the projected ICER in the initial indication, and the aggregate economic value is much greater as result of the second adjuvant indication. These results suggest that indication-specific models may have significant limitations for informing policy decisions for drugs with multiple indications licensed over time.

Applying a DLM approach to consider an innovative therapeutic agent highlights the importance of perspective. A short-term perspective focusing on indication-specific cost-effectiveness offers a different view than a longer-term perspective that recognizes the interdependence of drug development decisions and investment incentives. US private payers are generally expected to take a short-term perspective, considering value for money spent on an indication-by-indication basis. Reimbursement systems, particularly in Europe, do not adjust prices or reimbursement rates in response to new information on cost-effectiveness with new indications. In the United States, manufacturers can and do often increase the prices of branded products greater than the rate of general inflation. The net economic surplus would be lower in real terms if the real price of trastuzumab increases over this period. In any case, establishing evidence- and value-based reimbursement systems in this situation, particularly with fixed prices across indications at any point in time, is difficult.

Reimbursement systems that do not account for changing value across indications or over time may produce suboptimal, long-term societal outcomes. Under the patent system, drug prices for branded products can also be seen as a reward for innovation, with potentially far-reaching implications for incentives to undertake future innovative research and development. For example, the calculations of Philipson and Jena [15] for HIV drugs suggest that manufacturers are receiving only about 5% of the social surplus they create. They argue that this low proportion provides a much smaller incentive for innovation. In this trastuzumab case, the estimated range of surplus reward to the manufacturer varied between 24% and 71%, depending on the threshold value for a QALY. A role for the dynamic, long-term perspective is exemplified in oncology where early clinical trials are focused on the sickest patients, followed only many years later in patients diagnosed earlier where the potential benefit is greater. This has significant implications for reimbursement systems, the value of information generated by additional research, and incentives for investment. The recent debate between Claxton and Towse [16,17] about the “value-based” pricing proposal of the UK Office of Fair Trading report [18] has raised the possibility of ex post payments based on performance, including potentially different payments for different indications. The potential positive effects on incentives for investment have been noted by Thornton [19]. Lundin and Ramsberg [20] have recently argued theoretically that a “dynamic cost-effectiveness rule” could improve incentives for research and development.

There are several limitations to our analysis. First, these calculations are based on projections and may not represent actual practice in terms of either cost-effectiveness achieved or future volumes of use. The model is only from a US perspective, although a more global view would be appropriate for considering returns to R&D. The model also does not estimate any impact on volume of trastuzumab use or a revaluation of consumer surplus if there are follow-on competitive compounds (including the entry of branded competitors). If such a competition were to lower use, then the aggregate surplus generated by trastuzumab would be lower, although if price competition lowered prices, then the share of surplus going to the manufacturer would be reduced. And the model does not include any consideration of the use of trastuzumab in other indications, including “off-label” indications such as third- or fourth-line metastatic treatment, where formal testing has not been conducted. An oncology drug could be less cost-effective in such indications, which would raise its overall life-cycle cost-effectiveness ratio, although potentially still increasing aggregate net economic value.

Also, the time horizon to 2016 was chosen to represent a plausible period of limited competition from follow-ons which would also require years of testing to establish efficacy in both of these indications. A slightly shorter effective patent protection would reduce the gross economic value generated, but would have limited impact on the life-cycle ICER or the share of surplus reward, which varies more with the threshold value for a QALY. In addition, our estimates took a payer perspective, which did not include any indirect cost savings or time costs. Presumably, improved cancer survival improves the labor force participation and contributions of these patients, increasing the overall social surplus. This analysis was conducted in real terms (2006 US dollars) discounted to 1998. As is customary, “inflation” in either drug or other medical prices is not included in the projections. Also, we did not consider the impact of any fall in the real price of trastuzumab if there were any competition from a biosimilar product during this period, or of any real increase in price because, as discussed above, manufacturers in the United States often increase prices of branded products over time. Furthermore, we did not model the impact of the potential competitor lapatinib, an oral chemotherapeutic agent recently approved for HER2-positive patients whose metastatic disease has progressed after receiving regimens including trastuzumab.

Our calculations to consider the impact of double-counting suggest that our projection of the life-cycle ICER could be too conservative (i.e., biased upward) since we assume that all recurrent MBC patients receive trastuzumab. Note that although we project that 84% of metastatic use in 2016 would be for previously diagnosed patients, the vast majority of overall use by then would be for EBC patients receiving adjuvant therapy. Nonetheless, this could be an important area for future research, as it would be useful to have a reliable estimate of the impact of improved adjuvant outcomes on metastatic incidence and costs.

Conclusions

  1. Top of page
  2. ABSTRACT
  3. Introduction
  4. Study Data and Methods
  5. Results
  6. Discussion
  7. Conclusions
  8. References

Indication-specific cost-utility or cost-effectiveness models do not account for the important interdependence of drug development and expanded indications over time: the development and approval of subsequent indications is contingent upon success in those developed first. Average ICERs for multi-indication compounds can increase or decrease over the product life cycle. This is especially true in oncology where initial research on safety and efficacy occurs in the most critically ill patients: only after efficacy is demonstrated in this situation can new therapies be tested at earlier stages of disease. The field of pharmacoeconomics and reimbursement policies should give greater attention to dynamic, long-term aspects of drug pricing and reimbursement policies and how they affect incentives for innovation and drug development.

The authors wish to thank Deborah Lubeck, Deepa Lalla, and Carolina Reyes for their help with this research, and Marlene Gyldmark and Jamie Cross for useful comments on earlier drafts.

Source of financial support: This research was supported by unrestricted funding to the University of Washington from Genentech, Inc. This research was presented in part at The International Society for Pharmacoeconomics and Outcomes Research, European Meeting, October, 2006.

References

  1. Top of page
  2. ABSTRACT
  3. Introduction
  4. Study Data and Methods
  5. Results
  6. Discussion
  7. Conclusions
  8. References