Breast cancer is the most common cancer among women, and there are over a million new breast cancer cases in the world each year. In Finland, it is the leading cause of cancer deaths in women, and accounts for 16 percent of all female cancer deaths (25). Approximately 12–30 percent of breast cancer cases overexpress human epidermal growth factor receptor 2 (HER2) (Reference Joensuu, Isola and Lundin6;Reference Owens, Horten and Da Silva14;Reference Slamon, Godolphin and Jones21). These HER2-positive (HER2+) cases are associated with an aggressive disease, and the prognosis is poorer compared with HER2-negative cases (Reference Drucker, Skedgel and Virik2;4). Recently, new targeted therapies have offered new alternatives in the treatment of HER2-positive patients, however, with increasing costs.
Trastuzumab is a monoclonal antibody used for the treatment of HER2-positive breast cancer patients (4). At first, trastuzumab was used only in the metastatic disease and, until recently, it has been the only targeted therapy for HER2-positive breast cancer. There are several studies concerning trastuzumab in adjuvant treatment of early breast cancer (Reference Joensuu, Kellokumpu-Lehtinen and Bono7;Reference Perez, Romond and Suman15–Reference Romond, Perez and Bryant17;Reference Slamon, Eiermann and Robert22–Reference Smith, Procter and Gelber24). Today, it is widely used in this indication. Together with the indication extension, the number of eligible patients grew substantially. This, in turn, increased concern about the affordability of trastuzumab, because the healthcare budget and resources are limited.
Trastuzumab is used in addition to existing therapies and does not replace them. Because it is an expensive drug, it causes a substantial impact on the healthcare budget. Recent reviews show that in most analyses, trastuzumab is considered to be cost-effective, despite the diversity of the results (Reference McKeage and Lyseng-Williamson10;Reference Younis and Skedgel26). However, cost-effectiveness analyses alone do not provide information on the drug's impact on the total healthcare budget, because this is dependent on the number of treated patients. Thus, the expected budget impact of a new treatment should be explicitly estimated, in addition to the traditional cost-effectiveness. A few of the cost-effectiveness studies of adjuvant trastuzumab have included evaluation of its budgetary impact (Reference Liberato, Marchetti and Barosi8;Reference Neyt, Huybrechts, Hulstaert, Vrijens and Ramaekers11;Reference Shiroiwa, Fukuda, Shimozuma, Ohashi and Tsutani19). Furthermore, the cost burden of monoclonal antibodies has been evaluated from the Canadian healthcare perspective (Reference Drucker, Skedgel and Virik2). However, a comprehensive budget impact analysis of trastuzumab that includes the estimated effectiveness of the treatment and probabilistic analysis, has not, to our knowledge, been previously published.
In this study, we created an evaluation tool for estimating the budget impact of new cancer treatments. This was done on the request of the consortium of the Finnish Office for Health Technology Assessment (Finohta) and local hospital districts. In Finland, the health care system is publicly funded, and hospital districts are allowed to operate rather independently. Thus, there are concerns regarding the cost burden of trastuzumab, both at the national level and in hospital districts. In this study, the new tool was used to explore the budget impact of trastuzumab in early breast cancer in a single Finnish hospital district. Furthermore, we analyzed how different treatment protocols and changes in the number of patients affect the estimated budget impact.
MATERIALS AND METHODS
Perspective and Time Horizon
This study was conducted from the perspective of a single hospital district with ca. 250,000 inhabitants (payer's perspective). In the target organization (Hospital district of Northern Savo), all trastuzumab treatments are given in one hospital (Kuopio University Hospital). Only direct costs allocated to the hospital district were included in the study. The maximum follow-up time was set to be 4 years, because a longer perspective was not considered relevant for the budget holder. All results are depicted cumulatively from 1 to 4 years.
Population
Population-based epidemiological data were obtained from the Finnish Cancer Registry, which covers the vast majority of cancer cases in the country. The 5-year average yearly number of new breast cancer cases (ICD-10 C50) in the hospital district was 176. The age-adjusted incidence of breast cancer in the whole country in 2006 was 86.6/100,000, which is slightly higher than in the target hospital district (3). The prevalence of HER2+ was 18 percent, which reflects the 3-year average HER2+ incidence rate in the local district (1). Thus, there are approximately thirty-two new HER2-positive breast cancer cases in this district yearly. Previously diagnosed breast cancer cases were not included in the analysis. Ten percent of the HER2-positive patients were assumed not to receive trastuzumab because of other diseases and poor general condition, or because of a very small tumor size.
Model
A spreadsheet model was created to assess the budget impact of trastuzumab. The model took into account the number of patients, HER2+ prevalence, the length and cost of treatment, and its effectiveness. Estimates concerning distant disease free survival (DDFS) were derived from the published data of a Finnish trial (Reference Joensuu, Kellokumpu-Lehtinen and Bono7). These estimates were applied to the model to include the effectiveness of the treatment. Mortality was not included due to the short time horizon. The effectiveness of trastuzumab was assumed not to depend on the length of the treatment. The model used is a simple stage transition model typically used in cost-effectiveness analyses. The mutually exclusive health stages in the model are “free of distant recurrence” and “distant recurrence.” The number of new cases was used as the population entering the model each year. Yearly new cases were followed throughout the model, depending on the chosen time span. Each additional follow-up year increased the number of eligible patients and prolonged the time spent in the model. The results of the model apply to periods from 1 to 4 years. Consistent with the current recommendations for budget impact analyses, discounting was not incorporated in the model (Reference Mauskopf, Sullivan and Annemans9).
Treatment Mix and Costs
The treatment mix and costs of breast cancer therapy are based on current clinical practice in the treating hospital. Activity-based costing was used as the costing method (in 2008 euros, VAT excluded). Patient copayments attributable to hormonal cancer therapy and oral cytostatics were not included, because in Finland they are reimbursed by the Social Insurance Institute (SII). In the target hospital district, chromogenic in situ hybridization (CISH) is used to test the patient's HER2 status. Because all new cases are tested similarly, the cost of testing was not included in the analysis.
Treatment Cost of HER2+ Early Breast Cancer. The first year after diagnosis is more treatment-intensive than the following years. Average first-year adjuvant treatment without trastuzumab constituted a total cost of €9540 per patient for the hospital district (Supplementary Table 1, which is available at www.journals.cambridge.org/thc2010013). The treatment includes hormonal treatment (tamoxifen or aromatase inhibitor), but their cost is not allocated to the hospital district. Trastuzumab is given in addition to standard breast cancer treatment. In local clinical practice, adjuvant trastuzumab treatment in early breast cancer includes two treatment lengths—9 weeks and 1 year. All patients initially receive 9 weekly doses of trastuzumab (first dose 4 mg/kg, then 2 mg/kg) together with chemotherapy. Subsequently, 40 percent of the patients receive an infusion every third week (first dose 8 mg/kg, then 4 mg/kg) for up to 1 year. The cost of a subsequent month with trastuzumab was €2,800, which adds up to €35,000 a year. All trastuzumab costs include the administration and preparation costs covered by the treating organization. Trastuzumab drug wastage in the hospital is marginal due to the concentrated preparation practices, and is, therefore, not included. Because stable use of trastuzumab does not reflect real clinical practice, the developed model was adjusted to take into account variability in treatment lengths. Every tenth patient was assumed to discontinue the trastuzumab treatment at 6 months due to adverse events or other reasons. All adverse events were assumed to be fully reversible, and thus not to cause any additional costs to the hospital district.
Treatment Cost of HER2+ Metastatic Breast Cancer. The use of trastuzumab in metastatic disease was included in the model according to local clinical practice. All HER2-positive patients with metastatic breast cancer were assumed to receive trastuzumab, even if they had received it earlier in the adjuvant setting. In the treatment of metastatic breast cancer, trastuzumab is given in 3-week cycles for 1 year, which constitutes an average cost of €33,600 per year. The average treatment mix and cost are described in Supplementary Table 2, which is available at www.journals.cambridge.org/thc2010013.
Sensitivity Analyses
Sensitivity analyses were performed to evaluate uncertainty attributable to the applied assumptions and model parameters. Uncertainty related to the assumptions was explored by performing a series of “what-if” scenarios, which consisted of the following one-way sensitivity analyses. In scenario A, all patients were assumed to receive trastuzumab according to the 1-year treatment schedule. In scenario B, a short, 9-week treatment schedule was applied to all patients. The prevalence of HER2+ was assumed to be 12 percent in scenario C and 25 percent in scenario D. In scenario E, the cost of the trastuzumab treatment decreased by 40 percent. Finally, in scenario F, the effectiveness of the treatment was discarded from the model. These extreme scenarios were assumed to reflect the range of circumstances that the budget holders may face.
In addition to the traditional deterministic sensitivity analyses, a probabilistic sensitivity analysis was also performed. It takes into account the degree of variability and uncertainty related to these parameters simultaneously. In the model, probability distributions were assigned to all key parameters (i.e., number of patients, HER2+ prevalence, transition probabilities and treatment costs).
RESULTS
Base–Case Results
The introduction of trastuzumab in the treatment of HER2-positive early breast cancer caused substantial costs to a rather small hospital district with 250,000 inhabitants and around 29 patients receiving adjuvant trastuzumab each year. In a 4-year follow-up period, the net budget impact was €1,302,000, and in 1 year the figure was approximately €474,000. The additional cost per treated patient was around €16,000 in the first year. The total net costs included adjuvant treatment, the cost of the follow-up period, and treatment of metastatic disease. Most of the additional costs were accrued from the acquisition costs of adjuvant trastuzumab. However, there were also savings related to adjuvant trastuzumab treatment due to improved cancer recurrence rates. Thus, the acquisition costs were partially offset by the reduction in costs associated with the treatment of cancer recurrence and metastatic disease. However, the majority of these savings were due to the reduction in late-stage trastuzumab use. Figure 1 shows the contribution of each cost type to the total net budget impact when trastuzumab is added to the treatment of early breast cancer.
Figure 1. Cumulative changes in different cost types by time, along with the net budget impact of trastuzumab in early breast cancer.
Sensitivity Analyses
The sensitivity of the results was studied through alternative case scenarios (Figure 2). The short treatment schedule (scenario B) gave a smaller budget impact than the other scenarios. The budget impact of this 9-week treatment decreased over time when effectiveness, in terms of disease progression, was included in the model. The difference between short (B) and prolonged (A) therapy was around €750,000 in 1 year, and €3,000,000 in 4 years. In scenario E, the treatment cost of trastuzumab was reduced by 40 percent. This constituted a 46 percent reduction in the 4-year budget impact compared with the base case. When the prevalence of HER2+ was altered from 12 percent to 25 percent, the 4-year budget impact varied between €868,000 and €1,808,000. As assumed, the results of the study were fairly sensitive to the length of the trastuzumab treatment and the number of patients. In addition, when the effectiveness of the treatment was excluded from the model (scenario F) the budget impact was notably higher than in the base case.
Figure 2. Cumulative net budget impact of trastuzumab in the selected what-if scenarios and base case analysis (in a population with 176 annual breast cancer cases).
The results of probabilistic sensitivity analysis are presented as affordability curves (Figure 3), which depict the probability of staying within a certain budget (Reference Sendi and Briggs18). The information provided by these curves can be expected to be useful for the budget holders who make decisions under uncertainty. For example, if the annual budget constraint is €500,000, there is around a 70 percent probability that the budget will not be exceeded in the first year. In 4 years, the same level of confidence is reached when the budget constraint is €1,350,000. However, these particular results hold only with the base–case assumptions. Uncertainty related to the results grows along with the longer time horizon.
Figure 3. Affordability curves showing the probabilities that trastuzumab is affordable as a function of the budget constraint.
DISCUSSION
Evaluation of both cost-effectiveness and budget impact of new treatments is necessary due to limited healthcare resources. Economic evaluations are needed, not only to ascertain the best possible value for money, but also to assess to what extent the treatments are affordable. Nevertheless, the means for performing timely economic evaluations are currently limited. In this study, an evaluation model was developed to assess the budget impact of new cancer treatments. We found that adjuvant trastuzumab causes substantial costs even for a rather small hospital district. The net budget impact was greatly influenced by the number of treated patients, and the length and cost of the treatment.
Trastuzumab has proven its clinical efficacy in HER2-positive breast cancer. The aim of adjuvant treatment is to destroy micrometastases, and thus to prevent recurrences and improve the survival of the patients. Most clinical evidence regarding adjuvant therapy is from a 1-year treatment (Reference Perez, Romond and Suman15–Reference Romond, Perez and Bryant17;Reference Slamon, Eiermann and Robert22–Reference Smith, Procter and Gelber24), but a shorter 9-week treatment has also been studied (Reference Joensuu, Kellokumpu-Lehtinen and Bono7). However, the optimal duration or dosing frequency of trastuzumab is still not known. In the future, longer follow-up data of adjuvant trastuzumab treatment will be available from ongoing trials. Survival estimates are also dependent on patient characteristics, including age and disease stage. However, currently available information did not enable us to do a patient subgroup analysis. In this budget impact analysis, the effectiveness of the trastuzumab treatment was estimated according to the published data of the FinHer (Finland Herceptin) trial (Reference Joensuu, Kellokumpu-Lehtinen and Bono7). The health benefit gained from the treatment was assumed to be the same despite the length of the treatment. However, if the 1-year treatment would be clinically better than the 9-week treatment, it would also lead to bigger savings during the follow-up than evaluated in this analysis. Furthermore, treatments given together with trastuzumab, and possible synergism with certain chemotherapy regimens, can be of significance to the treatment outcomes. Nevertheless, the most reliable data concerning the best way to give the trastuzumab treatment in early breast cancer will be acquired from randomized studies comparing short and long trastuzumab treatment.
Effectiveness of treatment is seldom included in budget impact models, although it has been considered to be important (Reference Skedgel, Rayson and Younis20). In devastating and progressing diseases such as cancer, improved outcomes should be accounted for to depict real treatment practice. In local clinical practice, the trastuzumab treatment was continued beyond disease progression. This was also included in the analysis. The financial impact of late-stage trastuzumab may, however, be easily discarded from the total net budget impact. A budget impact analysis that is solely based on acquisition costs ignores the effects of the treatment. It simply describes how much money is spent on the drug within a certain time horizon, but, at the same time it fails to include other possible costs or savings related to the treatment. After adjuvant treatment, patients without relapsed disease will be able to continue their everyday life and return to work. However, productivity losses or time costs were not included in the study. The complex nature of reality can not be completely captured in economic evaluations. However, using the best available information and modeling techniques, this complexity may, to some extent, be taken into account.
Hospitals are not always able to manage in the pressure of increasing costs of new pharmaceuticals if costs grow faster than the budget assigned for the drug. Unrestricted reimbursement of expensive and relatively widely used pharmaceuticals, such as trastuzumab, may lead to inequality in treatment access. In the Netherlands, it has been already demonstrated that trastuzumab was unevenly distributed among patients in different hospitals, largely due to limited drug budgets. Nevertheless, budget impact estimations may play a role in preserving patient equality (Reference Niezen, Stolk, Steenhoek and Uyl-De Groot12;Reference Niezen, de Bont, Busschbach, Cohen and Stolk13).
Budget impact analyses are usually performed from the perspective of a target organization (e.g., a hospital district in this study). However, it is important to note that cost shifting between a hospital and other financing bodies, such as municipalities and SII in Finland, may lead to suboptimal decisions from the social perspective or the perspective of overall health care planning (Reference Häkkinen5). This budget impact model was created to depict the current situation in one Finnish hospital district. Thus, the results are not necessarily applicable to other jurisdictions due to possible differences in treatment practices and related costs. Furthermore, treatment practices may change over time, and thus affect the applicability of the results of economic evaluations. In addition, the results may need to be updated if new products are introduced for the same indication, because they may affect both the proportion of patients receiving the treatment and the actual price of the treatment. By using a relatively short time horizon, the number of assumptions does not grow to be excessive. The markets are volatile and may change when new products are introduced. Long-term budget impact analyses do not necessary depict the true future situation, which leads to instability in long-term results. In our analysis, the market diffusion rate was considered to be stable, although the possibility to vary the annual diffusion rate was enabled in the model. Possible market changes should be recognized when budget impact estimations are performed.
When new products are introduced, they usually slowly replace older treatments. This releases resources to fund the new treatment to some extent. However, in respect of add-on medicines, all costs will have to be compensated with additional funding or sacrifices made elsewhere. Due to the large potential financial impact of trastuzumab, there was an increased need for more comprehensive budget impact analysis. Despite the favorable incremental cost-effectiveness rations concerning trastuzumab, the eventual economic consequences of the treatment are substantial.
Practical evaluation tools offer a great help for decision makers and budget holders—especially with respect to new and expensive targeted therapies that add significant costs to the health care system. Alternative scenarios are an essential part of results in budget impact analyses. The model developed in this study was used in only one hospital district. However, when local epidemiological and treatment data are obtained, the model may be used similarly in other jurisdictions. The treatment costs of different stages of breast cancer were those of a single hospital and were derived from an average treatment protocol practiced in the target organization. To provide real-time estimates of the economic impact of novel treatments and to allow fluent data collection, more attention should be paid to improving the usability and coverage of electronic databases.
CONCLUSIONS
The present analysis found that the budget impact of trastuzumab is considerable, from the perspective of a Finnish hospital district. However, when the effectiveness of the treatment is taken into account, there are also savings related to adjuvant trastuzumab treatment. The length of the treatment has a strong effect on the eventual budget impact. Future trials will show to what extent the duration of trastuzumab treatment affects its effectiveness and the cost-effectiveness of the therapy.
SUPPLEMENTARY MATERIAL
Supplementary Table 1
Supplementary Table 2
CONTACT INFORMATION
Timo T. Purmonen, MSc (Pharm.), MSc (Econ.) (timo.purmonen@uef.fi), Researcher, Department of Social Pharmacy, Center for Pharmaceutical Policy and Economics, University of Kuopio, P.O. Box 1627, Kuopio FIN-70211, Finland
Päivi K. Auvinen, MD, PhD (paivi.auvinen@kuh.fi), Specialist in Oncology, Department of Oncology, Kuopio University Hospital, P.O. Box 1777, Kuopio 70780, Finland
Janne A. Martikainen, PhD (Health econ.) (janne.martikainen@uef.fi), Research Director, Department of Social Pharmacy, Centre for Pharmaceutical Policy and Economics, University of Kuopio, P.O. Box 1627, Kuopio FIN-70211, Finland
