The threshold should not be increased.
There are some obvious problems with the arguments for increasing the
threshold for cost effectiveness in the article by Towse (1) that argued
for increasing the cost effectiveness threshold.
The DH commissioned study on which much of the argument rests draws more
nuanced conclusions than the assertion by Towse that this presents
evidence for policy makers to double the current threshold for
reimbursement of treatments in the NHS.
Firstly, Mason et al (1) highlighted that willingness to pay might be
dependent on what type of QALY is being bought. There may be a marked
difference between the taxpayers willingness to pay for life saving
interventions in an emergency scenario (the so called Rule of Rescue) and
interventions that are given in as part of a planned course of care.
Although there is usually consensus with regard to the application of rule
of rescue in individual clinical circumstances, the debate is far from
settled when it comes to the public’s attitude towards applying the rule
of rescue on macro resource allocation decisions. This is an important
nuance, and the difference between individuals and population resource
allocation rules should not be overlooked.
Willingness to pay might also be highly dependent on sample selection and
who is being asked. There are large differences in peoples’ valuation of
lives when age, remaining life expectancy and current (and future) health
state is taken into account. Willingness to pay is usually reflective of
ability to pay.
Thus a simple assertion that there is enough evidence to simply double the
threshold seems an overstatement of the arguments.
There are more fundamental issues that should be addressed prior to
increasing the NICE threshold.
Opportunity cost. The health care system can barely afford to fully
implement all the health technologies that are currently appraised as cost
effective within the current NICE threshold. This problem may become
increasingly acute in a health system with reduced funding as a result of
the recession. An increased threshold will mean a large swathe of new
technologies will be introduced, offering poorer value for money than the
There are three alternative scenarios that might occur as a result of an
increased threshold and introduction of new, less cost effective
treatments. Commissioners would need increased budgets to fund new (and
existing) technologies. This seems unlikely in the current economic
climate – as unemployment rises tax receipts will fall and expenditure on
other social welfare systems will increase – most likely restricting the
increase of funds to health care. The second option is that commissioners
will need to make increasingly difficult choices about which programmes to
cut. It is possible that some of the soft underbelly services would be
high on the hit list for funding cuts. Such services might include
prevention, primary care and end of life care. It is not clear that such
an outcome is the one that the public desire. The taxpaying public may not
even be aware of the consequence of the introduction of a higher
threshold. A third alternative might be that taxpayers can be persuaded to
pay more tax to pay for increased investments in expensive new health
technologies that would fall within a higher cost effectiveness threshold.
Again, this seems unlikely.
The policy objective of increasing the threshold for routine reimbursement
would need to be clearly explained to the taxpayer. It seems likely that
treatments with an ICER of between £30,000 and £70,000 per QALY will not
achieve substantial improvements in life expectancy in a population. The
marginal return on additional investments of such expensive treatments
might be very small indeed. This raises the question of whether society
considers it worth introducing treatments with such high ICERS into
routine funding, or whether we would be entering into the arena of
diminishing marginal returns. It is instructive to compare a treatment
with an ICER of between £30,000 and £70,000 per QALY and the estimated
ICERs of current NHS services. A recent study (3) highlighted that the
cost per QALY in two high impact programmes has been estimated to be lower
than the current NICE threshold (22k for Ca, 14k for CVD). This has widely
been interpreted as evidence that the current threshold for new
technologies is on the high side.
There are well established mechanisms for approving the use of treatments
with a very high ICER. Ultimately these are at national level (through
NICE appraisal process, and the process for determining whether to approve
a treatment with an ICER in excess of the commonly quoted threshold), at
PCT level (though Exceptional Case Panels), and at the level of the
individual patient. There is undoubtedly scope for improving the process
at all three levels and making the criterion on which such decisions are
taken more explicit. Ultimately such decisions will create tension between
the ethical framework of the clinician (patient perspective) and the
commissioner (population perspective). Difficult ethical judgments are
required; politics is never far from the surface. It is exceedingly
difficult to see how the tension of how much efficiency to trade for
making new treatments available will be resolved by economics alone. There
is sometimes inconsistency in the principals of resource allocation in
clinical decisions (4); moreover there seems to be little agreement about
the morally relevant values upon which societal health care resource
allocation decisions should be taken at societal level. Perhaps this is a
key area for further research with the public, rather than continual
refinement of economic analysis models. The latter is a technical
solution, it will not bring social values into a more central place in
resource allocation and the trade off between efficiency and making
expensive treatments available for some.
There is vast scope for reducing inefficiency and increasing the
value that patients (and taxpayers) get out of the current mix of health
care treatment. This might be best achieved within a Programme Budgeting
and Marginal Analysis framework where clinicians and other important
stakeholders systematically identify (and reduce investment in) treatments
that offer low value for money and increase investment in high value
treatments. Within such a framework, assuming it is well controlled,
clinicians would be able to introduce more costly treatments; but they
would be equally responsible for determining which treatments to cut from
the mix to make way for any additions, and – importantly – responsible for
any overspend on the treatment budget for their population.
One might ask whose interests are served by increasing the threshold at
which interventions would be reimbursed. Clearly the pharmaceutical
industry and device manufacturers have much to gain, as has the patients
who might benefit from treatment that might not otherwise be reimbursed.
It is not clear, however, that the taxpayer would benefit from the
introduction of treatments that offer low value for money. Further clarity
is needed that those taxpayers whose views are elicited and say they are
willing to pay significantly more for health gain are doing so in the full
knowledge of the opportunity costs; or they are prepared to pay more tax
to increase the overall budget of the health care system.
(1) Towse A. Should NICE’s threshold range for cost per QALY be
raised? Yes. BMJ 2009;338:b181
(2) Mason H, Jones Lee M, Donaldson C. Modeling the monetary value of a
QALY: a new approach based on UK Data. Health Economics 2008.
(3) Link between spending and health outcomes. York Centre for Health
Economics Research Paper 42. 2008.
http://www.york.ac.uk/inst/che/pdf/rp42.pdf (accessed Feb 09).
(4) Persad G, WertheimerA , Emanue lE J. Principles for allocation of
scarce medical interventions. Lancet 2009; 373: 423–31
Competing interests: No competing interests