NICE: Colorectal cancer (metastatic) - bevacizumab: Evidence Review Group Response
ERG Comments On Manufacturers Response
Summary
The
ERG agrees that the cumulative effect of the three revisions to the
model parameters in the manufacturer’s response did not
significantly change the ICER values. The ERG believes that even
with the modifications in the manufacturers’ response the
ICERs are still associated with some uncertainty. The ERG comments
on each of the areas of uncertainty discussed in the
manufacturer’s response are included below.
1) The NHS resource cost of operating APAS and
the subsequent effect on the ICER
The
manufacturer’s response details research with pharmacists,
NHS business managers, and NHS finance and operations experts. This
research identified the employee time required to set up and
administer the APAS. The details were provided in a worksheet in
appendix C. This worksheet is unclear and includes links to values
in external worksheets so the ERG was unable to check these
calculations. The analysis spreads the costs of initial set-up
activities over 3 years. It is not clear to the ERG what the
appropriate timeframe is over which to spread these set-up
costs.
The
ERG can verify that based on an estimated cost per patient of
operating the APAS over years 1 to 3 is of £57 and £67
for B-XELOX and B-FOLFOX respectively the marginal ICERs increase
by £164 and £113 respectively.
2) The operation of the APAS in the context of an
intermittent treatment strategy
No
comments.
3) Bevacizumab treatment duration in clinical
practice and its effect on the ICERs
Clinical Advisory Board
The
manufacturer’s response states that “As recognised by
the committee, intermittent treatment, such as that specified in
the COIN study (Adams, 2009) is becoming more prevalent within UK
clinical practice. This typically leads to a shorter treatment
duration compared to the NO16966 study.”
The
model reflects the N016966 study in which both the protocol for the
bevacizumab and comparator arms was continuous treatment. If
intermittent treatment +- bevacizumab were modelled then we may see
a shorter treatment duration with bevacizumab compared with the
N016966 study but we would likely also see a shorter treatment
duration with XELOX/FOLFOX. As this would reduce the costs in both
intervention and comparator arms the effect on the ICER is unclear.
It is also unclear how intermittent treatment would effect the time
spent in the PFS and OS states.
The
manufacturers response stated that “It was considered that
should bevacizumab be given positive NICE guidance it is likely to
be added to the treatment strategy that is currently being
employed, either intermittent treatment or continuous.” In
contrast, one of the ERG clinical advisors suggested that although
they currently use intermittent treatment they would administer
B-XELOX of B-FOLFOX continuously because that is where the evidence
base lies. 2
The manufacturer’s response
states that “It would also be expected that if treatment with
oxaliplatin was stopped due to either a planned break or
unacceptable toxicity then treatment with bevacizumab would also
typically be stopped at this time.” The ERG note that this
differs from the trial protocol. This clinical debate is drawn to
the attention of the committee.
Real world evidence of bevacizumab treatment
duration
The
ERG group comment that although real world evidence may demonstrate
shorter treatment duration than was seen in the trial. There is an
inevitable tension between modelling the internal consistency of
the trial and the external reality of real world clinical practice.
However, the conventional approach to economic modelling would be
to reflect the trial situation as this enables the source for both
costs and efficacy to be the same.
4) The incremental pharmacy and administration
cost associated with adding bevacizumab to XELOX or
FOLFOX
The
manufacturer’s response includes a time and motion study to
provide a more precise estimate of the cost of preparing and
administering bevacizumab. Appendix B states that the study
included only 3 patients and was undertaken at one institution (the
Mount Alvernia Hospital in Guildford). The ERG suggests that such a
small sample size which just includes one hospital may not
accurately reflect the time required. The manufacturers response
suggest reducing bevacizumab preparation and administration costs
from £42 to £31 but it is unclear precisely where this
was altered within the model. When the model was modified by the
ERG to reflect this change the marginal effect on the ICERs was
slightly different to that in the manufacturer’s
response.
5) The health state utility values used in the
economic model
The
ERG agrees that varying the post progression utility value has
little effect on the ICER. The ERG agrees that reducing the PFS
post treatment utility value may make the model more realistic. The
marginal effect on the ICERs of changing this utility values has
been verified by the ERG.
Other points
The
manufactures response states that:
“ In section 4.14 the ACD states:
“Nevertheless, the
Committee understood from the ERG that the ICERS for both B-XELOX
and B-FOLFOX increased if bevacizumab treatment continued beyond
that of oxaliplatin.” It appears that this conclusion was drawn based
on the results of the sensitivity analysis conducted by the ERG
(section 3.25, ACD). Additionally it is not clear as to whether the
committee considered the impact of the price cap on the cost of
increasing treatment duration. ”
The
ERG reported that the ICERs for both B-XELOX and B-FOLFOX increased
significantly if bevacizumab treatment continued beyond that of
oxaliplatin and if survival was still assumed to be as in the trial
(i.e. the additional month of bevacizumab treatment did not alter
survival). The ERG analysis did not originally include the 12 month
price cap within the calculations. This analysis has now been
repeated to include the price cap and for an additional 1 month of
bevacizumab the ICERs increase from £36,354 to £45,958
for B+XELOX vs. XELOX and from £40,090 to £31,452 for
B-FOLFOX vs FOLFOX. The ERG wishes to emphasize that this was an
exploratory analysis which assumes no difference in survival is
caused by the additional month of bevacizumab, and as such is
likely to over-estimate the true ICER. This exploratory analysis
highlights that due to the structure of the APAS the incremental
cost of continuing bevacizumab after oxaliplatin cessation is
almost three times the incremental cost of adding bevacizumab to
oxaliplatin.
Posted: August 2010
