Converting international cost effectiveness data to UK pricesBMJ 2002; 325 doi: https://doi.org/10.1136/bmj.325.7358.275 (Published 03 August 2002) Cite this as: BMJ 2002;325:275
All rapid responses
Gosden and Torgerson present an interesting discussion about the
difficulty of applying international cost-effectiveness analyses (CEA) to
the local setting.1 Not only are variations in currency values and
clinical practice important, but the authors point to another potential
problem: the impact of time.
Guidelines for the conduct of CEA in health care recommend the
discounting of both costs and benefits.2,3,4 This allows for the impact of
time preferences but not for changes in prices. Furthermore, while some
guidelines may go so far as to suggest unit prices for use in CEA of
pharmaceuticals, they do not allow for prices to change over time. This
may complicate CEA where an economic model is required to extrapolate
beyond the available trial data. Where such an economic model extends
beyond one year, guidelines suggest that all costs be expressed in
constant prices. This means that prices valued in today’s terms should be
used for the entire period of an economic model, not that the same price
level should be applied throughout.
This presents a challenge in terms of how to conduct and interpret
CEA. In the presence of pharmaceutical price inflation, the use of
constant price levels may serve to underestimate future costs included in
an economic model, potentially presenting a more favourable cost-
effectiveness ratio than will eventuate.
But, what if there is pharmaceutical price deflation? Lower market
prices often result from increased competition such as occurs when a
pharmaceutical is no longer under patent.5 If treatment patterns remain
the same, as the post patent price falls then so too the cost of treatment
falls. Unless these lower costs are factored into the economic analysis,
it may result in a less favourable cost-effectiveness ratio than may
To maintain the same price level throughout the period of an economic
model in the face of a post patent price fall is to in effect use an
inflated price, expressed in today’s prices, for the post patent period.
Not only does this go against accepted guidelines, but it may have a
considerable impact on the results of the CEA if a significant proportion
of the modelled costs due to the acquisition of pharmaceutical products
arise in what would otherwise be the post patent period. We believe that
the impact of price changes in post patent periods on modelled economic
evaluations should be investigated further.
1 Gosden TB, Torgerson DJ. Converting international cost
effectiveness data to UK prices. BMJ 2002; 325: 275-276.
2 Drummond MF, Stoddart GL, Torrance GW. Methods for the Economic
Evaluation of Health Care Programmes. Oxford Medical Publications. Oxford.
3 Australian Department for Health and Ageing. Guidelines for the
Pharmaceutical Industry on Preparation of Submissions to the
Pharmaceutical Benefits Advisory Committee including major submissions
involving economic analyses. 1995. Online.
4 Ontario Ministry of Health and Long Term Care. Ontario Guidelines for
Economic Analysis of Pharmaceutical Products. 1994. Online.
5 Reekie WD. How competition lowers the costs of medicines.
Pharmacoeconomics 1998; 14 Suppl 1: 107-113.
Competing interests: No competing interests
In your editorial you mention that currency conversions make the use
of "foreign" pharmacoeconomic analyses of questionable value. I agree that
currency conversion is a difficult task, particulairly when we have seen
the Euro: US Dollar exchange rate go from 1:$0.80 to 1: 1.08 in less than
two years. However, the problem of external validity or usefulness in
other settings goes far beyond currency conversion issues. Even within the
same country, published analyses tend to be of little value because the
underlying prices or costs of the medications and other resources
themselves tend to vary dramatically between institutions. Drugs that were
prohibitly expensive in one analysis can be dominant (i.e. both more
effective and less expensive) from another institution's perspective.
In the United States, some authors try to get around this problem by
using published "Average Wholesale Prices" ("AWP") supplied by the
manufacturer; however, these are usually a dramatic overstatement of the
amounts actually paid by any real purchaser (including the retail or
The remedy to this problem probably lies in authors publishing (or
making available on software) both the description and number of each type
of health care resource used, and the price or cost assigned to it. In
that way, someone in Britain, or Poland, or even in another State in the
USA could substitute either their own unit price or number of resources
they consumed. Rather than stating a given surgery had a cost of $10,000,
if I knew that was comprised of 4 days of stay at a cost of $2,500 per
day, I could open up a computer spreadsheet and rerun the analysis using my
own assumptions, based on the standard of care in my institution or
setting and my cost per unit consumed in that setting.
Unfortunately, such information is often lacking. In unpublished
study of mine, in which 125 studies (1996 - 2001) of cost-effective
devices were described in surgery, only 5 articles actually listed the
unit costs, 2 listed the number of units consumed, and none listed both.
The best information that can typically be found in even a great analysis
is the cost per unit consumed (even that is often concealed for the sake
of preserving "proprietary" information).
Authors and sponsors of pharmacoeconomic research need to realize
that decision-makers require an analysis that is sufficiently transparent;
it can be customized for their own setting, and include access to the
underlying costs and resource consumption. Until then, I will continue to
have doubts as to the extent to which such published material is actually
being used by decision-makers in ANY country.
Lorne Basskin, PharmD
Competing interests: No competing interests