A systematic review of the effectiveness of adalimumab

TABLE 32 TNF inhibitors as last active therapy
Similarly, the results for first line therapy are
shown in Table 33.
A limited analysis of sequential use was carried
out, assuming that the properties of second or
third TNF inhibitors were equivalent to the use of
the same treatment as first TNF inhibitor. The
results were similar to the results for the
equivalent therapy as sole TNF inhibitor.
The main aim of the analysis was to assess the
cost-effectiveness of adding a TNF inhibitor to an
existing treatment pathway for RA compared with
the same pathway without that TNF inhibitor. The
costs are from an NHS perspective.
The analysis was conducted using an updated
version of the BRAM, 2 which was further
developed starting from the most recent previous
version (used in the assessment of anakinra 3 ).
The BRAM is an individual sampling model. The
model was devised to reflect the typical real
clinical patient pathway. A large number of virtual
patient histories is simulated with the
accumulation of costs and QALYs. The basic
model structure is shown in Figure 46. A complete
description of the model structure follows here.
Patients are assumed to follow a sequence of
treatments (single or combination therapy), which
© Queen’s Printer and Controller of HMSO 2006. All rights reserved.
Health Technology Assessment 2006; Vol. 10: No. 42
Sensitivity analyses (£)
TNF inhibitor Comparator Cost per QALY (£) Lowest Highest
Adal (no MTX) Base strategy of 40,000 27,000 64,000
Etan (no MTX) DMARDs with no 24,000 18,000 33,000
Adal (with MTX) TNF inhibitors 30,000 22,000 43,000
Etan (with MTX) 24,000 18,000 34,000
Infl (with MTX) 38,000 26,000 61,000
TABLE 33 TNF inhibitors as first-line therapy
Sensitivity analyses (£)
TNF inhibitor Comparator Cost per QALY (£) Lowest Highest
Adal (no MTX) Base strategy of 53,000 27,000 122,000
Etan (no MTX) DMARDs with no 49,000 23,000 119,000
Adal (with MTX) TNF inhibitors 171,000 38,000 Dominated
Etan (with MTX) 78,000 28,000 Dominated
Infl (with MTX) 654,000 45,000 Dominated
involves: starting a treatment, spending some time
on that treatment, quitting the treatment if it is
toxic or ineffective, and starting the next
treatment. The pattern is then repeated. Any
patient who has started and had to quit all the
active treatments moves on to palliation. Patients’
HAQ scores are assumed to improve (decrease) on
starting a treatment; this improvement is lost on
quitting the treatment, which may be for reasons
of either toxicity or loss of effectiveness. HAQ
scores can range from 0 (best) to 3 (worst) and are
constructed such that the smallest measurable
change in disability is 0.125 (see Appendix 1 for
further details of the HAQ). Reflecting observed
data, while on any treatment, a patient’s condition
is assumed to decline slowly over time; this is
modelled as periodic increases of 0.125 in HAQ
score.
All patients are followed through to death.
Mortality risk is assumed to depend on current
HAQ score, as well as age and gender.
There are two important improvements from
previous versions of the BRAM. First, there is
individual variation in HAQ improvement on
starting treatment. Secondly, time on treatment
includes explicit consideration of early quitting,
with early quitting owing to lack of effectiveness
being correlated with poor HAQ improvement on
starting treatment.
87