Cost-effectiveness of prophylaxis treatment strategies for febrile neutropenia in patients with recurrent ovarian cancer.

YGYNO-975433; No. of pages: 8; 4C: Gynecologic Oncology xxx (2014) xxx–xxx

Contents lists available at ScienceDirect

Gynecologic Oncology journal homepage: www.elsevier.com/locate/ygyno

Cost-effectiveness of prophylaxis treatment strategies for febrile neutropenia in patients with recurrent ovarian cancer Kelly Fust a,⁎, Xiaoyan Li b, Michael Maschio c, Richard Barron b, Milton C. Weinstein a,d, Anju Parthan a, Marjan Walli-Attaei c,1, David B. Chandler b, Gary H. Lyman e,2 a

OptumInsight, Cambridge, MA, USA Amgen Inc., Thousand Oaks, CA, USA OptumInsight, Burlington, ON, Canada d Harvard School of Public Health, Boston, MA, USA e Duke University, Duke Cancer Institute, Durham, NC, USA b c

H I G H L I G H T S • Primary prophylaxis with pegfilgrastim was cost effective compared to secondary prophylaxis in recurrent ovarian cancer patients receiving docetaxel. • Primary prophylaxis with pegfilgrastim dominated other strategies in recurrent ovarian cancer patients receiving docetaxel. • Primary prophylaxis with pegfilgrastim dominated all comparators in recurrent ovarian cancer patients receiving topotecan.

a r t i c l e

i n f o

Article history: Received 8 January 2014 Accepted 6 March 2014 Available online xxxx Keywords: Febrile neutropenia Prophylaxis Pegfilgrastim Filgrastim Ovarian cancer Cost effectiveness

a b s t r a c t Objective. Evaluate the cost-effectiveness of primary prophylaxis (PP) or secondary prophylaxis (SP) with pegfilgrastim, filgrastim (6-day and 11-day), or no prophylaxis to reduce the risk of febrile neutropenia (FN) in patients with recurrent ovarian cancer receiving docetaxel or topotecan. Methods. A Markov model was used to evaluate the cost-effectiveness of PP vs SP from a US payer perspective. Model inputs, including the efficacy of each strategy (relative risk of FN with prophylaxis compared to no prophylaxis) and mortality, costs, and utility values were estimated from public sources and peer-reviewed publications. Incremental cost-effectiveness was evaluated in terms of net cost per FN event avoided, incremental cost per lifeyear saved (LYS), and incremental cost per quality-adjusted life-year (QALY) gained over a lifetime horizon. Deterministic and probabilistic sensitivity analyses (DSA and PSA) were conducted. Results. For patients receiving docetaxel, the incremental cost-effectiveness ratio (ICER) for PP vs SP with pegfilgrastim was $7900 per QALY gained, and PP with pegfilgrastim dominated all other comparators. For patients receiving topotecan, PP with pegfilgrastim dominated all comparators. Model results were most sensitive to baseline FN risk. PP vs SP with pegfilgrastim was cost effective in 68% and 83% of simulations for docetaxel and in N 99% of simulations for topotecan at willingness-to-pay thresholds of $50,000 and $100,000 per QALY. Conclusions. PP with pegfilgrastim should be considered cost effective compared to other prophylaxis strategies in patients with recurrent ovarian cancer receiving docetaxel or topotecan with a high risk of FN. © 2014 Elsevier Inc. All rights reserved.

Introduction Febrile neutropenia (FN) is a serious side effect of myelosuppressive chemotherapy that often requires hospitalization and treatment with

⁎ Corresponding author at: OptumInsight, One Main Street, Suite 1040, Cambridge, MA 02142, USA. E-mail address: [email protected] (K. Fust). 1 Present address: McMaster University, Hamilton, Ontario, Canada. 2 Present address: Fred Hutchinson Cancer Research Center, University of Washington, Seattle, WA, USA.

intravenous (IV) antibiotics. FN is associated with significant morbidity, mortality, and costs [1], as well as reduced chemotherapy relative dose intensity (RDI), which may adversely affect long-term outcomes such as survival [2–4]. Granulocyte colony-stimulating factors (G-CSFs) such as filgrastim (NEUPOGEN®) and pegfilgrastim (Neulasta®) have been shown to reduce FN risk when used as primary prophylaxis (PP) with the first and every chemotherapy cycle [5,6]. Filgrastim is approved for daily administration up to 14 days per chemotherapy cycle, until the absolute neutrophil count (ANC) has reached 10,000/mm3 [7]. Though 10–11 days have been found to be effective in randomized

http://dx.doi.org/10.1016/j.ygyno.2014.03.014 0090-8258/© 2014 Elsevier Inc. All rights reserved.

Please cite this article as: Fust K, et al, Cost-effectiveness of prophylaxis treatment strategies for febrile neutropenia in patients with recurrent ovarian cancer, Gynecol Oncol (2014), http://dx.doi.org/10.1016/j.ygyno.2014.03.014

2

K. Fust et al. / Gynecologic Oncology xxx (2014) xxx–xxx

A) Decision Tree No prophylaxis Survive 6-day filgrastim On chemotherapy, FN Primary prophylaxis

11-day filgrastim Die from FN

Prophylaxis strategy

Pegfilgrastim 6-day filgrastim On chemotherapy, no FN; survive Secondary prophylaxis

11-day filgrastim Pegfilgrastim

FN events in cycle 1 are captured in this decision tree; chemotherapy cycle length = 3 weeks. The shaded triangles lead to Markov Phase 1.

B) Markov Phase 1 No FN; Survive On chemotherapy without history of FN

Survive FN Die from FN

M

No FN; Survive On chemotherapy with history of FN

Surviving patients re-enter Markov 1 “on chemotherapy with history of FN” or “on chemotherapy without history of FN” until chemotherapy is complete.

Survive FN Die from FN

FN events in cycles 2 & 3 are captured in Markov Phase 1; Markov cycle length = chemotherapy cycle length = 3 weeks. Following completion of chemotherapy, surviving patients move to Markov Phase 2.

C) Markov Phase 2

Survive

Completed chemotherapy (RDI < 85%)

Die from cancer Die from other causes

M

Survive Die from cancer Die from other causes

Post-chemotherapy phase is captured in Markov Phase 2; Markov cycle length = 1 year. A half-cycle correction was used (i.e., all events were assumed to occur in the middle of each Markov cycle).

Fig. 1. Model structure.

clinical trials [8,9], filgrastim is often administered for 4–6 days in clinical practice, albeit with reduced effectiveness [10]. Pegfilgrastim, a pegylated form of filgrastim, is approved for administration once per chemotherapy cycle [11]. The American Society for Clinical Oncology (ASCO) and the National Comprehensive Cancer Network (NCCN) recommend G-CSFs as PP in patients receiving myelosuppressive chemotherapy associated with N20% FN risk and as secondary prophylaxis (SP) following an FN event [12, 13]. An individual patient's FN risk depends on demographics (e.g., age and comorbidities), disease-specific factors (e.g., tumor stage and bone marrow involvement), and treatment-related factors (e.g., chemotherapy type and intensity) [12]. Cost-effectiveness analysis is increasingly being used to compare the costs and health outcomes of different interventions to inform policy decisions. A previous cost-effectiveness analysis of pegfilgrastim in epithelial ovarian carcinoma patients receiving taxane/platinum-based

chemotherapy reported that PP with pegfilgrastim dominated (i.e., resulted in better outcomes at lower costs) SP and no prophylaxis in terms of incremental cost per FN hospitalization [14]. Despite these results, the clinical benefit of pegfilgrastim to reduce infection-related mortality and support chemotherapy dose intensity may have been underestimated, as methods used to derive FN risk and efficacy parameters were not transparent and mortality was not modeled. Further, the incremental cost per quality-adjusted life-year (QALY), a common measure used in healthcare decision-making, was not evaluated. The objective of this study was to evaluate the cost-effectiveness of PP and SP with pegfilgrastim or filgrastim (6-day and 11-day) and no prophylaxis to reduce the risk of FN in recurrent ovarian cancer patients receiving docetaxel or topotecan from a US payer perspective. This study focuses on docetaxel and topotecan because these regimens are recommended by the NCCN for treatment of recurrent ovarian cancer [15] and are associated with an FN risk N 20% [12,16–18].



3

Table 1 Clinical input parameters. Model parameters

Base-case value

DSA range

PSA distribution

References/notes

FN risk in cycle 1 (docetaxel) FN risk in cycle 1 (topotecan) RR of FN for pegfilgrastim vs no G-CSF

22.7% 32.7% 0.08

13.0%, 34.1% 18.4%, 48.8% 0.034, 0.175

[16] [17] [6]a

RR of FN for 11-day filgrastim vs no G-CSF

0.08

N/A

RR of FN for 6-day filgrastim vs no G-CSF

0.50

N/A

Beta (SE = 0.0541) Beta (SE = 0.0782) Lognormal (log mean = −2.5606 SD = 0.4185) Lognormal (log mean = −2.5606 SD = 0.4185) Beta (SE = 0.050)

RR of FN in cycles ≥2 vs cycle 1, no FN history

0.21

0.152, 0.289

RR of FN in cycles ≥2, FN history vs no FN history

9.09

6.07, 13.09

Percentage of FN events requiring hospitalization (docetaxel) Percentage of FN events requiring hospitalization (topotecan) FN inpatient case fatality rate FN hospitalization length of stay (days)

83.6%

79.5%, 87.3%

Lognormal (log mean = −1.5621 SD = 0.1635) Lognormal (log mean = 2.1878 SD = 0.1961) Beta (SE = 0.020) b

Assumed equivalent to pegfilgrastim [8,9] Base-case value was linearly interpolated between RR for 11-day filgrastim (RR = 0.08) and RR for no prophylaxis (RR = 1.0, assumes 0 days of prophylaxis). PSA standard error value was assumed to be 10% of base-case value. [19,20]a [19,20]a [22]

100.0%

84.8%, 100.0%

Beta (alpha = 15.5, beta = 0.5)

[17]

8.4% 9.0

6.4%, 10.3% 8.2, 9.7

Beta (SE = 0.0099) N/A

[1] [1]

DSA: deterministic sensitivity analysis; PSA: probability sensitivity analysis; FN: febrile neutropenia; N/A: not applicable; SE: standard error; RR: relative risk; G-CSF: granulocyte colonystimulating factor; SD: standard deviation; OR: odds ratio; HR: hazard ratio. a Estimate based on breast cancer data. b Because the base-case percentage was 100%, Jeffreys prior interval approach was used to generate a suitable confidence interval (Brown LD, Cai TT, and DasGupta A. (2001) Interval estimation for a binomial proportion. Statistical Science 16:101–117).

Methods Model structure The target population in the model was a hypothetical cohort of 59-year-old recurrent ovarian cancer patients receiving docetaxel (100 mg/m2) via 1-hour IV infusion every 21 days for 3 cycles [16] or topotecan (1.25 mg/m2/day) for 5 consecutive days via 30-minute IV infusion every 21 days for 3 cycles [17]. A Markov cycle tree was developed in Microsoft Excel 2007 (Fig. 1), which included a decision tree tracking the initial chemotherapy cycle and associated FN events (Fig. 1A) and a Markov model with 2 phases for modeling repetitive, predictable events over time (Fig. 1B, C). The first phase tracked FN events in chemotherapy cycles 2 and 3 (cycle length = 3 weeks); the second phase tracked long-term cancer-related survival (cycle length = 1 year). Model inputs Model parameters were estimated from peer-reviewed publications and publicly available sources (Tables 1 & 2). FN risk The baseline FN risk for docetaxel was estimated to be 30% over a median of 3 chemotherapy cycles using data from Rose et al.; cycle 1 FN risk was calibrated to be 0.227 (Appendix A) [16]. The course-level baseline FN risk for topotecan was 41.7%. Using the same approach as for docetaxel, baseline cycle 1 FN risk for topotecan was estimated to be 0.327 [17]. The relative risk (RR) of FN in chemotherapy cycles 2–3 compared to cycle 1 without FN history (RR = 0.21), and the RR of FN in chemotherapy cycles 2–3 with FN history relative to no FN history (RR = 9.09) were obtained from Whyte et al. [20]. These RRs were based on breast cancer data reported by von Minckwitz et al., which is the only known study reporting individual cycle-specific FN events [19]. Efficacy of pegfilgrastim and filgrastim Pegfilgrastim efficacy was derived from a phase 3, double-blind, placebo-controlled study of breast cancer patients [6]. The RR of FN for

pegfilgrastim vs no G-CSF (RR = 0.08) was calculated by dividing FN incidence on pegfilgrastim (1.3%) by FN incidence on placebo (16.8%). In this study, we assumed that pegfilgrastim was given the next day following chemotherapy, as was done in the pivotal clinical trials [6,8,9]. Efficacy of 11-day PP with filgrastim was assumed to be equal to the efficacy of PP with pegfilgrastim (RR = 0.08) [8,9]. Efficacy of filgrastim dosing schedules between zero and 11 days was linearly interpolated between the RR for 11 days of filgrastim (RR = 0.08) and the RR for zero days of filgrastim (RR = 1.0). The risk of FN in chemotherapy cycles 2–3 was calculated as the baseline cycle 1 probability of FN, reduced when applicable by the relative efficacy of prophylaxis, and multiplied by the cycle 2+ risk reduction (0.21). For those with FN history, this value was further multiplied by the FN history risk multiplier (9.09). Mortality risk Median survival from initiation of docetaxel or topotecan treatment among the target population of the model is 6–13 months [16–18]. During the on-chemotherapy phase of the model, risk of death was assumed to be from FN only (and not from cancer or other causes); the case-fatality rate per FN hospitalization was estimated to be 8.4% [1]. Upon chemotherapy completion, patients receiving docetaxel or topotecan were subject to ovarian cancer-specific annual probabilities of death estimated from Kaplan–Meier survival curves provided by Verschraegen et al. [18] and Swisher et al. [17]. These data were fit to an exponential curve, and the estimated constant annual mortality probabilities (52.5% for docetaxel, 57.0% for topotecan) were used in the model. Long-term (i.e., post160 weeks for docetaxel, post-108 weeks for topotecan) survival was calculated by applying the constant annual mortality up to 20 years. See Appendix A for details. The small fraction of patients surviving 20 years after diagnosis were considered “cured” and were subject to the same all-cause mortality rates as the general population [21]. Costs Pegfilgrastim and filgrastim drug acquisition costs reflect average sales prices (ASP) obtained from the Centers for Medicare & Medicaid Services (CMS). The daily filgrastim cost was calculated as a weighted average of ASP by dosage, with 19% and 81% of patients expected to receive 300 and 480 μg, respectively (IntrinsiQ Research Market Data).


4


Table 2 Costs and utilities. Model parameters

Base-case value

DSA range

PSA distribution

References/notes

ASP of 6 mg pegfilgrastim

$2692

$2423, $2961

N/A

ASP of 300 μg filgrastim

$252

N/A

N/A

ASP of 480 μg filgrastim Percentage of filgrastim patients receiving 480 μg filgrastim G-CSF administration costb

$399 81% $24

N/A N/A N/A

CBC costb

$18

Centers for Medicare & Medicaid Services. 2012 ASP Drug Pricing Filea Centers for Medicare & Medicaid Services. 2012 ASP Drug Pricing Filea CMS 2012 ASP Drug Pricing Filea IntrinsiQ Research Market Data (2009–2010) CMS Physician Fee Schedule – PFS Relative Value File Items – July 2012 Releasec CMS 2012 Clinical Laboratory Fee Scheduled

b

Chemotherapy cost per cycle (docetaxel)

$1408

N/A N/A $21.60, $26.40 $16.20, $19.80 $1267, $1549

Chemotherapy cost per cycle (topotecan)b

$1073

$966, $1180

N/A

Hospitalization cost for FNe

$22,221

N/A

Post-hospitalization cost for FN Outpatient cost for FN

$4040 $10,101

N/A N/A

Estimated as 20% of FN hospitalization cost [23] Estimated as 50% of FN hospitalization cost [23]

Utility value for patients with recurrent ovarian cancer on chemotherapy Utility value for patients with recurrent ovarian cancer during FN hospitalization Utility value for patients with ovarian cancer post-chemotherapy

0.55

$19,925, $24,518 $3623, $4458 $9057, $11,144 0.03, 0.99

CMS 2012 ASP Drug Pricing Filea CMS Physician Fee Schedule – PFS Relative Value File Items – July 2012 Releasec CMS 2012 ASP Drug Pricing Filea CMS Physician Fee Schedule – PFS Relative Value File Items – July 2012 Releasec [1,23]

[24]

0.55

0.03, 0.99

0.83

0.13, 1.00

Beta (SE = 0.30) Beta (SE = 0.30) Beta (SE = 0.25)

N/A N/A

[24] [24]

All costs are in 2012 US dollars. DSA: deterministic sensitivity analysis; PSA: probability sensitivity analysis; ASP: average sales price; N/A: not applicable; CMS: Centers for Medicare & Medicaid Services; G-CSF: granulocyte colony-stimulating factor; RVU: relative value unit; CBC: complete blood count; FN: febrile neutropenia; SE: standard error. a Available at: http://www.cms.gov/Medicare/Medicare-Fee-for-Service-Part-B-Drugs/McrPartBDrugAvgSalesPrice/2012ASPFiles.html. b G-CSF administration cost was estimated using CPT code 96372 (subcutaneous or intramuscular injection); CBC cost was estimated using CPT codes 36415 (collection of venous blood by venipuncture) and 85025 (CBC, automated and automated differential count); chemotherapy administration cost was estimated using CPT code 96413 (chemotherapy IV infusion, 1 h). c Available at http://www.cms.gov/Medicare/Medicare-Fee-for-Service-Payment/PhysicianFeeSched/PFS-Relative-Value-Files.html. d Available at http://www.cms.gov/Medicare/Medicare-Fee-for-Service-Payment/ClinicalLabFeeSched/clinlab.html. e FN hospitalization cost was inflated by 10% to account for additional physician fees [23].

We assumed that pegfilgrastim was administered once per chemotherapy cycle and that patients received one complete blood count (CBC) the day before each chemotherapy cycle began. Filgrastim prescribing information recommends one CBC prior to chemotherapy and twice per week during filgrastim therapy; accordingly, 6-day and 11-day filgrastim were associated with 3 and 5 CBCs, respectively [7]. The administration cost of G-CSFs and the CBC cost were estimated using CMS data. The percentage of FN cases requiring hospitalization was estimated to be 83.6% for docetaxel [22] and 100% for topotecan (since only FN events requiring hospitalization were reported) [17]. The base-case FN hospitalization cost reflected the cost of FN-related hospitalization episodes for ovarian cancer patients identified in a hospital database [1]. Consistent with Lyman et al., the FN hospitalization cost was increased by 10% to account for additional physician fees [23]. The post-hospitalization cost, reflecting ambulatory services, was calculated as 20% of the initial hospitalization cost [23]. FN events not requiring hospitalization were assumed to be 50% of the initial FN hospitalization cost [23]. All cost estimates were adjusted to 2012 US dollars using the medical care component of the US Bureau of Labor Statistics Consumer Price Index. Chemotherapy cost estimates included drug acquisition and infusion costs. Costs of long-term treatment of recurrent ovarian cancer were excluded from the model. Utility values To calculate QALYs, life-years were adjusted by utility values obtained from Havrilesky et al. that reflect health-related quality of life during and following chemotherapy, FN, and FN-related hospitalization [24]. The

utility values for chemotherapy were obtained from interviews with members of the general public with no personal history of ovarian cancer; interviews were conducted by trained research nurses using the time trade off (TTO) method. The base-case utility value for chemotherapy was calculated by pooling the mean values for two recurrent ovarian cancer health states: responding to chemotherapy/grade 3–4 toxicity (0.61) and responding to chemotherapy/grade 1–2 toxicity (0.50). The utility value for FN hospitalization was elicited from general public volunteers and ovarian cancer patients [24]. The base-case value was calculated by pooling the mean values for volunteers (0.56) and patients (0.54) and applied for the duration of FN hospitalization (9.0 days [1]) in the model. Post-chemotherapy utilities, reflecting ovarian cancer clinical remission, were obtained through TTO interviews of study volunteers [24] and were applied following chemotherapy completion. Analyses Model outputs included total costs, number of FN events, lifeyears, and QALYs for each strategy (no prophylaxis, PP and SP with pegfilgrastim, PP and SP with 6-day and 11-day filgrastim). Cost-effectiveness was assessed in terms of incremental cost per FN event avoided, incremental cost per life-year saved (LYS), and incremental cost per QALY gained. If a more costly strategy provided no additional benefit compared to an alternative strategy, then it was “dominated” by the alternative strategy. If a more costly strategy provided additional benefit, then the two strategies were compared by dividing the incremental cost by the additional benefit. QALYs and life-years



5

Table 3 Base-case cost-effectiveness results. Prophylaxis strategy

Total costs

Total FN events

Total LYs

Total QALYs

ICER (Δ Cost/Δ FN)

ICER (Δ Cost/Δ LYS)

ICER (Δ Cost/Δ QALYs)

Docetaxel SP with pegfilgrastim PP with pegfilgrastim SP with 11-day filgrastim SP with 6-day filgrastim No prophylaxis PP with 6-day filgrastim PP with 11-day filgrastim

$12,877 $13,048 $13,824 $14,746 $15,801 $16,234 $18,158

0.316 0.027 0.316 0.398 0.495 0.205 0.027

1.381 1.414 1.381 1.372 1.361 1.394 1.414

0.894 0.916 0.894 0.888 0.881 0.902 0.916

Reference $600 Dominateda Dominateda Dominateda Dominateda Dominateda



Topotecan PP with pegfilgrastim SP with pegfilgrastim PP with 11-day filgrastim SP with 11-day filgrastim PP with 6-day filgrastim SP with 6-day filgrastim No prophylaxis

$12,440 $16,726 $17,547 $17,983 $18,788 $20,969 $24,435

0.040 0.449 0.040 0.449 0.321 0.616 0.812

1.275 1.232 1.275 1.232 1.246 1.216 1.197

0.822 0.794 0.822 0.794 0.803 0.783 0.770

Reference Dominateda Dominateda Dominateda Dominateda Dominateda Dominateda



FN: febrile neutropenia; LYs: life-years; QALYs: quality-adjusted life-years; ICER: incremental cost-effectiveness ratio; Δ: incremental; LYS: life-year saved; SP: secondary prophylaxis; PP: primary prophylaxis. a Compared to PP with pegfilgrastim.

were discounted at 3% annually [25]. Since all treatments and FN-related events occurred during the first year of the model, the number of FN events and all treatment and FN-related costs were not discounted. Sensitivity analyses Deterministic and probabilistic sensitivity analyses (DSA and PSA, Tables 1, 2) were performed to assess how changes in key model parameters and parameter uncertainty impacted cost-effectiveness results. In the DSA, parameters were varied using 95% confidence intervals (CIs) and standard errors derived from published literature, with the exception of G-CSF acquisition and administration, CBC, and chemotherapy costs, which were based on 90%–110% of base-case values. In the PSA, uncertainty in cost-effectiveness results was assessed using a second-order Monte Carlo simulation. Uncertainty in key model parameters (e.g., efficacy, risk of FN, and utilities) was characterized by probability distributions around each base-case value, as derived from published literature. PSA results are presented as cost-effectiveness acceptability curves (CEAC), which show the fraction of the 1000 simulations in which examined strategies were incrementally cost effective compared to reference strategies over a range of incremental cost-effectiveness ratio (ICER) thresholds. In addition to the DSA and PSA, three additional analyses were conducted assuming that (1) 50% of FN episodes could be treated in the outpatient setting, (2) RDI had an impact on mortality, and (3) additional disutility was associated with FN-related hospitalization [26,27]. Relative dose intensity In this sensitivity analysis, an RDI threshold of 85% was selected because this is considered a clinically meaningful reduction in delivered chemotherapy dose intensity [2,3]. Probabilities of RDI b85% and ≥85% were calculated for subgroups defined by age and FN history [28,29]. Data from Crawford et al. [28] were used to estimate the odds ratio (OR) of having RDI b 85% among patients ≥65 years vs 18–64 years with no FN history (OR = 1.41). Probabilities of RDI b85% for patients with FN history were calculated by applying an OR of 1.58 [29]. The mortality hazard ratio (HR) for RDI ≤85% (HR = 1.71) was obtained from Hanna et al. [2]. Separate annual probabilities of death were calculated for those with RDI N85% and RDI ≤85%, such that the weighted averages were equal to the overall mortality probabilities [17,18]. Results The cost-effectiveness of PP and SP with pegfilgrastim, PP and SP with 6-day and 11-day filgrastim, and no prophylaxis was evaluated for

docetaxel and topotecan. For patients receiving docetaxel, the ICERs for PP with pegfilgrastim vs SP with pegfilgrastim were $600 per FN event avoided, $5200 per LYS, and $7900 per QALY gained (Table 3). All other comparators were dominated by PP with pegfilgrastim. For patients receiving topotecan, PP with pegfilgrastim dominated all comparators on FN events avoided, LYs saved, and QALYs gained. For both docetaxel and topotecan, less than 1% of the cohort is still alive after 20 years. Sensitivity analyses DSA results are described only for PP vs SP with pegfilgrastim because PP with pegfilgrastim dominated all other strategies in the base case for both docetaxel and topotecan. Results of the DSA for docetaxel indicated that model results were most sensitive to baseline FN risk in cycle 1, pegfilgrastim cost, and RR of FN in cycles 2–3 compared to that in cycle 1 for patients with no FN history (Fig. 2). Varying baseline FN risk in cycle 1 to the lower and upper bounds of the 95% CI yielded ICERs of $278,000 per QALY and “dominant,” respectively. The cycle 1 FN risks corresponding to ICERs of $50,000 and $100,000 per QALY were 20.3% and 18.1%, respectively. Results of the DSA for topotecan indicated that model results were most sensitive to the baseline risk of FN in cycle 1; varying baseline FN risk in cycle 1 to the lower and upper bounds of the 95% CIs yielded ICERs of $64,100 per QALY and “dominant,” respectively. The cycle 1 FN risks corresponding to ICERs of $50,000 and $100,000 per QALY were 18.9% and 17.2%, respectively. For all the remaining parameters, results at one extreme of the plausible ranges included “dominant” in each case. PP with pegfilgrastim was cost effective compared to SP with pegfilgrastim in 68% and 83% of PSA simulations for docetaxel at thresholds of $50,000 and $100,000 per QALY, respectively (Fig. 3A) and in N 99% of the PSA simulations for topotecan at the same thresholds (Fig. 3B). All other comparators had low probabilities of being cost effective and are not displayed. When 50% of FN episodes were assumed to be managed in the outpatient setting, the ICERs for PP with pegfilgrastim vs SP with pegfilgrastim increased to $6000 per FN event avoided, $52,500 per LYS, and $80,200 per QALY gained for docetaxel (Fig. 2); PP with pegfilgrastim dominated all other comparators. For topotecan, PP with pegfilgrastim dominated all comparators in all outcome measures. Only modest effects were seen when we assumed that RDI b85% had an effect on mortality; the ICER for PP with pegfilgrastim vs SP with pegfilgrastim was $5600 per QALY (Fig. 2). Similarly, little effect was seen when we reduced the utility value of FN-related


6

K. Fust et al. / Gynecologic Oncology xxx (2014) xxx–xxx Risk of FN in Cycle 1a High (0.34) - Low (0.13) Pegfilgrastim Priceb Low ($2,423) - High ($2,961) Relative Risk of FN, Cycles 2+, No Previous FNa High (0.29) - Low (0.15) Hospitalization cost for FNa High ($24,518) - Low ($19,925) Percentage of FN Events Requiring Hospitalizationa High (87%) - Low (79%) Post-Chemotherapy Utility Valuea High (1.0) - Low (0.13) Efficacy of Pegfilgrastim (RR of FN versus no G-CSF)a Low (0.03) - High (0.18) Relative risk of FN, Cycles 2+, Previous FNa High (13.09) - Low (6.07) Post-Hospitalization Costs for FNa High ($4,458) - Low ($3,623) Outpatient Cost of FNa High ($11,144) - Low ($9,057)

Percentage of FN events requiring hospitalization (50%)c Hazard ratio for mortality for RDI < 85% (1.71)d Utility value for FN hospitalization (0.42; Brown 1998)e

$0

$50,000 $100,000 $150,000 $200,000 $250,000 $300,000

Cost per QALY Fig. 2. Deterministic sensitivity analysis and alternative analysis results: PP vs SP with pegfilgrastim for patients receiving docetaxel. Only the 10 variables that had the greatest impact on ICERs are depicted. The vertical line represents the base-case ICER, and the horizontal bars represent the ranges of ICERs generated when the model was run using parameter values between the high and low values of the plausible range. Dominant results are represented by arrows, which indicate that primary prophylaxis was more effective at a lower cost than secondary prophylaxis. aRange based on reported 95% confidence intervals or estimated 95% confidence intervals derived from data reported in source. bRange based on plausible range (90%–110% of base-case value). cAlternative analysis assuming that 50% of FN events could be treated in the outpatient setting [27]. dAlternative analysis assuming that RDI had an impact on survival. eAlternative analysis using a less conservative utility value for FN-related hospitalization (0.42).

hospitalization (Fig. 2). In both scenarios, cost per QALY for PP vs SP with pegfilgrastim remained under $10,000 per QALY for docetaxel, and PP with pegfilgrastim remained dominant for topotecan. Discussion Based on assumptions described in the Methods, our results suggest that PP with pegfilgrastim is associated with ICERs of approximately $600 per FN event avoided, $5200 per LYS, and $7900 per QALY gained relative to SP with pegfilgrastim for recurrent ovarian cancer patients receiving docetaxel. PP with pegfilgrastim dominated (i.e., produced better outcomes at lower costs) all filgrastim strategies and no prophylaxis. For topotecan, PP with pegfilgrastim dominated all comparators in all conducted analyses. Based on commonly used US willingness-to-pay thresholds of $50,000 [30] and $100,000 per QALY [31], base-case and PSA results indicated that PP with pegfilgrastim should be considered cost effective relative to other prophylaxis strategies. Model results were most sensitive to baseline FN risk. However, these results were generated from one-way DSAs based on specific assumptions made within the base-case scenario for the chemotherapy regimens examined here. Other chemotherapy regimens for recurrent ovarian cancer patients may have different cycle-level FN risks, different numbers of cycles, and different chemotherapy costs; thus, these data cannot be easily applied to other regimens. Interestingly, the results were not sensitive to FN mortality. When the mortality rate was varied from 6.4% to 10.3%, the ICER for docetaxel remained below $8200 per QALY and the ICER for topotecan remained dominant. Reducing the FN mortality rate resulted in a decrease in both incremental costs and QALYs; accordingly, little change in the ICERs was observed. Several additional sensitivity analyses were conducted. First, carefully selected patients with FN may be safely managed in the outpatient setting [32]. An alternative analysis assuming 50% of FN episodes were managed in the outpatient setting yielded ICERs for PP vs SP with pegfilgrastim of $52,500 per LYS and $80,200 per QALY gained for docetaxel. Depending on the willingness-to-pay threshold, the percentage of

FN events assumed to be managed in an outpatient setting may impact whether PP with pegfilgrastim is considered cost effective compared to SP for docetaxel. Second, maintaining high RDI is associated with improved overall survival in non-Hodgkin's lymphoma (NHL) and early-stage breast cancer [3], but conflicting results have been reported for ovarian cancer [2,4,33]. To account for this uncertainty, we performed an alternative analysis assuming that RDI b85% was associated with lower survival rates; results were similar to the base-case analysis. Finally, an alternative analysis was performed that allowed for additional disutility associated with FN-related hospitalization (0.42) [26]; this adjustment had only a modest impact on the results. Results of this study are similar to those of other US costeffectiveness models for ovarian cancer in terms of incremental cost per FN hospitalization [14] and for NHL, early-stage breast cancer, and other solid tumors [23,34–37]. Though differences in model structures and inputs exist, PP with pegfilgrastim was found to be cost effective compared to other prophylaxis strategies for high-risk patients in these studies. This study contributes to previously published literature [14] by evaluating cost-effectiveness within specific regimens and across a comprehensive set of prophylaxis strategies. Furthermore, mortality, RDI, and health-related quality of life measures were incorporated, and results were reported in terms of cost per QALY, the most commonly used measure in cost-effectiveness analyses and decisionmaking based on those analyses. Though our study and previous studies found that PP with pegfilgrastim was cost effective compared to other prophylaxis strategies, SP with pegfilgrastim may be appropriate in some situations not reflected in the population-wide averages used in our analysis. Ultimately, the decision regarding which prophylaxis strategy to use resides with the physician and should be tailored to the individual patient and their risk factors. Results of this study should be interpreted in light of several limitations. The target population was limited to recurrent ovarian cancer patients receiving docetaxel or topotecan. Docetaxel and topotecan were included in this analysis because both regimens have a high risk of FN



7

A 100% 90%

Probability

80% 70% 60% 50% 40% 30% 20% 10% 0% $0

$100,000

$200,000

$300,000

$400,000

$500,000

$600,000

$500,000

$600,000

Willingness-to-pay Threshold ($/QALY)

B 100% 90%

Probability

80% 70% 60% 50% 40% 30% 20% 10% 0% $0

$100,000

$200,000

$300,000

$400,000

Willingness-to-pay Threshold ($/QALY) Primary pegfilgrastim

Secondary pegfilgrastim

Fig. 3. Cost-effectiveness acceptability curves for docetaxel (A) and topotecan (B).

and are recommended by NCCN for the treatment of platinum-resistant recurrent ovarian cancer; however, topotecan and docetaxel are not as commonly used as platinum-based regimens for the treatment of recurrent ovarian cancer, which may limit the applicability of this study. The percentage of topotecan patients experiencing FN may be underestimated, as Swisher et al. only reported FN events treated in the inpatient setting [17]. Several parameters were obtained from breast cancer data (e.g., pegfilgrastim efficacy, RRs of FN in subsequent cycles, percentage of FN events requiring hospitalization) or data from other tumor types (e.g., costs of post-hospitalization FN events and FN events not requiring hospitalization). New studies of patients with ovarian cancer are needed to determine whether these factors differ among tumor types. Survival statistics were extrapolated from 160 weeks [18] and 108 weeks [17] to 20 years to capture long-term ovarian cancer survival. Finally, indirect costs associated with FN and G-CSF administrations (e.g., lost productivity, care giver costs, travel costs, and healthcare providers' human resource costs) [38–40] were not considered, as the analysis was performed from a US payer perspective. The clinical benefit of PP with pegfilgrastim, in comparison to daily G-CSF and no G-CSF, can translate into cost savings for ovarian cancer patients on high-risk regimens such as docetaxel and topotecan. Assuming commonly accepted willingness-to-pay thresholds of $50,000 and $100,000 per QALY, PP with pegfilgrastim was found to be cost effective compared to SP with pegfilgrastim.

Conflict of interest statement XL, RB, and DBC are employees of and stockholders in Amgen Inc. KF, MM, AP, MW-A, and MCW were paid consultants to Amgen at the time the study was conducted. GHL is a principle investigator on a research grant from Amgen Inc. to Duke University.

Acknowledgments The authors thank April Teitelbaum (medical oncologist and consultant with OptumInsight at the time the study was conducted) and Holly Watson (Amgen Inc.) for thoughtful insight. Medical writing support was provided by Kerri Hebard–Massey (Amgen Inc.). This study was funded by Amgen Inc. Appendix A. Supplementary data Supplementary data to this article can be found online at http://dx. doi.org/10.1016/j.ygyno.2014.03.014. References [1] Dulisse B, Li X, Gayle JA, Barron RL, Ernst FR, Rothman KJ, et al. A retrospective study of the clinical and economic burden during hospitalizations among cancer patients with febrile neutropenia. J Med Econ 2013. http://dx.doi.org/10.3111/13696998.2013. 782034. [2] Hanna RK, Poniewierski MS, Laskey RA, Lopez MA, Shafer A, Van Le L, et al. Predictors of reduced relative dose intensity and its relationship to mortality in women receiving multi-agent chemotherapy for epithelial ovarian cancer. Gynecol Oncol 2013;129(1):74–80. http://dx.doi.org/10.1016/j.ygyno.2012.12.017. [3] Wildiers H, Reiser M. Relative dose intensity of chemotherapy and its impact on outcomes in patients with early breast cancer or aggressive lymphoma. Crit Rev Oncol Hematol 2011;77(3):221–40. http://dx.doi.org/10.1016/j.critrevonc.2010.02.002. [4] Fauci JM, Whitworth JM, Schneider KE, Subramaniam A, Zhang B, Frederick PJ, et al. Prognostic significance of the relative dose intensity of chemotherapy in primary treatment of epithelial ovarian cancer. Gynecol Oncol 2011;122(3):532–5. http:// dx.doi.org/10.1016/j.ygyno.2011.05.023. [5] Crawford J, Ozer H, Stoller R, Johnson D, Lyman G, Tabbara I, 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 1991;325(3):164–70. http:// dx.doi.org/10.1056/NEJM199107183250305.


8


[6] Vogel CL, Wojtukiewicz MZ, Carroll RR, Tjulandin SA, Barajas-Figueroa LJ, Wiens BL, et al. First and subsequent cycle use of pegfilgrastim prevents febrile neutropenia in patients with breast cancer: a multicenter, double-blind, placebo-controlled phase III study. J Clin Oncol 2005;23(6):1178–84. http://dx.doi.org/10.1200/JCO.2005. 09.102. [7] NEUPOGEN® (Filgrastim) prescribing information. Amgen; 2013. [8] Green MD, Koelbl H, Baselga J, Galid A, Guillem V, Gascon P, et al. A randomized double-blind multicenter phase III study of fixed-dose single-administration pegfilgrastim versus daily filgrastim in patients receiving myelosuppressive chemotherapy. Ann Oncol 2003;14(1):29–35. [9] Holmes FA, O'Shaughnessy JA, Vukelja S, Jones SE, Shogan J, Savin M, et al. Blinded, randomized, multicenter study to evaluate single administration pegfilgrastim once per cycle versus daily filgrastim as an adjunct to chemotherapy in patients with high-risk stage II or stage III/IV breast cancer. J Clin Oncol 2002;20(3):727–31. [10] Weycker D, Hackett J, Edelsberg JS, Oster G, Glass AG. Are shorter courses of filgrastim prophylaxis associated with increased risk of hospitalization? Ann Pharmacother 2006;40(3):402–7. http://dx.doi.org/10.1345/aph.1G516. [11] Neulasta® (pegfilgrastim) prescribing information. Amgen; 2010. [12] NCCN Guidelines. Myeloid, growth factors, v1; 2013. [13] Smith TJ, Khatcheressian J, Lyman GH, Ozer H, Armitage JO, Balducci L, et al. 2006 update of recommendations for the use of white blood cell growth factors: an evidence-based clinical practice guideline. J Clin Oncol 2006;24(19):3187–205. http://dx.doi.org/10.1200/JCO.2006.06.4451. [14] Numnum TM, Kimball KJ, Rocconi RP, Kilgore LC, Straughn Jr JM. Pegfilgrastim for the prevention of febrile neutropenia in patients with epithelial ovarian carcinoma —a cost-effectiveness analysis. Int J Gynecol Cancer 2007;17(5):1019–24. http:// dx.doi.org/10.1111/j.1525-1438.2007.00915.x. [15] Guidelines NCCN. Ovarian Cancer, v2; 2013. [16] Rose PG, Blessing JA, Ball HG, Hoffman J, Warshal D, DeGeest K, et al. A phase II study of docetaxel in paclitaxel-resistant ovarian and peritoneal carcinoma: a Gynecologic Oncology Group study. Gynecol Oncol 2003;88(2):130–5. [17] Swisher EM, Mutch DG, Rader JS, Elbendary A, Herzog TJ. Topotecan in platinumand paclitaxel-resistant ovarian cancer. Gynecol Oncol 1997;66(3):480–6. http:// dx.doi.org/10.1006/gyno.1997.4787. [18] Verschraegen CF, Sittisomwong T, Kudelka AP, Guedes E, Steger M, Nelson-Taylor T, et al. Docetaxel for patients with paclitaxel-resistant Mullerian carcinoma. J Clin Oncol 2000;18(14):2733–9. [19] von Minckwitz G, Kummel S, du Bois A, Eiermann W, Eidtmann H, Gerber B, et al. Pegfilgrastim +/− ciprofloxacin for primary prophylaxis with TAC (docetaxel/ doxorubicin/cyclophosphamide) chemotherapy for breast cancer. Results from the GEPARTRIO study. Ann Oncol 2008;19(2):292–8. http://dx.doi.org/10.1093/ annonc/mdm438. [20] Whyte S, Cooper KL, Stevenson MD, Madan J, Akehurst R. Cost-effectiveness of granulocyte colony-stimulating factor prophylaxis for febrile neutropenia in breast cancer in the United Kingdom. Value Health 2011;14(4):465–74. http://dx.doi.org/ 10.1016/j.jval.2010.10.037. [21] Arias E. United States life tables, 2007. Natl Vital Stat Rep 2011;59:1–60. [22] Weycker D, Edelsberg J, Kartashov A, Barron R, Lyman G. Risk and healthcare costs of chemotherapy-induced neutropenic complications in women with metastatic breast cancer. Chemotherapy 2012;58(1):8–18. http://dx.doi.org/10.1159/000335604. [23] Lyman GH, Lalla A, Barron RL, Dubois RW. Cost-effectiveness of pegfilgrastim versus filgrastim primary prophylaxis in women with early-stage breast cancer receiving chemotherapy in the United States. Clin Ther 2009;31(5):1092–104. http:// dx.doi.org/10.1016/j.clinthera.2009.05.003. [24] Havrilesky LJ, Broadwater G, Davis DM, Nolte KC, Barnett JC, Myers ER, et al. Determination of quality of life-related utilities for health states relevant to ovarian cancer

[25] [26] [27]

[28]

[29]

[30]

[31]

[32]

[33]

[34]

[35]

[36]

[37]

[38]

[39]

[40]

diagnosis and treatment. Gynecol Oncol 2009;113(2):216–20. http://dx.doi.org/ 10.1016/j.ygyno.2008.12.026. Gold MR, Siegel JE, Russell LB, Weinstein MC. Cost-effectiveness in health and medicine. New York: Oxford University Press; 1996. Brown RE, Hutton J. Cost-utility model comparing docetaxel and paclitaxel in advanced breast cancer patients. Anticancer Drugs 1998;9(10):899–907. Lathia N, Isogai PK, De Angelis C, Smith TJ, Cheung M, Mittmann N, et al. Costeffectiveness of filgrastim and pegfilgrastim as primary prophylaxis against febrile neutropenia in lymphoma patients. J Natl Cancer Inst 2013;105(15):1078–85. http://dx.doi.org/10.1093/jnci/djt182. Crawford J, Dale DC, Kuderer NM, Culakova E, Poniewierski MS, Wolff D, et al. Risk and timing of neutropenic events in adult cancer patients receiving chemotherapy: the results of a prospective nationwide study of oncology practice. J Natl Compr Canc Netw 2008;6(2):109–18. Shayne M, Crawford J, Dale DC, Culakova E, Lyman GH. Predictors of reduced dose intensity in patients with early-stage breast cancer receiving adjuvant chemotherapy. Breast Cancer Res Treat 2006;100(3):255–62. http://dx.doi.org/10.1007/s10549006-9254-4. Grosse SD. Assessing cost-effectiveness in healthcare: history of the $50,000 per QALY threshold. Expert Rev Pharmacoecon Outcomes Res 2008;8(2):165–78. http://dx.doi.org/10.1586/14737167.8.2.165. Cutler DM, Rosen AB, Vijan S. The value of medical spending in the United States, 1960–2000. N Engl J Med 2006;355(9):920–7. http://dx.doi.org/10.1056/ NEJMsa054744. Flowers CR, Seidenfeld J, Bow EJ, Karten C, Gleason C, Hawley DK, et al. Antimicrobial prophylaxis and outpatient management of fever and neutropenia in adults treated for malignancy: American Society of Clinical Oncology clinical practice guideline. J Clin Oncol 2013;31(6):794–810. http://dx.doi.org/10.1200/JCO.2012.45.8661. Repetto L, Pace M, Mammoliti S, Bruzzone M, Chiara S, Oliva C, et al. The impact of received dose intensity on the outcome of advanced ovarian cancer. Eur J Cancer 1993;29A(2):181–4. Hill G, Barron R, Fust K, Skornicki ME, Taylor DC, Weinstein MC, et al. Primary versus secondary prophylaxis with pegfilgrastim for the reduction of febrile neutropenia risk in patients receiving chemotherapy for non-Hodgkin's lymphoma: cost-effectiveness analyses. J Med Econ 2013. http://dx.doi.org/10.3111/13696998.2013.844160. Eldar-Lissai A, Cosler LE, Culakova E, Lyman GH. Economic analysis of prophylactic pegfilgrastim in adult cancer patients receiving chemotherapy. Value Health 2008;11(2):172–9. http://dx.doi.org/10.1111/j.1524-4733.2007.00242.x. Lyman G, Lalla A, Barron R, Dubois RW. Cost-effectiveness of pegfilgrastim versus 6-day filgrastim primary prophylaxis in patients with non-Hodgkin's lymphoma receiving CHOP-21 in United States. Curr Med Res Opin 2009;25(2):401–11. http://dx.doi.org/ 10.1185/03007990802636817. Ramsey SD, Liu Z, Boer R, Sullivan SD, Malin J, Doan QV, et al. Cost-effectiveness of primary versus secondary prophylaxis with pegfilgrastim in women with earlystage breast cancer receiving chemotherapy. Value Health 2009;12(2):217–25. http://dx.doi.org/10.1111/j.1524-4733.2008.00434.x. Calhoun EA, Chang CH, Welshman EE, Fishman DA, Lurain JR, Bennett CL. Evaluating the total costs of chemotherapy-induced toxicity: results from a pilot study with ovarian cancer patients. Oncologist 2001;6(5):441–5. Fortner BV, Okon TA, Zhu L, Tauer K, Moore K, Templeton D, et al. Costs of human resources in delivering cancer chemotherapy and managing chemotherapyinduced neutropenia in community practice. Community Oncol 2004;1:23–8. Fortner BV, Tauer K, Zhu L, Okon TA, Moore K, Templeton D, et al. Medical visits for chemotherapy and chemotherapy-induced neutropenia: a survey of the impact on patient time and activities. BMC Cancer 2004;4:22. http://dx.doi.org/10.1186/ 1471-2407-4-22.


Comparison of anti-anaerobic antimicrobial strategies in cancer patients with febrile neutropenia and gastrointestinal symptoms.

Secondary infections in cancer patients with febrile neutropenia.

Clinical practice in secondary prophylaxis and management of febrile neutropenia in Poland: results of the febrile neutropenia awareness project.

Clinical practice in febrile neutropenia risk assessment and granulocyte colony-stimulating factor primary prophylaxis of febrile neutropenia in Poland.

Treatment of febrile neutropenia and prophylaxis in hematologic malignancies: a critical review and update.

Granulocyte colony-stimulating factors as prophylaxis against febrile neutropenia.

Febrile neutropenia in chemotherapy treated small-cell lung cancer patients.

Risk of chemotherapy-induced febrile neutropenia in cancer patients receiving pegfilgrastim prophylaxis: does timing of administration matter?

History of chronic comorbidity and risk of chemotherapy-induced febrile neutropenia in cancer patients not receiving G-CSF prophylaxis.

Growth factor therapy may have a clinical benefit for patients with cancer with established febrile neutropenia.

Risk factors for multidrug-resistant bacteremia in hospitalized cancer patients with febrile neutropenia: a cohort study.

Risk of chemotherapy-induced febrile neutropenia with early discontinuation of pegfilgrastim prophylaxis in US clinical practice.

The Effectiveness of Febrile Neutropenia Prophylaxis with Lipegfilgrastim in Routine Clinical Practice.

Risk factors for febrile neutropenia among patients with cancer receiving chemotherapy: A systematic review.

Response of first-line antibiotic therapy in patients with febrile neutropenia during treatment of hematological disorders.

Comparing the efficacy of ceftazidime and meropenem in treatment of febrile neutropenia in pediatric patients with cancer.

Results of a prospective dose intensity and neutropenia prophylaxis evaluation programme (DIEPP) in cancer patients at risk of febrile neutropenia due to myelosuppressive chemotherapy.

Management of the febrile patient with cancer and neutropenia.

Factors associated with hospital length of stay among cancer patients with febrile neutropenia.

tazobactam in cancer patients with hematological malignancies and febrile neutropenia after chemotherapy.

Febrile neutropenia in children treated for malignancy.

Outpatient management of febrile neutropenia associated with cancer chemotherapy: risk stratification and treatment review.

Evaluation of follow-up strategies for patients with epithelial ovarian cancer following completion of primary treatment.

Coronaviruses in children with febrile neutropenia.