Health economic evaluation of bevacizumab for metastatic colorectal cancer
Health technology assessment
|Updated
Key message
We have undertaken a model based analysis of the cost effectiveness of bevacizumab for first line treatment of metastatic colorectal cancer in Norway. We have investigated the effect of adding bevacizumab to standard chemotherapy. The analysis does have a considerable uncertainty, and the results should therefore be interpreted carefully.
In our main calculation we have estimated the cost per life year gained, by adding bevacizumab to standard first line treatment with irinotecan or oxaliplatin based chemotherapy, to 668 000 NOK and 549 000 NOK – seen from a health care and a societal perspective, respectively. The cost per life year gained is, independent of perspective, higher than several suggested pragmatic thresholds for assessment of cost effectiveness in the health care sector. The main part of the incremental costs by adding bevacizumab to standard treatment is related to the acquisition cost of bevacizumab. The health effect estimate, measured as life year gained, is 0,41 year (undiscounted) in our main calculation. The incremental cost per average patient is 247 000 NOK and 203 000 NOK, in the health care and societal perspective, respectively.
The sensitivity analysis shows that the cost per life year gained is very dependant on the health effect estimate, dependant on the bevacizumab price and the assumed number of bevacizumab treatment cycles, but to a lesser extent dependant of non drug costs, like treatment administration and evaluation. The uncertainty of the treatment effect estimate is due to the estimate being derived from only one relevant clinical study. The study has limited period of follow up (11-31 months) and compares bevacizumab with another, and presumably less effective irinotecan regime, than what is standard in Norway.
Also similar analyses from Denmark and England indicate relatively high costs per life year gained, of bevacizumab for colorectal cancer.
The evidence basis for health economic evaluations of bevacizumab in Norway (and other countries) would have been considerably better if there had been more relevant clinical studies, with longer follow up, and where bevacizumab where compared with standard treatment in Norway (and other countries).
Conclusion
In line with analyses from Denmark and England, the Norwegian analysis indicates that adding bevacizumab to standard irinotecan based first line treatment for metastatic colorectal cancer in average leads to 0,41 life years gained, and that the costs for this gain is higher than several suggested pragmatic thresholds for assessment of cost effectiveness in the health care sector. There are however no clear decisions on what thresholds should apply for cost effective treatment in Norway, and if these thresholds are independent of disease and patient life situation.
Thus, we have no basis for concluding that adding bevacizumab to standard first line treatment of metastatic colorectal cancer is cost effective. The calculations are however uncertain. Especially, the treatment effect estimate is uncertain, being based on only one clinical study, with a presumably sub optimal irinotecan regime and limited follow up period.
Summary
Background
Colorectal cancer is the second most commonly diagnosed cancer in Norway. About 30 % of patients presents with metastases at the time of diagnosis, while another 20 % develops metastases during the course of the disease. Five year survival rates for patients with distant metastases are reported at around 5 % during the 1990's according to the National Cancer Registry. For most patients with irresectable metastases, palliative chemotherapy is appropriate; to postpone progression of disease, prolong survival and to maintain a satisfactory quality of life. Today, 5-FU and calcium folinate combined with oxaliplatin or irinotecan is the first line treatment for most patients younger than 75 year of age. The standard Nordic treatment regimes with irinotecan and oxaliplatin are called FLIRI and FLOX, respectively. These regimes are used in Norway and are viewed as being equally effective.
Bevacizumab (Avastin®) is a relatively new monoclonal antibody targeted against VEGF (Vascular Endothelial Growth Factor), which is involved in the creation of new veins (angiogenese) in cancer tumours. By binding to VEGF bevacizumab blocks the binding of VEGF to its receptors at the surface of the endotel cells. In Norway bevacizumab is registered for first line treatment in combination with 5-FU/calcium folinate or an irinotecan based regime. It is also used in combination with oxaliplatin based regimes.
Monoclonal antibodies, like bevacizumab, are amongst the most costly drugs. According to the Norwegian patient right law, the balance between treatment costs and health gains should be reasonable to get funded. The objective of this analysis is to quantify the incremental costs and health effects by adding bevacizumab to standard treatment of metastatic colorectal cancer. The treatment effect is measured in life years gained as this is a relevant outcome and because there are no good, relevant data on quality of life effects.
Methods
We assessed the cost effectiveness of bevacizumab by two approaches:
- Analysis by our own health economic model ("model analysis") – for bevacizumab in first line treatment of metastatic colorectal cancer in Norway.
- Literature review of Norwegian and international cost effectiveness studies of bevacizumab for metastatic colorectal cancer.
1. Model analysis
We identified no randomised clinical effect studies of bevacizumab for metastatic colorectal cancer in a Norwegian setting. Therefore, we developed a health economic simulation modell adapted to our research question. The model is based on a typical metastatic colorectal cancer patient. We investigated two main treatment pathways, one with bevacizumab (plus FLIRI or FLOX) in first line treatment, and one without bevacizumab (but with FLIRI or FLOX) in first line treatment. In second and third line, the treatment is equal in the two pathways. We compared costs and health effect (life years) for the two pathways. The main outcome was cost per life year gained. We analysed costs and health effect from two perspectives: The health care perspective (incl VAT) and the societal perspective (excl. VAT).
In cooperation with the research group conducting the Norwegian Knowledge Centre for the Health Services' review of clinical effect of bevacizumab, we searched for randomised phase III studies that measured the effect of bevacizumab in combination with an irinotecan or oxaliplatin based regime. We identified only one published study: Hurwitz et al. In this study bevacizumab + IFL (irinotecan, bolus flurouracil and leucovorin) is compared with IFL alone for patients with previously untreated metastatic colorectal cancer. In our main scenario we based our effect estimates on the overall survival curves with and without bevacizumab in the Hurwitz study. We adapted a Weibull function to the observed survival curves and used this to estimate survival and life year gained beyond the 31 months period reported by the Hurwitz study.
The estimated costs for the health care sector includes costs for the drugs/chemotherapy, pharmacy services (final preparation of the drugs/chemotherapy), administration of the treatment (by nurses), CT-scanning for evaluation, outpatient/policlinic consultations to hospital physicians, prevention and treatment of adverse effects, and reimbursement of patients' travel expenses, food and stay at patient hotel. The costs depend on the number of treatment cycles. For the first line treatment we used data from the Hurwitz study. The number of bevacizumab treatment cycles (one each second week) is 15 in our main scenario.
2. The literature review
We carried out a systematic search for Norwegian and international studies. We searched in relevant, electronic databases, at Nordic HTA web sites, and asked the producer of bevacizumab for studies. We used a checklist to assess the quality of the included studies and collected data from the studies.
Results
In our main scenario the life year gained was 0,41 year in average per patient, and the costs for bevacizumab treatment was 668 000 NOK and 549 000 NOK per life year gained, seen from a health care and societal perspective, respectively. The main part of the incremental costs by adding bevacizumab to standard first line treatment is related to the acquisition costs of bevacizumab. Other incremental costs occurs because the number of treatment cycles in the bevacizumab pathway are higher than the number of treatment cycles in the non bevacizumab pathway. This leads to incremental costs to the cytostatika in combination with bevacizumab, and to administration and preparation of treatment, cycle evaluations etc. In our model the incremental costs with bevacizumab treatment occur in the first year. The incremental costs in our main scenario are 247 000 NOK per average patient, seen in the health care perspective.
The sensitivity analysis shows that the cost per life year gained is very dependant on the health effect estimate, dependant on the bevacizumab price and the assumed number of bevacizumab treatment cycles, but to a lesser extent dependant of non drug costs, like treatment administration and evaluation.
In the literature review we found two analyses similar to our. The evaluations from Denmark and England indicates relatively high costs per life year gained for bevacizumab for colorectal cancer, 564000 NOK and 601000 NOK, in the health care perspective.
Discussion
Our analysis is uncertain and thus, the results must be interpreted carefully. Especially there is uncertainty related to the health effect estimate, and to the number of treatment cycles required to achieve the health effect, i.e. the life years gained. Our analysis is based on effect data from only one clinical study. In the study bevacizumab is compared with an irinotecan based regime which is somewhat different, and presumably less effective, than current treatment regimes in Norway. The clinical study results were published in 2004 with 11-31 months follow up, and there are not published any results beyond this follow up period. The published study paper did not include any information on the number of cycles and other data for resource use, and we have not got access to such information.
The costs per life year gained might be too low in our main scenario, because our health effect estimate might be too optimistic. The effect reported by clinical studies might be greater than what is achievable in actual clinical practice. It also might be possible that parts of the bevacizumab life year gain in the Hurwitz study already has been achieved in Norway, due to the standard use of a presumably better treatment than what is given to the control group in the Hurwitz study. There are no published studies on the overall survival effect of adding bevacizumab to an oxaliplatin based regime as first line treatment. Existing preliminary data for such a treatment is difficult to interpret, but the life year gain seems to be less than what is achieved by adding bevacizumab to an irinotecan based regime.
Conclusion
In line with analyses from Denmark and England, the Norwegian analysis indicates that adding bevacizumab to standard irinotecan based first line treatment for metastatic colorectal cancer in average leads to 0,41 life years gained, and that the costs for this gain is higher than several suggested pragmatic thresholds for assessment of cost effectiveness in the health care sector. There are however no clear decisions on what thresholds should apply for cost effective treatment in Norway, and if these thresholds are independent of disease and patient life situation.
Thus, we have no basis for concluding that adding bevacizumab to standard first line treatment of metastatic colorectal cancer is cost effective. The calculations are however uncertain. Especially, the treatment effect estimate is uncertain, being based on only one clinical study, with a presumably sub optimal irinotecan regime and limited follow up period.