REVIEW ARTICLE Pharmacoeconomics 1999 Apr; 15 (4): 357-367

1170-7690/99/0004-0357/$05.50/0

© Adis International Limited. All rights reserved.

Recent Advances in the Methods of

Cost-Benefit Analysis in Healthcare
Matching the Art to the Science
Emma McIntosh,1 Cam Donaldson1,2 and Mandy Ryan1
1 Health Economics Research Unit, Department of Public Health, University Medical Buildings,
Foresterhill, Aberdeen, Scotland
2 Departments of Community Health Sciences and Economics, University of Calgary, Calgary,
Alberta, Canada

Contents
Abstract . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 357
1. Recent Criticisms of Economic Evaluation in Healthcare: Implications for
Cost-Benefit Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 358
2. Balance-Sheet (or Opportunity Cost) Approach . . . . . . . . . . . . . . . . . . . . . . . . . . . . 359
3. Recent Developments in Monetary Valuation: Willingness to Pay, Conjoint Analysis
and Valuation of Production Gains . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 361
3.1 Willingness to Pay . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 361
3.2 Conjoint Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 362
3.3 The Debate on Production Gains (or Indirect Costs) . . . . . . . . . . . . . . . . . . . . . . . . 363
4. Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 365

Abstract This paper outlines recent advances in the methods of cost-benefit analysis
(CBA). Economic evaluations in healthcare can be criticised for, amongst other
things, the inappropriate use of incremental cost-effectiveness ratios and the re-
porting of benefits in terms of cost savings, such as treatment costs averted. Many
such economic evaluations are, according to the ‘scientific’ definition, CBAs.
The ‘balance-sheet’ (or opportunity cost) approach is a form of CBA which can
be used to identify who bears the costs and who reaps the benefits from any
change. Whilst the next stage in a CBA, as defined in health economics, would
require that all costs and benefits be valued in monetary terms, the balance-sheet
approach, however, advocates that available monetary values can be augmented
by other measures of cost and benefit. As such, this approach, which has a theo-
retical basis, is proposed as a practical prescription for CBA and highlights the
notion that unquantified benefits are important and can be included within CBAs
even when monetarisation is not possible.
Recent methodological developments in monetary valuation for use in CBA
are the development of the technique of willingness to pay, the use of conjoint
analysis (CA) to elicit willingness-to-pay (WTP) values and advances in the
debate on the inclusion of production gains in CBAs. Whilst acknowledging that
there have been developments in each of these areas, it is claimed there has also
358 McIntosh et al.

been progress in using CBA as a framework for evaluation, as reflected by the

balance-sheet approach.
The paper concludes by stating that almost all types of economic evaluation
have an element of the ‘cost-benefit’ approach in them. The important issue is to
focus on the policy question to be addressed and to outline the relevant costs and
benefits in a manner which assists the evaluation of welfare changes resulting
from changes in healthcare delivery. The focus should not be on moulding a
question to fit a hybrid definition of an analytical technique.

Recently, the literature on economic evaluation
of healthcare has been criticised for not adhering
strictly to some basic economic principles. [1-4] The
outcome of this debate has been that not only
should more studies be recognised as cost-benefit
analyses (CBAs), but that CBA is the preferred
form of analysis in many situations.[5]
Despite this, the view remains that CBA is dif-
ficult to carry out: ‘CBA offers the most compre-
hensive and theoretically sound form of economic
evaluation, but its application in the context of
healthcare is fraught with practical problems.’[6]
Such ‘conceptual, ethical and practical prob-
lems’[7] are largely associated with valuing the ben-
efits of healthcare in monetary terms. However, in
recent years there has been substantial growth in
the number of studies estimating such monetary
values.[8] This reflects methodological advance-
ments in the area, some of which overcome the
ethical problems referred to above. Furthermore,
one issue that has been mentioned in recent criti-
cisms of economic evaluation, but not followed-up
in much depth, is that CBA does not have to involve
monetary valuation of benefits.
The aims of this paper are 2-fold. First, based
on the notion that unquantified benefits are impor-
tant and should be reported, to describe the bal-
ance-sheet approach which was proposed many
years ago as a mechanism to deal with this, but
which has been used hardly at all in the intervening
period.[9] Secondly, to outline methodological de-
velopments in monetary valuation for use in CBAs.
In the following section, the recent criticisms of
economic evaluation in healthcare and the implica-
tions of this for the technique of CBA are outlined
briefly. The paper then discusses the important ad-
vances in the ‘art’ of CBA. These advances include:
the balance-sheet approach, which allows useful
CBAs to be carried out even when benefits cannot
be valued in monetary terms; and developments in
valuing benefits in monetary terms (as this can still
have an important role). Finally, the issue of why
the points raised in this paper are of practical sig-
nificance is addressed.
1. Recent Criticisms of Economic
Evaluation in Healthcare: Implications
for Cost-Benefit Analysis
There are two basic criticisms of economic eval-
uation in healthcare that have implications for
CBA. The first centres around the notion of the
incremental cost-effectiveness ratio (ICER). If this
ratio is made up of data on costs and effectiveness
for a new treatment which are incremental to the
status quo, then no matter how low the extra cost
per unit of effectiveness gained, implementation of
the new treatment will require more resources to be
allocated to the area of care concerned. It is no
longer a question of technical efficiency but alloca-
tive efficiency. An opportunity cost will be in-
curred because those resources will not be avail-
able to pursue some other activity on behalf of
some other group of patients. Whether the pro-
gramme should be implemented will depend on its
size, from where the resources to fund it are ex-
pected to come and on the benefit associated with
other potential uses of those resources. Thus, it is
inappropriate to recommend funding of an inter-
vention solely on the basis of it having a low ICER;
such recommendations have been made by several
authors in the past.[10-13]
Indeed, the term cost effectiveness, within the

 Adis International Limited. All rights reserved. Pharmacoeconomics 1999 Apr; 15 (4)
Methods of CBA in Healthcare
notion of an ICER, is often a misnomer. If costs are
incremental to the status quo, a decision will have
to be taken as to whether these costs are worth in-
curring. Such decisions could be taken with or
without the aid of economic techniques of benefit
valuation. Whichever way decisions about ‘worth-
whileness’ and allocative efficiency are made,
ICERs really provide information for use in CBAs.
This view that ICERs provide information for use
in CBAs, has some theoretical justification.[14,15]
The second criticism is that many economic
evaluations report benefits in terms of cost savings
such as treatment costs averted.[16-19] This clearly
involves only a comparison of costs, i.e. the costs
of the intervention are compared with the costs of
what would otherwise have happened (the costs
averted). No consideration is given to the difficult
issue of valuing health improvements and other
benefits, a point made by Birch and Donaldson[20]
over 10 years ago, and rather disappointingly, one
which still had to be made 10 years later.[16]
According to Donaldson and Shackley,[21] items
that should be included in the cost side of an eco-
nomic evaluation include any resources that have
an opportunity cost as a result of being associated
with the healthcare programme under considera-
tion. Correspondingly, the benefit side would con-
sist of health effects and other impacts on peoples’
well-being. Thus, all effects on resource use, whe-
ther positive or negative, are netted out on the cost
side, whilst all effects on well-being are counted
on the benefit side. Table I provides a summary of
such costs and benefits.
The omission of welfare changes (i.e. valuation
of health and other benefits) renders many studies
with CBA labels as nothing more than cost analy-
ses. At best, if these welfare changes can be as-
sumed to be equal across options, such studies
amount to cost minimisation analyses, looking at
the least costly way of achieving a given objective.
In 1974, Williams[22] also noted this tendency, stat-
ing that far too many recommendations are based
on assertions that x is cheaper than y (without ad-
equate consideration of relative benefits), or that x
is more effective than y (without adequate consi-
 Adis International Limited. All rights reserved.
359
Table I. Costs and benefits
Organisations incurring Form of benefits to
resource use/costs patients/public
National Health Service Gains in health
Other public sector Other nonhealth benefits, e.g.
information and reassurance
Patients and their families Process benefits, e.g. reduced
waiting times
deration of relative costs). Williams[22] notes that,
‘Accountants are prone to the former fallacy, and
medical men the latter, and if the cost-benefit ap-
proach did no more than keep these errors in check
it would have made a valuable contribution to
clearer thinking!’[22]
But, if benefits (or welfare effects) are to be
included, how can they be dealt with within the
CBA framework? This is the main subject of this
paper. It involves consideration of a type of bal-
ance-sheet approach to CBA, examination of the
developments in monetary valuation (in particular,
willingness to pay and valuation of gains in pro-
duction) and a technique called conjoint analysis
(CA), which is new in health economics.
2. Balance-Sheet
(or Opportunity Cost) Approach
Whilst there has been some progress in methods
of monetary valuation of the benefits of healthcare,
there has also been progress in using CBA as a
framework for evaluation. This section examines
the balance-sheet approach, a proposed, practical
prescription for the application of CBA. The fol-
lowing section will discuss the developments in
monetary valuation.
The balance-sheet approach is a form of CBA
that can be used to identify who bears the costs and
who reaps the benefits from any change. Costs and
benefits can be measured in physical units which
seem both natural and appropriate. This approach
adopts the definition of costs and benefits outlined
in section 1, whereby all effects on resource use are
counted on the cost side, and all effects on patients’
well being are counted on the benefit side.[20,21]
Whilst the next stage in a CBA, as defined in health
economics, would require that all costs and bene-
Pharmacoeconomics 1999 Apr; 15 (4)
360 McIntosh et al.

fits be valued in monetary terms, this is often not Table II. Hypothetical example of the balance-sheet approach in
valuing the introduction of an initiative to reduce waiting times
feasible or practical.
The balance-sheet approach, however, advo- Costs (+ve or –ve) Benefits (+ve or –ve)
1 whole time consultant Short waiting time of 1wk
cates that available monetary values can be aug- (compared with 7wk)
mented by other measures of cost and benefit, £2000 annual equivalent Increased well-being due to
namely measures of quantity (e.g. numbers of re- resource cost to develop and avoidance of emergency
update guideline admissions (by 20%)
ferrals) and measures of time (e.g. time spent wait-
20-hour ad hoc preparation, Improved well-being or
ing for a consultation). This further highlights the travel time and administration discomfort (or utility) from
role of CBA as an aid to decision-making rather having to attend one less
than as the sole criterion for those decisions.[23] It appointment
Reduced emergency Improved well-being through
also supports the view of Culyer[24] whom, while admissions by 20% (cost earlier treatment and
recognising the imperfections of CBA in practice, savings) reassurance
also recognised its importance as a framework for One less outpatient Improved patient satisfaction
appointment to attend per
decision-making: ‘... a good CBA will: identify rel- patient (£41 per patient)
evant options for consideration; enumerate all 300 freed-up clinic slots
costs and benefits to various relevant social groups; (resource saving)
quantify as many as can be sensibly quantified; not Patient time and travel cost
savings of 50% (resource
assume the unquantified is unimportant; use dis- saving)
counting where relevant to derive present values; +ve = positive; –ve = negative.
use sensitivity analysis to test the response of net
benefits to changes in assumptions; and look at the
distributive impact of the options.’ idea of their significance ... he will certainly want
Gramlich[25] also supports this approach to to avoid spurious quantification.’ Using this horse
CBA, stating: ‘Benefits and costs should be quan- and rabbit stew analogy, the balance-sheet ap-
tified when they can be and not when they cannot proach can be seen as a means of quantifying the
be, but whether quantified or not they should never ‘horseflesh’ in a realistic and practical manner.
be ignored. Even when they cannot be quantified, In table II, a hypothetical example of the bal-
perhaps because they involve weighty matters of ance-sheet approach is presented. In this example,
life and death, there are ways of setting up the the costs and benefits of introducing an initiative
analysis to focus public decisions properly.’ to reduce waiting time are outlined. Included are
Mishan[14] comments that the outcome of all too costs in terms of staff units and resources. Other
many CBAs follows that of the classic recipe for cost implications in terms of preparation and freed-
making horse and rabbit stew on a strictly 50/50 up clinic slots are reflected in their natural units
basis, 1 horse to 1 rabbit. No matter how carefully because ‘monetary savings’ would not necessarily
the scientific rabbit is chosen, the flavouring of the reflect their true opportunity cost. Benefits are re-
resulting stew is sure to be ‘swamped by the horse- ported in terms of reduced waiting times, well-
flesh’. In Mishan’s analogy, the ‘horseflesh’ repre- being improvements from reduced emergency ad-
sents those other less easily measured spill-over missions and reduced numbers of appointments, as
effects. Mishan notes the strong temptation for well as improved satisfaction. Indeed, the costs and
economists to produce firm quantitative results, benefits of all parties involved such as friends and
and ignore the less easily measured spill-overs. In families could be included in a balance sheet.
stressing that economists should resist this tempta- Busschbach et al.,[26] in their cost-effectiveness
tion, Mishan offers a solution: ‘After measuring all analysis (CEA) outline of drugs for Alzheimer’s
that can be measured with honesty, he can provide disease, show the problems occurring in benefit as-
a physical description of the spillovers and some sessment where not only are there many potential

 Adis International Limited. All rights reserved. Pharmacoeconomics 1999 Apr; 15 (4)
Methods of CBA in Healthcare
costs and effects, but also where they affect the
caregiver as well as the patient. Many of the effects
of Alzheimer’s disease on patients’ and caregivers’
utility are not easily valued using health outcomes
and are often ‘care-related’ process-type outcomes.
Whilst economists have had a preference for more
general measures of quality of life, Busschbach et
al.[26] note that ‘... by generalising the effects of
healthcare programmes, specific information and,
therefore, sensitivity is likely to be lost.’
The study by Busschbach et al.[26] is an example
of the problems associated with using standardised
quality-of-life questionnaires to quantify a multi-
tude of health, nonhealth and process effects on
both patients and carers. Furthermore, such quan-
tification and, in a bid to adhere to convention,
collapsing of such effects into a cost-effectiveness
ratio may be done without consideration of the loss
of useful information accompanying this act.
Busschbach et al.[26] note that ‘... the scores of
these questionnaires are not appropriate to formu-
late a cost-effectiveness ratio, which is necessary
for an economic evaluation.’ This is an important
observation, but it seems to miss the point slightly.
An economic evaluation can constitute a list of
costs and benefits of alternative options and still
provide useful information, arguably more useful
than a cost-effectiveness ratio which has not cap-
tured the majority of the effects occurring. The
study by Busschbach et al.[26] is an example of
where a description of the various effects on all the
affected parties in a balance-sheet format may pro-
vide a more realistic and informative alternative to
trying to estimate a cost-effectiveness ratio which,
in fact, says nothing about the true effects or their
diversity. A similar approach has been advocated
by Mauskopf et al.[27] to evaluate drug treatments.
Whilst the balance-sheet approach can be seen
as a type of CBA in its own right, it can also be seen
as the first stage in a CBA, i.e. as a means of out-
lining the benefits before monetary valuation.
Whichever it is used for, it can be seen as a useful
decision-making framework. Section 3 examines
recent developments in the monetary valuation of
benefits.
 Adis International Limited. All rights reserved.
361
3. Recent Developments in
Monetary Valuation: Willingness
to Pay, Conjoint Analysis and
Valuation of Production Gains
3.1 Willingness to Pay
The technique of willingness to pay is based on
the premise that the maximum amount of money
an individual is willing to pay (sacrifice) for a com-
modity is an indicator of the utility or satisfaction
to them of that commodity. Furthermore, it can be
argued that when an individual is considering their
maximum willingness to pay, they will take ac-
count of all the attributes of the service of impor-
tance to them, not just health gains.
Despite the problems with monetary valuation,
there have been advances in using willingness to
pay as a measure of benefit.[8] However, it is worth
noting that most of the published healthcare will-
ingness-to-pay (WTP) studies are experimental in
nature, attempting to ‘... explore measurement fea-
sibility issues rather than being full programme
evaluations using CBA.’[8] This is apparent, for ex-
ample in the number of studies in the health eco-
nomics literature in which various methods of eli-
citing WTP values have been compared.[28,29]
Other research has been carried out to examine
whether WTP values conform to a priori (theore-
tical) expectations such as being sensitive to scale
effects.[30] Willingness to pay has been used to ad-
dress allocative issues such as the setting of priori-
ties within fixed public sector budgets.[31]
The technique of willingness to pay is often
criticised for attempting to assign a monetary value
to things which are considered by many to be in-
commensurate with monetary valuation, e.g. the
relief of suffering or the saving of a human life.
Another criticism of willingness to pay is that it is
inevitably a function of ability to pay, which it is
argued, could have implications for equity. Ability
to pay undoubtedly affects absolute willingness to
pay, but there is some evidence to suggest that it
may not affect relative willingness to pay.[32] There
is now a small amount of literature on to how to
test for this.[31,32] Donaldson et al.[32] used willing-
Pharmacoeconomics 1999 Apr; 15 (4)
362
ness to pay to elicit community values for child
health services. By asking questions about prefer-
ences, they were able to examine how willingness
to pay for options could be distorted by respon-
dents’ incomes. In this case, both preferences and
willingness to pay appear to follow a stable pattern
across both social class and income groups, i.e.
preferences are not affected by income levels.
Olsen and Donaldson[31] point out that similar
analyses are possible in studies in which WTP val-
ues for multiple broad policy options are elicited
from individuals. Hence, it is possible, by various
methods of disaggregation of data to, at the very
least, show how ability to pay might be taken ac-
count of when using the results of WTP studies.
This approach is useful when distributional issues
are shown not to have an effect. However, what
happens in situations where distributional effects
are shown to matter? In this situation, weights can
be attached to the valuations obtained by different
income groups.[33] However, there are several pro-
blems associated with using such weights.[34,35]
One way around this is to use sensitivity tests to
show the effects of using different weights and
leave it to the decision-maker to establish how im-
portant the distributional issue is.[36]
In response to concern that the growing number
of WTP studies in healthcare were not consistent
with the principles of CBA, O’Brien and Gafni[37]
provide a useful framework for WTP studies in this
area of application. Comprehensive guides on the
use of contingent valuation surveys to value public
goods have been developed by Mitchell and Car-
son[38] and Arrow et al.[39]
Although there are many criticisms of willing-
ness to pay, it has to be remembered that such val-
uations are being made anyway, arising from the
‘implicit’ judgements of decision-makers. Wil-
liams[22] states that it is important to recognise that
benefit measurement is ‘... necessary and valuable
whether or not we go the final step of attaching
explicit monetary values to health states.’ As we
have seen in this section, all that many health econ-
omists are doing is simply examining whether such
valuation is feasible, valid and reliable.
 Adis International Limited. All rights reserved.
McIntosh et al.
3.2 Conjoint Analysis
CA is a relatively new technique in the eco-
nomic evaluation of healthcare that has also been
used to elicit WTP values indirectly.[40] CA is a
technique used for establishing the relative impor-
tance of attributes in the provision of a good or
service. By including different amounts of money
as an attribute (known as the cost attribute) in a CA
study design, estimates of willingness to pay for
changes in the levels of the attributes of importance
can easily be derived using regression techniques.
These estimates are based on an interpretation of
the coefficient on the cost attribute as being equal
to the marginal utility of income.
Ryan[41] highlights a number of potential advan-
tages of using CA over the more direct WTP ap-
proach. Such advantages include the potential to
overcome scope effects, the potential ability to ad-
dress the closed-ended WTP ‘yes-saying’ problem
and, because of its less direct approach, its poten-
tial to appear less politically objectionable. These
are now addressed in turn.
As mentioned in section 3.1, there is evidence
that the direct WTP approach is insensitive to the
scale or the scope of the programme being evalu-
ated. For example, Boyle et al.[42] found that peo-
ple’s willingness to pay to take action to prevent
the death of 2000 migratory birds in an oil spill
is the same as their willingness to pay to prevent
20 000 and 200 000 such deaths. Ryan[41] suggests
that CA may overcome this scale effect as it may
be more specific in its approach to eliciting the
value of particular aspects of a service. Arrow et
al.[39] have recommended the closed-ended WTP
approach be used when directly estimating willing-
ness to pay, arguing that it more accurately reflects
the types of decisions individuals make every day.
However, there is evidence from the environmental
economics literature that, using the close-ended ap-
proach, individuals tend to say ‘yes’ to amounts
above their maximum willingness to pay, resulting
in an overestimation of willingness to pay.[43,44]
Since CA asks individuals to choose between alter-
native goods with different levels of cost, it may
Pharmacoeconomics 1999 Apr; 15 (4)
Methods of CBA in Healthcare
overcome the ‘yes-saying’ problem whilst at the
same time maintaining one of the claimed advan-
tages of the closed-ended approach, namely that
the questions asked resemble the type of decisions
individuals make on a daily basis.
Using willingness to pay in the area of
healthcare is politically sensitive and often results
in a number of protest bids, i.e. respondents not
answering the question because they have an ob-
jection to being asked their willingness to pay for
healthcare. It is possible that estimating willing-
ness to pay within a CA framework may result in
fewer protest answers as individuals see directly
what they are getting.
Proving these potential advantages, however, is
ultimately an empirical task. A number of studies
have already used CA to estimate willingness to
pay indirectly, all of which could potentially be
used within the CBA framework.[40,45-49] These
studies have also shown evidence of the theoretical
and convergent validity of CA-derived WTP esti-
mates. Theoretical validity assesses the extent to
which the results of a study are consistent with
economic theory, or sometimes, more generally, a
priori expectations. Ryan and Hughes[40] showed
the theoretical validity of the CA-derived WTP es-
timates. As would be expected in theory, women
on higher incomes valued a reduction in time spent
in hospital receiving treatment more highly than
those on lower incomes.
Convergent validity, which is a type of construct
validity, measures the extent to which the results
are consistent with other measures that are held to
measure the same construct. In a study looking at
the convergent validity of willingness to pay and
CA, willingness to pay for assisted reproductive
techniques was estimated to be £4690 from the
closed-ended WTP technique and £4760 from the
discrete choice CA technique.[50] The similarity of
these results suggests a high level of convergent
validity between the techniques, although further
empirical work is required in this area. Further
methodological issues in the application of CA to
health economics need to be addressed before it
 Adis International Limited. All rights reserved.
363
becomes an established tool for the monetary
measurement of benefits.[51]
Finally, in addition to using CA as a method for
estimating willingness to pay indirectly, CA is a
potential method for valuing those attributes of im-
portance to patients in the provision of goods or
services that have been identified and measured
using the balance-sheet approach. For example, in
table II, the attributes of importance could possibly
be: cost; waiting time; probability of emergency
admission; number of outpatient appointments; re-
assurance; and satisfaction.
CA can also be used to establish a total utility
score for a good or service by summing the indi-
vidual utility gained from each of the various attri-
butes or characteristics that make up the provision
of a good or service.[40,52] This may help to allevi-
ate the obvious disadvantage of the balance-sheet
approach; that it, quite deliberately, does not in-
volve a summary valuation of benefits and, there-
fore, leaves it up to the decision-maker to compare
programmes with disparate outcomes. However,
given the recent advances in the monetary valua-
tion of benefits, and the obvious decision-making
appeal of such ‘monetarisation’, it may be the case
that the practical limitations which historically re-
stricted such monetary valuation of benefits are no
longer such an issue in CBA.
3.3 The Debate on Production Gains
(or Indirect Costs)
There has been recent debate in the literature
regarding the inclusion of indirect costs or produc-
tivity gains in economic evaluation. Traditionally,
the human-capital method has been used to incor-
porate these production gains into economic eval-
uations. The human-capital method places mone-
tary weights on gains in healthy time resulting
from healthcare programmes. These weights are
based on market wage rates (discounted back to
present value). However, the reliance of this tech-
nique on earnings data has the potential to bias
resource allocation towards diseases that affect
Caucasian, middle-class men.
Pharmacoeconomics 1999 Apr; 15 (4)
364
Furthermore, using the economic principle of
opportunity cost, it can be questioned whether the
magnitude of the indirect costs of illness is often
overestimated.[53] It cannot be denied that prema-
ture retirement and loss of life are tragic. However,
in many cases, avoidance of such retirement and
loss of life may result in little by way of production
gains for society if those who retire or die can be
replaced from the pool of unemployed labour,
which exists in most countries. There will of course
be a gap between the loss of people from the work-
force and their subsequent replacement. However,
to deal with this, it has been suggested that the hu-
man-capital element of economic evaluations
should include only this frictional element, i.e. pro-
duction loss avoided which would have occurred
only during the gap between loss of a person and
their replacement and not over a person’s remain-
ing worklife.[54]
However, many subsequent economic evalua-
tions have not used this methodological develop-
ment and this may reflect the argument by Johan-
nesson and Karlsson[55] and Liljas[56] that the
friction-cost method is based on implausible as-
sumptions not supported by neoclassical economic
theory or empirical observations, and that the hu-
man-capital approach does in fact correctly esti-
mate indirect costs. The argument is that firms
hire people up to the point at which the marginal
cost of labour equals the marginal value of its prod-
uct. Thus, absenteeism, representing a marginal
loss of labour, is accurately valued using the hu-
man-capital approach, as this approach will reflect
the marginal value.
A similar argument is applied to the valuation of
internal labour reserves which, the friction-cost
proponents would claim, are used to offset absen-
teeism. Johannesson and Karlsson[55] go on to state
that the inclusion of the indirect costs of mortality
in an economic evaluation, as argued by the fric-
tion-cost method, has no foundation in economic
theory. They also discuss how the friction-cost
method assumes that sickness and absence from
work will reduce the overall unemployment rate in
the economy in the long term. However, there ap-
 Adis International Limited. All rights reserved.
McIntosh et al.
pears to be no empirical evidence which shows that
sickness causing absence from work actually re-
duces the unemployment rate in the long term.
In recently published guidelines, the Washing-
ton Panel[57] also provides recommendations on
how to incorporate productivity costs into CEA us-
ing a quality-adjusted life-year (QALY) approach.
This approach recommends incorporating produc-
tivity costs as health effects in the denominator of
the cost-effectiveness ratio. The Panel states that a
choice between inclusion of the productivity costs
in the numerator or the denominator should be
made, because to count both quality-of-life (QOL)
effects of decreased productivity and all costs nor-
mally measured with the friction-cost or the hu-
man-capital method would lead to double count-
ing. This stems from the belief that the loss of
income is assumed to be incorporated by the res-
pondents in a QOL measurement.
Liljas[56] has rejected the friction-cost method,
and along with Brouwer et al.[58] has refuted the
recommendations of the Washington Panel’s QALY
approach[57] on theoretical grounds. The main ob-
jection by Brouwer et al.[58] is that the effect com-
ponent of the cost-effectiveness ratio should be for
intrinsic health effects only and income is not be-
lieved to be part of health-related quality of life.
Liljas[56] describes the components that should
be included as indirect costs to be consistent with
economic theory, namely reduced paid and unpaid
production due to the individual’s disease and in-
direct costs accrued to the family and/or friends for
taking care of the individual. He states that the fric-
tion-cost approach excludes many of these compo-
nents and rests on very strong, unrealistic assump-
tions about the individual’s valuation of leisure and
the labour market. Liljas[56] also states that the
QALY approach is unrealistic in that all indirect
costs cannot be assumed to be included in the indi-
viduals reported utility weight for a health state.
Neither the QALY approach nor the friction-cost
approach, according to Liljas,[56] are consistent
with economic theory and cannot be recommended
over the traditional human-capital approach.
Pharmacoeconomics 1999 Apr; 15 (4)
Methods of CBA in Healthcare
Koopmanschap et al.[59] defend their friction-
cost approach on practical grounds by admitting
that although the friction-cost method is not yet
perfected, it is a sound concept and preferable to
an instrument based on the ‘totally unrealistic as-
sumptions’ of neoclassical theory. These are im-
portant points which need to be carefully consi-
dered but as yet remain unresolved. Williams[22]
addressed this in 1974, although not to the same
extent. He noted, ‘... if some re-arrangement of
work is envisaged that releases labour which is oth-
erwise unemployable, then the gain to the commu-
nity will be minimal and will certainly be over-
stated by the savings in expenditure on wages.’ At
this point in time, it is unclear where the debate on
indirect costs will lead.
4. Conclusions
This paper asserts that almost all types of eco-
nomic evaluation have an element of the cost-ben-
efit approach in them. The important issue is to
focus on the objectives and underlying theoretical
rationale of studies. By focusing in such a manner,
this will lead to more relevant, realistic economic
evaluations. It will also prevent current practice
from not only being perceived as of little use but
also leading to what some have called the ‘undoing
of economic evaluation’.[5] As Culyer[24] stated, we
must ‘... not assume the unquantified is unimport-
ant.’ Johansson[60] has also said that, ‘... a social
cost benefit analysis ... can capture and express in
a single monetary dimension (monetary units)
many, but not all, of the effects caused by public
sector programmes. They enlarge as well as im-
prove the information set available to decision
makers.’
Using these broader definitions, the majority of
economic evaluations can be termed CBAs. The
important point is not about moulding a question
to fit the production of a ratio that gives the appear-
ance of leading to ‘a decision’. Rather, it is about
relating the study design to the policy question and
outlining the relevant costs and benefits in a man-
ner that assists the evaluation of a welfare change.
 Adis International Limited. All rights reserved.
365
There have been substantial improvements in
the methods of CBA in recent years, especially in
the use of willingness to pay and CA to estimate
the monetary value of benefits. The debate on val-
uation of production gains is, as yet, unresolved,
but has involved an important re-examination of
this issue. We have also attempted to show that the
balance-sheet approach to CBA, which may not
involve monetary valuation of benefits or costs,
provides a useful decision framework. In 1974,
Williams[22] outlined the basis of the cost-benefit
approach in health and healthcare. The challenge
is to ensure that, in the future, the art of CBA re-
flects the science that Williams[22] outlined. The
recent developments outlined in this paper may
help this to happen.
Acknowledgements
The Health Economics Research Unit is funded by the
Scottish Office Department of Health. Cam Donaldson holds
the Svare Chair in Health Economics and is Alberta Heritage
Senior Scholar at the University of Calgary. Mandy Ryan is
an MRC Senior Research Fellow. The views expressed in
this paper are those of the authors and not the funding bodies.
References
1. Birch S, Gafni A. Cost effectiveness/utility analyses: do current
decision rules lead us to where we want to be? J Health Econ
1992; 11: 279-96
2. Birch S, Gafni A. Cost-effectiveness ratios: in a league of their
own. Health Policy 1994; 28: 133-41
3. Gerard K, Mooney GH. QALY league tables: handle with care.
Health Econ 1993; 2 (1): 59-64
4. Donaldson C. The (near) equivalence of cost-effectiveness and
cost-benefit analyses: fact or fallacy? Pharmacoeconomics
1998; 13 (4): 389-96
5. Johannesson M, Jonsson B. Economic evaluation in health care:
is there a role for cost-benefit analysis? Health Policy 1991;
17: 1-23
6. Robinson R. Cost-benefit analysis. BMJ 1993; 307: 924-6
7. Hutton J. Cost-benefit analysis in health care expenditure deci-
sion making [editorial]. Health Econ 1992; 1: 213-6
8. Diener A, O’Brien B, Gafni A. Health care contingent valuation
studies: a review and classification of the literature. Health
Econ 1998; 7: 313-26
9. Lichfield N. Economics in town planning. Town Planning Rev
1968; (38): 5-20
10. Michel BC, Seerden RJ, Rutten FFH, et al. The cost effective-
ness of diagnostic strategies in patients with suspected pul-
monary embolism. Health Econ 1996; 5: 307-18
11. Kuntz KM, Tsevat J, Goldman L, et al. Cost-effectiveness of
routine coronary angiography after acute myocardial infarc-
tion. Circulation 1996; 94: 957-65
Pharmacoeconomics 1999 Apr; 15 (4)
366
12. Johannesson M, Jonsson B, Kjekshus J, et al. Cost effectiveness
of simvastatin treatment to lower cholesterol levels in patients
with coronary heart disease. N Engl J Med 1997; 336: 332-6
13. Mark DB, Hlatky MA, Califf RM, et al. Cost effectiveness of
thrombolytic therapy with tissue plasminogen activator as
compared with streptokinase for acute myocardial infarction.
N Engl J Med 1995; 332: 1418-24
14. Mishan EJ. Cost benefit analysis. London: Unwin Hyman Ltd,
1971
15. Dreze J, Stern N. The theory of cost-benefit analysis. In: As-
uerbach AJ, Feldstein M, editors. Handbook of public eco-
nomics. Vol. II. Amsterdam: Elsevier Science Publishers, 1987
16. Zarnke KB, Levine MAH, O’Brien BJ. Cost-benefit analyses in
the health care literature: don’t judge a study by its label. J
Clin Epidemiol 1997; 50 (7) 813-22
17. Jackson LA, Schuchat A, Gorsky RD, et al. Should college stu-
dents be vaccinated against meningococcal disease? A cost-
benefit analysis. Am J Pub Health 1995; 85: 843-5
18. Cookson ST, Stamboulian D, Demonte J, et al. A cost-benefit
analysis of programmatic use of CVD 103-HgR live oral chol-
era vaccine in a high risk population. Int J Epidemiol 1997;
26 (1): 212-9
19. Legg RF, Sclar DA, Nemec NL, et al. Cost benefit of sumatripan
to an employer. J Occup Environ Med 1997; 39 (7): 652-7
20. Birch S, Donaldson C. Applications of cost-benefit analysis to
health care: departures from welfare economic theory. J
Health Econ 1987; 6: 211-25
21. Donaldson C, Shackley P. Economic evaluation. In: Detels R,
Holland WW, Mcewen J, et al., editors. Oxford textbook of
public health. Vol. 2: the methods of public health. 3rd ed.
Oxford: Oxford University Press, 1997: 849-71
22. Williams A. The cost-benefit approach. Br Med Bull 1974; 30
(3): 252-6
23. Sugden R, Williams A. The principles of practical cost-benefit
analysis. Oxford: Oxford University Press, 1978
24. Culyer AJ. Economics. Oxford: Blackwell, 1985
25. Gramlich EM. A guide to benefit-cost analysis. 2nd ed. Engle-
wood Cliffs (NJ): Prentice Hall, 1997
26. Busschbach JJV, Brouwer WBF, van der Donk, et al. An outline
for a cost-effectiveness analysis of a drug for patients with
Alzheimer’s disease. Pharmacoeconomics 1998; 13: 21-34
27. Mauskopf JA, Paul JE, Grant DM, et al. The role of cost-con-
sequence analysis in healthcare decision-making. Pharmaco-
economics 1998; 13 (3): 277-88
28. Donaldson C, Thomas R, Torgerson DJ. Validity of open-ended
and payment scale approaches to eliciting willingness to pay.
Appl Econ 1997; 29: 79-84
29. Johannesson M, Jonsson B, Borgquist L. Willingness to pay for
antihypertensive therapy: results of a Swedish pilot study. J
Health Econ 1991; 10: 461-74
30. Neumann PJ, Johannesson M. The willingness to pay for in vitro
fertilisation: a pilot study using contingent valuation. Med
Care 1994; 32 (7): 686-99
31. Olsen JA, Donaldson C. Helicopters, hearts and hips: using
willingness to pay to set principles for public sector health
care programmes. Soc Sci Med 1998; 46 (1): 1-12
32. Donaldson C, Farrar S, Mapp T, et al. Assessing community
values in health care: is the willingness to pay method feasi-
ble? Health Care Anal 1997; 5 (1): 7-29
33. Phillips PR, Russell IT, Jones-Lee MW. The empirical estima-
tion of individual valuation of safety: results of a national
sample survey. In: Jones-Lee MW, editor. The economics of
safety and physical risk. Oxford: Basil Blackwell, 1989
 Adis International Limited. All rights reserved.
McIntosh et al.
34. Little IMD. A critique of welfare economics. London: Clar-
endon Press, 1957
35. Mishan EJ. Cost benefit analysis. 4th ed. London: Unwin Hy-
man, 1988
36. Boardman AE, Greenberg DH, Vinning AR, et al. Cost-ben-
efit analysis: concepts and practice. New Jersey: Prentice
Hall, 1996
37. O’Brien B, Gafni A. When do the ‘dollars’ make sense? Toward
a conceptual framework for contingent valuation studies in
health care. Med Decis Making 1996; 16: 288-99
38. Mitchell RC, Carson RT. Using surveys to value public goods:
the contingent valuation method. Washington, DC: Johns
Hopkins University Press, 1990
39. Arrow K, Solow R, Portney P, et al. National Oceanic and At-
mospheric Administration (NOAA): report of the NOAA
panel on contingent valuation. Fed Regist 1993; 58: 4602
40. Ryan M, Hughes J. Using conjoint analysis to assess women’s
preferences for miscarriage management. Health Econ 1997;
6 (3): 261-73
41. Ryan M. Conjoint: what’s the point? A Health Economists Study
Group meeting. The National University of Ireland; 1933 Jul
13-15; Galway
42. Boyle KJ, Desvouges WH, Johnson FR, et al. An investigation
of part-whole biases in contingent valuation studies. J Environ
Econ Manage 1994; 27: 64-83
43. Ready R, Buzby J, Hu D. Differences between continuous and
discrete contingent valuation estimates. Land Econ 1996; 72:
397-411
44. Brown T, Champ P, Bishop R, et al. Response formats and pub-
lic good donations. Land Econ 1996; 72: 152-66
45. Bryan S, Buxton M, Sheldon R, et al. The use of magnetic
resonance imaging (MRI) for the investigation of knee inju-
ries: a discrete choice conjoint analysis exercise. Health Econ-
omists Study Group Meeting. University of York; 1997 Jul
2-4; York
46. Propper C. Contingent valuation of time spent on NHS waiting
list. Econ J 1991; 100: 193-9
47. Propper C. The disutility of time spent on the United Kingdom’s
National Health Service waiting lists. J Hum Resources 1995;
30: 677-700
48. Ryan M. Assessing the benefits of health interventions: a role
for conjoint analysis? A Labelle Lectureship in Health Service
Research. McMaster University; 1997 Oct 16; Hamilton
49. San Miguel F, Ryan M, McIntosh E. Establishing women’s pref-
erences for the treatment of menorrhagia using the technique
of conjoint analysis. Aberdeen: University of Aberdeen, 1997.
Health Economics Research Unit discussion paper no.: 06/97
50. Ryan M. Establishing the convergent validity of willingness to
pay and conjoint analysis for eliciting preferences. Health
Economists Study Group Meeting. University of Brunel;
1996 Jul 3-5; Brunel
51. Ryan M, Scott A, Farrar S, et al. Using conjoint analysis in
health care: unresolved methodological issues. Aberdeen: Uni-
versity of Aberdeen, 1996. HERU discussion paper no.: 02/96
52. Ryan M. Using consumer preferences in health care decision
making: the application of conjoint analysis. London: Office
of Health Economics, 1996
53. Drummond MF, O’Brien BJ, Stoddart GL, et al. Methods for
the economic evaluation of health care. 2nd ed. Oxford: Ox-
ford University Press, 1997
54. Koopmanschap MA, Rutten FFH, van Ineveld, et al. The fric-
tion cost method for measuring indirect costs of disease. J
Health Econ 1995; 14: 171-89
Pharmacoeconomics 1999 Apr; 15 (4)
Methods of CBA in Healthcare 367

55. Johannesson M, Karlsson G. The friction cost method: a com- 60. Johansson PO. An introduction to modern welfare economics.
ment. J Health Econ 1997; 16: 249-55 Cambridge: Cambridge University Press, 1991
56. Liljas B. How to calculate indirect costs in economic evalua-
tions. Pharmacoeconomcis 1998; 13: 1-7
57. Gold MR, Siegel JE, Russell LB, et al. Cost-effectiveness in
health and medicine. Oxford: Oxford University Press, 1996
58. Brouwer WBF, Koopmanschap MA, Rutten FFH. Productivity
costs measurement through quality of life? A response to the Correspondence and reprints: Ms Emma McIntosh, Health
recommendation of the Washington Panel. Health Econ 1997;
Economics Research Unit, Department of Public Health,
6: 253-9
59. Koopmanschap MA, Rutten FFH, van Ineveld BM, et al. Reply University Medical Buildings, Foresterhill, Aberdeen AB25
to Johannesson’s and Karlsson’s comment. J Health Econ 2ZD, Scotland.
1997; 257-9 Email: e.mcintosh@aberdeen.ac.uk

 Adis International Limited. All rights reserved. Pharmacoeconomics 1999 Apr; 15 (4)