Full text of DOI:10.1017/S0266462300103058

International Journal of Technology Assessment in Health Care, 16:4 (2000), 976–986.
c
Copyright 2000 Cambridge University Press. Printed in the U.S.A.
WHERE ARE THE ECONOMIC
GUIDELINES COMING FROM?
Alistair McGuire
Stephen Morris
Maria Raikou
City University, London
Abstract
Economic guidelines recommend methods that should be employed in conducting economic evaluations
of healthcare programs. The nature of the efficiency or equity goal underpinning economic guidelines
is unclear. What is also unclear is how the methods recommended in the guidelines are linked to the
underlying efficiency or equity goal being targeted. If it is unclear what efficiency/equity objectives are
being pursued, then it is unlikely that even full implementation of economic guidelines will improve
resource allocation.
Keywords: Guidelines, Efficiency, Cost
The notion of economic guidelines is relatively new. It stems from a general awareness that
the rationing of health care is unavoidable. This is the case for most goods and services
and merely reflects that demand normally exceeds supply. Rationing in health care is more
complex because the normal mechanism that allocates commodities, the price system, fails
to operate with its expected elegance in this area. With most goods and services, if the
demand exceeds supply, the price rises. This rise then reduces demand as individuals who
value the commodity in question less than the now higher price leave the market. Only those
who are willing to pay at least the increased market price are left. The market equilibrium
of demand matching supply is restored as prices ration on the basis of willingness to pay,
where this reflects the value individuals place on goods and services. While some healthcare
demand does respond to price signals, the impact is limited (11).
One reason that the impact is limited is the role of third-party payment. Healthcare
insurance, both public and private, has increased dramatically. This increase subsequently
introduces a more aggregate payer viewpoint. The payer is considering individual treatments
from a population perspective. The payer becomes an agent for covered individuals in terms
of purchasing decisions on both the financing and the valuing levels. Again, this situation
emphasizes the need for priority setting. It leads to two other tendencies: a) toward a
requirement to specify what is and what is not to be covered; and b) toward a comparison
of interventions in a commensurate manner.
A second reason health care fails to respond to demand signals is the existence of a
peculiar relationship between the supplier of health care (the physician) and the consumer
(the patient). As health care increases in complexity, consumers become ever more reliant
976

Where are the economic guidelines coming from?
on physicians to specify their demands. The sector becomes dominated by the suppliers of
health care. The main action is between the funders and the suppliers, with the consumer a
necessary but fringe player.
To understand the impact of these structural issues on the production of health care, it is
useful first to consider the funding of health care in more detail. The rationale for economic
guidelines is stated to stem from the peculiar relationships among funders, physicians,
and consumers that characterize this market. It is argued that economic guidelines are
a manifestation of a desire to control costs, which is not necessarily consistent with the
pursuit of economic efficiency. The guidelines’ role is limited, however, not only because
they deal with a relatively small proportion of total expenditure, but also because they deal
predominantly with methodology rather than practice. Moreover, even this concentration
on methodology is confused because it is not clear how any of the proposed guidelines
actually relate to economic efficiency, since in most cases the definition of efficiency is
either missing or left to the imagination. Without well-identified objectives and definitions,
any system of guidelines is open either to abuse or to capture by the very parties they are
intended to regulate.
RELATIONSHIPS, BEHAVIOR, AND GUIDANCE
In a conventional market the three functions of the consumer’s “cost-benefit calculus”—
decision making, benefit receipt, and cost bearing—are performed by the individuals them-
selves. In the healthcare sector the decision making is largely delegated to the physician.
The benefit receipt is also largely influenced by the information gained from the physician
through specification of treatment and prognosis. Much of the cost-bearing at the point of
consumption rests with the third-party payer. The structure of the healthcare sector is conse-
quently much more complex than most. The behaviors of the individuals within this sector
are defined by their various interactions. Central to the whole process are the predominance
of the third-party payer and the relationship formed by the healthcare supplier (the physi-
cian) and the consumer (the patient). These central elements are integrally interrelated, as
discussed below.
From an individual perspective, insurance contracts are bought as a means of reducing
the risk faced. In the healthcare sector two risk conditions are faced by the contractors: the
state of health, and the expenditure incurred. If the insurance contract is specified on the
basis of the state of health (i.e., disease-contingent), then a pre-specified amount is paid
if a particular disease is experienced. Health, however, is difficult to define and measure,
and insurance contracts at the individual level are therefore difficult to specify in terms of
precise medical treatment for different medical conditions. As a result of this difficulty,
most private healthcare insurers reimburse healthcare expenditure by having first specified
what is included and what is not included in the coverage. Even here difficulties present
themselves, because the relationship between health care and health is not straightforward.
Given that it is the providers who determine the health care to be matched to the illness, the
treatment coverage must be specified. Moreover, the insurance contract is normally widened
to include the provider, which allows incentive structures to be defined between the insurer
and the provider.
These incentive structures are largely built on disease-contingent contracts in which re-
imbursement is based on either disease-defined payments (e.g., through diagnostics-related
groups [DRG]) or on merging the insurance and provision of health care (e.g., through
health maintenance organizations). Merged insurance contracts reimburse in kind. This
arrangement moves the private funding of health care away from any actuarially based
insurance system toward a reimbursement system that is based on service delivery. The
specification of the service delivery can be undertaken on a retrospective basis, giving rise
INTL. J. OF TECHNOLOGY ASSESSMENT IN HEALTH CARE 16:4, 2000
977

McGuire, Morris, and Raikou
to low-powered incentive structures because the provider is reimbursed for care delivered,
or on a prospective basis that induces greater movement toward cost containment.
It has also been argued that, even in systems characterized as privately financed systems,
a normative proposition has arisen that health care ought to be widely available to all sections
of the population (16). Private financing of health care is therefore either heavily subsidized
or supplemented with governmental resources. In the United States government funding
of health care now runs at more than 55% of total healthcare expenditures (12). Public
funding of health care explicitly incorporates redistribution into the funding objectives. This
large-scale intervention by governments leads financing further away from an actuarially
based insurance concept toward alternative means of specifying both the coverage and the
distribution of healthcare resources.
Saying that all healthcare systems focus on populations rather than on individual health-
care requirements is not, in itself, revealing much. It does, however, highlight why there has
been considerable interest in aggregate methods of coverage and in their ability to provide
incentives in keeping with cost containment. Economic guidelines are one manifestation of
this aggregate control of costs.
COST CONTAINMENT
Cost-containment policies can operate on the demand or the supply side of the market.
On the demand side of the market such policies operate by shifting financial penalties
to the consumer/patient. They operate through four main mechanisms: exclusion policies,
coinsurance, copayment, and deductibles. The first mechanism is sometimes justified on
efficiency grounds as it relates to adverse selection. Essentially, given the asymmetry of
information between consumer and supplier in the health insurance market, the insurer can
find it difficult to distinguish between high- and low-risk individuals. This difficulty is not
a problem unless the high-risk individuals distort the market through nonrevelation of risk,
which allows them topurchase insurance at lessthanthe actuariallydeterminedrate. Insurers
would then experience losses. In an attempt to isolate such losses, high-risk individuals may
be excluded. Since the difficulty in identifying high-risk individuals remains, insurers may
exclude individuals who have characteristics that correlate with high risk of ill health. Being
elderly and having a low income, given their high correlation with ill health, become criteria
for exclusion.
While exclusion criteria are common in private insurance, exclusion is also used in
the public sector and is increasingly related to coverage specified in terms of treatments.
Implementation of explicit criteria for exclusion in Oregon and specification of acceptable
treatment interventions by the Dutch government are two examples. Exclusion criteria
comprise the most important demand-side cost-containment policy and may or may not be
consistent with efficiency. One way to consider the exclusion of particular treatments is to
assess whether the benefits from these treatments exceed the costs. If they do, the treatment
is said to be efficient, in a particular sense to be discussed below, and it should be allocated
funding. This notion of efficiency, as it relates to coverage, is examined in greater detail
below, since this form of rationing appears to be the main rationale for the subsequent focus
on guidelines.
After who or what is to be covered is determined, the three other demand-side cost-
containment policies are invoked. They operate by restricting the financial liability of the
insurer. By implication they assume that the individual consumer/patient has some control
over the purchase decision. Coinsurance shares costs between the insurer and the individual
through payment of a set proportion of the cost of each service; copayment relates to a
fixed amount paid by both parties of the insurance contract. A deductible is a fixed amount
to be paid by the insured individual before any insurance benefit is paid. All represent
978
INTL. J. OF TECHNOLOGY ASSESSMENT IN HEALTH CARE 16:4, 2000

Where are the economic guidelines coming from?
significant amendments to the price of insurance coverage, and empirical evidence confirms
that healthcare demands are affected by these amendments. Such amendments are deemed
necessary, since in a world of full insurance coverage individuals would have no incentive
to restrict treatment. Moral hazard, where there is overtreatment accompanied by increased
costs, would occur.
The financial incentives imposed on consumers are limited, however, because of con-
sumers’ ignorance of health states, of the availability of treatments, and of likely treatment
effectiveness. As a result, consumers have to rely on the supplier, normally the physician, to
assist them in determining their appropriate demands. The potential monopoly exploitation
that could result is ameliorated through an agency relationship that is formed between the
physician and the patient, in which the producer (physician) determines the outcome for
the consumer (patient).
To operate this agency relationship, the medical profession has recognized that certain
expectations are imposed on their behavior. The general response is self-regulation of entry
to the profession and imposition of standards of behavior after entry. The first response has
two implications: a) it serves the public interest by ensuring minimum standards of quality
through certification; and b) it reinforces the monopoly position of the physician. Conse-
quently, emphasis is placed on the importance of ethical behavior. The primary importance
of ethical behavior is that it replaces demand as a constraint on supply. It is an internalized
check on suppliers.
Normally consumers can acquire information at low cost on product quality and will
also have an indication of the expected satisfaction (utility) to be gained from consumption.
Such information formulates individual demands, and supply responds via price signals to
these demands. With health care the individual is unable to formulate demands on both
counts. Moreover, not only are there prohibitive costs in gaining such information, but
also consumers may wish to actively default on the decision-making process completely.
Given their circumstances they may want the physician to decide. For all these reasons
the consumer willingly engages an agent. Medically ethical conduct, as regulated by the
profession, serves to sustain the agency relationship, providing reassurance that the supplier
(physician) is acting in the consumer’s (patient’s) best interest.
Supply-side cost-containment becomes a necessary and increasingly important adjunct
to demand-side cost-sharing. There are complications, however. Professional conduct in
this sector is defined with regards to medical rather than economic objectives. Indeed,
professional conduct serves to reassure the consumer (patient) that there is a separation of
medical from economic considerations. This separation is one characteristic that underlies
the agency relationship. It is enhanced by the peculiar pricing practices that have been
pursued by the medical profession. Price discrimination by income is common even in
private healthcare systems, operating through cross-subsidy and bounded in the extreme by
charity provision.
An indirect consequence of the arms-length reimbursement by a third-party payer is
support of this separation of medical and economic objectives. Moral hazard, realized as
overconsumption, may stem from physician behavior, however. To control this hazard,
financial incentives are aimed at the suppliers of health care. These incentives range from
reimbursement incentives to proscriptive measures that again may limit the coverage of
treatments (i.e., implementation of managed care). The payers of health care increasingly
use their power in the sector to aim cost-containment policies at the supplier of health
care. To the extent that such cost-containment measures undermine the agency relationship,
physicians may collectively respond.
One such response is becoming widespread across different areas of health care. This
spread stems partly from implementation of changes in the reimbursement of healthcare
providers, including physicians. As noted above, incentive mechanisms are increasingly
INTL. J. OF TECHNOLOGY ASSESSMENT IN HEALTH CARE 16:4, 2000
979

McGuire, Morris, and Raikou
related to disease-specific reimbursement (for example, DRG-based systems). Such reim-
bursement attaches funding to expenditure related to some predefined event. In the case of
the DRG system, the event is a defined set of days in hospital. The reimbursement system
operates through implementation of a benchmark or yardstick, where the average level of
healthcare provision is taken to be the criterion against which other levels of provision are
judged. The definition of the average or the benchmark becomes critical.
Clinical guidelines, aided by increasing empirical evidence on the impact of different
interventions, are becoming a common manifestation of defining clinical standards. Such
clinical guidelines are an attempt to specify best medical practice on the basis of existing
evidence. They reinforce the self-regulatory nature of the medical profession and the agency
relationship between physician and patient. They are not concerned with cost, but with the
maximization of health benefit regardless of resource implication. If a treatment is shown
to be beneficial in terms of its effect on an individual, then clinical guidelines may defend
its use even if it is prohibitively expensive. In this respect clinical guidelines may be at odds
with the exclusion criteria operating on the demand-side of the healthcare market, if these
criteria are based on cost alone.
Clinical guidelines are largely concerned with medical practice, with which treatments
are effective, and therefore with which treatment packages to implement according to clini-
calevidence. Theyarelessconcernedwiththemethodologyofhowsuchevidenceisreached,
in the sense that it is taken for granted that the methods used to assemble, disseminate, and
scrutinize the scientific basis of the evidence are acceptable. As such, clinical guidelines aid
performance in at least two ways: a) through increasing the application of scientific knowl-
edge; and b) by acting as minimum standards through which physicians may be assessed.
They help define quality in the clinical setting. Within a managed care environment, such
a definition may tend toward a commitment to providing minimum assured quality. Even
here, however, the commitment may not be sustainable in the long run if quality, as defined
through clinical guidelines, remains in the control of the medical profession.
It is in this environment that the evolving economic guidelines have been introduced.
One presumption is that, by extending clinical guidelines to include economic evaluations,
we would move decision making towards efficient outcomes. This presumption raises ques-
tions about the definitions of efficiency and the relationship that economic guidelines have
to these definitions.
GUIDELINES OF AN ECONOMIC KIND
It is unclear what the underlying objectives of economic guidelines are, but presumably
they are aimed at the pursuit of economic efficiency of some sort. It would be logical,
given the discussion above, for such guidelines to provide incentive structures that are
consistent with clinical guidelines and aid decisions on which treatment to implement on
cost-effective grounds. In this way economic guidelines would tend to complement clinical
guidelines. Of course, implementation of economic guidelines is more difficult than the
implementation of clinical guidelines because economics is concerned with identification
of the marginal gainers and losers arising from treatment. Given treatments may be more
or less cost-effective across different patient populations, and one possible effect of this
variability is a higher demand on the information required to sustain economic guidelines.
In fact, economic guidelines are generally not concerned with such matters. Although
logically stemming from a desire by the third-party payers for cost control, these guidelines
are predominantly concerned with methodologic issues. That is, they are not so much
concerned with which treatments can be supported on cost-effective grounds, but more
with which methods should be employed in establishing the economic criteria used in
particular studies.
980
INTL. J. OF TECHNOLOGY ASSESSMENT IN HEALTH CARE 16:4, 2000

Where are the economic guidelines coming from?
For example, one of the most comprehensive sets of guidelines is that established
by Gold et al. (6). These guidelines recommend the use of cost-effectiveness analysis
and the use of quality-adjusted life-years (QALYs) as a measure of outcome. They also
recommend how to calculate the numerator and the denominator in the cost-effectiveness
ratio, which discount rate to use, and how to deal with uncertainty. Surprisingly, and this
is generally true of all the economic guidelines, while concern is with methods there is
little discussion of the relationship of the various forms of evaluation to the underlying
maximization principle being employed. There is, for example, little discussion of the
efficiency objective being pursued or of the underlying assumptions that must be made to
reconcile the cost-effectiveness ratio with the assumed efficiency goal.
In discussing the prominence of cost-effectiveness analysis in the Gold et al. (6) rec-
ommendations, Kenkel (8) states that, while Gold et al. realize that this form of analysis
has to be related to cost-benefit analysis to be supported by conventional welfare-based
notions of efficiency, they give no justification for this support. Instead, they argue that
they concentrate on cost-effectiveness analysis because it commands widespread approval
as opposed to the “skepticism that greets cost-benefit analysis.” It is one thing to give cus-
tomers what they want under a caveat emptor condition. It is quite another to mislead. If
cost-effectiveness relies on cost-benefit analysis for its use in pursuing efficiency, we must
be clear that the customers know both the relationship between these types of evaluation
and what is meant by efficiency.
Economists have a number of efficiency definitions concerned with the primary objec-
tive of maximizing welfare, where welfare is determined by the value (ultimately defined
in terms of happiness, satisfaction, or utility) attached to resource use. We have already
mentioned technical efficiency: the pursuit of the highest level of output, given available
resources. This definition is consistent with demonstrated efficacy and effectiveness. If
physicians are concerned with maximizing health, it is also consistent with the use of
evidence-based medicine and clinical guidelines. In this way economic guidelines ought to
be complementary to clinical guidelines.
Technical efficiency is not sufficient to ensure that maximum value is attained from
resource use. To attain maximum value, we have to also ensure that productive efficiency
is attained. Productive efficiency is defined as minimizing cost for a given output (or the
corollary, maximizing output for given resource cost). A further level of efficiency is alloca-
tive efficiency that maximizes the value, strictly the value defined in terms of the utility an
individual receives from the output, given resource cost. That value is different from output
is easily seen—the value of an extra day of life will differ among individuals. If outputs are
then allocated among individuals such that no further allocation can make anyone better
off without making someone else worse off, the highest form of efficiency is gained, called
Pareto efficiency. Under certain circumstances, not attained in the healthcare sector as a
result of the market failures discussed above, competitive market forces result in technical,
productive, allocative, and Pareto efficiency. As discussed above, the healthcare market fails
to achieve such efficiency if left to its own devices. This failure is why various regulatory
measures are invoked. These regulatory measures have their own failures in trying to correct
the market’s flaws in pursuing the notions of economic efficiency.
It may be noted that this welfare-based definition of efficiency, Pareto optimality, is held
by some to be rather narrow. In particular, a number of analysts have argued that to define
efficiency purely in terms of individuals’ utilities (or welfare) fails to capture dimensions
that may be relevant when comparing different states. For example, Sen (15) argues that
welfarist notions of efficiency disregard the process through which utility is maximized.
Thus, efficiency is consistent with coercion. Dasgupta (3) goes further, stating that welfare
definitions of efficiency can be consistent with hunger, illness, and unemployment across
large sections of the population. Such economists have argued that a wider definition of
INTL. J. OF TECHNOLOGY ASSESSMENT IN HEALTH CARE 16:4, 2000
981

McGuire, Morris, and Raikou
efficiency is required so that it captures fuller aspects of welfare, not least the notion of “basic
capabilities.” These capabilities may be thought of as the initial natural endowments with
which individuals enter the world. Of primary importance are the initial and the subsequent
health states enjoyed by individuals.
Economic guidelines are therefore supposedly concerned with the pursuit of economic
efficiency. But we must be clear as to what is meant by efficiency. Are we dealing with
productive efficiency, allocative efficiency, or efficiency definitions that go beyond such
narrowly based welfarist concepts? Before turning to such matters, note that one rationale
for concern over resource allocation, namely cost containment, need not necessarily be
consistent with economic efficiency, whether defined as welfarist or extra-welfarist. Cost
containment may pursue a purely budgetary objective without concern for either technical
or productive efficiency. So let us make the grand assumption that decision makers are in
fact concerned with pursuing economic efficiency. If this is taken as their starting point, then
the various guidelines that have been implemented appear to take cost-effectiveness/cost
utility as the basis of their recommendations, the implication being that these instruments
are somehow consistent with efficiency.
The normal justification for the use of either cost-effectiveness or cost-utility analysis
is that they are consistent with the notion of Pareto efficiency and the related concept
of cost-benefit analysis. This concept of cost-benefit analysis is related to the exclusion of
treatment coverage in publicly funded health plans discussed above. Treatment is considered
efficient if the benefits derived, valued in monetary terms, exceed the monetary value of the
costs. The monetary value of the treatment benefits is defined as the amount of money the
beneficiaries of the treatment are willing to pay to ensure that the treatment is undertaken.
This willingness to pay may not be the actual cost to the patient of the treatment, as it
is meant to represent the utility value of treatment to the individual. The decision rule in
cost-benefit analysis is that the treatment should be undertaken if there is a net benefit.
All other things being equal, this rule would determine the size of the healthcare budget;
resources would be allocated until there was no longer a net benefit to be gained from
treatment packages. If there is a budget constraint, the decision rule becomes more complex
and based on comparison of incremental benefit/cost ratios. Treatment packages should be
implemented in order of benefit/cost ratio as long as the packages are mutually independent
and exhibit constant returns to scale. If not, the process becomes even more complex, and
nonlinear programming is required for a comparison across treatment packages.
Cost-benefit analysis, however, remains largely outside the scope of most economic
guidelines. The Pharmaceutical Benefits Scheme of the Australian government actually
excludes cost-benefit analysis, largely on a narrow representation of the means of valuing
health benefits based on a notion of human capital. Most other schemes, again recognizing
the difficulties in attaching a monetary value to health benefits, demote discussion of this
form of economic evaluation. Cost-effectiveness and cost-utility come to the fore.
Arguably, cost-effectiveness and cost-utility analyses address the same issue. They are
both attempts to measure productive efficiency. The only difference is that so-called cost-
utility analysis measures health outcome in two dimensions, normally length and quality of
life, while cost-effectiveness analysis measures only one dimension. On the other hand, cost-
utility analysis moves the decision maker toward allocative efficiency, if the analysis is taken
to encompass preference-based weights on outcomes. Under this second interpretation,
quality of life is not merely an additional dimension of outcome—it encompasses some
utility value attached to various states of health. Considerable literature shows that only
under extremely restrictive conditions is the latter position tenable.
If quality-of-life weights are to reflect health preferences in a manner compatible with
the conventional economic notion of efficiency (Paretian or welfare efficiency), then they
must conform to the axiomatic notions defined by that doctrine. The three most important
982
INTL. J. OF TECHNOLOGY ASSESSMENT IN HEALTH CARE 16:4, 2000

Where are the economic guidelines coming from?
axioms are normally specified in the context of expected utility theory, and are: a) that
preferences exist and are transitive; b) a condition of mutual utility independence that
gives rise to consistency in the face of risky choices; and c) continuity of preferences. It is
important to note that this axiomatic approach to the specification of quality-of-life weights,
and therefore QALYs gained, is a normative theory. It outlines how individuals ought to act
under conditions of uncertainty. There is ample evidence that individuals do not really act in
this way (14). In other words, cost-effectiveness league tables based on cost per QALY are
premised on a normative, rather descriptive view of allocation under uncertainty. If quality
of life is measured across a number of dimensions, an additional axiom is required: utility
independence across the measured dimensions (7).
Johannesson (10) shows that any comparison using health states worse than death
violates the condition of mutual utility independence. Even when we limit considerations
to choices across health states that are better than death, we must impose a number of
restrictions when we examine how QALYs relate to the above axioms. In particular, the
specific relationship between quality of life and length of life must be stated. Normally,
this statement is undertaken through the imposition of a specific functional form given to
the general function U(Y,Q), where U is the utility or welfare derived from the quality (Q)
and length (Y) of life. A number of different specific forms exist; for some of the most
well known see Pilskin et al. (13), Loomes and Mckenzie (9), and Broome (1). In all cases,
particular inferences can be drawn. Both Pilskin et al. (13) and Loomes and Mckenzie (9)
show that constant proportionality over time has to hold and that the risk parameter must be
the same in all health states. The former condition implies that an individual who regards
16 years in perfect health as equivalent to 20 years in a predefined, less-than-healthy state
is assumed to regard 4 years in a healthy state as equivalent to 5 years in the less-than-
healthy state. These conclusions are reached using multiplicative forms of utility function
that assume that quality of life is constant over time. Broome (1) uses an additive utility
function that can take account of quality of life varying through time. Such a function
implies, however, that the utility gained from the quality of life in one period does not
depend on the quality of life in any other period. If this assumption is not adopted, health
profiles that combine quality of life and life expectancy each year throughout an individual’s
life must be calculated if we wish to take account of nonconstancy of quality of life.
If these restrictions hold in one form or another, then QALYs are consistent with
welfare-based definitions of efficiency (5). In principle, they become measures of cardinal
utility, that is, measures that depict the strength of preference with which given health
states are defined. Cardinal measures are required in answering the question, “How much
more effective is this intervention than that intervention?” Two common methods are used
to address this issue: the standard gamble and the time trade-off. The standard gamble
method might give a cardinal approach because under certain assumptions it will rank
gambles by strength of preference. The standard gamble approach is limited, however,
because it assumes that individuals are risk-neutral. Similarly, the time trade-off method
may give cardinal measures of utility. Again, this method is limited in that individuals
must be assumed not to be applying a discount rate to the trade-off. Some have argued
that the standard gamble is more consistent with the axioms of expected utility since it
relates to choice under conditions of uncertainty, while the time trade-off is concerned
with choice under conditions of certainty. Notwithstanding this argument, both instruments
must be used under additional assumptions to make their claim to be cardinal scales that
are consistent with conventional definitions of efficiency. Thus, given restrictions on the
normative manner in which individuals ought to decide under uncertainty, which we know
are extensively violated, and given the adoption of specific instruments that will measure
these preferences under certain assumptions, QALYs get us where we (economists) want:
they move us to Pareto optimality.
INTL. J. OF TECHNOLOGY ASSESSMENT IN HEALTH CARE 16:4, 2000
983

McGuire, Morris, and Raikou
Some have argued that by adopting a wider definition of efficiency, the maximization
of health, QALYs are cardinal representations of utility in some unrestricted sense (4).
This claim is not fully elucidated, however, and the theory underpinning it has not been
outlined. Certainly, QALYs may be consistent with the maximization of health, if this is the
efficiency criterion. They are certainly not, in this sense, consistent with welfare definitions
of efficiency. But neither is it clear, when economists are generally concerned with value
and utility as it relates to value, why health maximization alone should be an appropriate
aim.
Some have gone further. Culyer (2), for example, states that not only are QALYs not
consistent with the conventional Paretian definition of welfare, but also they encompass
an extra-welfarist perspective because they measure the dimension of basic capabilities. In
this sense, the QALY is not a utility measure at all but simply a measure of an individual’s
health characteristics. Thus, a QALY is a QALY no matter whom this QALY belongs to.
That it does not take a different value dependent on the recipient shows, in itself, that we
are not dealing with utility valuation or strength of preference.
Broome (1) reaches the same conclusion, arguing that, whatever the preferences elicited
fromtheQALYrepresent, theycannotbeusedtojustifycardinalrankingsofhealthoutcomes
because they take no account of strength of preference. But QALYs could be assessed in
terms of the amount of “good” (or benefit) derived, not utility that is the utility value placed
on that benefit, but merely the benefit itself. In other words, QALYs could be assessed as an
outcome measure rather than as a utility measure. There may be a relationship between the
outcome and the value of that outcome under certain circumstances. Indeed, if the benefit is
a linear transform of the value function, the outcome measure may be a cardinal measure of
benefit. But this need not, and generally will not, be the case. QALYs, therefore, generally
become merely another dimension of output. Or, to put this another way, cost-utility analysis
collapses into cost-effectiveness analysis.
Cost-effectiveness can then be justified either because it relates to the conventional
notion of efficiency via its relationship to cost-benefit analysis, or simply because the aim
is not to maximize the value of health but merely to maximize health itself.
Letusconsiderthefirstjustification. GarberandPhelps(5)notethat, forthisequivalence
tohold, thepreferencestructureunderlyingcost-effectivenesshastobesimilartothatincost-
benefit analysis. This argument merely returns us to the discussion above; this situation will
only be the case under severe restriction. Moreover, Garber and Phelps make the arguable
point that cost-benefit analysis is concerned with evaluation of public (rival and excludable)
goods, while cost-effectiveness analysis is concerned with the provision of private goods.
As this is the case, an average ratio value is applied in a population perspective, which
drives this form of evaluation away from the conventional notion of welfare efficiency. As
noted above, this notion rests on the principle that if benefits exceed costs then the program
should be undertaken. The rationale is that with this allocation of resources the gainers
can, at least in principle, compensate the losers. Once we start dealing with averages, this
results is no longer true. At the calculated average level, some individuals are gainers and
some are losers. Compensation can no longer take place, as the average is based precisely
on there being gainers and losers. We are no longer in a world of conventional welfare
maximization.
Now consider the second justification, that cost-effectiveness, because it pursues pro-
ductive efficiency, is consistent with health maximization. This statement is an obvious
truth. It leaves the issue of how to value health to one side. In fact, the Canadian guide-
lines explicitly highlight that if cost-effectiveness is undertaken through the calculation of
a cost per QALY, then it becomes necessary to calculate the societal willingness to pay
in order to make any resource allocation decision. This decision is internalized and left
with the decision maker. This decision-making process is bound to be influenced by the
984
INTL. J. OF TECHNOLOGY ASSESSMENT IN HEALTH CARE 16:4, 2000

Where are the economic guidelines coming from?
budgetary conditions. Under such circumstances, we do not get the true societal valuation
but a constrained valuation based on prior funding decisions.
In other words, under these circumstances economic guidelines that support the use
of cost-effectiveness analysis avoid the real issue—how to value healthcare interventions.
The guidelines, through pushing methods that rest upon cost-effectiveness analysis, leave
the valuation decision to others, unless some explicit cut-off that defines a threshold for
cost-effective and cost-ineffective treatments is specified.
Let us assume, however, that the decision maker is truly only concerned with health
maximization. To maximize health, as defined by the QALY, resources are allocated to those
who benefit most from treatment. This approach has certain equity implications, of course.
Yet economic guidelines have little to say on issues of equity. Perhaps this silence is not
surprising, since they do not make it clear what efficiency criteria are being invoked.
CONCLUSIONS
Under circumstances in which resources are limited, choice is inevitable, and it seems
misplaced to attempt to get the most from scarce resources. Economic concerns in the
healthcare sector are legitimate. However, to get the most from resources one has to be
clear about what one wants to maximize: is it health or the value of health? To add to this
confusion, QALYs are used by different analysts to measure both health per se and the
value of health. The extent to which they do the latter is open to dispute and certainly only
occurs under very restrictive conditions if traditional notions of value, essentially expected
utility, are used. The notions of efficiency and the objectives being pursued are less than
clear in the healthcare sector.
These are fundamental methodologic concerns, but they are not the concerns with
which most economic guidelines grapple. These guidelines do deal with methodology,
but in a superficial manner because they do not relate methods to notions of economic
efficiency. In this sense the guidelines do not discuss the methodologic issues associated
with the supportive use of cost-effectiveness via its technical relationship with cost-benefit
analysis. Neither do they recognize the methodologic concerns over the use of QALYs as
preference measures. They do not address the definition of efficiency being pursued by the
various forms of economic evaluation. Rather, economic guidelines as currently formulated
are concerned with descriptions of method after the decision to undertake an economic
evaluation of healthcare technologies has been made.
In most cases, and interpreting the QALY as a measure consistent with health maxi-
mization, guidelines may be justified as dealing with productive efficiency and minimizing
cost for given output. It is unclear whether the practitioners of this method, let alone the
decision makers, fully understand the implications and the theoretical relationships invoked
in these calculations. So why are the guidelines popular? Presumably because decision
makers and practitioners believe that, given their concern with methods, these guidelines
will result in the appropriate analysis being undertaken. Yet, as is implied from the above,
the definition of appropriate is rather limited.
Adopting the guidelines uncritically will not necessarily lead to improved efficiency.
Even if the outcome measures and methods are decided on, unless it is clear what the objec-
tives to be pursued are, it is not possible to improve efficiency. Given that most guidelines
are partial, in that they are generally applied to one part of the healthcare sector, if the
definition of efficiency itself is far from clear and decision makers rarely pursue unidimen-
sional objectives in any case, it is unlikely that even full implementation of these guidelines
will necessarily improve resource allocation. That said, the specification of guidelines does,
in itself, help decision makers address appropriate questions within resource-constrained
systems.
INTL. J. OF TECHNOLOGY ASSESSMENT IN HEALTH CARE 16:4, 2000
985

McGuire, Morris, and Raikou
REFERENCES
1. Broome J. QALYs. Journal of Public Economics. 1993;50:149-163.
2. Culyer A. The normative economics of health care finance and provision. In: McGuire A,
Fenn P, Mayhew K, eds. Providing health care. Oxford: OUP; 1990:65-98.
3. Dasgupta P. Positive freedom, markets and welfare state. Oxford Review of Economic Policy.
1986;2:25-36.
4. Dolan P, Gudex C, Kind P, Williams A. Valuing health states: A comparison of methods. Journal
of Health Economics. 1996;15:209-233.
5. Garber AM, Phelps CE. Economic foundations of cost-effectiveness analysis. Journal of Health
Economics. 1997;16:1-33.
6. Gold M, Siegel J, Russell L, Weinstein M. Cost-effectiveness in health and medicine. New York:
OUP; 1996.
7. Keeney H, Raiffa H. Decisions with multiple objectives: Preferences and value trade-offs.
New York: Wiley; 1976.
8. Kenkel D. On valuing morbidity, cost-effectiveness analysis, and being rude. Journal of Health
Economics. 1997;16:749-757.
9. Loomes G, Mckenzie L. The scope and limitations of QALY measures. Soc Sci Med. 1988;27:299-
308.
10. Johannesson M. Theory and methods of economic evaluation of health care. Amsterdam: Kluwer;
1996.
11. Newhouse J. Reimbursing health plans and health providers: Efficiency in production versus
selection. Journal of Economic Literature. 1996;34:1236-1263.
12. Pauly MV. Taxation, health insurance and market failure in the medical economy. Journal of
Economic Literature. 1986;24:629-675.
13. Pilskin J, Shepard D, Weinstein M. Utility functions from life years and health status. Operations
Research. 1980;28:206-224.
14. Schoemaker P. The expected utility model: Its variants, purposes, evidence and limitations. Jour-
nal of Economic Literature. 1982;20:529-563.
15. Sen A. Personal utilities and public judgments: Or what’s wrong with welfare economics? Eco-
nomic Journal. 1979;589:537-558.
16. Weisbrod B. The health care quadrilemma: An essay on technological change, insurance, quality
of care and cost containment. Journal of Economic Literature. 1991;24:553-572.
986
INTL. J. OF TECHNOLOGY ASSESSMENT IN HEALTH CARE 16:4, 2000