|
Advertisement |
|
|
||
 |
|
|||||
|
|||||
|
 |
||||||
|
||||||
|
||||||
|
Journal of Clinical Oncology, Vol 26, No 2 (January 10), 2008: pp. 290-296 © 2008 American Society of Clinical Oncology. DOI: 10.1200/JCO.2007.13.0898 Modeling the Cost Effectiveness of Secondary Febrile Neutropenia Prophylaxis During Standard-Dose Chemotherapy
From the Departments of Medical Oncology, Pulmonary Diseases, and Epidemiology, Biostatistics & Health Technology Assessment, Radboud University Nijmegen Medical Centre, Nijmegen; Department of Health Organisation, Policy, and Economics, Maastricht University; Department of Clinical Epidemiology and MTA; and Department Internal Medicine, Division Medical Oncology, University Hospital Maastricht, Maastricht, the Netherlands Corresponding author: J.N.H. Timmer-Bonte, MD, 452 Department of Medical Oncology/Department of Pulmonary Diseases, Radboud University Nijmegen Medical Centre, PO Box 9101, 6500 HB Nijmegen, the Netherlands; e-mail: J.Timmer{at}onco.umcn.nl
Purpose Current guidelines (ie, by the American Society of Clinical Oncology and the European Organisation for Research and Treatment of Cancer) do not recommend secondary infection prophylaxis, whereas, in contrast, caregivers prefer secondary prophylaxis to chemotherapy dose reduction after an episode of febrile neutropenia (FN). Because granulocyte colony-stimulating factor (G-CSF) is expensive, this study investigates the economic consequences of secondary prophylactic use of different prophylactic strategies (antibiotics, antibiotics plus G-CSF, and a combined sequential approach) in a population at risk of FN, using a Markov model.
Methods The input for the model is mainly based on the clinical outcome and patient-based cost data set (adopting the health care payer's perspective for the Netherlands) derived from a randomized study on primary prophylaxis in small-cell lung cancer (SCLC) patients; establishing mean cost of an episode FN of
Results Secondary prophylaxis with antibiotics was the least expensive strategy (mean, Conclusion This model-based economic analysis demonstrates that in the Netherlands and most likely also in the United States, if secondary prophylaxis is preferred, the strategy with antibiotics is recommended.
Febrile neutropenia (FN) is a potentially fatal chemotherapy-related complication leading to morbidity and consumption of medical resources, primarily because of hospitalization and intravenous antibiotics use. Primary prophylaxis with antibiotics or with granulocyte colony-stimulating factor (G-CSF) can significantly decrease the incidence of FN.1-3 Recently, we reported the results of a randomized phase III clinical trial assessing the impact of primary G-CSF prophylaxis in small-cell lung cancer (SCLC) patients treated with standard-dose chemotherapy and an estimated baseline risk of FN in the first cycle of at least 20% despite the use of primary antibiotic prophylaxis. Patients receiving prophylactic antibiotics plus G-CSF developed significantly less FN (18% v 32%).4
Because of considerable expenses (G-CSF) and antimicrobial resistance (antibiotics), current guidelines recommend restraint in use of primary prophylaxis and advise it only in patients considered at risk of FN.5-7 A prospective economic analysis performed alongside our randomized trial demonstrated that, the addition of primary G-CSF prophylaxis to antibiotics was not cost saving.8 The mean incremental cost of adding G-CSF per patient in the first cycle (with the highest incidence of FN) amounted to A prior episode of FN is considered a risk factor for developing FN in further cycles, with recurrences reported in 50% to 66% of patients.4,10,11 The administration of prophylaxis after a prior episode of FN is called secondary prophylaxis. The placebo-controlled G-CSF registration trial by Crawford et al1 provided some evidence on the efficacy of secondary prophylaxis because patients in the control arm who developed FN received open-label G-CSF in subsequent cycles of chemotherapy. Such patients who were subsequently treated with open-label G-CSF had an incidence of (only) 23% in cycle 2, despite receiving the same doses of chemotherapy. Because of missing high-level evidence data, present guidelines do not recommend secondary prophylaxis. However, in clinical practice, caregivers prefer secondary prophylaxis to chemotherapy dose reduction after an episode of FN in 50% to 90% of patients, as several surveys in 1996, 2003 and 2005 have demonstrated.12-14 The previously reported prospective economic evaluation alongside our randomized trial on primary FN prophylaxis resulted in a patient-based cost data set leading to a better understanding of the cost variance in production processes of primary prophylaxis.8 As an extension to our primary prophylaxis trial, using the data from this trial as input, we planned a modeling study investigating the economic consequences of secondary prophylaxis comparing three secondary prophylaxis strategies. In this prospective economic analysis, consumed medical resources and consequences with regard to less morbidity are modeled in monetary terms and FN-free cycles using a Markov model taking into account the probability of premature chemotherapy cessation and the stage-variant probability of FN.
Sources This analysis is, to a large extent, based on input from the clinical outcome and cost data set from our randomized phase III study in SCLC patients.4 That study included patients at risk of FN defined as 60 years of age or older, extensive disease, a Karnofsky performance status of 40% to 70%, and/or having received prior chemotherapy. Patients were treated with five cycles of what was at the time standard-dose chemotherapy (cyclophosphamide, doxorubicin, and etoposide) every 3 weeks and received primary prophylaxis with antibiotics (ciprofloxacin and roxithromycin) or with antibiotics plus G-CSF. Primary outcome was the incidence of FN and treatment-related cost. The combination of G-CSF and antibiotics in primary prophylaxis more effectively reduced the incidence of FN compared with antibiotics, with the effect being largest in the first cycle. However, this gain in clinical effect came with substantial incremental costs (Table 1). Not all necessary input for the model could be derived from these study results, and, consequently, data from other published sources were used.
Markov Model The clinical process and the morbidity and mortality that a hypothetical patient went through during a maximum of five chemotherapy cycles was modeled by means of a probabilistically analyzed Markov chain process (Fig A1, online only), which is a multistate transitory model that allows patients to make event-driven transitions among various health states, at different transition probabilities, over certain periods. After each chemotherapy cycle, the patient may be in different states that are mutually exclusive (Fig 1). We assumed that an episode of FN without prophylaxis always leads to modification of therapy and that, in all subsequent cycles, prophylaxis was administered (Fig A1). Three secondary prophylactic strategies are considered: (1) antibiotics, (2) antibiotics plus G-CSF, or (3) antibiotics after the first episode of FN and antibiotics plus G-CSF after another episode of FN (termed "sequential"). The passage of time is divided into intervals representing a complete chemotherapy cycle (up to five cycles). During each cycle, each hypothetical patient may remain in the same state of health or move to another state ("FN" and "no FN"; Fig 1), except when chemotherapy is discontinued (absorbing state, "stop"; Fig 1). The consequences of FN vary per strategy (in cost and effect). The probability of making a transition from one state to another during a given cycle is called a transition probability. The costs of FN, full- or reduced-dose chemotherapy, prophylactic antibiotics and/or prophylactic G-CSF were assigned to health states and driven by events that occurred after transition probabilities. All of this information was analyzed using software from TreeAge (Williamstown, MA).
Input for the Model Costs. The clinical trial on primary prophylaxis provided in a patient-based cost data set adopting the health care payer's perspective in the Netherlands, as described elsewhere in detail.8,15 For the Markov model, the original sample of cost data related to FN was bootstrapped and resulted in a normal distribution of costs of FN treatment (Fig A2, online only).16 The model was probabilistically analyzed using combined first- and second-order Monte Carlo simulation. This procedure runs a fixed number (1,000) of individual hypothetical patients randomly through the model and simultaneously samples from the distributions assigned to variables in the model. Our overall analysis contained two multidimensional Monte Carlo simulations: one considering costs as outcome parameter only, and one considering cost effectiveness. Transition probabilities. The probabilities were derived from the randomized phase III trial, except for the probabilities of the first episode of FN. The efficacy of the sequential prophylactic strategy has never been clinically investigated to our knowledge; therefore, we used a combination of probabilities as found in our clinical trial.4 Table 2 provides an overview of critical probabilities in the model and their sources. The probability of a first episode of FN was based on the pivotal G-CSF trial by Crawford et al,1 and, as such, on a similar patient population as in our primary prophylaxis study. If a patient stopped treatment after an episode of FN, transition costs were calculated. The stop probability in patients who did not experience FN was based on a subgroup of SCLC patients treated in an EORTC trial with standard-dose CDE (cyclophosphamide, doxorubicin, and etoposide) without prophylaxis (0.05, 0.09, 0.03, 0.44, and 0.36 in cycles 1, 2, 3, 4, and 5 respectively).3,17 The stop probabilities in patients who experienced FN -and thus received prophylaxis were based on the probabilities found in our trial and varied per prophylactic strategy (0.04, 0.09, 0.19, and 0.20 in cycles 1, 2, 3, and 4 with antibiotics; 0.11, 0.1, 0.11, and 0.09 with antibiotics plus G-CSF; 0.09, 0.1, 0.11, and 0.09 in sequential prophylaxis).
Effects. The underlying study on primary prophylaxis did not find a difference in mortality rates between treatment groups. However, the different strategies resulted in different FN rates, and, consequently, the extra costs might be traded off by an extra effect. Therefore, effect was defined as an FN-free cycle. We assigned to each cycle without FN a value of 1, and if FN occurred, a transition reward of –1 was applied. The incremental cost-effectiveness ratio (ICER) was expressed as the mean incremental cost per FN-free cycle. A cost-effectiveness plane and a derived acceptability curve graphically represent cost effectiveness.
Exploration of Nonparametric Uncertainty: Sensitivity Analyses In the baseline model, we used recurrence rates as found in our primary prophylaxis trial.4 Note that in this study we observed, based on a small number of patients, higher recurrence rates in the patients receiving antibiotics plus G-CSF compared with the antibiotics arm. Because variations in recurrence rates of FN depend on patient characteristics, we subjected the recurrence rates to sensitivity analysis. We identified one article where patients who experienced FN in cycle 1 continued full-dose chemotherapy with secondary G-CSF prophylaxis, and they reported an FN recurrence rate of 0.16 and 0.10 in cycles 2 and 3 respectively.10
The cost of an episode of FN found in our trial is relatively low (mean,
Baseline Model of Secondary Prophylaxis The first multidimensional Monte Carlo simulation focusing only on costs showed that secondary prophylaxis with antibiotics was the least expensive strategy, with mean costs of  4,496 (95% CI, 3,637 to 5,436) per patient (Table 3). The most expensive strategy was antibiotics plus G-CSF, with a mean cost of 8,998 (95% CI, 6,286 to 8,969). The sequential strategy resulted in mean additional costs of 5,970/patient (95% CI, 4,867 to 7,254). From this, we conclude that the reduction in FN-related cost did not compensate for the extra cost associated with the addition of prophylaxis (G-CSF).
Cost Effectiveness Including effects as well as costs demonstrated that the mean incremental cost compared with antibiotics were 5,824 and 2,156, respectively and the mean incremental effects, defined as FN-free cycle, were 0.02 and –0.11 for the antibiotics plus G-CSF strategy and sequential strategy, respectively (Table 4). The results are graphically presented in Figure 2, and this implies an ICER of, respectively, 343,110 per FN-free cycle for the strategy of antibiotics plus G-CSF. Figure 2 shows the acceptability for a given willingness to pay (WTP) for an FN-free cycle. Figure 2 demonstrates clearly that, in the Dutch context, this will not be reached for any reasonable WTP (the WTP 300,000 for an FN-free period is approximately 0.45).
Sensitivity Analyses Sensitivity analyses demonstrated that, for the clinically important threshold of the probability of FN in cycle 1 of 0.2, prophylaxis with antibiotics remained the most favorable alternative with respect to costs (Table 4). Compared with the baseline model (FN-probability cycle 1, 0.57), a lower FN probability in cycle 1 results in lower incremental costs of antibiotics plus G-CSF. But still, both strategies are dominated by the antibiotics strategy, because the effects between strategies are rather similar (Fig 2).
Haim et al10 reported less frequent FN-recurrences than we found in our primary prophylaxis study, this affected both the incremental effect as the incremental costs (Fig 2), resulting in an ICER for antibiotics plus G-CSF and sequential strategy of
When using the lower outpatient FN-related cost as determined by Bennett et al19 or the higher FN-related cost in the US as determined by Kuderer et al18 as input for the model, the mean costs were
To our knowledge, this is the first study comparing the economic consequences of the use of different secondary prophylactic strategies in preventing FN by means of a Markov model based on a prospectively derived patient-based cost data set in SCLC patients at risk of FN. Secondary prophylaxis is not advocated by international guidelines, but is preferred and commonly used by many caregivers. The cost analysis adopted the health care payer's perspective for the Netherlands and investigated three different strategies (antibiotics only, antibiotics plus G-CSF, and a combined sequential approach). Secondary prophylaxis with the combination of antibiotics plus G-CSF was more costly than was treatment with antibiotics alone. The cost effectiveness of antibiotics plus G-CSF depends on the willingness to exchange money for effectiveness. Figure 2 shows that a reasonable WTP is not reached for the strategy of antibiotics plus G-CSF in the baseline model, and the sequential strategy is dominated by the antibiotics strategy (less effect, additional costs). Combining these results with accompanied sensitivity analyses allow the conclusion that the cost of G-CSF prophylaxis is the crucial factor with regard to the economics of secondary prophylactic strategy. Baseline risk of FN, recurrence risk, or the cost of FN have, given a realistic range of outcomes, no influence on the economics of secondary prophylaxis. We could not demonstrate differences in survival or response to antitumor treatment in the primary prophylaxis trial. Still, good clinical practice may require examining more since strategies may differ in their clinical effects without impact on survival and as a consequence the preferred approach may be to use multiple economic analyses. As discussed elsewhere,8,20 cost per FN-free period may seem artificial, but can be a useful assessment. Policy makers should decide on the acceptable amount of money to spend for supportive care where there is no impact on survival. Today, with expanding treatment options but limited financial resources, rational choices have to be made. Phase III clinical trials investigating new treatment strategies preferably should incorporate economic analyses. However, for different reasons, cost-effectiveness studies are not always feasible alongside clinical trials; in those cases, modeling cost effectiveness might be the answer.21 If a detailed presentation of the model is provided, as in our model, those interested can assess the transferability of the model to their specific situation.22 Our model demonstrates that the cost (and effects) of secondary prophylaxis with antibiotics plus G-CSF move toward the antibiotics strategy if there is a low probability of a first FN (Fig 2). When caregivers have implemented the recently updated guidelines, selecting patients with a risk of 20% or higher for use of primary G-CSF prophylaxis, only patients with a low baseline risk are eligible for secondary prophylaxis. In such a patient population, the absolute economic differences between the different secondary prophylactic strategies are less distinct but, as a result of comparable efficacy, antibiotics are still the less costly alternative. We assumed in our model that the relative recurrence reduction of FN by secondary prophylaxis is similar to that seen in our study on primary prophylaxis. However, a high incidence of FN is consistently reported in the first cycle of chemotherapy. In most studies, approximately half of all patients experiencing FN during their treatment do so in the first cycle.3,23-25 In cycles 2 to 5 (Table 1), both the incidence of a first episode of FN as well as recurrence seemed to be low and rather similar in both the arm with primary antibiotics and the arm with primary antibiotics plus G-CSF prophylaxis. Put otherwise, this suggests that the addition of G-CSF may not influence recurrence rates, at least not to the degree seen in the first cycle with primary prophylaxis. These clinical findings support the use of primary prophylaxis rather than of secondary prophylaxis.
Two other strategies to prevent recurrent FN (no prophylaxis but chemotherapy dose reduction and secondary prophylaxis with G-CSF) were not included in our model because they were not investigated in our primary prophylaxis trial and literature does not provide sufficient data on these strategies to incorporate them in the model. Most likely, secondary prophylaxis with G-CSF alone will result in comparable cost to the strategy of antibiotics plus G-CSF because antibiotics are associated with little extra costs. Using the results of 38 SCLC patients treated in an EORTC trial with standard CDE chemotherapy receiving no prophylaxis and our cost data, a global calculation results in mean cost of approximately In conclusion, this model-based economic analysis demonstrated that, for a realistic range of FN baseline risk and FN-related costs, secondary prophylaxis with antibiotics is least expensive compared with antibiotics plus G-CSF or a sequential approach. This study demonstrated that the relatively high price of administering G-CSF is the determining factor because effects are less distinct. Taking into account that there is no high-level evidence of the clinical efficacy of secondary prophylaxis, at this point, we cannot recommend secondary prophylaxis. Essentially, the clinical efficacy of different secondary prophylactic strategies compared with placebo needs to be prospectively assessed with accompanying economic analyses.
The author(s) indicated no potential conflicts of interest.
Conception and design: Johanna N.H. Timmer-Bonte, Johan L. Severens, Vivianne C.G. Tjan-Heijnen Administrative support: Evelien Termeer Provision of study materials or patients: Johanna N.H. Timmer-Bonte, Eddy M.M. Adang, Vivianne C.G. Tjan-Heijnen Collection and assembly of data: Johanna N.H. Timmer-Bonte, Eddy M.M. Adang, Evelien Termeer, Vivianne C.G. Tjan-Heijnen Data analysis and interpretation: Johanna N.H. Timmer-Bonte, Eddy M.M. Adang, Johan L. Severens, Vivianne C.G. Tjan-Heijnen Manuscript writing: Johanna N.H. Timmer-Bonte, Eddy M.M. Adang, Johan L. Severens, Vivianne C.G. Tjan-Heijnen Final approval of manuscript: Johanna N.H. Timmer-Bonte, Eddy M.M. Adang, Vivianne C.G. Tjan-Heijnen
We thank the responsible investigators from participating hospitals.
Supported by a research grant from the Dutch Healthcare Insurance Board (OG 99 053). Authors' disclosures of potential conflicts of interest and author contributions are found at the end of this article.
1. Crawford J, Ozer H, Stoller R, et al: Reduction by granulocyte colony-stimulating factor of fever and neutropenia induced by chemotherapy in patients with small-cell lung cancer. N Engl J Med 325:164-170, 1991[Abstract] 2. Trillet-Lenoir V, Green J, Manegold C, et al: Recombinant granulocyte colony stimulating factor reduces the infectious complications of cytotoxic chemotherapy. Eur J Cancer 29A:319-324, 1993[CrossRef] 3. Tjan-Heijnen VC, Postmus PE, Ardizzoni A, et al: Reduction of chemotherapy-induced febrile leucopenia by prophylactic use of ciprofloxacin and roxithromycin in small-cell lung cancer patients: An EORTC double-blind placebo-controlled phase III study. Ann Oncol 12:1359-1368, 2001 4. Timmer-Bonte JN, de Boo TM, Smit HJ, et al: Prevention of chemotherapy-induced febrile neutropenia by prophylactic antibiotics plus or minus granulocyte colony-stimulating factor in small-cell lung cancer: A Dutch randomized phase III study. J Clin Oncol 23:7974-7984, 2005 5. Hughes WT, Armstrong D, Bodey GP, et al: 2002 guidelines for the use of antimicrobial agents in neutropenic patients with cancer. Clin Infect Dis 34:730-751, 2002[CrossRef][Medline] 6. Smith TJ, Khatcheressian J, Lyman GH, et al: 2006 Update of recommendations for the use of white blood cell growth factors: An evidence-based clinical practice guideline. J Clin Oncol 24:3187-3205, 2006 7. Adams JR, Lyman GH, Djulbegovic B, et al: G-CSF as prophylaxis of febrile neutropenia in SCLC. Expert Opin Pharmacother 3:1273-1281, 2002[CrossRef][Medline] 8. Timmer-Bonte JN, Adang EM, Smit HJ, et al: Cost-effectiveness of adding granulocyte colony-stimulating factor to primary prophylaxis with antibiotics in small-cell lung cancer. J Clin Oncol 24:2991-2997, 2006 9. Tjan-Heijnen VC, Caleo S, Postmus PE, et al: Economic evaluation of antibiotic prophylaxis in small-cell lung cancer patients receiving chemotherapy: An EORTC double-blind placebo-controlled phase III study (08923). Ann Oncol 14:248-257, 2003 10. Haim N, Shulman K, Goldberg H, et al: The safety of full-dose chemotherapy with secondary prophylactic granulocyte colony stimulating factor (G-CSF) following a prior cycle with febrile neutropenia. Med Oncol 22:229-232, 2005[CrossRef][Medline] 11. Chouaid C, Bassinet L, Fuhrman C, et al: Routine use of granulocyte colony-stimulating factor is not cost-effective and does not increase patient comfort in the treatment of small-cell lung cancer: An analysis using a Markov model. J Clin Oncol 16:2700-2707, 1998[Abstract] 12. Timmer-Bonte JNH, Tjan-Heijnen VC: Prevention of chemotherapy-related episodes of febrile neutropenia (FN) in small-cell-lung-cancer (SCLC) patients: In practice not theory. Eur J Cancer 5:s279-s2003 (suppl) 13. Bennett CL, Smith TJ, Weeks JC, et al: Use of hematopoietic colony-stimulating factors: The American Society of Clinical Oncology survey—The Health Services Research Committee of the American Society of Clinical Oncology. J Clin Oncol 14:2511-2520, 1996[Abstract] 14. Danova M, Rosti G, De Placido S, et al: Use of granulocyte colony-stimulating factor: A survey among Italian medical oncologists. Oncol Rep 14:1405-1412, 2005[Medline] 15. Oostenbrink JB, Koopmanschap MA, Rutten FFH: Manual for cost analysis, methods and guideline prices in health care economic evaluations [in Dutch], 2000 16. Hesterberg T, Moore DS, Monaghan S, et al: Bootstrap Methods and Permutation Tests (ed 2). New York, NY, W.H. Freeman, 2005, http://bcs.whfreeman.com/ips5e/content/cat_080/pdf/moore14.pdf 17. Ardizzoni A, Tjan-Heijnen VC, Postmus PE, et al: Standard versus intensified chemotherapy with granulocyte colony-stimulating factor support in small-cell lung cancer: A prospective European Organization for Research and Treatment of Cancer-Lung Cancer Group Phase III Trial-08923. J Clin Oncol 20:3947-3955, 2002 18. Kuderer NM, Dale DC, Crawford J, et al: Mortality, morbidity, and cost associated with febrile neutropenia in adult cancer patients. Cancer 106:2258-2266, 2006[CrossRef][Medline] 19. Bennett CL, Calhoun EA: Evaluating the total costs of chemotherapy-induced febrile neutropenia: Results from a pilot study with community oncology cancer patients. Oncologist 12:478-483, 2007 20. Ferretti G, Papaldo P: Can the addition of prophylactic filgrastim be considered cost effective in early breast cancer patients treated with epirubicin and cyclophosphamide? J Clin Oncol 24:5615-5616, 2006 21. Bennett CL, Smith TJ, George SL, et al: Free-riding and the prisoner's dilemma: Problems in funding economic analyses of phase III cancer clinical trials. J Clin Oncol 13:2457-2463, 1995 22. Welte R, Feenstra T, Jager H, et al: A decision chart for assessing and improving the transferability of economic evaluation results between countries. Pharmacoeconomics 22:857-876, 2004[CrossRef][Medline] 23. Gómez H, Hidalgo M, Casanova L, et al: Risk factors for treatment-related death in elderly patients with aggressive non-Hodgkin's lymphoma: Results of a multivariate analysis. J Clin Oncol 16:2065-2069, 1998[Abstract] 24. Meza LA, Green MD, Hackett JR, et al: Filgrastim-mediated neutrophil recovery in patients with breast cancer treated with docetaxel and doxorubicin. Pharmacotherapy 23:1424-1431, 2003[CrossRef][Medline] 25. Crawford J, Dale DC, Wolff D, et al: Risk of neutropenic events during the first cycle of systemic cancer chemotherapy: Results from a prospective study. Support Care Cancer 13:425, 2005 (abstr 06-049) Submitted June 19, 2007; accepted October 1, 2007.
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
|
|||||||||||
|
|||||||||||
|
|
Copyright © 2008 by the American Society of Clinical Oncology, Online ISSN: 1527-7755. Print ISSN: 0732-183X
|
TOP
ABSTRACT
INTRODUCTION
