Ethics and effectiveness: rationing healthcare by thresholds of minimum effectiveness

The authors believe that basing treatment funding decisions on
clinical effectiveness rather than cost-effectiveness would be more
ethical, fairer and more acceptable to the public. They propose the
employment of predefined thresholds of direct clinical effectiveness as
the cut-offs for healthcare technology approval.

However, the method they propose would not meet the requirements of
an agency faced with budget constraints and charged with decision making
and ensuring the efficient allocation of limited resources at population
level. Basing decisions predominantly on clinical effectiveness without
sufficient consideration of costs ignores the reality of the need to
ration resources to ensure that value for money is maximised across the
sector; this requires the joint consideration of cost and effectiveness.

In the example given by the authors, Erlotinib would not be approved
because it would not reach the (example) minimal clinical effectiveness
threshold of 3 months life extension (and the authors also add that it is
relatively expensive). Would the decision be the same if there were a
treatment available that was a quarter of the price but typically only
gave 2 months life extension? Is there a price at which such a treatment
would be considered value for money compared to other treatments available
or funded across the health sector? Basing decisions on clinical
effectiveness alone would make it difficult to answer such questions.

Initially it seems the authors advocate removing cost from the
healthcare technology funding decision process. Subsequently it is stated
that rationing by the newly proposed minimal effectiveness method would
contain costs. It is not clear in what way costs would be introduced into
the decision making process and what levels of costs would be considered
acceptable or whether there would be any relationship to the effects
(value) of the treatments under consideration. In this sense, the proposed
method is less fair than that currently employed by NICE which, despite
its drawbacks, is at least explicit and relatively transparent in the
methods and benchmarks it employs.

Presumably it is hoped that the decision process would eliminate less
effective treatments early in the process on the basis of specific
clinical effectiveness thresholds (for which decision makers are not
responsible or culpable) thereby negating economic-based decisions.
Consequently, rejections are more likely to be palatable to the public by
stating that cost is only a final consideration. The decisions made using
the proposed method would not be made fairer simply by being less well
defined or by being based on fuzzy economic thresholds that are withheld.

Instead of employing the quality adjusted life year (QALY) as the
metric to measure the benefit of treatments, the proposed decision process
would involve separate evaluations of the health-related quality of life
and survival benefits of a technology under consideration. In life
extending treatments the authors seem to suggest three thresholds should
operate - an explicit one each for survival and for health-related quality
of life and an implicit one relating to costs. Not only would this be
considerably more complex than the current system used by NICE but it
would also be less transparent and less amenable to comparative analyses.

The minimal clinical effectiveness method would require the
derivation of thresholds - for every therapeutic area - to determine
whether a treatment is effective or not. Establishing the thresholds and
their validity and equitability (across all therapeutic areas) would be a
substantial and challenging undertaking. In essence, the derived
thresholds may be arbitrary rather than representing a meaningful measure
of clinical effectiveness. Generally, when asked to state the minimum
acceptable levels of things for which they have ill-defined preferences,
individuals tend to use rules of thumb and heuristics. This may explain
why the authors suggest an improvement of 10% in effect as a suitable
threshold and why responses to the reported example of minimum acceptable
life extension by funded treatments appear to be clustered at 6 and 12
months. Although the NICE QALY threshold range of 20,000-30,000 Pounds
itself may be based on arbitrary valuations, this at least now reflects
coherent arbitrariness [1] since it has been used for several years and
has set a precedent which can be used as an anchor on which to base future
value judgements.

If new thresholds were to be introduced then the use of meaningful
definitions of clinical benefit such as minimal importance differences for
changes in health-related quality of life measures [2] would, at least, be
less arbitrary. Regardless of the meaningfulness of the thresholds,
deriving unique benchmarks for every therapeutic area precludes the
ability to compare the treatment benefits across the sector.
Incomparability would make informed and consistent distribution of
resources across the health sector - a key requirement of the proposed
value-based pricing system in the UK [3] - very difficult.

While the authors should be applauded for attempting to address some
of the shortcomings of the cost-effectiveness paradigm, it is unlikely
that the minimal clinical effectiveness method would meet the needs of a
national health care system agency trying to maximise value for money from
a fixed budget.

Attempting to increase acceptability to the public and perceptions of
fairness at the expense of transparency, simplicity and comparability is
not desirable or responsible. Worse still, such a move may threaten
fairness by precluding effective distribution of resources.

1. Ariely D, Loewenstein G, Prelec D (2009). "Coherent
arbitrariness": stable demand curves without stable preferences. In
Loewenstain G (Ed) Exotic Preferences: Oxford University Press.

2. Revicki D, Hays RD, Cella D, Sloan J. Recommended methods for
determining responsiveness and minimally important differences for patient
-reported outcomes. J Clin Epidemiol. 2008 Feb;61(2):102-9.

3. Department of Health. A new value-based approach to the pricing of
branded medicines: A consultation. 16th December 2010.