Intended for healthcare professionals

Analysis

NICE and the challenge of cancer drugs

BMJ 2009; 338 doi: https://doi.org/10.1136/bmj.b67 (Published 13 January 2009) Cite this as: BMJ 2009;338:b67

Economic consequences of approval extensions: the case of rituximab

The recent development of numerous targeted treatments for oncology
patients has raised
well-known issues of economic sustainability [1]. The majority of
comments in this area have focused
on new chemical entities, and particularly on new high-cost agents such as
monoclonal antibodies [1,2]. However, one problem that has so far attracted
little attention is the economic impact that can result from approval extensions (i.e. in situations where a
target
treatment, originally approved for its first clinical indication, receives
approval for additional indications, especially when large series of patients are involved by the approval
extension).

The case of rituximab is emblematic. The first approval of this agent
dates back to 1999.
After these initial approvals as induction therapy for some types of
non-Hodgkin's lymphoma (NHL) as well as for rheumatoid
arthritis, the drug has received in the past year the approval extension
also
for the maintenance treatment of NHL and
for first-line or second-line treatment of
chronic lymphocytic leukaemia (CLL).

What economic
impact can be predicted for the approval extension regarding CLL? A combined
analysis of epidemiological studies and administrative databases (Messori et al, unpublished observations,
2011) indicates that, in Italy, the
number of patients with CLL that will be treated with rituximab is
likely to be
around 2,000 per year. The cost
of this
treatment at standard dosages (500 mg/sqm for 6 cycles) is around EUR 14,000
per patient. This gives rise to a nationwide cost for the Italian health
care
system of EUR 28million per year.
If one carries out the same analysis considering the approval extension
for maintenance
treatment of NHL, the typical cost per
patient is around EUR 10,000 (at the dosage of 375 mg/sqm every
two months), while the total number of expected cases in Italy is around
1,500 per year (Messori et al. unpublished
observations, 2011). This translates into a total nationwide yearly
expenditure
for this new indication of EUR 15 milllion per year.

Figure 1 (that can also be seen from url www.osservatorioinnovazione.org/netma/rituximabfigure

.pdf)
shows the curve of predicted expenditure versus time for rituximab based
on a separate
analysis of these two new indications. The attainment of the maximum
number of patients has been modeled
as a gradual process. The equations governing this progressive attainment of
steady state are described in Appendix 1.

Overall, our data
indicate that, while the historical expenditure of rituximab in Italy has
been around EUR 140 million per
year, the approval of these two new indications
could imply an extra-budget of nearly EUR 45 million per year.
The price of rituximab has already been negotiated at the time of its
initial approvals (induction
treatment in NHL and rheumatoid arthritis), and so it is very unlikely that
this price can be renegotiated. Hence, the Italian agency will have very
limited opportunities to find a solution to appropriately govern the
economic
implications of this problem.

In conclusion, this example shows that, from the viewpoint of
expenditure predictions,
approval extensions can be, in economic terms, as relevant as approvals
of new
chemical entities. On the other hand renegotiating
the drug's price for approval extensions can be, in operative
terms, more difficult than the initial negotiation of firstly approved
indications.

References

1. Claxton K, Briggs A, Buxton MJ, Culyer AJ, McCabe C, Walker
S, Sculpher MJ. Value based pricing for NHS drugs: an opportunity not
to be missed? BMJ 2008;336:251-254 doi:10.1136/bmj.39434.500185.25

2. Raftery J. NICE and the challenge of cancer drugs.BMJ 2009
338:b67; doi:10.1136/bmj.b67

3. Gibaldi M, Perrier B. Pharmacokinetics. New York: Dekker, 1982;p.494.

4. Russo P, Mennini FS, Siviero PD, Rasi G. Time to market and
patient access to new oncology products in Italy: a multistep pathway
from European context to regional health care providers. Ann Oncol. 2010
Oct; 21(10):2081-7. Epub 2010 Mar 24.

5. Engelberg AB, Kesselheim AS, Avorn J. Balancing innovation,
access, and profits -- market exclusivity for biologics. N Engl J Med
2009;361:1917-1919[Erratum, N Engl J Med 2010;362:664.]

APPENDIX 1.

The predictions in terms of yearly
expenditure illustrated in Figure 1 are based on a specific modeling
(see Box 1
for symbols and abbreviations) in which the mathematic equations employed to
generate the curves are the following:

Equation 1:

PATIENTS =
TARGET x (1 - e-0.693/HLgrowth x TIME)

for cases
where
the value of time was between 0 and LEdrug.

Equation 2:

PATIENTS =
TARGET x e-0.693/HLobsolescence x (TIME - LEdrug)

for cases
where
the value of time was between LEdrug and infinity.

These curves were directly derived from
standard exponential equations that are identical to those commonly used for
pharmacokinetic modeling [3]. In the graphs, the starting point for the
curve
was set at a date equal to TIME0+LAGTIME and the curves were plotted
accordingly. In our analysis, LEdrug was set at 90 mos, HLgrowth at 3 mos,
HLobsolescence at 18 mos. The value of LAGTIME was set at 12 mos by assuming
that time zero was the approval by EMA [4].

The information on HLobsolescence
available
from the literature is very limited (e.g. see reference 5); furthermore, the
lack of adequate information affects not only HLobsolescence but also on
LEdrug, which are two strictly interrelated parameters influencing the
time-course of the "right tail" of the curves. For this reason, the model
reliability was thought to be insufficient for predictions extending
beyond the
achievement of a steady state; as a
result the curve represented in Figure 1 (large panel) was truncated at
the end
of 2015.

Box
1. Parameters employed in the budget impact predictions described in
Appendix 1.

Parameter

Explanation

Units

TARGET

Yearly population (at steady state) of
patients who are candidates to the new treatment

No. of patients

HLgrowth

Rate at which the number of patients
grows from 0 to TARGET; the process is assumed to be based on a
first-order
constant defined as 0.693/HLgrowth. Like in standard pharmacokinetic
calculations, the steady state is assumed to be reached after as much
time
has elapsed as five times the value of HLgrowth

Months.

LEdrug

Life expectancy of the drug (starting
from its marketing authorization until the beginning of its obsolescence
process)

Months

HLobsolescence

Rate at which the number of patients at
the end of LEdrug starts to decline from TARGET to zero; the process
of this
decline is assumed to be based on a first-order constant defined as
0.693/HLobsolescence. Like in standard pharmacokinetic calculations,
a value around
zero patients is assumed to be reached when as much time has elapsed
as five
times the value of HLobsolescence.

Months

TIME0

This parameter is defined as the
date of approval by the regulatory
agency.

Date (dd/mm/yyyy)

LAGTIME

Interval between TIME0 and the
time-point
where the curve of the expenditure starts; in practice, LAGTIME
expresses the
delay between TIME0 and the real availability of the product in the
market so
that this interval essentially includes the price determination (in
countries where
this negotiation takes place) and the inclusion in local
formularies.

Months

Competing interests: No competing interests

26 May 2011
Dario Maratea
research assistant
Laboratorio SIFO di Farmacoeconomia, ESTAV-Centro,Prato. Italy