A Markov model of the cost-effectiveness of human-derived follicle-stimulating hormone (FSH) versus recombinant FSH using comparative clinical trial data

Article (PDF Available)inFertility and Sterility 83(3):804-7 · April 2005with47 Reads
DOI: 10.1016/j.fertnstert.2004.08.024 · Source: PubMed
This study compared the cost and effectiveness of highly purified, human-derived follicle-stimulating hormone (FSH) (Bravelle) to recombinant FSH (Follistim) using Markov modeling and Monte Carlo simulation. One IVF treatment cycle resulted in costs of 11,584 dollars +/- 211 dollars for human-derived FSH and 12,762 dollars +/- 170 dollars for recombinant FSH, while three treatment cycles, holding the transition probabilities of the first cycle constant for the next two cycles, resulted in costs of 22,712 dollars +/- 1,107 dollars for human-derived FSH and 24,935 dollars +/- 1,205 dollars for recombinant FSH.
A Markov model of the cost-effectiveness of human-
derived follicle-stimulating hormone (FSH) versus
recombinant FSH using comparative clinical trial data
This study compared the cost and effectiveness of highly purified, human-derived follicle-stimulating
hormone (FSH) (Bravelle
) to recombinant FSH (Follistim
) using Markov modeling and Monte Carlo
simulation. One IVF treatment cycle resulted in costs of $11,584 $211 for human-derived FSH and $12,762
$170 for recombinant FSH, while three treatment cycles, holding the transition probabilities of the first
cycle constant for the next two cycles, resulted in costs of $22,712 $1,107 for human-derived FSH and
$24,935 $1,205 for recombinant FSH. (Fertil Steril 2005;83:804 –7. ©2005 by American Society for
Reproductive Medicine.)
Recently, there have been several published articles com-
paring the cost-effectiveness of human-derived
menotropins (hMG) and follicle-stimulating hormone
(hFSH) with that of recombinant FSH (rFSH) for IVF and
intrauterine insemination (1–3).
Using Markov modeling of clinical trials data, Silver-
berg et al. (3) calculated that rFSH was more cost-effective
than hFSH. However, the brand of hFSH (Fertinex, Serono
Inc., Rockland, MA) chosen for inclusion in the analysis is
no longer in common clinical use. More recently, a new,
highly purified form of hFSH (Bravelle, Ferring Pharma-
ceuticals, Inc., Suffern, NY) has been approved for use in
ovulation induction and IVF. Using current, state-of-the-art
purification methods, this highly purified FSH has been
shown to have consistent biological activity and to be
comparable to rFSH in both efficacy and safety. Therefore,
we performed a similar cost-effectiveness analysis compar-
ing medications that are commonly used in IVF programs
in the United States, hFSH (Bravelle) and rFSH (Follistim,
Organon, Inc., West Orange, NJ).
Ideally, a prospective, clinical, assisted reproductive
technologies trial specifically designed and powered to
measure cost-effectiveness would be performed. For sev-
eral reasons, this is not a practical option, hence economic
modeling of existing clinical data, supplemented with ad-
ditional data from the literature or from expert opinions,
has become the established method for examining the rel-
ative cost-effectiveness of the gonadotropins used to stim-
ulate follicular growth in IVF cycles (3, 4).
The present study compares the relative cost-effective-
ness of hFSH (Bravelle
) to rFSH (Follistim
) using data
from identically designed, recently completed, prospective
randomized multicenter comparative clinical trials compar-
ing hFSH with rFSH in IVF (5). A Markov model was
constructed based on these trials. This approach eliminates
bias and assumptions that are inherent in relying on the
published literature and “expert opinions.” The model was
run using one and three cycles of IVF using the transition
probabilities and costs associated with each of the mea-
sured health states in the trial.
The specific Markov model used here (Fig. 1) was cre-
ated (using Data Pro by Tree Age Software Inc., Williams-
town, MA) to examine the costs and effectiveness of hFSH
vs. rFSH. The transition probabilities used in the present
analysis were based on pooled outcomes data that were
collected from two recently completed, comparative clini-
cal trials (5). Drug costs for hFSH and rFSH were based on
actual doses consumed during the trial and used the most
recent wholesale acquisition costs for the two agents; all
other costs corresponding to the relevant health states were
taken from a previous model (3). The data from that anal-
ysis were obtained in 2000 and therefore had to be inflated
in the present study to 2003 dollars using the medical
services component of the consumer price index.
Effectiveness in the Markov model was measured as the
proportion of subjects that entered the health state of con-
tinuing pregnancy. After two positive
-hCG results, trans-
vaginal ultrasound was performed 2 weeks later to confirm
a clinical pregnancy and repeated 1 week later to document
a continuing pregnancy. In addition to the probabilistic
sensitivity analysis, we used adjusted probabilities from
national statistics that were also varied in the model to
approximate the resulting effectiveness and costs of using
either agent in a more nationally representative IVF patient
population. In this analysis, the transition probabilities
were adjusted to reflect the overall success rates seen in
data obtained from 54 Society for Assisted Reproductive
Technology (SART) participating centers (6).
Received December 19, 2003; revised and accepted August 2, 2004.
Supported by Ferring Pharmaceuticals, Inc., Suffern, New York.
Reprint requests: Dennis C. Marshall, R.N., M.S., Ph.D., Executive Director,
Medical Affairs, Ferring Pharmaceuticals, 400 Rella Boulevard, Suite
300, Suffern, New York 10901 (FAX: 845-770-2661; E-mail:
Fertility and Sterility Vol. 83, No. 3, March 2005 0015-0282/05/$30.00
Copyright ©2005 American Society for Reproductive Medicine, Published by Elsevier Inc. doi:10.1016/j.fertnstert.2004.08.024
The clinical data were taken from two identically de-
signed, recently completed, prospective, randomized, mul-
ticenter comparative clinical trials (5) that compared sin-
gle-cycle treatment with subcutaneously-administered,
purified hFSH (n 120) vs. rFSH (n 118) in infertile
female subjects undergoing IVF. Standard site-specific pro-
cedures were used for oocyte retrieval and IVF-ET. Intra-
cytoplasmic sperm injection (ICSI) and assisted hatching
were prohibited. A maximum of four embryos were al-
lowed to be transferred to each subject. Progesterone was
given for luteal phase support. The primary efficacy pa-
rameter for each single-cycle analysis was the number of
oocytes retrieved.
The mean age for the hFSH group was 32.0 3.9 years, vs.
32.5 3.7 years for the rFSH group. The number of subjects
undergoing their first IVF procedure was 87 (72.5%) in the
hFSH group and 91 (77.1%) in the rFSH group. Subjects
having one prior IVF attempt numbered 22 (18.3%) in the
hFSH group and 18 (15.3%) in the rFSH group.
Running the health-economics model with one treatment
cycle resulted in a probability (mean SD) of continuing
pregnancy of 0.40 0.05 and 0.37 0.05 for hFSH and
rFSH, respectively (Table 1). Corresponding costs were
$11,584 $211 and $12,762 $170 for the two agents,
respectively. The transition probabilities corresponding to
the health states are shown in Table 2. The costs calculated
at each health state for each of the two treatments are
presented in Table 3. Running the Markov model through
three treatment cycles resulted in ongoing success rates for
a clinical pregnancy of 0.78 0.05 and 0.75 0.05 for
hFSH and rFSH, respectively. The corresponding costs
were $22,712 $1,107 and $24,935 $1,205 for the two
agents, respectively (Table 1).
In terms of using the results to make an informed deci-
sion regarding treatment, it is useful to note that in looking
at the 10,000 cohorts of hFSH and rFSH side by side, as
recommended for probabilistic sensitivity analysis by
Briggs et al. (7), hFSH was less costly than rFSH in 100%
of the cohorts. In addition, hFSH was found to be more
effective than rFSH in 65% of the cases. Further, the results
indicate that selecting hFSH can (particularly in a patient
population similar to that in the clinical trials) result in
lower costs and better outcomes 65% of the time and lower
costs 100% of the time.
Since payers and decision makers are looking for evi-
dence to aid in their decision-making process, cost-effec-
tiveness data have become an important adjunct to data
establishing clinical efficacy. As such, it is not surprising to
see the emergence of papers dealing with the subject matter
(8). It is important to note that we held the initial proba-
bilities of success constant in our three-cycle model, based
on a previously published report (9). Since cost-effective-
ness is dependent on the rate of success and since the rate
of success obtained from the clinical trials data was much
higher than what was previously used (3), the cost-effec-
tiveness ratios after three cycles in the present study were
almost twice as high as what was previously reported.
To reflect a more realistic patient population, we also ran a
three-cycle model for the two agents with transition probabil-
ities adjusted downward to account for a potentially higher
rate of failure from one cycle to the next (6). The results from
this portion of the analyses indicated a modest reduction in
probabilities of success and cost-effectiveness, but the incre-
mental advantage with hFSH in terms of better outcomes and
lower costs remained the same.
In terms of generalizing from the results presented here,
one should note that the Markov model used in this study
is U.S. based. This issue obviously pertains more to the
costs of the transition states than to the transition proba-
bilities. It is interesting that the cost for an IVF cycle that
we reported comes very close to that reported in a recent
international survey conducted on the health economics of
IVF and ICSI (10). In 2002, the average cost per IVF/ICSI
Outline of Markov model illustrating transition
between health states for each treatment cycle.
OHSS ovarian hyperstimulation syndrome.
Hatoum. Cost-effectiveness of hFSH vs. rFSH. Fertil Steril 2005.
805Fertility and Sterility
cycle was $9,547 in the United States as compared with
$3,518 in 25 other countries; moreover, Collins (10) reported
average cost-effectiveness ratios per live birth at $58,394 in
the United States and at $ 22,048 in other countries.
In conclusion, economic modeling is increasingly used
in IVF in lieu of performing a head-to-head comparison in
the context of one or more clinical trials. The use of
analytic models is not limited to studies regarding the
relative cost-effectiveness of treatments for infertility and
is increasingly deployed as a means to supplement a variety
of clinical guidelines (11). The results of the Markov model
reported in this study capitalized on existing, newly avail-
able, randomized-treatment, comparative efficacy data for
hFSH and rFSH. The analysis was based on data accrued
by pooling results from two identically designed trials and
as such did not rely on estimates from experts as two recent
publications have done (3, 4).
The current study findings indicate that hFSH is a cost-
effective choice in relation to rFSH. Specifically, hFSH
resulted in better efficacy in 65% of the simulated cohorts
Results of Monte Carlo simulation of 10,000 cohorts and 1,000 subjects per cohort.
Cost ($)
hFSH, one cycle
Mean 11,584.19 0.92 0.48 0.42 0.40
SD 211.15 0.03 0.05 0.05 0.05
rFSH, one cycle
Mean 12,761.79 0.96 0.42 0.38 0.37
SD 170.41 0.02 0.05 0.05 0.05
hFSH, three cycles
Mean 22,712.05 1.81 0.94 0.83 0.78
SD 1,106.68 0.10 0.06 0.05 0.05
rFSH, three cycles
Mean 24,934.79 1.93 0.85 0.77 0.75
SD 1,205.29 0.10 0.06 0.05 0.05
Hatoum. Cost-effectiveness of hFSH vs. rFSH. Fertil Steril 2005.
Transition probabilities based on the comparative clinical trials.
No. of patients
in hFSH arm
patients in
No. of patients
in rFSH arm
patients in
Canceled ovum pick-up 7 113 3 115
No oocytes recovered 0 113 0 115
No fertilization 2 111 2 113
No embryo transfer 0 111 0 113
No chemical pregnancy 53 58 63 50
No clinical pregnancy 7 51 5 45
No continuing pregnancy 3 48 1 44
No live birth 6 42 6 38
No multiple birth 28 14 23 15
No premature birth 34 8 29 9
No ovarian hyperstimulation syndrome 114 6 115 3
Note: Transition probabilities were calculated directly from data obtained from the comparative clinical trials. Note that
these data were used as is and were not supplemented with expert opinions, or data from the published literature.
Hatoum. Cost-effectiveness of hFSH vs. rFSH. Fertil Steril 2005.
806 Hatoum et al. Correspondence Vol. 83, No. 3, March 2005
used in the analysis and was less costly in 100% of the
Hind T. Hatoum, Ph.D.
William R. Keye, Jr., M.D.
Richard P. Marrs, M.D.
Surrey M. Walton, Ph.D.
Dennis C. Marshall, R.N., M.S., Ph.D.
Hind T. Hatoum and Company, Chicago, Illinois;
William Beaumont Hospital, Royal Oak, Michigan;
California Fertility Partners, Los Angeles, California;
University of Illinois at Chicago, Chicago, Illinois;
Ferring Pharmaceuticals Inc., Suffern, New
1. Attia AM, Soliman EM, El-Sadek MM, Hamza HS, Salem ME,
Younis AS. Effectiveness and cost-effectiveness of urinary FSH
versus recombinant FSH in clomiphene citrate–resistant anovulatory
infertile patients: opposing the policy of abandoning the use of
urinary FSH. Middle East Fertil Soc J 2003;8:172–9.
2. Filicori M, Cognigni GE, Pocognoli P, Tabarelli C, Ferlini F, Perri T,
et al. Comparison of controlled ovarian stimulation with human
menopausal gonadotropins or recombinant follicle-stimulating hor-
mone. Fertil Steril 2003;80:390–7.
3. Silverberg K, Daya S, Auray JP, Duru G, Ledger W, Wikland M, et
al. Analysis of the cost effectiveness of recombinant versus urinary
follicle-stimulating hormone in vitro fertilization/intracytoplasmic
sperm injection programs in the United States. Fertil Steril 2002;77:
4. Sykes D, Out HJ, Palmer SJ, van Loon J. The cost-effectiveness of
IVF in the UK: a comparison of three gonadotropin treatments. Hum
Reprod 2001;16:2557– 62.
5. Dickey RP, Nichols JE, Steinkampf MP, Gocial B, Thornton M,
Webster BW. Highly purified human-derived follicle stimulating hor-
mone (Bravelle
) has equivalent efficacy to rFSH (Follistim
infertile women undergoing in vitro fertilization. Reprod Biol Endo-
crinol 2003;1:63.
6. Meldrum DR, Silverberg KM, Bustillo M, Stokes L. Success rate with
repeated cycles of in vitro fertilization–embryo transfer. Fertil Steril
7. Briggs AH, Goeree R, Blackhouse G, O’Brien BJ. Probabilistic
analysis of cost-effectiveness models: choosing between treatment
strategies for gastro-esophageal reflux disease. Med Decis Making
2002;22:290 –308.
8. Sonnenberg FA, Beck JR. Markov models in medical decision mak-
ing: a practical guide. Med Decis Making 1993;13:322–38.
9. Doubilet P, Begg CB, Weinstein MC, Braun P, McNeil BJ. Probabi-
listic sensitivity analysis using Monte Carlo simulation. A practical
approach. Med Decis Making 1985;5:157–77.
10. Collins J. An international survey of the health economics of IVF and
ICSI. Hum Reprod Update 2002;8:265–77.
11. Mol BW, Bonsel GJ, Collins JA, Wiegerinck MA, van der Veen F,
Bossuyt PM. Cost-effectiveness of in vitro fertilization and embryo
transfer. Fertil Steril 2000;73:748–54.
Procedures and related costs ($) at each
health state modeled in one IVF cycle.
Health state hFSH ($) rFSH ($)
Start of cycle cost 3,172.08 3,709.93
Oocyte retrieval 3,723.72 3,723.72
Embryo success 3,844.89 3,844.89
Embryo failure 2,876.02 2,876.02
Chemical pregnancy 960.96 960.96
Clinical pregnancy 806.52 806.52
Ovarian hyperstimulation
1,155.44 1,155.44
Note: Figures were taken from corresponding values
reported in a previous study (3) and inflated to 2003
Hatoum. Cost-effectiveness of hFSH vs. rFSH. Fertil Steril 2005.
807Fertility and Sterility
Show more

Recommended publications

Discover more